UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Environmental heterogeneity produced by termitaria in Western Uganda with special reference to mound… Melton, Derek Arthur 1975

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1975_A6_7 M44_7.pdf [ 24.83MB ]
Metadata
JSON: 831-1.0058259.json
JSON-LD: 831-1.0058259-ld.json
RDF/XML (Pretty): 831-1.0058259-rdf.xml
RDF/JSON: 831-1.0058259-rdf.json
Turtle: 831-1.0058259-turtle.txt
N-Triples: 831-1.0058259-rdf-ntriples.txt
Original Record: 831-1.0058259-source.json
Full Text
831-1.0058259-fulltext.txt
Citation
831-1.0058259.ris

Full Text

ENVIRONMENTAL HETEROGENEITY PRODUCED BY TERMITARIA IN WESTERN UGANDA WITH SPECIAL REFERENCE TO MOUND USAGF BY VERTEBRATES by DEREK ARTHUR MELTON B.Sc. University of Leicester, 1972 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE i n the Department of Zoology We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA In presenting th is thes is in p a r t i a l fu l f i lment of the requirements for an advanced degree at the Un ivers i ty of B r i t i s h Columbia, I agree that the L ibrary shal l make it f r ee ly ava i lab le for reference and study. I fur ther agree that permission for extensive copying of th is thes is for scho la r ly purposes may be granted by the Head of my Department or by his representat ives . It is understood that copying or pub l i ca t ion of th is thes is fo r f inanc ia l gain sha l l not be allowed without my wri t ten permission. Department of The Univers i ty of B r i t i s h Columbia Vancouver 8, Canada r ABSTRACT i Much e c o l o g i c a l work has d e a l t w i t h the ways e x t r i n s i c f a c t o r s (weather, f o o d , o t h e r a n i m a l s , a p l a c e i n w h i c h t o l i v e ) a f f e c t a s p e c i e s ' d i s t r i b u t i o n and abundance. I n t h i s s t u d y I chose t o examine a v e r y c o n s p i c u o u s h a b i t a t f e a t u r e o f A f r i c a n savanna, i n o r d e r t o ask two q u e s t i o n s r e l a t e d t o such e n v i r o n m e n t a l f a c t o r s . F i r s t , i n what ways do t e r m i t a r i a change the environment f o r v e r t e b r a t e s p e c i e s ? Second, what t e r m i t a r i a f e a t u r e s a r e u t i l i z e d by v e r t e b r a t e s and a r e any s p e c i e s a f f e c t e d i n t h e i r l o c a l d i s t r i b u t i o n by them? The d i s t r i b u t i o n and d e n s i t y o f e p i g e a l t e r m i t e s t r u c t u r e s was d e s c r i b e d f o r one a r e a , Ruwenzori N a t i o n a l P a r k (R.N.P.) Western Uganda, u s i n g ground and a e r i a l methods. E n v i r o n m e n t a l changes caused by t e r m i t a r i a were i n v e s t -i g a t e d i n f o u r a r e a s : the e f f e c t s o f t e r m i t e mounds on l a n d r e l i e f ; the i n t e r i o r s o f mounds; the s o i l o f mounds; and the e f f e c t s o f mounds on v e g e t a t i o n . The f o l l o w i n g c o n c l u s i o n s were drawn as t o the p o t e n t i a l i m p o r t a n c e o f changes found t o v e r t e b r a t e s . The h i g h p e r c e n t a g e o f ground s u r f a c e c o v e r e d by mounds s u g g e s t s t h a t any p l a n t o r s o i l changes o c c u r r i n g a t mounds ar e l i k e l y t o be o f s i g n i f i c a n c e t o h e r b i v o r e s . Mounds were i d e n t i f i e d as b e i n g d i r e c t l y o r i n d i r e c t l y i n v o l v e d i n w a l l o w f o r m a t i o n . Both l i v e and dead mounds can o f f e r a r e a s o f complex s u b d i v i d e d space a v a i l a b l e t o s m a l l and medium-s i z e d v e r t e b r a t e s . Mounds ar e o f t e n c o n c e n t r a t e d s o u r c e s o f m i n e r a l s d i r e c t l y a v a i l a b l e t o a n i m a l s v i a geophagy, o r i n d i r e c t l y v i a mound v e g e t a t i o n . The m a j o r i t y o f mounds i n R.N.P. were i n f a c t v e g e t a t e d , and i t was found t h a t mounds i n c r e a s e d t h e range o f n u t r i t i v e v a l u e o f g r a s s e s a v a i l a b l e t o a n i m a l s by s e l e c t i v e g r a z i n g . V e r t e b r a t e usage o f mounds was s t u d i e d u s i n g b o t h d i r e c t and i n d i r e c t methods. The f o l l o w i n g c o n c l u s i o n s were drawn. Mounds were used by many v e r t e b r a t e s p e c i e s as r a i s e d p l a t f o r m s . S p e c i e s u s i n g mounds as t e r r i t o r i a l markers may be a f f e c t e d i n t h e i r d i s t r i b u t i o n by them. Mound i n t e r i o r s were u s u a l l y o c c u p i e d by a number o f v e r t e b r a t e s p e c i e s when a c c e s s i b l e . D u r i n g f l o o d s and f i r e s t e r m i t a r i a a f f e c t the l o c a l d i s t r i b u t i o n o f c e r t a i n s m a l l mammal s p e c i e s i n R.N.P. L i t t l e geophagia o c c u r r e d a t mounds o r el s e w h e r e i n R.N.P. P r e f e r e n t i a l g r a z i n g o f mound g r a s s e s was demonstrated i n R.N.P. T h i s may complete the n u t r i e n t r e q u i r e m e n t o f a n i m a l s t h r o u g h o u t t h e y e a r , o r be m a i n l y i m p o r t a n t as a dry season f o r a g e . O r y c t e r o p u s a f e r was the o n l y v e r t e b r a t e r e g u l a r l y d i g g i n g i n t o mounds f o r i n s e c t f o o d i n R.N.P. A l t h o u g h t h i s a n i m a l d i d n o t p r e f e r e n t i a l l y f o r a g e a t mounds, t h e mound b u i l d i n g M . s u b h y a l i n u s was the main s p e c i e s e a t e n . Mounds may i n f l u e n c e the d i s t r i b u t i o n o f 0 . a f e r . The i m p o r t a n c e t o v e r t e b r a t e s o f the mound p r o c e s s e s d e s c r i b e d was a l s o d i s c u s s e d f o r o t h e r a r e a s than R.N.P. Emphasis was l a i d on the v a r i a t i o n found t o o c c u r between the mounds o f d i f f e r e n t t e r m i t e s p e c i e s , and between a r e a s . I t was c o n c l u d e d t h a t i t was p o s s i b l e t h a t t e r m i t e s a r e i n c r e a s i n g the c a r r y i n g c a p a c i t y o f many a r e a s o f A f r i c a , t h r o u g h the d i v e r s e mound p r o c e s s e s d e s c r i b e d . i i i TABLE OF CONTENTS Page I. INTRODUCTION 1 I I . STUDY AREA 7 I I I . ENVIRONMENTAL CHANGES CAUSED BY TERMITARIA . . . . 16 1. A PRELIMINARY SURVEY 16 2. METHODS 27 i ) The d i s t r i b u t i o n and density of t e rmi ta r i a 27 i i ) Environmental changes introduced by t e rmi ta r i a 3 4 Land r e l i e f 34 The i n t e r i o r of dead mounds 34 The s o i l of termite mounds 41 The ef fects of mounds on vegetat ion 43 3. RESULTS 47 i ) The d i s t r i b u t i o n and density of t e rmi ta r i a 47 Macrotermes subhyalinus 47 Macrotermes be l l i co sus 52 Large mounds of Odontotermes kibarens i s 54 Odontotermes sp. subterranean nests with shafts 57 A e r i a l v i s u a l transect re su l t s . . . . 59 i i ) Environmental changes introduced by t e rmi ta r i a 59 Land r e l i e f 59 i v Page Mound i n t e r i o r s 67 The microclimate within termitaria 73 The s o i l of termite mounds 81 The vegetation of termite mounds . 92 Chemical composition of grasses .. 97 4. DISCUSSION AND CONCLUSIONS 105 i) Termitaria i n R.N.P.: t h e i r structure and d i s t r i b u t i o n 105 Conclusion i 113 i i ) Environmental changes introduced by termitaria ;. .. 114 The e f f e c t s of mounds on land r e l i e f 114 Conclusion i i 116 Accessible space within termitaria 117 The microclimate of mound space .. 119 Conclusion i i i 124 The s o i l of termite mounds ...... 124 Conclusion i v 130 The vegetation of termite mounds . 130 Conclusion v 137 IV. THE USE OF .TERMITARIA BY VERTEBRATES 139 1. METHODS: THE USE OF MOUNDS BY VERTEBRATES , ... 139 i) Mounds as fixed elevated points 139 i i ) The occupation of mound i n t e r i o r s by vertebrates 140 V Page The species involved 140 The use of mounds as den s i t e s by a small carnivore 140 The use of large burrows i n mounds and elsewhere by medium-sized mammals 141 Termitaria and small mammal d i s t r i b u t i o n 142 i i i ) The use of moundssoil and vegetation 145 iv) The use of termi t a r i a as a concentrated source of food 146 2. RESULTS: THE USE OF MOUNDS BY VERTEBRATES . . 150 i) Mounds as fixed elevated points 150 i i ) The occupation of mound i n t e r i o r s by vertebrates 161 The species involved 161 The use of mounds as den s i t e s by a small carnivore 17 2 The use of large burrows i n mounds and elsewhere by medium-sized mammals 173 Termitaria and small mammal d i s t r i b u t i o n 173 Crater Region 173 Ishasha 191 i i i ) The use of mound s o i l and vegetation 195 iv) The use of termitaria as a concentrated source of food 197 Orycteropus afer 198 The d i s t r i b u t i o n of orycteropus digs 202 v i Page Orycteropus foraging behaviour and use of large burrows from direct observation 206 Nightly distance travelled by orycteropus 207 Reaction to predators 209 The diet of orycteropus 211 3. DISCUSSION: THE USE OF MOUNDS BY VERTEBRATES 215 i) Mounds as fixed elevated points 215 i i ) The occupation of mound interiors by vertebrates 219 The use of large burrows in mounds and elsewhere by medium-sized mammals 222 Termitaria and small mammal distr ibution .'"223 Crater Region . . 223 Rodents, termitaria, and f i re in the Crater Region 224 Mounds, flooding, and rodents . . 226 i i i ) The use of mound s o i l and vegetation . . 227 iv) The use of termitaria as a concentrated source of food 232 V. CONCLUSIONS: THE USE OF MOUNDS BY VERTEBRATES 237 1. MOUNDS AS FIXED ELEVATED POINTS 237 2. THE OCCUPATION OF MOUND INTERIORS 237 3. THE USE OF MOUND SOIL 238 4. THE USE OF MOUND VEGETATION 239 5. THE USE OF TERMITARIA AS A CONCENTRATED SOURCE OF FOOD 23 9 v i i Page 6. CONCLUSION 240 VI. TABLES 242 VII. LITERATURE CITED 32 7 VIII. APPENDICES 3 37 1. Nearest neighbour calculations 337 2. A comparison of mound density figures obtained from ground transects and a e r i a l v i s u a l transects 339 3. Mound volume estimation 340 4. Density of termite mound s o i l 340 5. The s o l d i e r and worker caste termite mandibles found i n droppings of Orycteropus afer 342 v i i i LIST OF TABLES Table Page I. Between area comparison of M. subhyalinus mound densities 243 II . Dispersion of M. subhyalinus mounds within study areas 244 III . Dispersion of M. subhyalinus mounds within two Crater Region aggregations 245 IV. Dispersion of M. subhyalinus mounds from a e r i a l photographic transects 246 V. Ishasha ground transect r e s u l t s 247 VI. Variance to mean analyses for mound densities i n transects only where M. be l l i c o s u s occurred 248 VII. Dispersion of mounds at Ishasha from two a e r i a l photographic transects 249 . VIII. Comparison of M. be l l i c o s u s and non-Macrotermes mound dispersions from two a e r i a l photographic transects 250 IX. 0. kibarensis mound dispersion from three a e r i a l photographic transects 251 X. 0. f u l l e r i nest densities from ground transects run a f t e r a burn i n the Crater Region 252 XI. M. subhyalinus d i s t r i b u t i o n i n the Crater Region 253 XII. 0. f u l l e r i nest dispersions from Crater Region trapping grids 254 XIII. Odontotermes and Macrotermes r e l a t i v e dispersion patterns i n the Crater Region using data from small mammal trapping g r i d maps 255 XIV. Comparison of a e r i a l and ground transect results for per hectare Macrotermes mound densities 256 XV. Comparison of a e r i a l v i s u a l density estimates (adjusted) with ground transect results for four M. subhyalinus mound classes 25 7 i x T a b l e Page XVI. Mound Dimensions 258 X V I I . Comparison o f M. s u b h y a l i n u s mound volumes and n e a r e s t n e i g h b o u r d i s t a n c e s 259 X V I I I . Mound s u r f a c e a r e a and the i n c r e a s e i n s u r f a c e a r e a caused by mounds 260 XIX. The a s s o c i a t i o n between O. k i b a r e n s i s mounds and w a l l o w s , from two a e r i a l p h o t o g r a p h i c mosaics - 261 XX. The n a t u r e and d i s t r i b u t i o n o f h o l e s i n the f o u r t e r m i t a r i a t y p e s s t u d i e d 262 XXI. D e t a i l s o f a l l h o l e s i n one h e c t a r e o f c r a t e r b ottom examined a f t e r the January 1973 b u r n .. 264 X X I I . Between a r e a comparison o f the d e n s i t y o f M. s u b h y a l i n u s mounds w i t h c a v e - i n s 265 X X I I I . Between a r e a comparison o f the d e n s i t y o f M. s u b h y a l i n u s mounds w i t h l a r g e burrows 26 5 XXIV. Between a r e a comparison o f the d e n s i t y o f M. s u b h y a l i n u s mounds w i t h s m a l l burrows 266 XXV. Between a r e a comparison o f the d e n s i t y o f M. s u b h y a l i n u s mounds w i t h h o l e s from w e a t h e r i n g i 267 XXVI. Between a r e a comparison o f the d e n s i t y o f M. s u b h y a l i n u s mounds w i t h any e n t r y p o i n t ... 26 8 XXVI I . W i t h i n a r e a comparisons o f mound dime n s i o n s .. 269 X X V I I I . Volumes o f mounds w i t h n o n - t e r m i t e made e n t r y h o l e s 2 71 XXIX. Volumes w i t h i n s h a f t s o f O. f u l l e r i n e s t s .... 272 XXX. D e n s i t y and volume o f f i v e l a r g e burrows i n the n o r t h o f R.N.P 273 XXXI. A n a l y s i s o f mound and c o n t r o l s o i l samples ... 274 X X X I I . Comparison o f s u r f a c e s o i l from M. s u b h y a l i n u s mounds w i t h t h a t o f a d j a c e n t c o n t r o l a r e a s u s i n g the S t u d e n t ' s t - t e s t f o r p a i r e d v a r i a t e s 2 77 X X X I I I . Comparison o f m i n e r a l w e i g h t s i n M. s u b h y a l i n u s mounds w i t h m i n e r a l w e i g h t s i n t o p s o i l , 1. Nyamagasani 2 78 X T a b l e Page XXXIV. Comparison o f m i n e r a l w e i g h t s i n M. s u b h y a l i n u s mounds w i t h m i n e r a l w e i g h t s i n t o p s o i l , 2. Mweya 2 79 XXXV. Comparison o f m i n e r a l w e i g h t s i n M. s u b h y a l i n u s mounds w i t h m i n e r a l w e i g h t s i n t o p s o i l , 3. C r a t e r Region 280 XXXVI. Comparison o f m i n e r a l w e i g h t s i n M. b e l l i c o s u s mounds a t I s h a s h a w i t h m i n e r a l w e i g h t s i n t o p s o i l 2 81 XXXVII. Comparison o f m i n e r a l w e i g h t s i n O. k i b a r e n s i s mounds a t I s h a s h a w i t h m i n e r a l w e i g h t s i n t o p s o i l 282 X X X V I I I . P e r c e n t a g e o f M. s u b h y a l i n u s mounds i n f o u r v e g e t a t i o n c o v e r c l a s s e s 283 XXXIX. P e r c e n t a g e o f M. b e l l i c o s u s mounds i n f o u r v e g e t a t i o n c o v e r c l a s s e s 283 XL. P e r c e n t a g e o f O. k i b a r e n s i s mounds i n f o u r v e g e t a t i o n c o v e r c l a s s e s 283 X L I . V e g e t a t i o n d e t a i l s o f 0. k i b a r e n s i s mounds from t h r e e a e r i a l p h o t o g r a p h i c t r a n s e c t s ..... 284 X L I I . Between a r e a comparison o f M. s u b h y a l i n u s v e g e t a t i v e c o v e r . 1. Number o f ba r e mounds p e r h e c t a r e 285 X L I I I . Between a r e a comparison o f M. s u b h y a l i n u s v e g e t a t i v e c o v e r . 2. Number o f mounds w i t h b a s a l g r a s s p e r h e c t a r e 285 XLIV. Between a r e a comparison o f M. s u b h y a l i n u s v e g e t a t i v e c o v e r . 3. Number o f g r a s s c o v e r e d mounds p e r h e c t a r e 2 86 XLV. Between a r e a comparison o f M. s u b h y a l i n u s v e g e t a t i v e c o v e r . 4. Number o f t h i c k e t e d mounds p e r h e c t a r e 2 86 XLVI. W i t h i n a r e a mound d i m e n s i o n s , w i t h a comparison between bare and v e g e t a t e d mounds .. 287 X L V I I . W i t h i n a r e a v a r i a t i o n i n mound s i z e and v e g e t a t i o n c o v e r a t Mweya 289 X L V I I I . S i g n i f i c a n c e o f d i f f e r e n c e s between t h e means o f mound and t h i c k e t c h a r a c t e r i s t i c s between t h r e e t o p o g r a p h i c a r e a s a t Mweya 290 X I Table Page XLIX. Detailed vegetation description of ten thicketed M. subhyalinus mounds near Mweya .... 291 L. Grass biomass on four M. subhyalinus mounds and four control areas at Mweya 292 LI. Comparison of analysis means from i n d i v i d u a l mound and control samples 29 3 LII. Animal a c t i v i t y at termite mounds i n R.N.P. ... 295 LIII. Defecation by medium sized mammals on termite mounds i n R.N.P 296 LIV. Baboon faeces counts i n 40 meter X 40 meter area at Ishasha containing four 0. kibarensis mounds and two small bare Trinervitermes mounds 297 LV. Baboon faeces analysis 298 LVI. Germination t r i a l s for shrub seeds from baboon faeces and shrub seeds from ripe f r u i t at Ishasha 299 LVII. The rubbing of mounds by large mammals 300 LVIII. Vertebrate species occupying the i n t e r i o r of holed M. subhyalinus mounds 301 LIX. Vertebrate species occupying the i n t e r i o r of M. be l l i c o s u s mounds 302 LX. Summary of res u l t s from 0. f u l l e r i nest excavations 30 3 LXI. Details of excavation r e s u l t s from four 0. f u l l e r i nests with signs of vertebrate occupa-tion 304 LXII. Results of 0. kibarensis mound excavations .... 305 LXIII. Details of 2 3 banded mongoose dens on Mweya peninsula 306 LXIV. a. Burrow use by six classes of medium-sized mammals on Mweya Peninsula: Nov. '72-June '73 30 8 b. Monthly burrow usage expressed as percent of burrow-days or burrow-nights monitored ... 309 LXV. Results of small mammal removal trapping from "shafted" and "non-shafted" grids i n the Crater x i i Table Page Region, expressed as the number of animals caught per 120 trap-nights 310 LXVI. Correlation between the numbers of small mammals caught and the numbers of Odontotermes shafts occurring on 16 trapping grids i n the Crater Region '. 311 LXVII. Small mammal trapping: monthly comparisons of cumulative percent captures on traps ranked with respect to distance from Odontotermes shafts, and expected cumulative percent captures assuming equal chance of capture at each trap point 312 LXVIII. Significance of p o s i t i v e difference between observed and expected number of small mammals caught i n traps nearer to shafts than the distance coinciding with the maximum po s i t i v e difference i n cumulative percent*small mammal captures 313 LXIX. Snap-trapping results i n the Crater Region immediately a f t e r the January 19 73 burn 314 LXX. Number of rodents captured during 36 trap-days and nights each month, live- t r a p p i n g i n the Crater Region a f t e r the January 1973 burn .... 315 LXXI. Small mammal stomach analyses. 316 LXXII. 0. kibarensis mound rodent droppings transects: A p r i l 1973 317 LXXIII. 0. kibarensis mound rodent droppings transects: comparison between holed and non-holed mounds -March 1973 318 LXXIV. Heights of Sporobolus pyramidalis l e a f blade on 0. kibarensis mounds and control areas at Ishasha 319 LXXV. D i s t r i b u t i o n of 0. afer digs and large burrows i n termitaria from ground transects 320 LXXVI. Results of orycteropus dig transects 321 LXXVII. Details of 12 successful orycteropus watches .. 323 LXXVIII. Details of 16 stops which an animal made during 35 minutes of observation on 19th July 1973 324 LXXIX. Analysis of orycteropus droppings 325 x i i i LIST OF FIGURES Figure Page 1. The p o s i t i o n of Ruwenzori Nat ional Park i n East A f r i c a 8 2. Ruwenzori Nat ional Park showing the s i x study-areas 10 3. R a i n f a l l at Nyamagasani 11 4. R a i n f a l l at Mweya . . . 11 5. R a i n f a l l i n the Crater Region 12 6. R a i n f a l l at Kamulikwezi 12 7. R a i n f a l l at Ishasha 13 8. M. subhyalinus mound near Mweya 17 9. Excavated M. subhyalinus mound at Nyamagasani . . 18 10. M. b e l l i c o s u s mound at Ishasha 20 11. Mound shown i n Figure 10 p a r t l y excavated 21 12. Turret of fresh growth on a M. b e l l i c o s u s mound . 22 13. 0. k ibarens i s mound i n seasonal f loodland at Ishasha 2 3 14. Generalized sect ion through an O. k ibarens i s mound 23 15. O. f u l l e r i turre t s above subterranean nest 25 16. Generalized sect ion through an 0. f u l l e r i nest . . 26 17. Ruwenzori Nat ional Park showing the p o s i t i o n of a e r i a l transects 29 18. Wing s t ru t c a l i b r a t i o n for v i s u a l a e r i a l transects 31 19. The Ishasha Camp area showing pos i t ions of three a e r i a l photographic transects 33 20. a. A cave-in probably caused by large mammal trampling 36 x i v Page b. An orycteropus burrow i n a M. subhyalinus mound at Kamulikwezi 36 c. Small mammal burrows i n a worn M. subhyalinus mound 37 d. Weathered M. be l l i c o s u s mound showing hole .. 37 e. M. be l l i c o s u s mound rubbed by large mammals .. 38 21. The placement of three temperature probes i n the f i r s t microclimate examination 40 22. The placement of two pairs of temperature probes i n the second microclimate examination 4Q 23. The placement of two pairs of temperature probes i n the t h i r d microclimate examination 42 24. The placement of two pairs of temperature probes i n the fourth microclimate examination 42 25. Comparison of M. subhyalinus mound densities with non-Macrotermes mound densities 48 26. Comparison of M. bell i c o s u s mound densities with non-Macrotermes mound densities 5 3 27. Part of a photographic transect at Ishasha 56 28. Comparison of M. subhyalinus mound volumes with nearest neighbour distances; Kamulikwezi 61 29. Comparison of M. subhyalinus mound volumes with nearest neighbour distances; Kasenyi 62 30. Comparison of M. subhyalinus mound volumes with nearest neighbour distances; Crater Region 63 31. Comparison of M. subhyalinus mound volumes with nearest neighbour distances; Mweya 64 32. Comparison of M. subhyalinus mound volumes with nearest neighbour distances; Nyamagasani ....... 65 33. Comparison of M. subhyalinus mound volumes with nearest neighbour distances; a l l f i v e northern study areas 66 34. O. kibarensis workers bu i l d i n g up t u r r e t i n the wet season at Ishasha 68 35. Nature of the space inside a dead M. subhyalinus mound 71 X V Page 36. Nature of the space ins ide a dead M. b e l l i c o s u s mound 71 37. Temperatures ins ide a l i v e mound, ins ide a dead mound and i n shade at the ground surface 74 38. Temperature i n a non-mound large burrow and i n shade at the ground surface 75 39. Relat ive humidity i n a non-mound large burrow and i n shade at the ground surface 76 40. Relat ive humidity i n a dead mound burrow, a non-mound burrow and at the ground surface 77 41. Rela t ive humidity i n a l i v e mound large burrow, a cont ro l burrow, and at the ground surface . . . . 78 42. Temperature i n a dead mound large burrow and a cont ro l burrow 79 43. Temperature i n a l i v e mound large burrow and a cont ro l burrow 80 44. Mound and contro l surface s o i l samples; ex-changeable cations 82 45. Mound and contro l surface s o i l samples; n i t r a t e nitrogen 84 46. Mound and contro l surface s o i l samples; ava i lab le phosphorus 84 47. Mound and contro l surface s o i l samples; conduct iv i ty 85 48. Pos i t ions of e ight s o i l samples from two mounds at Mweya 86 49. S o i l analyses for two mounds at Mweya; exchangeable cations 87 50. S o i l analyses for two mounds at Mweya; n i t r a t e nitrogen 88 51. S o i l analyses for two mounds at Mweya; ava i lab le phosphorus 88 52. S o i l analyses for two mounds at Mweya; conduct iv i ty 89 53. M. subhyalinus grass cover at Mweya 95 54. 0. k ibarens i s grass cover at Ishasha 96 xvi Page 55. Chemical composition of grasses from mounds and control areas; calcium 98 56. Chemical composition of grasses from mounds and control areas; magnesium 99 57. Chemical composition of grasses from mounds and control areas; sodium 100 58. Chemical composition of grasses from mounds and control areas; potassium 101 59. Chemical composition of grasses from mounds and control areas; s i l i c a - f r e e ash 102 60. Chemical composition of grasses from mounds and control areas; crude protein 103 61. Fish eagle eating T i l a p i a on a M. subhyalinus mound 151 62. Baboons use mounds as lookout posts and rest areas at Ishasha 152 63. Kob using mound as a lookout post 155 64. a. Four banded mongooses on M. subhyalinus mound 156 b. Banded mongoose anal marking mound 157 c. Banded mongoose chin marking mound 157 d. Banded mongoose defecating on mound 15 8 65. Graph of percentage bare Macrotermes mounds against percentage mounds rubbed by large mammals 160 66. Excavated M. subhyalinus mounds 163 67. Excavated M. be l l i c o s u s mounds 166 68. Excavated 0. f u l l e r i nests 169 69. Excavated 0. kibarensis mound 171 70. Four large burrows excavated at Mweya 174 71. Graph - of t o t a l small mammals captured i n Crater Region against number of 0 don to te rme s shafts ... 178 72. Graph of weight against length for M. natalensis 179 x v i i Page 73. Graph of weight against length for A. n i l o t i c u s .. 180 74. Graph of weight against length for L. s t r i a t u s .. 181 75. Graph of weight against length for L. sikapusi .. 182 76. Graph of weight against length for M. t r i t o n .... 18 3 77. Graph of weight against length for Z_. hildegardeae 183 78. Graph of weight against length for M. minutoides 183 79. Results of snap-trapping i n the Crater Region; mound grids ... 184 80. Results of snap-trapping i n the Crater Region; non-mound grids 184 81. Graph of percentage cumulative captures ( a l l species) against ranked trap points 186 82. Graph of percentage cumulative captures (M. natalensis) against ranked trap points 18 7 83. Graph of percentage cumulative captures (L. s triatus) against ranked trap points 188 84. Graph of percentage cumulative captures (nocturnal species) aginast ranked trap points .. 189 85. Graph of percentage cumulative captures (diurnal species) against ranked trap points .... 190 86. Results of monthly snap-trapping at Ishasha ..... 192 87. Results of live- t r a p p i n g at Ishasha 194 88. M. b e l l i c o s u s mound s h e l l and adjacent s o i l used as a wallow by a medium sized mammal 196 89. M. subhyalinus mound showing orycteropus food digs i n side 199 90. Orycteropus defecation scrape 200 91. Position of eight large burrows near transect IJ at Ishasha 203 92. 0. afer digging on Mweya Peninsula 208 93. Map showing the routes followed by 0. afer during twelve successful watches on Mweya Peninsula .... 210 x v i i i Page 94. The as soc ia t ion between 0. afer d i e t and r a i n f a l l , at Mweya 212 95. A graph of r a i n f a l l index against percentage ants i n 0. afer d i e t . . 213 96. Pathways of mineral movement inf luenced by vegetated termite mounds 2 33 ACKNOWLEDGEMENTS I would l i k e t o acknowledge t h e h e l p o f t h e R o y a l S o c i e t y and t h e Leverhulme T r u s t e e s who g r a n t e d a R o y a l S o c i e t y Leverhulme S t u d e n t s h i p w h i c h f i n a n c e d a l l f i e l d w o r k i n t h i s s t u d y . My thanks go t o t h e D i r e c t o r o f Uganda N a t i o n a l P a r k s f o r a l l o w i n g t h e p r o j e c t t o t a k e p l a c e , and t o t h e Uganda I n s t i t u t e o f E c o l o g y f o r p r o v i d i n g accomodation and l a b o r a t o r y f a c i l i t i e s . P a r t i c u l a r thanks go t o S a l l y f o r many t h i n g s , n o t t h e l e a s t b e i n g becoming a f i e l d b i o l o g i s t f o r a y e a r . Thanks a l s o go t o Dr. K e i t h E l t r i n g h a m , t h e n t h e D i r e c t o r o f t h e Uganda I n s t i t u t e of E c o l o g y , f o r p r a c t i c a l h e l p t h r o u g h o u t my s t a y a t Mweya i n c l u d i n g p i l o t i n g t h e Cessna 182 d u r i n g a e r i a l t r a n s e c t s . F i n a l l y I would l i k e t o thank Dr.I.McT.Cowan f o r s u p e r v i s i n g t h e w r i t i n g - u p o f t h i s work. 1 I. INTRODUCTION It has long been recognized that there is an intimate relationship between the two fundamental questions of what factors determine the distri-bution of a species, and what factors influence the abundance of a species. Howell (1924) and Grinnell (1928) concluded that it was the totality of the biotic and abiotic environment, which determined a species presence and density. They also emphasized the difficulty of identifying the important environmental factors for an isolated group of individuals, let alone producing generalizations at the species level, or for higher taxa. Macfadyen (1963) reviewed the limitations of an autecological approach in examining distribution phenomena. He pointed out that although the presence or absence of a species, on both the zoogeographical scale and the local scale, was a result of interactions between the past distribution of the species and the biological demands of the species, an examination of populations to explain relationships between animal species was also required. An under-standing of these relationships being important in fully accounting for the often patchy distribution of animals within their geographical range. Lotka (1925) and Gause (1934) gave a base to such work by demonstrating theoretically that competing species can only coexist if they are each uniquely specialized. Since then field studies have been made on many classes of animals, with the aim of describing the environment in a way which is meaningful in understanding animal/environment interactions in both space (distribution) and time (abundance). 2 Two main types of investigation have been used to explain the patterns of animal numbers occurring in an area through time. The first places emphasis on identifying mechanisms which produce stability or fluctuations in populations of a single species. A review of the general theories of population regulation will not be given here. They range from an emphasis on extrinsic factors, such as weather, food, and predators, to behavioural patterns and genetic processes. The second type of investigation examines how similar species are able to co-exist in the same general habitat. Both of these approaches are therefore mainly concerned with the thesis brought together by Andrewartha and Birch (1954) that the environment, abiotic and biotic including animals of the same and different species, affects an animal's ability to survive and reproduce, and that this in turn produces observed patterns of distribution and density. A review of a few specific studies will show that the environmental factors identified as accounting for variations in distribution and abundance, lie on a continuum from quantified environmental parameters of unknown biological significance, to discrete habitat components of known importance. Examples from the former end of the scale include those of MacArthur (1961) who found that the number of vertical layers of vegetation could account for the number of breeding bird species present; Edington and Edington (1972) who described the horizontal, vertical, and temporal separation of insectivorous bird species in a woodland; Rosenzweig and Winakur (1969) who described how the densities of various species of granivorous desert rodents varied in response to the presence or absence of 3 vegetation of certain physiognomies; and Pianka (1966) who showed that both horizontal and vertical components of spatial heterogeneity were cor-related with the number of lizard species in a desert habitat. The reasons for the above associations are largely unclear, but it has been suggested that microhabitat segregation is often an important deter-minant of the food eaten (Pianka 1966, Edington and Edington 1972), and may be achieved by behavioural and morphological specializations equipping each species to utilize one particular spatial area (MacArthur 1958, Edington and Edington 1972). The recognition of possible species-specific requirements for discrete habitat units for particular uses, such as breeding site, shelter, etc., was part of Burt's (1943) clarification of the home range concept. Thus Brown (1972) found that the density of woodrats in southern California was signifi-cantly correlated with the density of cacti. The cacti supplied most resources required by the woodrats; food, water, materials for den construction, and a means for avoiding predation. Ross et al (1968) described how in North-western Minnesota the jumping mouse (Zapus hudsonius) and a number of other vertebrates, were more likely to be found in Mima mounds than away from them. Some workers have identified highly specific habitat requirements, or very specific features of the habitat, which affect distribution and density. These may be spatially discrete or may represent a level of quality or quantity of a resource. Dale (1939) found that an area was only occupied by the kangaroo rat (Dipodomys) if in addition to more general habitat 4 requirements, there were areas of very fine sand for dusting; a process believed to have behavioural significance as well as a purpose of removing parasites. Aumann (1965) and Aumann and Emlin (1965) found that the peak densities attained by microtine rodents were correlated with the level of soil sodium, and suggested that the concentration of this mineral was a critical factor limiting vole populations in many areas. Weir (1971, 1972) showed that the distribution of sodium rich water had a major influence on the distribution of elephants (Loxodonta africana) during the dry season in semi-arid areas of Wankie National Park, Rhodesia. The environment does influence the distribution and abundance of animals. It can be classified into broad categories such as food, a place to live, other animals, etc., as was done by Andrewartha and Birch (1954). Each of these may in turn be studied in terms of distribution in space and time, in quantity, and in quality. The features thus found to be determining distribution and density will range from broad details defining zoogeographic range, to finer measurements explaining distribution on a smaller scale, both spatially and temporally. All the studies cited above, apart from that of Ross on Mima mounds, have concentrated on a species or groups of species, rather than a particular habitat feature. The use of the reverse method, concentrating on a spatially discrete habitat feature, is expected to be most productive in studying animal/environment relationships, when the feature has a number of characteristics different from those of the surrounding area. The African termite mound is such a feature. Studies on mounds, particularly those 5 of the genus Macrotermes, have shown them to be of large size, up to several metres high, and the same in basal diameter; to often be present in high densities over extensive areas; to have a complex interior space sometimes opening to the exterior; to be made of soil differing in many properties from the surrounding soil; with a vegetation sometimes very distinct from surrounding vegetation; and to possibly persist for long periods, the oldest recorded being 700 years (Harris 1971, Lee and Wood 1971, Watson 1969). A number of studies have shown the epigeal termite structures are utilized by vertebrates for a variety of activities ranging from small vertebrates entering them for refuge (Cowles 1930, Mitchel 1965, Goodland 1965, Neal 1970), to medium sized mammals digging into them for food (Kingdon 1971), to larger mammals using them as lookout posts, rubbing areas, and salt licks (Burt 1942, Hediger 1948, Hesse 1955, Jackson and Gartlan 1965). The aim of this study was to investigate the ecology on termite mounds with respect to vertebrate usage, but without emphasis of any particular vertebrate species. The areas of study were as follows: first, to quantitatively define the contribution to environmental heterogeneity given by epigeal termite structures in one area, Ruwenzori National Park (R.N.P.) western Uganda, with emphasis on those aspects of apparent potential importance to vertebrates; second to find out to what extent these c features were being utilized by vertebrates, and particularly if they were affecting the distribution of any species. 6 The project was c a r r i e d out i n R . N . P . between September 1972 and Augus t 1973, while I was at the Uganda Institute of Ecology ( U . I . E . ) , Mweya . Studies were made i n areas representative of most soi l types and vegetation associations, with the aim of encountering al l types of epigeal termite s tructure present , and ident i fy ing al l poss ible vertebrate /mound interactions wi th in the P a r k . 7 II. STUDY AREAS Ruwenzori National Park (formerly Queen Elizabeth National Park) lies in the Western Rift Valley of Uganda, south of the Ruwenzori mountains and adjoining Lakes Amin (formerly Edward) and George (fig. 1). The equator passes through the northern part. The altitude is just over 900 meters, and the terrain generally flat or gently undulating, characterized by Euphorbia candelabrum, termitaria, and open to closed bush communities with local, scattered diminishing forest remnants (Thornton et al 1968). The only extensive area of forest is the closed semi-deciduous Maramagambo forest, which lies between the eastern escarpment of the Rift and Lake Amin, cutting the Park in two. The soils in the north of the Park are for the most part derived from volcanic deposits, and are generally rich in plant nutrients, having a high reserve of weatherable minerals and an upper horizon well provided with organic matter (Thornton et al 1968). In the south they are generally less mineral rich having been derived from Rift Valley sediments (Harrop 1960), but there are areas of hydromorphic soils with high mineral content developed under seasonal waterlogging, (Thornton et al 1968). The climate is of a dry equatorial type with two rainy seasons centered around April and September (U.N.P. handbook 1965). The annual rainfall varies from 1200 mm. near the escarpment to 600 mm. near the lakes, (D.L.S. 1962). Maximum air temperatures reach 35*C, with a daily variation usually between 11.0°C and 14.0°C (D.L.S. 1962). Most grassland areas of the Park burn at least once a year. Figure I . The p o s i t i o n of Ruwenzori N a t i o n a l Park i n East A f r i c a . 9 Six regions within R.N.P. were used as study areas. The study areas were Nyamagasani, Mweya, Crater Region, Kamulikwezi, Kasenyi, and Ishasha, (figure 2). Five were in the north of the Park and one in the South. They represent six of the eleven blocks described by Lock (1970a); with the exception of Ishasha they are generally ecologically uniform, and cover the main soil types and vegetation associations, apart from forest. No termite mounds were found on a number of exploratory visits to the Maramagambo forest. The mean rainfall, calculated using Park's records from 1968 to 1972, and the rainfall during the 1972-73 study period, are given for each study area, in figures 3 to 7: data are lacking for Kasenyi. The Nyamagasani area is a flat strip bordering Lake Amin. The soils are black clay loams of the Nyakatonzi series derived from volcanic ash (Harrop 1960). The study area lay between the riverine forest of the Nyamagasani river and the now abandoned Kayanja ranger post. This area supports bushed grassland for the most part, being a mosaic of Themeda  triandra grassland with Sporobolus pyramidalis (Beauv.), Chloris sp., and Bothriocloa sp.; these last three are typical of the more heavily grazed parts (Lock 1970a). Capparis tormentosa (Lam.) dominated thicket is found throughout. Mweya has soils similar to those of Nyamagasani. The area is dominated by Capparis tormentosa thicket, with Euphorbia candelabrum (Tremaut) and Euphorbia dawii (Tremaut) present near water. There is much overgrazing especially near the Kasinga Channel and lake edge (Lock 1970a). The common Figure 2. Ruwenzori N a t i o n a l Park showing the s i x study areas. 11 140. 120. 100. IT £ 80. _J 6a < Z 40. < C 20 1 1 O N F* M* M F i g u r e 3. Mean monthly r a i n f a l l a t Nyamagasani f o r 1968 t o 1972 ( h i s t o g r a m ) and monthly r a i n f a l l d u r i n g t h e stu d y p e r i o d (*). 200, 180. 160. 140. 120. 100. E E 80J 60J Z 40-| < C 20J I I I I 1_1 1 I 1 S O N D J F M A M J J A F i g u r e 4. Mean monthly r a i n f a l l a t Mweya f o r 1968 t o 1972 ( h i s t o g r a m ) and monthly r a i n f a l l d u r i n g t h e s t u d y p e r i o d ( x ) . 12 180. 160J 140. 120J 100. H 60. 4o: < DC 20. 1 Hf ± S O N D J F M A M J J A F i g u r e 5. Mean m o n t h l y r a i n f a l l i n t h e C r a t e r R e g i o n f o r 1968 t o 1972 ( h i s t o g r a m ) and m o n t h l y r a i n f a l l d u r i n g t h e s t u d y p e r i o d ( x ) . 120. 100J £ 8 0 < 60J 40. 20 IT S OT N D~ J " F M " A* M J J A F i g u r e 6. Mean m o n t h l y r a i n f a l l a t K a m u l i k w e z i f o r 1968 t o 1972 ( h i s t o g r a m ) a n d m o n t h l y r a i n f a l l d u r i n g t h e s t u d y p e r i o d ( x ) . 160. 140. 120. 100J £ , 80. H 60 < CC 40J 20 1_T 0 D M A M - * — r F i g u r e 7. Mean monthly r a i n f a l l a t I s h a s h a f o r 1968 t o 1972 ( h i s t o g r a m ) and monthly r a i n f a l l d u r i n g t h e s t u d y p e r i o d ( x ) . , 14 grasses include Spbrobolus pyramidalis, Bothriocloa sp. , Cenchus ciliaris (Linn.), Chloris sp., and Cynodon dactylon (Steud.). The Crater Region is a distinct area in the north of the Park, being made up of some 80 explosion craters believed to be about 10,000 years old (de Heinzelein 1957) . The soils are shallow black clay loams of the Kyamatoma catena (Harrop 1960). Much of the area is covered by Imperata  cylindrica (Beauv.) , a tall broadleaved pyrophillous grass, with varying amounts of Cymbopogon afronardus. Also present are areas of Themeda triandra grassland. Thickets are mainly confined to the outer Crater Region. In contrast to all other soils north of the Kazinga Channel, those of Kamulikwezi are derived from Rift Vally sediments, mainly being fine textured deep brown loams of the Kassese and Sebwe series (Harrop 1960). The vegetation cover is diverse with extensive swampy grasslands along the lakeshore, and much of the rest covered by short mixed grassland, with some Acacia sieberiana (Scheele) woodland, and extensive Capparis  tormentosa thickets (Lock 1970a). The Kasenyi area is similar to Nyamagasani with respect to soils and appearance of the vegetation. The soil, however, is much richer in clay than that of other areas on the Nyakatonzi series, becoming very sicky and temporarily flooded in wet weather (Lock 1970a). Hyparrhenia filipendula/ Heteropogon contortus (Beauv.) grassland with thicket clumps covers most of the area. It is probably the driest region of the Park (D.L.S. 1962). Soils of the Ishasha area are sandy loams of the Ishasha complex derived from Rift Valley sediments (Harrop 1960) . The vegetation of the area is 15 very varied; riverine forest, dense dry thicket, large swamps, and open grassland are all present. Two large areas of open grassland occur near Kikeri, and between Ishasha camp and Katoke. These are heavily grazed, but elsewhere grazing is light to moderate (Lock 1970a). A list of most of the vertebrate species found in the Park is given in the Uganda National Parks Handbook (Anon. 1965). The most conspicuous animals are elephant (Loxodonta africana Blumenbach) , and six species of ungulate; buffalo (Syncerus c. caffer Sparrman), hippopotamus (Hippopotamus  amphibius L . ) , waterbuck (Kobus defassa Ruppell) , Uganda kob (Adenota  kob thomasi Erxleben) , topi (Damaliscus korrigum jimela Ogilby), and warthog (Phacochoerus aethiopicus Pallas) . All are widespread throughout the Park except for topi which are confined to the southern end. Between them they play a major role in the ecology of the Park (Field 1968) . Other mammals occurring in the Park include baboon (Papio anubis Fischer), vervet monkey (Cercopithecus aethiops L . ) , civet (Viverra civetta Schreber) , genet (Genetta genetta L.) , banded mongoose (Mungos mungo Gmelin) , white-tailed mongoose (Ichneumia albicauda Cuvier), and aardvark or orycteropus (Orycteropus afer Pallas) . Murid rodents present include Arvicanthis niloticus (Deomarest) , Lemniscomys striatus (L.), Mastomys  natalensis (Smith), Zelotomys hildegardeae (Thomas) , Mus minutoides (Smith), Mus triton (Thomas) and Lophuromys sikapusi (Temminck) . 16 III. ENVIRONMENTAL CHANGES CAUSED BY TERMITARIA 1. A PRELIMINARY SURVEY Preliminary surveys were made in each of the six study areas, to examine all the types of epigeal termite structure. Samples of termites were taken from several of each mound type found and preliminary identifications made using the keys of Bouillon and Mathot (1965) , and Ruelle (1970). Samples were also sent to the British Museum (Nat. Hist.) for identification. The external appearance and dimensions of the mounds were recorded, and some were excavated so that records could be taken of internal structure. Two species of Macrotermes were found to build the large mounds found throughout the Park. Macrotermes subhyalinus (Rambur) occurred in all five study areas north of the Maramagambo forest, but was absent from the Ishasha area. Its mounds ranged in size from small bare domes ( < 50 cm high x <1 m. diameter) to large, often thicketed, mounds (> 2 m. high x > 5 m. diameter) . Figure 8 shows an average sized vegetated mound from the Mweya region. A section of a live mound from the Nyamagasani area is shown in figure 9, along with a generalized mound section. The thick outer casing to the mound has no termite-made openings connecting to the interior, nor any tunnels >1 cm. in diameter near to the surface. The fungus garden and central hive are diffuse with no hollow cellar region below them. Fresh growth occurred by the addition of compact low mounds of soil onto the surface. F i g u r e 8. M. s u b h y a l i n u s mound near Mweya. 18 19 The second species building large mounds was Macrotermes bellicosus (Smeathman) , which occurred only in the Ishasha study area. Figure 10 shows a ]VL bellicosus mound, and figure 11 shows a section of the same mound with an explanatory diagram. Mounds varied in size from a small (< 50 cm. high x < 1 m diameter) bare "turret", or group of "turrets", to large (> 2 m high x>4 m. diameter) conical, often bare structures. In contrast to the mound growth in M_. subhyalinus, in M_. bellicosus fresh growth was in the form of thin-walled turrets (figure 12); the hollow interior of which connected with a distinct cellar region below the nest. Basal holes up to 25 cm. in diameter were sometimes present and these also led to the cellar. The fungus garden and central hive were more distinct than those of 1VL_ subhyalinus with a definite space, occurring between them and the outer casing. A third large mound type occurred in areas of low-lying seasonal floodland in the Ishasha area. The mounds were always grass covered and often thicketed, being around 10 m. in diameter and up to 2 m. in height, (figure 13). Live mounds were inhabited by Odontotermes kibdrensis (Fuller) , but only a small portion of such a mound had termite galleries and associated ventilation chimneys, the rest being unstructured soil, (figure 14). Two other termite species were identified as building small mounds (<50 cm. high x < 80 cm. diameter) throughout the park. These were Pseudocanthotermes spiniger (Sjost.), constructing narrow based small bare mounds, and Trinerntermes gratiosus (Sjost.) building small bare domed mounds. 20 Figure 10. M. b e l l i c o s u s mound at Ishasha. CM Figure I I . Mound shown i n figure 10 partly excavated. 22 Figure 1 2 . Turret of f r e s h growth on a M . b e l l i c o s u s mound. Figure 13. 0. k i b a r e n s i s mound i n seasonal f l o o d l a n d a t Ishasha. bare term, soil unworked soil n,miui\iAiWMiv Figure 14. Generalized s e c t i o n through an 0. k i b a r e n s i s mound. The final type of epigeal termite structure identified was a group of turretted shafts above subterranean nests, found mainly in the Crater Region A l l specimens taken from these nests were identified as Odontotermes fulleri (Emerson). Figure 15 shows how these turrets appear in the field, and figure 16 represents a section through a live nest. Note that the large shafts do not connect directly with the active nest area. Worker termites were seen building up open turrets in the early morning of 24th April 1973 during the first rainy season and in subsequent excavations allates were found. These utilize the turret as a platform for swarming (Harris 1971). The main body of this study which follows concerns the mounds of the two Macrotermes species, the large 0_. kibarensis mounds and the turretted Odontotermes subterranean nests. Neither of the two small non-Macrotermes mound types were studied further, apart from recording their presence on transects, because of their small size and compact internal gallery systems. F i g u r e 15. O . f u l l e r i t u r r e t s above subterranean nest. Figure 16. Generalized s e c t i o n through an 0 . f u l l e r i nest. 2. METHODS i) The distribution and density of termitaria Mound densities and their dispersion patterns within the six study areas were investigated by means of ground and aerial transects. At least ten ground transects were run in each study area, except Kamulikwezi and Kasenyi; the lower numbers for these two areas (4 and 3 respectively) were owing to the fact that they were originally considered a single area because of vegetational and soil similarities (see section II). Ground transects were usually 500 m in length, occasionally shorter if a natural obstacle was reached. Prior to the placement of transects grid maps were made of each study area with a scale of 1:25,000, from 1:50,000 maps covering sections of R.N.P. (U.L.S.D., 1965). Grid squares representing one square kilometer were chosen randomly from these maps and one transect taken in each such grid square. This method was modified for the Ishasha study area where distinct topographical and vegetational regions could be identified; in this case grid squares were chosen until all these regions were represented. This method is essentially one of stratified random sampling in which biotic, edaphic and topographic differences both between and within study areas have been acknowledged. It was used not for an accurate estimate of total mound populations for the Park, but rather for sampling mounds in all possible environmental situations, and for the study of their density and condition. The individual ground transects were run as follows. When the grid area was reached where a transect was to be run, a compass bearing was taken at random and the transect covered by walking with an even 1 metre pace. A l l mounds with at least 50% of their basal area within ten metres either side of the observer were recorded and examined: the nature of qualitative details of mound condition which were taken are described in section III-2-ii. Two details which helped in density and dispersion calculation were recorded, apart from mere presence, these being the distance along the transect at which the mound occurred, and the distance from the mound to the nearest neighbour of the same species (N.N.D.). If there was any doubt about the termite species which built the mound, an attempt was made to collect a sample of the insects and observe the mound interior. Additional information on mound densities and dispersion was obtained from the small mammal trapping grids in the Crater and Ishasha Regions. In the Crater Region the grids were positioned with respect to Odontotermes shafts which were accurately plotted on maps of the grids. In the Ishasha area the large O^ kibarensis mounds were mapped on trapping grids, (see section I V - l - i i ) . Aerial visual and aerial photographic transects were made to assess the practicability of using this method for obtaining either absolute or com-parative mound density figures. A total of eighteen transects were flown using a Cessna 182 aircraft. The positioning of transects was carried out on a 1:125,000 scale map of R.N.P. (U.D.L.S., 1969) upon which a 1 cm north/south to east/west grid system had been superimposed, (see figure 17). The vertical and horizontal lines were numbered, and random pairs taken, Figure 17. RUwenzori N a t i o n a l Park showing the p o s i t i o n of. a e r i a l t r a n s e c t s . 30 one for each line direction. If the intersection of the two lines thus chosen was on land, then a north/south line representing a transect was drawn through this point, the ends being either the lake shore, the park boundary (if observable) or any obvious geographic feature. Pairs of numbers were taken until all large vegetation areas (see section II) apart from forest had been covered. Prior to running the aerial visual transects a linear scale was taped to the wing strut on the observer side of the aircraft, and after calibration pairs of streamers were attached, each pair allowing a given width of ground to be observed at a certain height. The calibration was carried out at altitudes of 400 feet and 500 feet by flying over two lines of white markers, first set 250 feet apart and then 500 feet apart. The pilot flew the aircraft as level and steadily as possible, and to one side of the pair of marker lines, so that the observer could note where on the scale the two lines were "cutting" the wing strut, (figure 18). A number of such runs were made for each height width combination and the mean scale values taken. In this way the distance apart to attach streamers on the strut for the four height width combinations was found.. After several trial runs counting Macrotermes mounds at each combination, a flying height of 400 feet was decided upon and a transect width of 250 feet. With this combination the observer (D.A.M.) did not feel overloaded by the number of mounds within a transect at the cruising speed of the aircraft which was approximately 100 m.p.h. Mounds within the transect were totalled on a hand held counter, and the transect length was measured from the map used in its placement, thus enabling the -7 calibrated wing strut. observer 0% / / / / I cairn 5 0 0 ft c a i n"> Figure 18. Wing s t r u t c a l i b r a t i o n f o r v i s u a l a e r i a l t r a n s e c t s . 32 mean density of mounds in the area covered to be calculated. Aerial photographic transects were made to obtain mound density figures and also mound dispersion details. The camera used was a Hasselblad 500 EL fitted with a 80 mm lens, attached by a frame to the observer's aircraft door with the windows open. The camera was set to take one frame per second when operated by a remote lead. Kodak Tri-X (400ASA) film was used, loaded in 70 frame cassettes. Unfortunately due to a shortage of film only a limited number of frames could be taken. Samples of five continuous frames were taken at random along visual transect lines in the Nyamagasani, Kasenyi and Rutanda areas. At Ishasha bellicosus mounds were sampled at random within a bushed area on high open grassland. Lastly three photographic transects were made as shown in figure 19, sampling Q. kibarensis mounds near Ishasha camp. Photographic transects were run at 500 feet, 600 feet and 1200 feet. This meant that in all cases consecutive negatives yielded prints with at least 50% overlap allowing definite identification of mounds by observing pairs of prints through a 3-D viewer. Mosaics were made of these sets of consecutive prints once mounds had been identified, and their scale cal-culated from the formula: -s focal length of lens (feet) x I C a 6 height above ground (feet) print enlargement factor given by Spurr (1960) . The area covered by each mosaic was calculated and the mean density of mounds per hectare found. Nearest neighbour measurements were made and dispersion characteristics calculated as described later, (secion III-3-i). 3^" Z A I R E < ^ * n ' pp ° * ••"4\*lX Pool A ± vv \v isCV. 0 ^ o <?•• 1 km z c0 O.kibarensis mounds ./' - Aerial transect Track niniiiiu Escarpment IsC Ishasha camp *-N» River jk. Swamp ,0°3 igure 19. The Ishasha Camp area showing positions of three a e r i a l photographic transects. 34 ii) Environmental changes introduced by termitaria  Land Relief Data concerning the ways termitaria alter land relief were obtained from the ground and aerial photographic transects previously described. The details taken on the ground which were used in these calculations, apart from mere presence of a mound, were mound height and two measure-ments of basal diameter. Additional information for O. kibarensis mounds obtained from aerial photographic mosaics was first an estimate of the per-centage area covered by these mounds; this was possible because of the large mound size. This percentage was calculated as a mean value of all samples by weighing the prints (W) , cutting out the mounds using a scalpel and weighing again (w). The percentage area covered by mounds then equals W-w/W, assuming even thickness of photographic paper. The second area of additional information taken from these mosaics was the possibility of associ-ation between 0_. kibarensis mounds and wallows. The measurements taken and calculations made are described later (section III-3-ii). The interior of dead termite mounds So far the methods described have been primarily concerned with mounds as points or areas on a two dimensional surface. This section deals with the types and amount of those dimensional space available inside dead termite mounds. Details were recorded during ground transects of all non-termite openings into termite nests. These openings were classified as to size and put into one of five classes denoting the probable primary cause of the opening. These were, 1) cave-ins, where obvious pressure from above had caused the mound surface to be pressed into the termite nest, usually causing small openings to occur around the edge of this depression; 2) large aardvark burrows; 3) small burrows, regular holes showing signs of small mammal usage; 4) irregular holes due to weathering; and 5) holes caused in the mound side by large mammal rubbing, (figure 20). In addition to this, details were taken of the number and size of the termite shaft openings in M_. bellicosus, Q. kibarensis and 0_. fulleri termitaria. The number and nature of all holes in a one hectare area in a crater bottom was also recorded after the January Crater Region burn. The nature of the internal space in mounds with entry points was obtained from mound excavations; at the same time notes were made on signs of vertebrate occupation, (see section IV-1) . The calculations made, using all these data plus details of mound densities and dimensions, are given along with the results in section III-3-ii. Additional results presented under this heading in section III-3-ii concern the space available in large aardvark burrows. The densities of large burrows in mounds and elsewhere were recorded for the Mweya and Crater areas during aardvark digging surveys (see section IV-1), and also for Mweya from large burrow monitoring work for vertebrate usage (see section IV-3). Five burrows were completely excavated in connection with the moni-toring, allowing the volumes of burrows per entrance to be calculated. The microclimate inside mounds, burrows in mounds and burrows elsewhere was investigated using two Cambridge temperature recorders , Cambridge Instruments Ltd., Cambridge, England. 36 Figure 2 0a. A cave-in probably caused by l a r g e mammal trampling. Figure 20b. An orycteropus burrow i n a M.subhyalinus mound at Kamulikwezi. 37 38 Figure 20e. M. b e l l i c o s u s mound rubbed by la r g e mammals, e s p e c i a l l y b u f f a l o ; holes are beginning to appear i n s i d e . 39 Each instrument had a pair of probes at the end of 3 metres of P.V.C. tubing. One probe could be kept moist using a water drip tube from a feed bottle, allowing wet and dry bulb temperatures to be taken, from which relative humidity could be calculated. The recorder continuously plotted the two temperatures on a one week circular chart. A Grant multi-channel thermistor temperature recorder was also used, set to record probe temperatures on a paper roll at 30 minute intervals.;. (Grant Instruments Ltd., London, U.K.). The first readings taken were of the temperature inside a live M. subhyalinus mound (A) and a dead mound (B) and also the air shade temperature; both mounds were thieketed, of similar size and on flat ground less than 100 m apart. A small hole was bored into the live mound until resistance ceased at about 20 cm. Thermistor probe 1 from the Grant recorder was placed in the space at the end of this shaft and the opening resealed with soil. Thermistor 2 was placed near the ground surface in the permanent shade of the bush covering the mound. A probe (3) from the Cambridge recorder was inserted 30 cm into the hollow interior of dead mound B, via a 10 cm diameter hole in the side of the mound, probably produced by weathering. Temperature records were taken over 3 days and nights , (figure 21) . The second set of readings were from the two Cambridge recorders. One pair of wet and dry bulbs were placed 2 metres into a large non-mound burrow on Mweya peninsula, and the other pair positioned in permanent shade near the ground surface adjacent to the burrow, (figure 22). Again readings were taken for three days and nights. ( F i g u r e 22. The placement of two p a i r s of temperature probes i n the second m i c r o c l i m a t e examination. 41 The third set of microclimatology data came from again using the two pairs of wet and dry bulbs of the Cambridge recorders. One pair was positioned 2 metres into a large burrow under a dead mound, and the other pair placed 2 metres inside a control burrow (figure 23). Readings were taken over a three day period. The final set of data was obtained from the equipment used in two burrows as just described above. In this case though the mound over the burrow was still inhabited by a M, subhyalinus colony (figure 24) . The observations were continued for four days. The Soil of Termite Mounds Ten soil samples were taken at a depth of 10 cm from mound surfaces and control areas at Nyamagasani, Mweya and the Crater Region. Between 1 and 6 pairs of similar samples were taken in the other study areas, including some from both mound types at Ishasha. Mound interior samples were also taken from dead mounds of the two Macrotermes species. Control samples down to a depth of 3 metres were taken in the Myewa and Ishasha areas. Soil was also collected from a M_. subhyalinus mound completely trampled and flattened by Hippopotamus (Hippopotamus amphibius L_.) , and from the bottom of a wallow inside the shell of a dead M_. belli cos us mound at Ishasha. Soil samples were analyzed individually except for the ten pairs taken in the three study areas, which were bulked as mound and non-mound samples for each area. The following tests were made on all samples, the methods used being those described by the B.C. Department of Agriculture (1973). 42 control burrow as in fig. 22 , 1M i n i * Figure 23. The placement of two p a i r s of temperature probes i n the t h i r d microclimate examination. Figure 24. The placement of two p a i r s ' of temperature probes i n the f o u r t h microclimate examination. 43 Soil texture was investigated using the "feel" method, and expressed on a six point scale, ranging from very coarse (sands, coarse loamy sands), to organic (> 30% organic matter by weight). pH was measured using a meter, recording the level in a 1: 2 soil/water mixture after thirty minutes equilibration time. Conductivity was measured on the same soil/water mixture using a Radiometer model CDM-2C conductivity meter and a CDC-104 glass dip cell; -1 o results were expressed in Mmhos.cm , corrected to 25 C. Nitrate nitrogen was extimated using a modified phenoldisulphonic acid (colorimetric) method, and results given in p.p.m. Phosphorus was measured using the Bray P-I colorimetric method; results being expressed as p.p.m. Potassium, calcium, magnesium and sodium were determined by atomic absorption spectrophotometry, using a Techtron-AA4 machine with an AA5 amplifier, on a neutral normal ammonium acetate soil extract solution; results were given as milliequivalents per lOOg of soil. The Effects of Mounds on Vegetation A description of the vegetation of mounds was made during the ground transects (see section IV-1). Mounds were classified as either bare; having good grass growth around the base but a bare top; having complete grass cover; or lastly as being thicketed. In addition to this the density of thickets without mounds was recorded for transects in the Mweya area. Further in-formation on the number of thicketed CK kibarensis mounds and also the total number of thickets in the Ishasha camp area, was obtained from aerial photo-graphic mosaics (see section III—2—i) . 44 A detailed vegetational survey was carried out on ten thicketed mounds in the Mweya area. Mound dimensions were taken and all species of tree, shrub, herb and grass identified using the U.I.E. herbarium with the assistance of the botanical staff. A more extensive survey was made to examine the species of grasses occurring on and off mounds. Samples were made on 55 M. subhyalinus mounds at Mweya and also 55 adjacent control areas, around each mound. A similar procedure was used for 20 0_. kibarensis mounds at Ishasha. The biomass of grasses present on and off of mounds was also investi-gated. This was carried out for four mounds and adjacent control areas at 2 Mweya, using a randomly placed 0.25 m quadrat, within the sample site. A l l vegetation, which consisted only of grasses, inside the quadrat was cut to ground level and sealed in a plastic bag. A wet weight reading was taken immediately, and a dry weight reading taken after 24 hours in a drying oven at 110°C. Two of the mounds sampled had been protected from large mammal grazing and trampling for over three years, two had not. One of the protected mounds, along with the surrounding half of the ditched enclosure in which it was located, had been burnt over three months prior to sampling. A l l four mounds were approximately 1 m high with a 3 m basal diameter, and all showed an active M^ subhyalinus nest upon digging. The final investigation of the possible effects mounds could have on vegetation was an analysis of the nutrient content of vegetation from mounds and from adjacent control areas. Mounds of M_. subhyalinus were sampled for this purpose in the Mweya area, and those of O. kibarensis sampled in the Ishasha area. Between 5 and 10 mound and control sample pairs were taken monthly from October 1972 until July 1973. Grass species were mainly sampled but some Capparis tormentosa leaf was also collected. Grass species were kept separate and mound and control samples were separately bulked for each month, except November 1972 for 0_. kibarensis samples and January 1973 for M. subhyalinus samples. During these months individual mound samples were analyzed to examine the variation present. Species samples were collected in paper bags, whole plants being cut off at ground level. The samples were dried for a minimum of 24 hours at HO^C, before being cut up and then milled in an electric hammer mill using a fine screen. The powdered sample was sealed in plastic prior to analysis. The procedure for analysis was as follows, and the methods used are those given in the A.O.A.C. handbook (1970), unless stated otherwise. An approximately one gram portion of a sample was accurately weighed to 1/10 microgram (W^), in a preweighed procelain crucible (W). The sample was ignited in a Lindberg heavy duty muffle furnace with pyrometer at 540°C until all ash was white. The crucible plus ash was reweighed when cool, (W2). One drop of distilled water was added to the ash prior to adding 5 ml of 6N HC1; the solution formed was evaporated to dryness. The salts de-posited on the crucible walls were redissolved in 2 ml of 6N HC1 and washed through ashless filter paper into a 100 ml volumetric flask. The flask solution was made up to the mark, shaken and stored in a plastic bottle for mineral analysis. The filter paper was replaced in the correct crucible and ignited in the muffle furnace at 540° C for four hours. The crucible was reweighed (Wo). The following calculations were then carried out: -Experimental error was calculated by running duplicates; blanks were also run and the appropriate corrections made , (section III— 3—ii) . The mineral extract was analyzed for calcium, potassium, magnesium, and sodium. A l l except calcium were measured using an E.E.L. atomic absorption flame photometer. Calcium concentration was tested for using the basic method of Bett and Fraser (1959) , used to determine calcium in blood serum. Calcium concentration was found by using the fluorescent dye calcein and titrating using an Oxford Titrator with E.D.T.A. to quantitatively chelate Ca++ ions. Al l mineral ions were expressed on a percentage dry weight basis. Crude protein was found using the micro-kjelkahl method described by Nelson and Sommers (1973). Experimental error was estimated from running duplicates, and blanks were also run (see section III-3-ii). Results were expressed on a percentage dry weight basis. weight of dry vegetation sample weight of ash in sample % ash in sample weight of silica in sample % silica in sample % silica free ash in sample w x - w = s W2 - W = A A/S x 100 W3 - W = SI SI/S x 100 A - SI/S x 100 3. RESULTS i . The distribution and density of termitaria  M. subhyalinus Mounds of this species occurred in all transects from the five northern study areas. The mean per hectare densities, the ranges in density and the significance of the differences between mean densities for these five areas, as found by ground transects, are given in Table I. A one way analysis of variance was used to find if all five samples were taken from the same population. The F-value of 5.86 shows that there is at least one pair of areas between which the mean number of mounds differ significantly (P< 0.01). The mean density of mounds considering all five areas was 10 per hectare, with a range of per transect local densities of 2.0 per hectare at Kamulikwezi to 23.3 per hectare at Mweya. The mean mound densities for the Crater Region, Kasenyi and Kamulikwezi were lower than the densities for Nyamagasani and Mweya. The relations between the densities of small non-Macrotermes mounds and M_. subhyalinus mounds are shown graphically in figure 25. Each point on a graph refers to results from a single transect; no data on small non-Macrotermes mounds were collected for the Nyamagasani study area. Con-sidering all areas together (figure 25c) these two mound types are found to be independently distributed. Only in the Kamulikwezi area is a high cor-relation coefficient recorded (r = +0.91), however this is not significant (P>0.1) because of the small sample size. 48 Craters r=0.212, t=0.379 14 P>0.05 12. 10. 8. Y 6 4. 2. 2 4 6 8 10 12 14 16 X Kamulikwezi . A r=0.910 , t=3.070 1 4 - P>0.05 12. 10. 8 Y a 4. 2. X X - I 1 1 r -2 4 6 8 10 12 14 16 X Figure 25. Comparison of M. subhyalinus mound d e n s i t i e s (Y=no. per hectare) w i t h non-Macrotermes mound d e n s i t i e s (X=no. per h e c t a r e ) . 49 Kasenyi 2 4 6 8 10 X Mweya 10 r = 0. 259 , t =0.760 8. P > 0 . 0 5 6. •x x , . 1 x x x . . , — x -2 4 6 8 10 12 14 16 18 2022 24 Y Figure 25. cont. 50 All areas r = 0.074 , t=0.330 P > 0.05 X X X ,XX X . •—X X X 2 4 6 8 10 12 14 16 18 20 22 2426 Y Figure 25. cont. The nearest neighbour measurements were used to examine within-study-area mound dispersion patterns using the method of Clark and Evans (1954) which is described in appendix 1. The results for M. subhyalinus mounds are given in Table II. Dispersion coefficients (R) significantly greater than unity indicate overdispersed mound populations in the two densest areas, Nyamagasani and Mweya. The three remaining areas also have R values greater than one, but not significantly so. This same method of analysis was used to study the spacing of mounds within denser aggregations occurring in two transects from the Crater Region. Both transects covered apparently uniform areas of flat ground. The results are shown in Table III. Each area produced R values indicating significantly overdispersed (at least P< 0.01) mound populations. This trend towards regular spacing of M. subhyalinus mounds is also indicated by a similar analysis of nearest neighbour distances taken from the aerial photographic mosaics (Table IV). Eight of the nine photographic samples showed mounds which were regularly spaced; four of these were significantly different from a random dispersion pattern (at least P<0.05). The single Kasenyi sample indicating aggregation was not significant. A further look at the variation between the per hectare densities in the five Kasenyi samples using the variance to mean ratio, indicated a clumped dis-tribution (£Z/x =2.4) but this was not significant (X 2 = 1.67, P>0.1). 52 M . bel l icosus M . bel l icosus mounds were i r r e g u l a r l y dis tr ibuted within the Ishasha study area where they o c c u r r e d . T h e mean density from al l eleven transects was 6.4 mounds per hectare , with a range from zero to twenty p e r hectare . Table V shows the results of the i n d i v i d u a l ground transects with respect to both Macrotermes and small non-Macrotermes mound densities; details are also g iven of surface soi l type , vegetation c o v e r , and local topography. T h e var iance to mean ratio for these transects indicates significant contagion of mounds ( 5 / x = 8.96, Xx - 89.6, P < 0.001). Note the low density or complete lack of mounds i n areas l iable to seasonal f looding , and also i n the dense d r y thicket area near the Ishasha r i v e r . T h e densities of _M. bel l icosus mounds and small non-Macrotermes mounds are compared graphica l ly i n f igure 26. A negative correlat ion exists between them (r = - 0 . 3 3 ) , but it i s not significant (t = 0.92, P > 0 .05) . Table VI gives the results of var iance to mean analysis for only those transects i n which Macrotermes mounds were found. A signif icant degree of aggregation st i l l o c c u r r e d , (P< 0.001). T h e d i spers ion patterns could not be investigated on a smaller scale us ing nearest ne ighbour measurements from ground transects since they were not r e c o r d e d . Instead the wi th in transect d i spers ion was seen from looking at the number of mounds wi th in 50 M units of three of the denser transects and calculat ing the var iance to mean ratio for these lengths , (Table V I ) . Significant aggregation was st i l l indicated (P<0 .01) even i n these most dense local areas . 53 r = -0.328, t =0.918 P >0.05 10 20 30 40 50 60 70 X Figure 26. Comparison of M.bell i c o s u s mound d e n s i t i e s (Y=no. per hectare) w i t h non-Macrotermes mound d e n s i t i e s (X=no. per h e c t a r e ) . 54 Nearest neighbour distances could be measured from two aerial photo-graphic mosaics, and were used to generate Clark and Evans dispersion coefficient R. The results of these calculations suggest local aggregation of mounds in both cases, being significant (P< 0.01) in one (Table VII). Also shown in Table VII are the density and dispersion calculations from the same two photographic mosaics, for small non-Macrotermes mounds. Neither mosaic shows a significant deviation from a random distribution. The high per hectare value of 83.3 small mounds per hectare for the Ishasha flats high ground, is caused by inclusion of all distinct bare circles of earth visible, whether there was a small bare mound in the middle or not (see section IV-2). Further information on the relative dispersion patterns of M. bellicosus and small non-Macrotermes mounds was also obtained from these two mosaics using the nearest neighbour method described in general by Pielou (1961) and utilized by Lee and Wood (1971) to examine the relation-ships between termite mounds of more than one species (see appendix 1). The results from both mosaics indicate that the two mound types are distributed independently of each other. (Table VIII), as has already been suggested from ground results. Large mounds of CL kibarensis These large mounds occurred only in two transects at Ishasha, ID and IG (Table V) . In both cases they were in low lying areas liable to seasonal flooding and accompanied by wallows. The most extensive group of mounds was near Ishasha camp in the zone outlined in figure 19. Because of this very confined distribution all density and dispersion calculations have been 55 made within this one area. The density of mounds found in the single ground transect was 12.0 per hectare. The mean density figure obtained from the nine mapped trapping grids (see section IV-1) was 22.3 per hectare. This figure however is an overestimation since because of the small grid size relative to the large mound diameters, mounds were included when much of their area was outside the mapping boundary. No attempt was made to compensate for this because accurate density figures were obtained from the aerial photographic results given below. The variance to mean ratio for the grids (o/x = 0.187) in-dicated a significant regular distribution of these mounds (X2 = 0.99, N-1 = 8, P<0.01). The aerial photographic mosaics gave a very clear picture of the density and dispersion of these large mounds (figure 27). The positions of the three transects run in the Ishasha camp area have been given in figure 19. The results of mean density calculations, and dispersion pattern calculations as found from the nearest neighbour analysis of Clark and Evans (1954) described above, are given in Table IX. In calculating density figures the areas of sloping ground to either side of the small stream where mounds did not occur (figure 27), were not included. The mean density from the three transects was 5.97 mounds per hectare, and all three gave dispersion coefficients significantly greater than unity (at least P<0.01) indicating regular spacing. Further details concerning the association of mounds and wallows as shown on these aerial photographs will be presented in section III-3-ii. Figure 27. Part of photographic t r a n s e c t 2-12-26 taken at a height of 600ft. at Ishasha, showing O.kibarensis mounds (M), wallows (W) , and t h i c k e t s (T). An example of the many small non-Macrotermes mounds shown i s arrowed. 57 Odontotermes _sp_. subterranean nests with shafts A very local distribution is indicated for the characteristic funnelled ventillation holes of this genus. Apart from nests located in the Crater Region only two others were found, both occurring together on a transect at Kamulikwezi. It is possible however that the distribution is more wide-spread, small shafts and shaftless holes being overlooked in denser vegetation. Within the Crater Region the density and dispersion details were calculated using only data from five ground transects run soon after burns went through the area. This was to negate possible underestimations in other transects in the tall Imperata cylindrica grassland, with its often thick ground layer of dead leaves. The results from these five transects are given in Table X, including the Macrotermes mound densities for comparison, and a general description of each transect area. The variance to mean ratio for these five Odontotermes densities, of 4.89 indicates a significantly patchy distribution (X2 =19.6, P< 0.001). Note that each of the sites lacking shafts were on crater slopes with bare rock showing, whilst the other sites with shafts were located on flat ground with no bare rock showing. Table XI shows that overall in the Crater Region, 1VL subhyalinus mounds were also significantly (P< 0.05) more numerous in crater bottoms than on crater sides. The correlation coefficient for the five "Odontotermes transects" was found to be +0.14, not significantly different from zero (t = 0.24, P>0.1), indicating independent dispersion. 58 Further Odontotermes subterranean nest dispersion information comes from the analysis of density and nearest neighbour distance measurements for turret groups occurring within the fifteen small mammal trapping grids which were accurately mapped in the Crater Region (see section IV-1). The grids were selected for high and low nest density, and so could not be used to generate variance to mean ratios. An analysis of nearest neigh-bour measurements as described above (appendix 1) however, first for a single high nest density grid and second for all grids, shows significantly regular spacing of nests (P< 0.001), with a mean per hectare density con-sidering nest and non-nest grids, of 12, and a maximum per hectare density of 39, (Table XII). The trapping grid maps also allow for a comparison of the relative dispersion patterns of M_. subhyalinus mounds and Odontotermes nests in the Crater Region. The calculations made have been described above (appendix 1) for Macrotermes and small non-Macrotermes mound dispersions. Table Xllla gives the 2x2 contingency table covering data from all grids. There is no trend towards one species having a higher or lower than expected number of nearest neighbours of either species, and the X 2 value of 1.57 (df = 1) indicates that the two species were distributed randomly with respect to each other. A similar analysis of data is shown in Table XHIb; this time results are only used from six grids laid out in the apparently uniform crater bottom. Again two independently dispersed sets of nests are shown. 59 Aerial visual transect results. The per hectare aerial visual transect results are given in Table XIV, along with the corresponding mean ground transect results and aerial photo-graphic results. One other set of density calculations is also presented, which shows adjusted figures based on the aerial visual results. These were produced by disregarding the percentage of the transect area covered by permanent swamp or forest when the area of the transect was calculated. The areas covered by such vegetation were found by superimposing the transects on a 1:50,000 detailed vegetation map of the Park. (Lock 1970b). These results are discussed in appendix 2. i i . Environmental changes introduced by termitaria  Land Relief The mean dimensions of the three mound types in all study areas are given in Table XVI. For subhyalinus mounds the largest size are between the taller narrower mounds at Kamulikwezi and all other areas except Kasenyi, which also has taller mounds than the other three areas. To investigate the relationship between mound dispersion and mound size, the volume of individual mounds was calculated (see appendix 2), and correlated with its nearest neighbour distance. For each area only data from transects in the more uniform (environmental) parts were used. The data from Mweya and Nyamagasani were pooled and the correlation coefficient found for this larger sample; also all data from the five study areas where M. subhyalinus occurred were pooled and the coefficient again calculated. The results are given in Table XVII. Four of the five areas showed positive correlations, these being significant (at least P<0.05) for Mweya and Nyamagasani. When data from all areas were pooled the positive correlation was significant at the 2% level. Combining data for the edaphically similar areas of Mweya and Nyamagasani produced a highly significant positive correlation, (P< 0.001). These relationships are plotted in figures 28 to 33. The percentage of each study area occupied by mounds was calculated using the mean density figures and mean mound dimensions. Using these same data the increase in surface area because of mounds was also calculated, assuming mounds to be conical with a surface area equal to (if x basal radius x slant height). The slant height was calculated knowing the mound height and basal radius. Since the ground transects gave unreliable density figures for the large O. kibarensis mounds, the above calculations for this mound type were made using the density figures found from aerial photographic transects. The large diameter of these mounds allowed their surface area to be taken directly from aerial photographs as described above (section III-2) . The results are given in Table XVIII. The mean percentage area covered by Macrotermes mounds was found to be 0.46%, whilst that of the large Odontotermes mounds was 5.97%. The mean increase in surface area caused by a single Macrotermes mound was 1.214 M 2 , giving a mean per hectare increase of 10.97 M 2 ; the corresponding figures for 0.kibarensis mounds were 3.47 M 2 and 20.68 M 2 . The possibility of association between 0_. kibarensis mounds and large wallows was investigated by using the three aerial photographic mosaics taken 60 X X 4 0 2 20 X* X X X X XX 10 20 30 MOUND VOLUME (M 3 ) 40 igure 28. Comparison of M.subhyalinus mound volumes w i t h nearest neighbour d i s t a n c e s ; Kamulikwezi. 60. 40 2 cs z 20 - I 1 1 r -10 20 30 40 MOUND VOLUME - (M ) Figure 29. Comparison of M.subhyalinus mound volumes with nearest neighb distances; Kasenyi. 63 40 4 8 12 MOUND VOLUME (M 3) Figure 30. Comparison of M.subhyalinus mound volumes w i t h nearest neighbour d i s t a n c e s ; C r a t e r Region. 64 32. 24 8 x X X X X X X X X -i 1 r~ 4 8 12 16 MOUND VOLUME (M 3) F i g u r e 31. Comparison o f M . s u b h y a l i n u s mound volumes w i t h n e a r e s t n e i g h b o u r d i s t a n c e s ; Mweya. 65 40. 2 20 XX X X x X X XX K X X X X XXX # X X X X 4 8 12 16 MOUND VOLUME (M 3) F i g u r e 32. Comparison of M.subhyalinus mound volumes with n e a r e s t neighbour d i s t a n c e s ; Nyamagasani. 60 , * X 4 0 20 X X Y XX X X X * X XX X X< x xx [<XX X x O C X X X X X X X X X X X X X X 10 20 MOUND VOLUME (M 3) 30 40 F i g u r e 33. Comparison of M.subhyalinus mound volumes with n e a r e s t neighbou d i s t a n c e s ; a l l f i v e northern study areas. 67 in the Ishasha camp area (figure 19). The method used was the nearest neighbour treatment of Pielou (1961) previously outlined, (see appendix 1). Nearest neighbour distances were measured from the centre of mounds and wallows, these points having been estimated by eye. The contingency tables produced indicate that there was a significant correlation between the two (at least P< 0.02) for each of the photographic mosaics (Table XIX). Mound Interiors The results from the ground surveys concerning the number and nature of holes occurring in the four termitaria types studied, are given in Table XX. Between 12.5% and 37.5% of KL subhyalinus mounds had some type of entry point. In two study areas the main cause was aardvark burrowing, in another two cave-ins predominated, and in the remaining one small vertebrate burrows. Aardvark burrows were the only entry points greater than 30 cm in diameter; all other entrances were less than 20 cm, usually less than 10 cm. At Ishasha over 50% of M. bellicosus mounds had openings other than basal termite shafts. The main cause of holes was attributable to weathering. 34.8% of ]VL bellicosus mounds had between 1 and 10 basal termite shafts, which were usually oval, of mean diameter 7.4 cm. No 0_. kibarensis large mounds had any other type of opening apart from the usually un-lipped shafts attributed to the action of the termites themselves. That these shafts were termite constructed was confirmed by excavations, and also by observation on November 26, 1972 of fresh turrets being built up around these holes, presumably prior to swarming •. (figure 34). 59.4% of mounds examined had between 1 and 14 such shafts, of average F i g u r e 34. O . k i b a r e n s i s workers b u i l d i n g up t u r r e t i n t h e wet season a t I s h a s h a . 69 diameter 5.4 cm. 40.6% of mounds also showed signs of small mammal usage of shafts in the form of runways and droppings, and it is possible that a number had actually been completely constructed or at least modified by small mammals. In the Crater Region the mean number of shafts per CL fulleri nest was 6.1, with a range from 1-22. The average diameter of shafts from the Craters and Kamulikwezi was 5.8 cm. Additional entry points were caused by aardvark burrowing, which were found in 9.8% of nests in the Crater. The results of examining all holes in a one hectare area of a crater bottom after a burn, show that Odontotermes shafts were the most numerous hole type, comprising over 90% of all holes (Table XXI). The difference between the numbers of M_. subhyalinus mounds per hectare with the various types of openings in the five northern study areas was examined using one way analyses of variance. The results are presented in Tables XXII and XXVI. Cave-ins occurred in all areas except Kasenyi; however, only three transects were run in this area. Large aardvark burrows occurred in mounds in all areas, giving per hectare ranges of 0 to 3 such mounds per hectare. Small burrows were found in mounds in all areas, with no significant difference between any pair of areas; the per hectare range of such mounds was 0 to 8. Between 0 and 2 mounds per hectare had holes directly attributable to weathering, again with no significant differences between any areas. The mean numbers of mounds with any entry point ranged from 1 per hectare at Kasenyi to 3.1 per hectare at Mweya. 70 The results of the analysis of height and diameter measurements for the three mound types, with respect to mounds with entry points, mound lacking entry points, and all mounds, are shown in Table XXVII. Holed 1VL subhyalinus mounds were shorter than intact ones in all areas except Kasenyi and Kamulikwezi. The only significant difference in mean height (P<0.02), however, was for lower holed mounds in the Craters. Apart from those in Nyamagasani, holed mounds had a larger diameter than intact ones, being almost significantly larger (P < 0.1) at Kasenyi. M. bellicosus mounds without basal termite shafts were shorter, almost significantly so (P<0.1), and had significantly narrower bases (P<0.02) than those with holes. Bare mounds with non-termite holes also had a larger diameter than those without such holes, but their heights were similar. The kibarensis mounds with no termite shafts were shorter and narrower at the base than holed mounds, but this difference was not significant. The types of interior space occurring within the two Macrotermes mound types, as shown by excavation of mounds with openings, are given in figures 35 and 36. Two general types of interior could be distinguished for M. subhyalinus mounds; those where the galleries had remained intact to give a complex series of small passages through the mound, and those where the galleries had fallen away from the outer casing of the mound to produce a large space above them. The same two types could just be distinguished for M. bellicosus mounds, however because of the more definite central hive with the parade surrounding it, the mound always had some dropping of the galleries from the wall. Figure 35. Nature of the space i n s i d e a dead M.subhyalinus mound. Figure 36. Nature of the space i n s i d e a dead M . b e l l i c o s u s mound. 72 The mean volumes per hectare of mounds with entry points are g iven for the three mound types and al l areas , i n Table XXVIII. T h e formulae used i n these calculations are explained i n A p p e n d i x 3. These figures probably do not represent a true spatial vo lume, but it can be cons idered a maximum va lue . Space often exists below a dead Macrotermes mound, but it is not b e i n g taken into account here . Between 3.4 M 3 and 10.2 M 3 of available mound space occurs p e r hectare wi th in dead _ M . subhyal inus mounds, and 6.5 M 3 i n dead ]VL bel l icosus mounds. The value of 237.6 M 3 for CK k ibarens i s mounds was calculated and inc luded for completeness, but does not even r e -present a potential spatial volume since as has been descr ibed above little of the mound is ever worked b y termites and most remains unstructured s o i l . F o r the volume calculations of Macrotermes mounds above only mounds with non-termite constructed holes have been cons idered , that i s dead mounds . However , over 1/3 of M . bel l icosus mounds had termite made shafts. T h u s up to two mounds p e r hectare may have openings into the complex cel lar region below the mound. L i k e w i s e , the volume enclosed wi th in O . fu l l er i shafts i s s ignif icantly large when cons idered on a per hectare b a s i s . T h e volumes were calculated from excavations, assuming the shafts to approximate to a cy l inder with a diameter equal to the mean of the surface and subsurface diameter readings . Table XXIX shows these resu l t s . A n interest ing comparison at this time is to present results on the volume of large aardvark b u r r o w s . The volumes were calculated b y approximating the cross-sect ion of a burrow to a semicircle, details of shape and dimensions are g iven later, (section IV-2). The volumes and number of entrances of 73 five burrows completely excavated out on Mweya peninsula are presented in Table XXX. The volume increased with the number of entrances, the mean per entrance volume being 0.59 M<*. Using this figure along with the densities of entrances per hectare, the mean per hectare volume of large burrows was found (Table XXXb) . These results show that 32.4% of all large aardvark burrows were in or under termite nests. The microclimate within termitaria Finally in this section of results, the details of the microclimate of the space in mounds, and space in burrows elsewhere are given. Figure 37 shows how the temperatures fluctuated inside a live mound, inside a dead mound and in a shaded area external to the mound. Within the live mound a 20.2 C external air shade temperature range was reduced to only 3 C. At the end of this experiment however the probe in the live mound, which was originally inserted into an air space, was covered with moist soil; these readings therefore probably do not apply to mound air temperature. The temperature inside the dead mound followed variations in the external air temperature but with a lag period of a few hours. There was also a reduction of maximum fluctuations down to a range of 8 . 7 ° C . Greater reductions in temperature fluctuation occur in underground burrows, whether under mounds or not, than inside a dead mound above ground (figures 38, 42, and 43). The external temperature range of up to 18°C was reduced to 4-5° C in all cases. The results presented in figures 39 and 40 show how fluctuations in relative humidity were reduced to the same extent in both dead mound burrows and control burrows. In March 1973 the 5-10-72 6-10-72 7-10-72 F i g u r e 37. Temperatures 20cm. i n s i d e a l i v e mound (•--), 20cm. i n s i d e a dead mound (—) , and i n shade a t the ground s u r f a c e . ( — ) . 75 Figure 38. Temperature i n a non-mound large burrow >(---)'.,a-nd i n shade at the ground surface (—). 76 F i g u r e 39. R e l a t i v e humidity in. a non-mound l a r g e burrow ;(---) and i n shade a t the ground s u r f a c e (—). 77 F i g u r e 40. R e l a t i v e h u m i d i t y i n a ideadi'mo^ a non-mound b u r r o w (—) , and a t t h e g r o u n d s u r f a c e (—) . 78 F i g u r e 41. R e l a t i v e humidity i n a l i v e mound l a r g e burrow (—) and a c o n t r o l burrow (---) , and a t the ground s u r f a c e (—) . 30, 23-5-73 24-5-73 25-5-73 26-5-73 2 7-5-73 gure 42. Temperature i n a dead mound large burrow (--) and a c o n t r o l burrow (—) . Figure 4 3 . Temperature i n a l i v e mound large burrow (—) and i n a c o n t r o l burrow (—) . 30. 26-3-73 27-3-73 28-3-73 29-3-73 30-3-73 81 external relative humidity varied between 55% and 87% daily, whilst that of a large burrow not under a mound remained between 70.0% and 80.0%. In May the external relative humidity varied between 60.0% and 91.0%, whilst that in a large burrow under a dead mound varied between 72% and 85%; over the same period that of a control burrow remained between 72% and 88%. In May 1973 readings were also taken in a control burrow and a burrow under a large live M_. subhyalinus mound (figure 41) . Once again both burrows showed a reduction in external relative humidity fluctuation, but the values for the mound burrow were consistantly higher than those of the control. Whilst the external air varied between 54% RH and 91% RH, the air in the live mound burrow stayed between 76% and 94%, and that in the control burrow ranged between 66% and 72%. The soil of termite mounds All results of soil analyses are given in Table XXXI. The results of total exchangeable cation (T.E.C.) analysis in bulked samples for the three mound types are graphed in figure 44, along with their respective pH values; mound and control samples are shown for each area. For M_. subhyalinus mounds the T.E.C. values were consistantly higher for mound surface soil as compared with control surface soil, with higher pH values as well. The highest percentage increases in mound soil cations occurred for sodium (Table XXXI). Because of the generally low number of non-bulked individual samples analysed, the significance of the differences between mound and control samples was only tested for using data from ]VL subhyalinus surface and control surface 82 o rt H-o cn H-fD S O 0) & O O rt hi O cn C H Hi 0) o fD CO O H-cn PJ 3 Ti H fD cn fD X o fD cr CATIONS (M.eq./100g.) o IV) O Craters •g Mweya NyamagasaniA B A B A B A B A B Ishasha Ishasha I I I , I • • i • . • . * .. i i i i i i I . I i i i i i i i i I I I 1 1 1 1 1 D n zz Si O 9 ZZ cr c p O f 1 fD 83 samples (Table XXXII). The method used was the Student's t-test for paired variates. Total cations were significantly higher in JVL subhyalinus mound surface soil (P<0.01), as are all individual cations tested for (at least P<0.05), except potassium. Conductivity was significantly higher (P<0.01) in mound soil. Nitrate nitrogen was higher in mound soil. Phosphorus was significantly less in mound soil (P<.0.05). Results for bulked samples from individual areas are shown in figures 45 to 47. Figure 48 shows sections of two mounds on Mweya peninsula, along with the positions of eight soil samples taken from these mounds and the nearby soil. The results of the analyses of these samples are given graphically in figures 49 to 52. Both freshly worked termite soil and the weathered mound base had higher T.E.C. values than the surrounding soil, even at a depth of 1 M where the control value was highest. The maximum T.E.C. value was for gallery soil from a dead mound interior; this sample contained over twice as many cations as any other sample. Nitrate nitrogen was over 100 times as con-centrated in the dead mound interior than in any other sample taken; the amount in the worn mound base was six times as great as that in either the fresh working or the surrounding surface soil. Over three times as much phosphorus was present in the fresh working as in either the worn mound base or the dead mound interior. The lowest mound phosphorus value was for the weathered base, which was however higher than any of the control samples. H-C hi (D 01 *1 H-iO r i fD hi CD 01 AVAILABLE NITRATE 3 H-rt i-i 0) rt CD H-rt t-i O iQ CD pj P-. < H-h-1 PJ tr CD 3" O cn o i-i cn 3 o c o o rt i-i O ro in hi •Hi Co O CD cn O H-M cn P> 3 cn Nyamagasani A B Craters Mvveya Ishasha Ishasha A B A B A B A PHOSPHORUS(ppm) NITROGEN (ppm) CJi (j) Oi ro O O O (Ji Cn CJi CONDUCTIVITY (hrnhos) o ^ In b Nyamagasani A B Craters A B Mweya A B Isasha A B Ishasha A B CO S5(50cm) 8.3 SQ1M) 8.3 SX2M)8.6 1. SAMPLE NUMBER 2. pH C:- FRESH WORKING ^ DEAD INTERIOR S83M)8.2 F i g u r e 48. P o s i t i o n s of e i g h t s o i l samples from two mounds at-Mweya. 87 40 30. * no data-control values used 20 Pa GO z O Mg \ Na. S1 S2 S3 S4 S5 S6 S7 S8 F i g u r e 49. S o i l a n a l y s e s f o r two mounds a t Mweya; e x c h a n g e a b l e c a t i o n s . W 10 E 6 j CL CL cr tz' z LU 325 SI S2 S3 S4 S5 S6 S7 S8 Figure 50. S o i l analyses f o r two mounds at Mweya; n i t r a t e n i t r o g e n . CL CL ^ 1 4 0 • 9 , 1 0 0 y 60 a. i 20 J Z Z I c SI S2 S3 S4 S5 . S6 S7 S8 Figure 51. S o i l analyses f o r two mounds at Mweya; a v a i l a b l e phosphorus. o 9.6 1.0. ^ 0 . 6 U Q O u 0.2 S1 S2 S3 S4 S5 S6 S7 S8 gure 52. S o i l analyses f o r two mounds at Mweya: c o n d u c t i v i t y . 90 Conductivity was lowest for the fresh working; the highest value was for the dead mound interior. Similar increases were found in M_. bellicosus mound samples, however a sample from a dead mound interior did not give high values as found above for _M. subhyalinus. Once again as with all M_. subhyalinus samples, the mound soil texture was equal to or finer than the adjacent surface or sub-soil. The unstructured O. kibarensis mounds showed increases in conductivity, all cations except magnesium, available phosphorus and nitrate nitrogen. Note the very fine soil texture of both mound and control soil in this lowland area. Finally the cation results allowed a comparison to be made of the weights of minerals in mounds as compared with the weights in the topsoil. The average above-ground mound volume was calculated for each area using mean mound dimensions in the formulas given in Appendix 3. The bulk density of mound and control soil was not measured, but was assumed to be 2.0 for both (i.e. 2 Kg per 1 M3soil) (see Appendix 4). The mean above ground mound weight was calculated from these volume and density estimates, and expressed as per hectare weights. The weight of soil in the top 15 cm soil was also calculated on a per hectare basis, allowing the mound soil weight to be expressed as a percentage of mound soil plus topsoil. Using mean area mineral content values for surface mound soil and surface control soil the weights and per-centage weights of minerals per hectare in all mounds, in bare mounds and in the topsoil were calculated. The results are given in Tables XXXIII to XXXVII. The mean weight of M. subhyalinus mounds per hectare at Nyamagasani was 40,364.1 Kg, compared with 3,040,364.1 Kg in mounds plus 91 topsoil. This means that 1.33% of mound and topsoil was in the form of mounds; however 1.55% of exchangeable calcium, 1.90% of magnesium, 4.66% of sodium, and 4.31% of nitrate nitrogen was in the mound soil. Similar figures were recorded for the Crater Region and Mweya; in both of these the highest cation percentage in mounds was reached by sodium, being 9.09% at Mweya with only 1.74% of the total soil being mound soil. In all areas the percentage potassium was less than the percentage of mound soil, reflecting the lower concentrations of potassium in mounds as compared to control areas. The high mound soil concentration of nitrate nitrogen in the Craters meant that over 20% of this mineral was in mounds, with mound soil only representing 0.99% of total soil. It should be remembered that the mound soil values used in these calculations were from surface samples which may be much less than those of mound interiors. For M_. belli cos us mounds 0.78% of topsoil and mound soil was in mounds, but the percentage of cations in mound soil as compared to mound and topsoil ranged from 1.20% for magnesium to 2.56% for sodium. 29.91% of nitrate nitrogen was contained in mound soil. The mean above ground mound weight of the large _0. kibarensis mounds at Ishasha was 177,936.0 Kg, over forty times the mean weight of Macrotermes mounds, and representing 26.15% of topsoil and mound soil in the area where they were locally abundant. The percentage of minerals showed high values for nitrate nitrogen, potassium and calcium, these being 94.44%, 74.21% and 66.66%, respectively. 92 The vegetation of termite mounds The mean per hectare densities of mounds with the four vegetation cover types were calculated for each study area. The significance of the differences in mean values for _M. subhyalinus mounds between the five northern study areas were analysed as described above for mound entry points. The results are given in Tables XXXVIII to XLV. The majority of M. subhyalinus mounds were thicketed or grass covered in all areas. The percentage of bare mounds was highest at Kamulikwezi and Kasenyi. These areas also stand out as having the lowest percentage of mounds with a distinct ring of grass around their bases. The highest percentage of grass covered mounds (32.6%) occurred at Nyamagasani; the number of such mounds per hectare varied significantly between at least one pair of areas (F=7.86, P< 0.001). The highest percentage of thicketed mounds occurred at Mweya (57.6%), giving a mean of 8.3 such mounds per hectare. The F-value of 14.21 suggests highly significant differences (P< 0.0001) between areas. A high proportion (45.6%) of M_. bellicosus mounds were bare and none were classified as having a growth of grass around the base. The large 0_. kibarensis mounds were either grass covered (15.6%) or thicketed (84.4%). The aerial photographic transects gave corresponding figures of 50.2% and 49.8%. These lower figures probably reflect the difficulty in distinguishing small thickets from Maerua dwarf shrub on photographs taken at 1200'; Maerua was not recorded as thicket. These photographs also demonstrated that 56.5% of thickets were on mounds if the whole surveyed area was considered, and 76.3% of thickets were on mounds if the slopes on either side of the small stream were not included. The mean dimensions of vegetated mounds and bare mounds were cal-culated for each mound type within each study area. Mean values were com-pared within each area using the Student's t-test. The results are given in Table XLVI. _M. subhyalinus bare mounds had a significantly narrower (P at least<0.05) base in all areas except Nyamagasani, and were shorter than vegetated mounds in all areas except Nyamagasani, significantly so at Kamulikwezi and Kasenyi (at least P< 0.01). Neither the mean height nor mean basal diameter were significantly different for_M. bellicosus bare and vegetated mounds. Grassed mounds of O. kibarensis had a lower mean height and narrower mean diameter than thicketed mounds, but these differences were not significant. In the Mweya study area eight of the transects fell into three definite topographic zones; low lying areas which were seasonally flooded; highland areas on level ground; and steep slopes bordering the Kazinga channel. The mean mound dimensions, mean number of mounds per hectare, mean number of thickets per hectare, the percentage of mounds with thickets, and the percentage of thickets with mounds were calculated for each zone. The significance of differences between all pairs of zones was examined using the student's t-test. The results are given in Tables XLVII and XLVIII. The highest mounds occurred in the seasonally flooded zone; the mean height of 0.96 M being significantly larger (P< 0.01) than the lowest mean height of 0.64 M on high ground. The mean diameter of mounds was also greatest in the seasonally flooded zone, but not significantly so. Significantly fewer (P<0.05) mounds occurred on the lowland and on the steep slopes, but 94 the mean number of thickets was least on the high ground. The lowest percentage of thicketed mounds occurred on the high ground, and was significantly different (P<0.01) from the higher percentages in the other two areas. The percentage of thickets with mounds ranged between 50.0% and 57.4%, there being no significant difference between areas. The results of the detailed vegetational survey of ten thicketed mounds at Mweya are given in Table XLIX. The majority of mound thickets were composed of Capparis tormentosa and Tarrena graveolens, with Euphorbia  candelabrum on seven of the ten mounds. All species of shrub, herbs and grasses recorded were also seen away from mounds as well, but no quanti-tative details were taken. The biomass per unit area of grasses on grazed and ungrazed mounds and adjacent control areas are shown in Table L. It was found that inside the enclosure, mounds had significantly greater (P ^ 0.001) amounts of grass biomass than control areas. Outside the enclosure however mounds had less grass biomass than control areas, but this difference was not significant. The analysis of mounds to identify species of grass present and compare with control areas, showed that the only significant difference (P< 0.001) was for a lack of Cynodon dactylon in control areas, and a lack of Themeda  triandra on mounds (figure 53). For the large 0_. kibarensis mounds there was also a significant lack of Themeda triandra on mounds (P< 0.05) . Bothriocloa sp. occurred on these large mounds more frequently (P<0.05) than control areas (figure 54). 95 S3 ZD Q 4 8 4 0 32 24 16. 8. CH LU m g 24. . 16. 8. x2=1.3 c P->0.o5 c xl3.2 p>0.05 xl1.1 -p>0.05 p>0.05^ xi24.1 p<0001 c c o as o c. CD p>0.05 a) XL A B A B A B A B A B A B x ^ 2 p<Q001 D) a LLI LO a O "D y X 5 I- H — - p O O b p > Q , Q 5 n f l n n n 5 c g V a p>0.05 A B A B A B A B A B F i g u r e 53. M.subhyalinus grass cover; presence or absence on 55 mounds (A) and 55 c o n t r o l areas (B) at Mweya. 20, CO a z o L L o Q : LU C Q ZD Z cr 1 a Q . a o u o '£_ JZ CG cn n ! CD a . i - CL LU cn c O 6 « X2='4.0 P < 0.05 H g c_ cO D . o -)-> c - o X u II •8 • * * p > o . 0 5 8 "c_ CJ n E Z5 E x o £ A B A B A B A B A B A B A B A B A B A B Figure 54. 0.kibarensis grass cover; presence or absence on 20 mounds (A) and 20 c o n t r o l areas (B) at Ishasha. 97 Chemical composition of grasses Means of the monthly results for each nutrient and each species were calculated, and the significance of the difference between means from mounds and control areas over the year calculated using the Student's t-test for paired variates. A similar analysis was undertaken for the individual mound results for the months of November 1972 and January 1973; 95% confidence limits were also calculated for mound and control results for each species. The results are shown graphically in figures 55 to 60. For ML subhyalinus mounds it can be seen that many of the cation results show higher concentra-tions in mound grasses than adjacent controls, and this is also reflected in the percentage dry weight silica free ash results. The significance of the differences between monthly pairs for each species over the period samples were taken are given on each graph. Sporobolus pyramidalis gave consistently higher values for all cation concentrations and silica free ash, the values being significantly different (P<0.05) for all analyses. Cenchus cilliars gave similar results, except that the differences for calcium were not significant. Bothriocloa sp. gave a number of higher values for control samples, which did not show any pattern with season or even between minerals; with no significant differences being recorded over the study period for any results. Chloris  guy ana showed mainly higher mound values, but these were only significantly consistent (P<0.05) for potassium concentration. The few results for Cynodon dactylon indicate high mineral content for this species, especially for sodium . The few results for the leaves of the deep rooting shrub Capparis tormentosa show a lack of correlation between mounds and higher mineral content: note 98 0.60 0.40 0.40 0 .30 0 .50 ir 0.30 O ^ 0 . 50 fe 9 0 . 30 0.60 J • 0.40 < 1.00 0.80 Sporobolus pyramidalis P<0.05 S. pyramidalis (Ishasha) P > 0 0 5 C. ciliaris P > 0.05 Bothriocloa sp. P > 0.05 C. guyana P>0.05 C. dactylon C. tormentosa * OCT NOV JAN FEB MAR APR MAY JUN , JUL F i g u r e 55. Chemical composition of grasses from mounds (—) and c o n t r o l areas (--) ; c a l c i u m . *shrub 99 I— X Q LU 0.3 0.1 0.15 0.05 0.3 >~ cn Q 0.1 0.2 5J | o . 1 5 O OA < 0.2 0 4 0.3 0.5 0.4 S.pyramidalis P<0.05 S. pyramidalis (Ishasha) P>0,05 C.cilbris P< 0O5 Bothriodoa sp. p > a o 5 C.guyana P>0.05 C.dactylon C.tormentosa * OCT NOV JAN FEB MAR APR MAY JUN JUL Figure 56. Chemical composition of grasses from mounds (—) and c o n t r o l areas (--.-) ; magnesium. *shrub 100 0.04 0.02 0.03 .002 x g 0.04 £ 0.02 Q ^ 0 . 0 4 ^ 0 . 0 2 ZD Q ! ^ O 0 5 0.03 0.09. 0.08. 0.04. S. pyramidal is \ P<0.05 S. pyramidalis (Ishasha) P > 0.05 C. ciliaris P< 0.05 Bothriocloa sp. P>0,05 C.guyana P>0.05 C. dactylon C. tormentosa * OCT NOV JAN FEB MAR APR MAY JUN JUL F i g u r e 57. C h e m i c a l c o m p o s i t i o n o f g r a s s e s f r o m mounds (—) and c o n t r o l a r e a s (--•) ; s o d i u m . * s h r u b I— X CD LU 2.30 1.50 070 •"1.1: 0.7 20 1.0] fe Q -° 1.6. o ~ 1,2 2 20 oo 1.6 < • •5 1 2 ^ 1.6 2.8 2.4J 2.4 2.0J S. pyramidal is P< 005 S.pyramidalis (ishasha) P>0.05 C.ciliaris P< 005 Bothriocloa sp. P>0.05 C.guyana P< 0.05 C. dactylon C. tormentosa C.tormentosa * OCT NOV JAN FEB MAR APR MAY JUN JUL Figure 58. Chemical composition of grasses from mounds (—) and c o n t r o l areas (—) ; potasium. *shrub X o LU 7 5 3 QE 4 CD ~ ° 3 o. 11. X CO 9 . < 7 . LU LU 5 . CH 3 i < 8 ( J 7 _ J 1 2 -CO 1 0 . 8 . S . p y r a m i d a l i s P < 0 . 0 5 S . p y r a m i d a l i s ( I s h a s h a ) P > 0 . 0 5 C . c i l i a r i s P < 0 0 5 B o t h r i o c l o a s p . P > 0 0 5 C . g u y a n a P > 0 0 5 C . d a c t y l o n C . t o r m e n t o s a * O C T N O V . . . . . . J A N F E B M A R A P R M A R J U N J U L Figure 59. Chemical composition of grasses from mounds (—) and c o n t r o l areas (—) ; s i l i c a - f r e e ash. *shrub •8 6] 4 6 4 2 10 8 _ 6 110 5 8 or 6 Q £ 10 8 z 5 4 S. pyramidalis P<0.01 S. pyramidalis (Ishasha) P<0.01 G.ciliaris P>0.05 Bothriocloa sp. P>0.05 C. guyana P<0.05 -*- C.dactylon u 3 0 . § 2 0 * — — — — — -x C.tormentosa * 20. - * -—*-C. tormentosa * (Craters) OCT NCV JAN FEB MAR APR MAY JUN JUL Figure 60. Chemical composition of grasses from mounds (—) and c o n t r o l areas (---) ; crude p r o t e i n . * shrub 104 however the much higher mineral content in these leaves compared to the grasses. The percentage crude protein was generally higher in M_. subhyalinus mound grasses than in control samples. As with the mineral content results, the data are too few to suggest trends with the season. The consistently higher values for Sporobolus pyramidalis and Chloris guy an a are significant in both cases (P at least<0.05). As with the mineral results no trend is seen with the Capparis leaf protein figures; note however their high values, as is also the value for Cynodon dactyIon. No significant differences were found for mineral content in Sporobolus  pyramidalis on 0_. kibarensis mounds and control areas at Ishasha. The con-sistently higher values for crude protein in mound samples were however significantly different from the lower control values. The results of the analysis of samples from individual mounds for both M. subhyalinus and 0_. kibarensis (Table XI) again show that most sample means were higher from mounds, and that the only significant differences occurred when this was so. In contrast to the overall results presented above, the individual data for Sporobolus pyramidalis at Ishasha, show no significant increase in protein content, but do indicate significant increases in all mineral contents apart from sodium. 105 4. DISCUSSION i) Termitaria in R.N.P. : Their structure and distribution All but one of the six termite species identified as constructing epigeal structures in Ruwenzori National Park belong to the fungus-growing Isopteran family Macrotermitinae, which includes all the main mound-building species of Africa. These termites differ from those of other families by building with sand and clay mixed with saliva, rather than digested plant material, excrement and saliva (Npirrot 1970; Harris 1971). The single non-Macrotermitinid mound builder identified, Trineritermes gratiosus, is a nasute termite using excrement to form small compact galleries. This species does, however, use more clay than other non-Macrotermitinae, and the outer walls of its mounds are formed completely of earth, (Harris 1971). In a previous survey of termites occurring in R.N.P., Lock (1970c) found only one other genus of mound building termite, Microcentermes, which can construct small mounds but often has only a subterranean nest or a nest in dead wood (Noint 1970). Other species recorded by Lock (1970c), from all vegetational areas except forest, were Odontotermes stercorivorus (Sjost.), Bifiditermes mutubue (Harris), Schedorhinotermes lamarianus (Sjost.), Microtermes sp., Basidentitermes aurivilli (Sjost.) , Protermes minutus (Grasse), Nasutitermes infucatus (Sjost.), and Nasutitermes latifrons (Sjost.). The internal design and general external appearance of the four nest types and associated epigeal structures concentrated on in this study were constant throughout their distributions within the park. This constancy pre-sumably reflects a balance between the design potential of the species and the 106 influence of a l l environmental conditions; edaphic , biotic ( inc luding zoogenic) , topographic , and cl imatic . Such a balance has been suggested before b y workers who have noted that a mound b u i l d i n g species with a wide geographical range can show large differences i n mound des ign . F o r example, H a r r i s (1956) descr ibed M_. subhyal inus mounds with a s ingle large open vert ica l ch imney, and Pomeroy (pers . comm.) reports a s imi lar type of mound i n the d r i e r areas of north-eastern U g a n d a . In a countrywide survey Pomeroy (pers . comm.) reported that apart from the type just mentioned, M_. subhyal inus mounds lack termite made openings wherever they were examined i n Uganda . Weir (1973) however , recorded basa l shafts i n M_. subhyal inus mounds i n T a n z a n i a . The structure of M_. bel l icosus mounds descr ibed above is s imilar to prev ious mound descriptions for this species (Ruelle 1964, H a r r i s 1956, No irro t 1970). It should be noted that the often present large open basa l canals do not communicate direct ly with the inner galleries and can therefore be cons idered external to the nes;t p r o p e r . L u s c h e r (1956) and others have demonstrated airflow through these canals and suggested that they function as a ventilation system. L u s c h e r (1956) further suggested differences i n the types of airflow o c c u r r i n g i n mounds with and without basa l holes , d e s c r i b i n g them as distinct mound types and ca l l ing them Uganda and Ivory coast mounds, re spec t ive ly . Ruel le (1964), however , concluded that the holes represent a stage of maturity of these mounds, the outer wa l l hav ing been worn away to expose the basa l canals; and as i n this study he observed both mound types i n the same a r e a . Noirrot (1970) descr ibed the growth of a M . bel l icosus nest from a copularium to a large epigeal s tructure and suggested the same method of hole formation. 107 Such a wearing away of the outer wall process is supported by my data on mound size which showed that mounds with basal holes were taller (almost significantly, P<0.1) and significantly wider at the base (P<0.02) than intact mounds, suggesting that smaller younger mounds lack basal holes. The large mounds of CL kibarensis fall into the category of usually thicketed "valley bottom mounds" described by Langdale-Brown (1964) as being common throughout Uganda on sites with impeded drainage. These large mounds can be caused by either M. subhyalinus (Pomeroy pers. comm., Sands pers. comm.) or M. bellicosus (Pomeroy pers. comm.) in other parts of Uganda, and Sands (pers. comm.) states that Pseudocanthotermes produces similar mounds in swamps around the northern shore of Lake Victoria. Al-though these different species create one mound type of similar appearance, the soil characteristics, internal structure, nature of external openings, and weathering properties, will depend upon the species in occupation. Odontotermes species usually construct subterranean nests with wide, often turretted, vertical canals above them; the turrets have been observed to be washed down in the rainy season and so form an unstructured mound of earth above a nest (Harris 1971). Active CL kibarensis mounds at Ishasha had such turrets during the wet season, but were unturretted with the holes surrounded by bare cracking soil in the dry season. It therefore seems that these mound slopes are the result mainly of prolonged Odontotermes inhabitation; they represent the genus expected by Sands (pers. comm.) to occur on the structurally weak clay soils of the Ishasha camp area: Macrotermes species having a poor tolerance of such soils. 108 The subterranean nests and above ground turrets of 0_. fulleri are described by other workers as typical Odontotermes nests (Harris 1971, Noirrot 1970). As was the case with the large canals of_M. bellicosus, the vertical shafts of this species do not connect with the central nest area. Open turrets were observed in both wet and dry seasons as were some turrets closed with a cap of soil. There was however no indication that all, or even most, turrets were closed in the dry season as implied by Jeffery (1973) for this same area. None of these four species had an even distribution throughout the Park. The presence of the large _0_. kibarensis mounds in a local situation has been explained in terms of a combination of edaphic and topographic features; the distributions of the other three species are more difficult to explain. In other areas of Uganda both M_. bellicosus and M_. subhyalinus mounds are found together (Pomeroy pers. comm.). It seems likely that the confinement of these species to the north and south of the park, respectively, is historical in cause rather than implying differences in tolerance to some environmental factor, with the Maramagambo Forest presenting an effective barrier to the intermingling of the species within the Park. Ruelle (1970) actually records M. bellicosus termites being collected from the "Kazinga flats", which would be within the Park just north of the Forest. The grid reference of 0°40' S, 30° E given for the collection site is just south of the Forest, however, and therefore makes its location uncertain. 109 It seems likely that the restricted distribution of O. fulleri nests may not be so clearcut as suggested in this survey. In a previous study Odontotermes species were collected at Mweya (Lock 1970c), and in this study Odontotermes species were found in aardvark dig spoil and aardvark droppings at Mweya. Also one Banded Mongoose den was believed to have been in a cavity beneath open Odontotermes shafts under dense bush at Mweya (see section IV-2). It is therefore certain that these nests do occur away from the Craters and Kamulikwezi and that they remained unlocated because either they were unturretted and inconspicuous during the surveys or more likely that they were confined to soil beneath the denser shrub vegetation. It also seems certain that the density of turretted Odontotermes nests is much higher in the Crater region, the reason for this not being known. Within this region its distribution parallels that of M_. subhyalinus with few nests being found on Crater sides, making interaction between these species unlikely. A number of environmental variables are different in the Crater region than for the rest of the Park and may be contributing to make the area favourable. The soil of the area is unique with vast stands of Imperta cylindrica dominated pyrophillous grassland and a very low density of shrubs. The highest rainfall for the Park falls in the Craters and it is at the highest altitudes within the Park. For most of the year there is a low density of large grazing mammals, this in-creasing temporarily only after burning. A number of factors could therefore be interacting to favour Odontotermes species in this area. Even within the north of the Park significant differences were found in mean densities of 1VL subhyalinus mounds. The lowest density was at Kamulikwezi on very fine textured sandy clay loams of alluvial origin. This may be a direct result of the low tolerance already mentioned for Macrotermes species of unstable fine clay soils. The same reason could also explain the second lowest density at Kasenyi, where although the soil is of the Nyakatonzi series as are those of Mweya and Nyamgasani, it is much richer in clay than the other areas and becomes very sticky and temporarily flooded in wet weather (Lock 1970b, Table XXXI). Another factor contributing to the low densities in both these areas may be increased spacing between mounds. The mean volumes of mounds at Kamulikwezi and Kasenyi were higher than those of the three other northern areas; as were their heights, significantly so in the case of Kamulikwezi when compared to Mweya or Nyamagasani, Hesse (1955) similarly reports taller mounds on finer soils in Kenya. The positive correlation found over all areas for M. subhyalinus mound volume and nearest neighbour distance may, therefore, partly explain the lower densities in these two areas. This is discussed further below when mound dispersion patterns on a small scale are considered. The third lowest M_. subhyalinus mound density, in the Crater region, can also be explained with reference to soils. The overall mean low density was due to a significantly lower mound population on crater sides, where the soil was shallow, often with bare rock being exposed. This soil feature was likely to have directly caused the uneven distribution, and so produced the low mean density. At Mewya and Nyamagasani where the soils are similar, the mean density figures were no different. I l l Although edaphic factors and topography appear to expla in local mound distr ibut ion it is l ike ly that other environmental factors are also contr ibut ing . That vegetation can act as a p h y s i c a l b a r r i e r has been seen b y the separation of the two Macrotermes species b y the Maramagambo Fores t . The difficulty of corre lat ing mound dis tr ibut ion with f iner details of vegetation has been d i scussed b y Boui l lon (1970), and no such exercise is attempted here . No evidence was found for an interaction between Macrotermes subhyal inus mounds and n o n - M aero termes mounds, to affect their d i s tr ibut ion . T h e above d iscuss ion shows that it is difficult to predic t w h i c h mound b u i l d i n g species w i l l be present i n an area , and what the details of mound structure w i l l b e . Both these facts must be k n o w n , along with details of all environmental var iab le s , i n order to expla in the cycle of mound b u i l d i n g and wear , i n c l u d i n g the associated effects on soi l and vegetation, a l l of which are of potential importance to vertebrates . T h e range i n mean densities of 6-15 per hectare for Macrotermes mounds is s imi lar to the mean density of 12.4 per hectare found b y Sands (1965) for M . bel l icosus mounds over three separate areas i n Niger ia ; however it is l arger than densities recorded b y most other workers (Hesse 1955, Boui l lon and K i d i e r i 1964, H a r r i s 1956) wh ich probably reflects upon my inc lus ion of v e r y w o r n vegetated mounds i n the counts . That mounds are not randomly dis tr ibuted wi th in aggregations has been noted before (Sands 1965, Wood and Lee 1971). M . subhyal inus mounds , _0 . k ibarens i s mounds, and CL fu l ler i nests a l l show a s ignif icantly regu lar d i spers ion pattern wi th in aggregations. M_. bel l icosus mounds , however , are randomly dispersed or significantly aggregated. In South Africa Nel (1968) found that Hodotermes mossambicus (Hagen) which builds subterranean nests, but forages for grass on the surface, had territorial areas above the nest. Sands (1965) showed that mounds of Trinervitermes geminatus (Wassmann) were aggregated even in high density populations; this was because mounds were predominantly constructed in non-shaded areas. It is likely that if over-dispersion occurs it is maintained by some sort of territorial behaviour or mutual reaction, but the mechanisms of such a system are not known. Waloff and Blackith (1962) described similarly over dispersed ant mounds of Lasius  flavus (Fabrius); they found no relation between mound size and nearest neighbour distance. Lee and Wood (1971) did find a positive correlation between these two variables for mounds of Nasutitermes exitiosus (Hill) in Australia, as was found in this study for M_. subhyalinus mounds. This suggests a relation between the size of the territory and the requirements of the colony. Mounds may not be overdispersed as was shown by Sands (1965) for T . geminatus, and in this study by M_. bellicosus, which had significantly aggregated mound populations even when studied on a very local scale in high density situations. No obvious relation existed between mound position and any environmental feature. Their general absence on the seasonally flooded clay flats, is likely to be a direct result of this soil type. Their absence in the dense thicket/woodland near the Ishasha river was probably caused by the vegetation having a direct physical effect preventing allates entering the area to form nests. Presumably, as found by Sands (1965), 113 their positioning is due to some environmental variable but it remains unknown. No significant correlation was found between local M_. bellicosus density and local non-Macrotermes mound density, nor in their relative spacing as seen on aerial photographs. This result is similar to that for all species pairs analysed and suggests that competition between species contructing epigeal structures, is having little effect on mound distribution and dispersion in R.N.P. Lee and Wood (1971) considered interspecific competition a signi-ficant factor in determining mound dispersions in a number of multispecies associations in Australia, and Bouillon (1970) records an area around a M. bellicosus mound unoccupied by Cubitermes sankurensis mounds which were common outside this area. Bouillon (1970) also suggests a three dimensional territory, where workers penetrate deep in search of suitable soil for mounds. From the above it appears that mounds will be evenly distributed throughout a uniformly favourable habitat, but that this could be disrupted by other termite species distributions. This includes subterranean nest builders and is therefore difficult to determine, and may be the cause of some of the aggregations noted in this study. Conclusion i The variation in structure, size and density of mounds between areas means that the potential importance of termite mounds to different size classes of vertebrates cannot be predicted from a knowledge of the species of builder alone. However, the often high densities recorded, and their often regular spacing, suggest that they will form a major part of the habitat for many species, ranging from small amphibia, reptiles and mammals, to large mammals. 114 T h e nature of the potential importance of mounds , and the actual use of specific mound features, are dealt with i n the fol lowing sections, i i ) Environmental changes introduced b y termitaria The results i n this section were partit ioned into environmental categories of potential importance to vertebrates . The idea of potential importance is s tressed s ince it i s hoped that the measurements taken and calculations made cover most propert ies of mounds, the quantification of w h i c h could help to explain mound vertebrate interact ions. One large area omitted was the qual i f i -cation of changes with respect to time, the study p e r i o d b e i n g short; however , the direct ion of such changes can be in ferred from the details of many mounds, recorded at the same time. T h e effects of mounds on land rel ief The results presented above are an attempt to quantify this aspect of mound presence . T o do this the s ingle density f igures are needed along with mean mound dimensions, i n c l u d i n g a volume estimate. The mean percentage area covered b y mounds has been used prev ious ly to judge the influence of mounds i n an area . The value of 0.46% for mean percentage area occupied b y Macrotermes mounds is s imi lar to the value g iven by Nye (1955) for JVL be l l icosus mounds i n Niger ian savanna with five mounds per hectare. Meyer (1960), however , describes an area of v e r y large M . bel l icosus mounds with approximately the same dens i ty , but o c c u p y i n g 30.0% of the area . Lee and Wood (1971) expected such h i g h percentage cover to have significant effects on soi l and vegetation, a n d , i f mounds were i n cult ivated areas , to change local agr i cu l tura l prac t i ces . T h i s i s most certainly 115 t r u e , s ince the large 0_. k ibarens i s mounds o c c u p y i n g only just under 6% of the area had a profound v i s u a l effect on the landscape (figure 13). Two statistics w h i c h further help to define the effects of mounds on land rel ief are the mound surface area and the increase i n surface area caused b y mounds . The importance of these figures are that they represent the area available for evaporation of ground water from a r a i s e d often vegeta-tionally different surface , accompanied b y the deposition of minera l s . T h e y are also the area from w h i c h so i l and minerals may be washed down and so enable quantification of any watershed effect. These aspects are d i scussed be low, both b e i n g known processes of soi l modification with subsequent effects on vegetation. T h e last part of this section of results dealt with the association of large Q . k ibarens i s mounds and wal lows . A s would be expected the wallows were mainly found i n the lower l y i n g areas where the mounds also o c c u r r e d . Within this area though they o c c u r r e d more often nearer to a mound than another wal low, than expected i n a random situation. T h e i r s teep-s ided often c i rcu lar form, and the lack of slopes i n the area r u l e against the poss ib i l i ty that these wallows or ig ina l ly formed i n natural depress ions . It is possible that over many y e a r s , run-of f from mounds d u r i n g the wet season could cause gouging and s i l t ing which init iated some of these pools , however they were no doubt always accentuated b y large mammal usage , especially wal lowing b y buffalo (Syncerus caffer Sparrmann) . It appears that this large mammal activity i s preferent ia l ly directed to areas near mounds but whether this i s because of increased water from r u n off or owing to the quality of the r u n off water i n terms of mineral 116 content, is not k n o w n . Jarman (1972) observed the development of salt l i cks into zoogenous pools i n the Middle Zambezi V a l l e y , but noted that some l i cks i n c l u d i n g those i n the sides of termitaria , d i d not become pools because of their lack of water ho ld ing proper t i e s . In Tanzania however , elephant and warthog (Phacochoerus aethiopicus Pallas) were reported to d i g into the base of termite mounds to obtain ground water made more available by termite activity penetrating a calcium carbonate concretion l a y e r . It was be l ieved that these digs enlarge through increased usage and were largely respons ible for the many waterholes associated with large termitaria i n the East Selo:us Game R e s e r v e , wh ich i n turn were important i n maintaining the h igh wildl i fe c a r r y i n g capacity of the area (Tanzania M i n . of A g . R e p . 1969). A s noted (section III-4-i) the nature of_0. k ibarens i s mound formation prec ludes such a process o c c u r r i n g h e r e , and those processes descr ibed above represent the possibi l i t ies i n the Ishasha camp area . Mounds may i n fact become wallows i n R . N . P . , owing to termite activity as descr ibed above. T h i s was suggested b y the f ind ing of severa l small wallows i n the Ishasha area formed either i n the worn she l l of a dead mound or near to a mound (figure 88). The few wallows noted of this type suggest that it is not an important process here . Chemical analysis of soi l from the bottom of s u c h a wallow showed s imilar conductivity levels and mineral content as surface control samples for the same area (Table X X X L ) . Conc lus ion . i i L a n d rel ief results have shown that a h i g h percentage of ground surface may be occupied b y mounds , emphasiz ing that mound effects come not only from 117 their presence as fixed points but also from their large surfaces. This means that any plant or soil changes occurring on or in mounds can be over con-siderable areas, and are likely to be of significance to grazing mammals. The identification of mounds being either directly involved in wallow formation, or in enhancing a local area for zoogenic wallow formation, is of obvious importance to large mammals. Accessible space within termitaria Section III— 2—ii shows that Macrotermes mounds possessing non-termite holes were generally of larger basal diameter than intact mounds, and that the only significant difference in heights was for mounds with holes being lower. This suggests that mounds were more likely to have non-termite access points as they were being worn down by weathering and other agencies, after they had been abandoned, the soil from the top of the mound being largely deposited around their bases. Ruelle (1964) emphasized how live M. bellicosus mounds survived erosion by rain, attacks by ants, and physical wear by large vertebrates, keeping a fairly uniform appearance, because of the unceasing work of the mound's many occupants. The actual erosion of mounds to produce holes has been noted by Nye (1955), whilst Williams (1959) reported large mammal trampling destroying nests. From the excavation results it was found that all ML subhyalinus mounds having non-termite holes through their thick outer walls had been abandoned by their original colony. If termites were present in holed mounds, the nest size was smaller than that of the original mound builders. M_. bellicosus, on the other hand, with much thinner walls could apparently be easily worn 118 down by large mammal rubbing and weathering to produce entry points which were tolerated by the inhabitants. So only for M. subhyalinus mounds does the percentage with entry points equal a minimum percentage of abandoned mounds. This figure ranged between 12.5% and 37.5% for the five study areas, and is similar to the lower estimates presented by other workers (Sands 1965, Lee and Wood 1971). The use of holes in termite mounds by animals has been recorded many times and is an area discussed in section IV-3, however, as far as is known no worker has previously presented details on the density of mounds with entry points. It is not known how many of the abandoned mounds described by other researchers had access points. As was observed in this study and by Morrison (1948) , abandoned mounds can be worn down to slight raises of wide diameter and still have no entry points. The main cause of holes into M. subhyalinus mounds were animal agencies, including here cave-ins which probably mainly stem from large mammal trampling. In I V L bellicosus mounds weathering was the main direct cause. It can, therefore, be seen how important the species of mound builder is to the weathering properties of a mound. This is again shown by 0_. kibarensis mounds, where, because of the local compact nature of the galleries no non-termite holes were formed, apart from rodent burrows which were probably usually associated with old termite shafts. The type of interior into which holes enter also depends on mound species. Holes into the side of a 1VL bellicosus mound were more likely to enter a large space, formed by the dropping of galleries from the outer casing, than holes into the side of a M. subhyalinus mound. Both however 119 always had areas of collapsed galleries which were intact, and partitioned some of the spaces into numerous complex passages. This is important since it could allow a number of small vertebrates to occupy a single mound at the same time. The natural termite shafts of live 1VL bellicosus mounds and Odontotermes species mounds and nests must also be considered as potential space for smaller vertebrates. The significance of the figures given for the volumes of dead mounds with entry points, and Odontotermes fulleri shafts, are difficult to appreciate with lack of comparative data. The main way that their importance is seen comes from the results of usage of mound space by vertebrates, with often a number of species occupying one mound at the same time, let alone sequentially. That the mean per hectare volume of dead M.subhyalinus mounds with entry points equalled 9.6 average sized aardvark burrows, over twice the mean number which occurred in any area, also em-phasizes the large volume represented by available mound interiors. Probably the only other natural features of the environment in R.N.P. that represent at least partially physically protected sites of similar proportions were deep erosion gullies on steep slopes. On flat ground bush vegetation would appear to be the only other dense cover for all but the smallest vertebrates. The microclimate of mound space. The temperature inside live termite nests has been found previously to vary diurnally and from day to day; however, especially inside large epigeal structures like the thick walled Macrotermes mounds, the temperature has been found to be very much damped in variation and always higher than the surrounding temperature, following variations in this with a lag (Holdaway and Gay 1948, Ruelle 1964, Noirot 1970). Ruelle (1964) recorded a mean central nest temperature of approximately 30 C in _M. bellicosus mounds in Central Africa, with a maximum daily fluctuation of 3 . 5 ° C , particularly under the influence of rain. These figures are very similar to those found in this study for the temperature in the galleries of a live M_. subhyalinus mound. The main cause of this high constant temperature is the metabolic heat produced by many thousands of termites within a nest (Luscher 1961, Noirot 1970). Ruelle (1964) and Luscher (1961) both suggested that the constancy observed cannot be completely explained by a passive termite role, and that termite activity to open and close tunnels may be carried out to directly affect temperature. No measurements were taken here of the temper-ature in available space in live nests, but the above studies indicate that these areas will have a more constant temperature than the exterior air shade temperature or the temperature in burrows elsewhere, and also probably higher. Whatever the precise mechanism of temperature regulation, it appears that the effect is largely lost once a nest is abandoned (Ruelle 1964). Holdaway and Gay (1948) recorded higher soil temperatures from probes in a compact live nest of Nasutitermes exitosus as opposed to control soil nearby; after application of arsenic to kill the colony the nest temperature followed the adjacent soil temperature at an equal depth. In this study it has been shown that the air temperature in an abandoned M. subhyalinus mound followed the external air-shade temperature, but was greatly damped and followed with a lag period of a few hours. It was also found that any subterranean burrow 121 showed greater damping than that inside an epigeal dead mound, but that mound burrows may have a more constant temperature, warming up more quickly in the morning than control burrows, and cooling down more slowly in the evening. Bradley (1971) similarly records damped temperature variation in warthog burrows in Nairobi National Park; the air temperature inside ranging between 16° C and 20*C, whilst the air-shade temperature varied between 14°C and 24°C. One process that would be expected to affect mound temperature in dead as well as live mounds was suggested by Ghilarov (1962) who described deep galleries below Anacanthotermes nests in the U.S .S .R. , descending to damp soil at depths of 10-15 M. These would allow moist air to rise to the nest and so cool as well as humidify them. A dead Macrotermes mound with entry points is therefore expected to offer a more damped temperature regime than air-shade temperatures, but probably not so damped as subterranean parts of the mound structure or burrows elsewhere. Mound shape, structure, including nature of entry points, mound position, mound vegetation, and adjacent vegetation will all affect temperatures inside mounds, dead ones more so than live ones. Termites need to live in conditions of high humidity since they are ex-tremely susceptible to dessication (Noirot 1970, Luscher 1961, Harris 1971, pers. obs.). Luscher (1961) found relative humidities of 96% and above in live nests of five species including M. bellicosus in the Ivory coast. No measurements were taken of relative humidity inside natural termite space within live mounds in this study, however, it seems that such space would 122 have a higher more constant humidity than the external air. A number of mechanisms have been suggested as the primary cause of this high humidity. In Macrotermitinae the fungus garden has been thought to produce sufficient water to affect humidity and it is known to have high moisture retaining coparisites. (Hesse 1957, Noirot 1970). Decomposition of cellulose which is the main food of termites results in the release of water. Such metabolic water is also considered important in moisture control. (Hesse 1955, Luscher 1961, Noirot 1970). During this study fresh mound growth, which was damp to the touch, was found on both M_. subhyalinus and _M. bellicosus mounds during the dry season, when both the older mound suface and the surface soil were extremely dry. Luscher (1961), Bodot (1967) and Noirot (1970) describe worker termites bringing up moist clay or purely water into mounds from considerable depths . This process would, therefore, also increase nest humidity. All these three processes for increasing humidity involve the activities of termites. It is, therefore, not surprising that in this work the interior galleries of live mounds always felt moist and were pliable, whereas the interior galleries of dead mounds were very dry. Hesse (1955) also reported consistantly lower soil moisture figures for abandoned mound interiors than for "living" mounds. It is possible that the high moisture retaining properties of clay used in mound construction along with the often deep shafts below mounds, which may reach the water table (Bodot 1967, Ghalarov 1962, Yakushev 1968), could cause moist air to rise from below (Ghalarov 1962), which would be 123 retained b y dead mounds . Equal ly w e l l , however , it is l ike ly that these galleries would promote v e r y good drainage and cause a dead mound to be p a r t i c u l a r l y dry. M y data for relat ive humidity i n a dead ML subhyal inus mound i n the wet season showed variat ions between 74% and 88%, which was at the h igher end of the external a ir range of approximately 60%-90%. F u r t h e r readings would have to be taken at var ious times of the year to f ind the signif icance of th i s . Unoccupied large u n d e r g r o u n d b u r r o w s showed a more damped relative humidity regime than i n the epigeal dead mound, with again a h igher mean humidity b e i n g r e c o r d e d . T h i s has been noted prev ious ly i n E . A f r i c a by Geigy (1955) who recorded relative humidities always close to 90% i n u n -occupied warthog b u r r o w s . The actual mean about which these damped b u r r o w humidities v a r y , w i l l depend upon climate, so i l and the posit ion and size of the b u r r o w . In this study a burrow under a dead mound showed equivalent humidity values and fluctuations as an adjacent control b u r r o w . A large burrow under a l ive M_. subhyal inus mound, however , was found to have a consistantly h igher relative humidity than a control b u r r o w ; which seems to show how such an environment can be modified to a large extent by a termite nest. Further records would have to be made to test for the generality of this f i n d i n g , however , especial ly as the relative humidity of the control b u r r o w was lower than expected from al l prev ious data, and may suggest that this burrow was for some reason non-representat ive . 124 Conclusion iii Both live and dead mounds can offer accessible refuge and breeding areas of complex subdivided space, available to small and medium sized vertebrates. Physical protection and a buffered microclimate are expected to favour mounds as den sites. As stated by Southern and Hook (1963).,the behavioural selection of suitable microclimates by small mammals is probably more important than the potential for physiological adaptation. In mounds, a well buffered temperature regime would combine with any behavioural adapta-tions such as huddling to produce a thermoregulation that is achieved at a lower energy cost to the animal. Mounds could also offer protection from extreme environmental fluctuations such as flooding and fire. The soil of termite mounds Soil particle size. Most studies on Macrotermes mounds have found mound soil to be similar to adjacent subsoil with respect to particle size, this area being assumed to be the site of soil collection for the mound (Hesse 1955, Nye 1955, Harris 1956). Some workers however did note a selection for finer fractions in a soil not rich in clay, and against finer fractions in a soil with much clay (Hesse 1955, Nye 1955, Stoops 1964). In R.N.P. M. subhyalinus mounds had soil textures equal to, or finer than, both surface soil and subsoil irrespective of the absolute textures of these soils. M_. bellicosus mounds had soil of texture equal to subsoil at a depth of one metre. Boyer (1958) also records ML subhyalinus bringing up much fine material into the mound and cementing it into cone-like structures. 125 Conductivity and Minerals The amounts of water soluble salts as measured by conductivity were higher in Macrotermes mounds than in the adjacent soil. Watson (1962) recorded increases in water soluble salts in a large dead mound in Rhodesia; the maxi-mum increases occurring in the central nest area and directly below the mound. Watson (1969) found much higher conductivity measurements in a dead mound as compared with the adjacent soil, whereas a live mound had conductivity values equivalent to those of adjacent soil. Similarly in this study, the con-ductivity of fresh termite worked mound soil was lower than older soil on the same mound and much lower than the soil from a dead mound. The values for total exchangeable bases parallel the conductivity measurements, as has been found in previous work. Weir (1969) recorded increased conductivity in termite mound soil during a wider survey of areas used by large mammals as salt licks; he was also able to attribute high con-ductivity to high concentrations of soluble sodium salts. In this study tlie high-est percentage increase in mound soil for any one exchangeable base was for sodium. Hesse (1955) found increased exchangeable bases in termite mounds as compared to the subsoil from which they were constructed, whether the mounds were alive or abandoned; he suggested that minerals entered mounds in ground water. Stoops (1964) recorded high total exchangeable base content in M_. bellicosus mound soil as compared with nearby topsoil or subsoil, most concentration occurring in the mound interior. New working and older mound surface soils were equivalent in exchangeable base content. He suggested that minerals from excreta and other organic matter attributable to termite activity 126 were responsible for these increases. Nye (1955) also noted maximum ex-changeable cation content in the central nest of Macrotermes mounds in West Africa, over that in other mound parts, adjacent topsoil, and adjacent subsoil. In this study soil from fresh termite working was found to be low in water soluble salts, but high in exchangeable bases, suggesting that there was leaching from the fresh working probably because because of its position on the top of the mound (Watson 1962). At Mweya the highest pH value, con-ductivity , total exchangeable bases and nitrate nitrogen figures were from the dead M_. subhyalinus mound interior. This was not so for the M. bellicosus dead mound interior at Ishasha. The main difference between these two mounds, apart from the nature of the subsoil already mentioned, was that the M. subhyalinus mound was liable to seasonal flooding. This finding, therefore, supports the basic ideas of Hesse (1955) and Weir (1973) that mineral accumu-lation in mounds can occur by the mounds acting as sites for the evaporation of mineral rich ground water, either from the mound surface or ventilation shafts. As we have seen the mineral rich subsoil will also cause accumulation, as will the differential leaching of the intervening soil to give an apparent increase in the mounds (Watson 1969). The importance of the activity of termites themselves adding to mineral content via organic material they bring into the mound is not known, (Lee and Wood 1971). It seems unlikely that such a process could contribute much to the often large accumulations of minerals, particularly calcium carbonate, which have been recorded, (Milne 1947, Pendleton 1941). 127 So far only Macrotermes mounds have been considered. The kibarensis mounds at Ishasha are different in structure because as we have seen they were formed from the accumulation of soil in an area of seasonal floodland, much of the soil never having been worked by termites. That some of the soil is attributable to washed down turrets which are probably made of sub-soil (Harris 1971) may partly explain the higher values for exchangeable bases, mineral nitrogen, available phosphorus, pH and conductivity found in them when compared with surface soil. However the subsoil nearby was low in minerals. It seems likely, therefore, that the process of seasonal iriundation in this case of not particularly mineral rich ground water, has caused these accumulations over a long period of time. That other agencies may have con-tributed cannot be ruled out, including minerals arising from the preferential defecation by mammals on mounds in this area (see section IV-3) . The importance of these results in the context of the present study is that mounds can offer localized sites of mineral accumulation when compared to adjacent topsoil and even subsoil. In the past most studies have emphasized that this means the minerals are lost from the soil-plant-animal-soil cycle, until the mounds are abandoned and eroded away (Goodland 1965, Lee and Wood 1971). Goodland (1965) stated that live large mounds in South America were bare, and assumed that dead ones eroded away quickly. Live mounds also erode, although probably less quickly than abandoned ones; the growth of a mound therefore is a net increase, the volume added being less than the volume of soil brought up by termites, the difference being that washed away. Watson (1969) studied two large Macrotermes mounds in Rhodesia, one was alive 1 2 8 and bare, the other was unoccupied with the interior crumbling and had dense tree vegetation on it. Hesse (1955) stated that it was rare to find vegetation on live M. subhyalinus mounds in Tanzania. The results presented here are in contrast to the above. Vegetated mounds were often taller and usually significantly larger in diameter than bare mounds, suggesting that only new mounds were bare, older live ones as well as dead ones having vegetation. Excavation results confirmed this; low, worn vegetated mounds sometimes having termites whereas large bare ones were sometimes abandoned; the former case probably being one of recolonisation. Therefore, as discussed in the next section, at least some of the minerals accumulated in mounds are directly avail-able to vegetation in R.N.P. , and even live bare mounds are subject to soil wash-off. The possibility of mound erosion causing a unique area around a mound will also be discussed in the next section with respect to vegetation. Some workers have considered mound erosion to be an important soil forming process in parts of Africa. In Uganda Harrop (1962) noted the occurrence of a conspicuous stone line 0.5-2.0 meters below the soil surface. This was explained as having been formed over a long period of time by termites bringing up only fine material for mound construction; the mounds were then eroded to form a fine layer over the ground surface. Nye (1955) also described this process and believed it to be more important than any mineral enriching due to the distribution of mound soil. He suggested that by continually lowering the topsoil it could help the soil from becoming too deeply weathered, and enable deeper rooting plants to keep in touch with de-composing mineral releasing nutrients in the subsoil. Nye (1955) estimated 129 that 0.5 tons of soil per acre, was deposited on the topsoil by eroding mounds in West Africa each year. This weight of soil equalled 1/10 of the standing crop of mounds, which is exactly the same figure given by Pomeroy (pers. comm.) from a more detailed study of mound decay in Uganda. Applying this fraction to the standing crop of minerals found for Macrotermes mounds in this study, suggests that an appreciable amount of minerals could be added to the topsoil annually. For instance, at Mweya this means that 19.1 kg of calcium, 4.4 kg of potassium, 4.9 kg of magnesium and 0.7 kg of sodium are added to a hectare of topsoil annually; values which would significantly con-tribute to plant annual mineral requirements, especially calcium. The importance of these figures must be estimated with reference to the status of adjacent top-soil. As we have found out (section II) much of the soil within R.N.P. is rich in minerals even in the topsoil, so the importance of the mounds in mineral cycling remains uncertain in R.N.P., as a process, however, it is shown to have potential importance especially in areas subject to leaching. Langdale-Brown et al (1964) have attributed the low fertility of many Ugandan soils to the process of leaching, and in such areas mound erosion is likely to be important in keeping minerals available in the topsoil. Considered here are only the direct effects of termite mounds on soil. The important role termites themselves play, by aerating soil with their passages away from mounds has not been covered, nor have the possible detrimental effects of the removal of surface litter and aiding of soil erosion (Harris 1971). 130 Conclusion iv Jarman (1972) stated that the main requirement for wallow formation by large mammals, was a fine textured soil that puddled well. The selection of finer grained soil by termites for mounds could be important in attracting large mammals to form wallows in the wash-off area around mounds. Mounds are often concentrated sources of minerals, directly available to animals via geophagy, or indirectly via plants growing on or around mounds. Wash off of soil from even live mounds may also significantly add to minerals available in the adjacent topsoil. The role of mounds in mineral cycling is expected to have the most effect on vegetation and animals in areas of soils with lower mineral content than found in R.N.P . , and especially where surface leaching occurs. The general opinion given above that mound minerals are largely unobtainable by vegetation appears to need reassessing in more specific situations; the whole process being more dynamic than has been previously implied. The vegetation of termite mounds In contrast to a number of previous studies, all but the new small M. subhyalinus mounds in R.N.P. were usually vegetated. Only the areas of Kamulikwezi and Kasenyi stand out as having higher percentages of bare mounds; these areas had taller thinner mounds built from particularly clay-rich soil. Since the mineral content of these soils was as high as those from other northern areas of the Park, it is likely that the steepsideness and imperviousness of the mounds to water are preventing plant colonization. Kamulikwezi, Kasenyi, and Ishasha, where a similar process probably accounts 131 for the high number of bare M. bellicosus mounds, lie in the centre of the rift valley and are subject to the lowest rainfalls within the Park. Glover (1964) recorded that crop plants apparently grew better on broken-up mound material, but that live mounds were usually bare; he also suggested that it was the soil's hardness and imperviousness to water, that were the important factors. Thomas (1941) recorded that in grassland on the Sese Islands in Lake Victoria, only the new mounds were bare, all others showing stages in a definite successional sequence of colonization from grasses to trees. In contrast, Murray (1938) emphasized the bare nature of live mounds in South Africa, believing the cause to be a reduced availability of minerals in mound soil, even when the absolute content of minerals was higher. The highest percentage of thicketed M. subhyalinus mounds occurred in the Mweya area, and here they were more abundant on mounds in a low lying area liable to seasonal flooding. Likewise, in seasonal floodland at Ishasha thickets were largely restricted to 0_. kibarensis mounds, except on the slopes bordering a small stream. Langdale-Brown et al (1964) noted this phenomenon as a distinct vegetational type, and attributed it to the better drainage of mound soil. Harris (1971) also described this feature and in-cluded it as an example of Troll's (1936) term "Termitensavannen", which covered grassland areas where thickets were confined to mounds. Such definite affects occurring in otherwise grassland areas are expected to have significant effects on the vertebrate fauna of the area. Jackson and Gartlan (1965) noted that in some areas where crops grew better around the base of mounds, it was because of increased drainage in these sites . 132 The high percentage of thicketed mounds on steep slopes near Mweya was probably in part due to the difficulty of colonizing thickets to survive the harsh run off occurring on these bare overgrazed areas. In such places mounds and their vegetation could be an important stabilizing factor resisting erosion. It should be remembered though that the termites themselves will probably contribute to maintaining bare soil by removing trampled vegetation once overgrazing has reached a certain level. The thicket and tree species found on mounds in R.N.P. were those of the surrounding grasslands. In other studies a distinct vegetation has been found associated with termite mounds, completely different from that of the adjacent area. Harris (1966) reported that in dry areas where Brachystegia woodland was favoured rather than forest, the presence of termite mounds led to the development of evergreen vegetation in advance of the main body of invading deciduous trees. The raising of vegetation above grass fires has been suggested to help cause this process (Griffith 1938, Harris 1966). In Rhodesia Wild (1952) recorded distinct vegetation on large Macrotermes mounds, and suggested that mound species were restricted to this habitat because they were outside the areas whose climatic factors favoured their optimum growth and enable them to compete successfully with the dominant species found there. The important environmental factors involved were be-lieved to be the increased water-holding capacities of clayey mound soil, and its increased mineral content. Thomas (1941) found increased mineral content in soil at the base of Macrotermes mounds on the Sese Islands of Lake Victoria; he believed this to be the factor allowing trees rather than grass to grow 133 around the base of mounds in those aras. Depending on the area, a number of processes could act to produce thicket vegetation on mounds. Protection from fire, better drainage, in-creased water holding capacity, and increased nutrients may allow thickets to establish preferentially on top of mounds. Increased moisture, increased drainage or increased nutrients from run-off, could account for a ring of thicket vegetation around mounds. In the present study this latter type of association was seen but was not distinguished from thickets on mounds. A number of workers have emphasized the importance of termite thickets to forest regeneration in previously cleared areas (Thomas 1942, Jackson and Gartlan 1965, Harris 1966). Such a major effect on vegetation would have equally large effects on the vertebrate fauna of such an area. As in this study, a number of other workers recorded mounds with distinct bare areas around their bases (Thomas 1942, Macfadyen 1950, Glover 1949, Glover et al 1964). Glover (1949) showed how wash-off of fine mineral rich material caused hard pans to form around the base of mounds, which prevented plant colonization. Glover et al (1964) noted that on sloping ground in the Loita Plains of Kenya the shape of the bare area was elongated downhill, with a tall grass zone uphill from the mound in the area of highest run-off and moisture penetration. Such a process is probably responsible for the bare areas with vigorous grass growth at their periphery found around mounds in R.N.P. Here the pattern is also accentuated by large mammals rubbing mounds, with associated trampling around the base. 134 A s expected low r u n n i n g grasses were more common on mounds than i n the s u r r o u n d i n g area where taller grasses dominated. Espec ia l ly s tanding out as a "mound species" was Cynodon dacty lon, a species noted b y Lock (1970a) to occur preferent ia l ly on mounds i n R . N . P . , especial ly i n a small variety; he also noted that this small variety was p a r t i c u l a r l y heavi ly g r a z e d . T h e only other significant difference was for a lack of Themeda t r i n d r a on mounds, a tall grass of low nutrient value not favoured by graz ing animals (Field 1968). It is poss ible that i n areas with a h igh density of vegetated mounds , their presence may cause beneficial changes i n the proport ion of palatable grasses present . It i s not known if the biomasses invo lved would be s ignif icant. In R . N . P . the biomass calculations suggest that i n the absence of large mammal trampl ing and g r a z i n g , mounds produce more grass than control areas , presumably because of one or a number of the factors mentioned above. Normal ly within the Park less biomass of grass on mounds seems the r u l e . T h i s i s poss ibly because of mammal graz ing and t r a m p l i n g , with termites then removing vegetation on and near mounds. T h e poss ibi l i ty that preferential g r a z i n g of mound grasses by mammals was contr ibut ing to the difference, is d i scussed i n section I V - 3 . T h e f inal way i n w h i c h mound vegetation could affect vertebrates is if it has h igher nutrient content than other vegetation. In R . N . P . the overa l l result is that mounds grasses often have significantly more crude prote in and a h igher mineral content than control samples , and that this increase occurs i n both wet and d r y seasons. The deeper rooted s h r u b s do not have 135 any significant differences in nutritive value between these two sites. The potential importance of these findings to grazers in R.N.P. can be seen by examining the absolute nutritive values which occurred. The crude protein values both on and off mounds were generally low; some, especially control samples, being close to or below the 5% figure suggested by Glover et al (1960) for maintenance, and below which digesti-bility of the whole food seriously declined. The protein figures of control samples were similar to those of the same species from R.N.P. analysed as whole plants by Field (1968). The mound figures are generally higher. Long et al (1969) recorded equally low protein values for many grass species collected in R.N.P. and other parts of Ankole. Because of these low figures they confined their discussion of quality and productive value to this one nutritive component. In view of this it seems likely that the increased protein in mound grasses is a useful increase available through selective grazing. The common mound grass Cynodon dactylon had a high protein content, as had previously been found in other Ugandan studies (Field 1968, Long et al 1969). Calcium content in grasses was sometimes low , especially in control samples, falling below the limit of 0.4% dry wt. set by Naik (1965) as the value needed by most large mammals for bare maintenance. All values were within or slightly above the range recorded by Long et al (1969) for grasses in Ankole, including R.N.P. ; which they considered adequate for livestock beef production if not milk production. The A . R . C . (1965) suggested 0.24 -0.56% calcium for beef cattle bare maintenance, depending on the calorific value of forage. 136 Potassium values for grasses on mounds and also control areas varied between 0.37% dry wt. and 2.16% dry wt., being similar to the range of 0.49-2.15% dry wt. recorded by Long et al (1969) for grasses in R.N.P. and Ankole. Neal (1941) suggested that plant growth would fail before pasture potassium became limiting for livestock. Sodium values ranged from 0.013-0.089% on mounds and from 0.009-0.048% away from mounds. Long et al (1968) reported a range of 0.030-0.089 away from mounds, which they considered low for dairy cattle. Du toit et al (1940) suggested 0.02% as the minimum dietary level for cattle maintenance. Although the requirement of the native game are not known, it seems probable that increased sodium in mound grasses in R.N.P. could be important to grazers in some areas, especially since sodium has been shown to increase the most in mound soil over other minerals. A mound sample of Cynodon  dactylon showed the highest sodium content. Weir (1973) believed that large mammals within R.N.P. could obtain adequate sodium from food and drinking water, without considering mound vegetation. The magnesium levels of both mound and control samples were generally above the range of 0.08-0.10% suggested by the A . R . C . (1965) for beef cattle maintenance. Long et al (1969) reported similar values to those in this study, and which they considered adequate for beef cattle. Phosphorus was not analysed for in grass samples from R.N.P. Long et al (1969) reported increasing phosphorus content in grasses as the amount of available phosphorus in the soil increased. In view of the very high values of available phosphorus found in the Park, it is expected that it is adequate in all vegetation. Long et al (1969) recorded the highest level of phosphorus in Setaria aequalis, a grass associated with termite mounds in another part of Ankole. In R.N.P. mound soil had significantly less available phosphorus than control soil. Most Park soil samples contained in excess of 20 ppm available phosphorus; Naik (1965) believed 0.1 ppm to be adequate to produce grass with a sufficiency of this mineral. Conclusions v Mounds range from being completely bare, to having thicket vegetation. It is expected that vertebrate/mound interactions will vary between these types For example, mounds could offer a major part of the cover available to small and medium-sized vertebrates if thicket vegetation was restricted to them. In contrast, bare mounds may make a better marker post or lookout post, and offer a superior surface for large mammal rubbing. That the majority of mounds are vegetated in R.N.P. means that mineral accumulations in mounds are largely available to plants. Results have shown that mound grasses do generally have higher protein and mineral levels, and the data on Cynodon dactylon, and also on plant biomass, both suggest that selective grazing of these grasses does occur. Mounds can therefore increase the range of nutritive value of grasses available to animals by selective grazing. Depending on the area, this proces could be important with respect to any mineral, or other nutritive chemical produced because of improved growth of a plant. In R.N.P. increased proteir and sodium are potentially significant. The importance of any increase will not only depend upon the absolute values found in control vegetation, but 138 also on how selective the animal is with respect to plant species, to in-dividual plants, and to plant parts. The availability of alternative sources of minerals, and the biomasses of mound vegetation are also important. The significance of mound vegetation may not be the same throughout the year. Mound grasses could become more beneficial to grazers during the dry season, when protein and mineral content of grasses usually falls (Howard et al 1962). These topics are discussed further in section IV-3. The pathways that minerals travel from the soil, to plants and then to animals is far from fully understood. Mineral increase in one soil will result in plant increases of that mineral, but in another soil will not; high values of one mineral in vegetation can interfere with the utilization of a second mineral when the vegetation is eaten by a herbivore (U.S.D.A. 1965). Such compli-cating factors as these mean that generalizations regarding the beneficial nature of mound vegetation to herbivores are very difficult to make; each local situation has to be assessed separately, taking into account soil deficiencies of the area. 139 IV. THE USE OF TERMITARIA BY VERTEBRATES 1. METHODS: THE USE OF MOUNDS BY VERTEBRATES, i . Mounds as fixed elevated points. The presence of vertebrates on termite mounds was recorded during the whole study period. Apart from some cases when the observation period was prolonged and behavioural notes were taken, the animal's activity was noted as one of the following eight classes. 1) Feeding on mound vegetation. 2) Feeding on animals found on or in the mound. 3) Feeding on food brought to the mound. 4) Using the mound as a lookout point; that is actively scan-ning the surrounding area. 5) Defecation and urination. 6) Display; defined as a repeated conspicuous behaviour pattern directed from a mound towards individual(s) of the same species. 7) Marking; rubbing of secretions onto a mound. Large mammals rubbing against mounds were also noted. 8) The final class was resting, where none of the other activities were taking place. Notes on birds were only taken if carrying out an activity other than resting. All other methods in this section on mounds as fixed points are indirect. During the ground transects (see section III—2) details were taken of mounds rubbed, horned, or trampled by large mammals. The presence of small carnivore and primate droppings was also recorded. The possibility of preferential defecation on mounds by grassland primates, as described by Jackson and Gartlan (1965), was investigated for baboon (Papio anubis) on large_0_. kibarensis mounds at Ishasha. A 1600 sq. meter study area was used which enclosed four partially thicketed CL kibarensis 140 mounds and two small bare Trinervitermes sp. mounds. The remaining ground was a mosaic of bare earth, Maerua edulis dwarf shrub, and patches of long grass particularly S. pyramidalis. Each month between November 1972 and July 1973 baboon droppings were mapped and then removed from the area. Seeds in the droppings were identified using a reference collection of shrub seeds collected nearby. Attempts were made to germinate seeds from faeces and from ripe fruit collected at the same time. This was carried out in the lab on moist filter paper in petri dishes. i i . The occupation of mound interiors by vertebrates. The species involved. The vertebrate species entering mounds in R.N.P. were identified through mound excavations. Mounds were only excavated which had some type of entry point, either termite-made or otherwise. Five M_. subhyalinus mounds, eight 1VL bellicosus mounds, thirteen O. fulleri shafted nests, and four O. kibarensis valley bottom mounds, were dug out. Measurements of the mounds were taken, including details of the nature and probable cause of openings. Drawings were made of mound interiors. Attempts were made to capture and identify all live animals found. Any other signs of vertebrate presence, such as droppings or bones, were collected and their position in the mound recorded. The presence or absence of an active termite nest was also noted. The use of mounds as den sites by a small carnivore. Neal (1970) identified banded mongoose as a regular user of dead Macrotermes mounds as dens, and stated that virtually all dens were in mounds. His method of locating dens was to search an area for piles of droppings which are deposited communally near the den entrance. This biased his results towards mound dens since at these the droppings are laid conspicuously on the mounds themselves. My method of den location did away with this problem. Packs were followed to their den in the early evening. The site was visited the next morning and the den examined. Details were taken of the number and size of entrances, and the placing of the den with respect to local topography. Twenty-three dens were located in this manner in the Mweya area, between December 1972 and May 1973. The use of large burrows in mounds and elsewhere by medium-sized mammals. Large burrows usually over 30 cm diameter were common in the Mweya area (see section III—3—ii) . Six mound burrows and eight burrows elsewhere on Mweya peninsula were monitored for 2 to 8 days each month, between November 1972 and June 1973. Black cotton was tied low across the entrance to each burrow between 17.30 hr. and 18.30 hr., and fine sand smoothed near the entrance. The burrows were checked again approximately one hour after dark at 20.00 hr., and also between 8.00 hr. and 9.00 hr. the next morning, the cotton being reset as necessary. This allowed nocturnal and diurnal occupants to be separated. Four large burrows were excavated on Mweya peninsula in addition to those under mound D3. Measurements were taken and particular attention paid to the ends of tunnels to attempt to confirm what species had dug them . 142 Termitaria and small mammal distribution. This subject was investigated by means of small mammal trapping in the Crater Region and at Ishasha. Craters In the Craters three questions were being asked. First, was any species of rodent restricted in its local distribution by a lack of Odontotermes shafts? Second, when foraging does any species tend to remain closer to shafts than would be expected on a random basis? Third, is there particular evidence of either one or two above, after an area is burnt? The basic method used to answer these questions was removal snap-trapping for three days and nights each month, from November 1972 until June 1973. Two trap grids were run each month, one placed in an area of high shaft density and the other in a similar topographical and vegetational situation, but locally lacking shafts. Each grid was accurately mapped with reference to shafts and different grids used each month. The grids consisted of five rows of eight trap points, each point five meters apart, with the rows ten meters apart. Because of the high number of traps in this small 0.014 hectare area, only one trap was placed at each point. Large Reporter traps and smaller Museum Special traps were positioned alternately. These two trap types were used in order to capture all size classes of rodent found in this area (Neal and Cook 1969) . Traps were baited with a mixture of maize meal and ripe bananas; a previously successful recipe (Neal 1970). Traps were checked and set between 8.00 hr. and 10.00 hr. and again between 16.30 hr. and 18.30 hr. Each capture was identified to species 143 and the trap site recorded. The animal was sexed, weighed, and measured with respect to head plus body length and tail length. A sample of stomachs was also taken each month for analysis of diet with emphasis on the termites taken. Additional snap-trapping was undertaken for four days and nights immediately after a 'cool burn' (one against the wind) had passed through the southern part of the Crater Region on the 26th of January 1973. Thirty traps were set around the edges of bushes, six near six Odontotermes shafts, and thirty-six controls in the open, over 100 m. from bushes or shafts. The traps were set and checked as described above. The only other method involved was live trapping. This was carried out on each of three months after the January burn. Two grids were used both in the non-shrubbed Imperata grassland in the floor of Naruzygoti crater. Each grid consisted of two rows of six Haverhart traps (type 1) each 5 meters apart, with the rows 10 meters apart. One grid was in an area of local high shaft density, while the other lacked shafts within the grid and for at least 30 meters around. Setting and checking of traps was as described above for snap traps. On first capture animals were toe clipped for individual identification, using the hind toes as units and the fore toes as tens, sexed, and weighed. On subsequent captures they were identified and weighed. The shaft grid was accurately mapped. Ishasha At Ishasha trapping was carried out to see if small mammals were more likely to occur on or near large CL kibarensis mounds than elsewhere 144 in an area liable to seasonal flooding. Snap-trapping was carried out between November 1972 and July 1973 as described above for the Crater Region. Each month two grids were trapped, one in a mound area and the other between 300 and 500 meters away from the nearest mound. Live trapping was also conducted at Ishasha, with the same arrangement of traps and general methods described above for the Crater Region. Trapping took place each month between February 1973 and July 1973, on two grids. One was positioned over a large 0_. kibarensis mound (No. LI) (see figure l:19)v..:. the other was in the adjacent grassland approximately 500 meters away. Both grids were between the snap-trapping area and the high ground near the Ishasha river forest (figure 19) . Another method employed at Ishasha was the use of small mammal dropping counts as a comparative estimate of area utilization. In March 1973 twenty-two mounds were sampled in a large mound area burnt on February 7, 1973. Notes were taken of the presence or absence of termite shafts, rodent runways, rodent droppings, and thicket vegetation. In April 1973 five mounds in the same area, not previously sampled, were examined. Details were taken of mound height, mean mound diameter, number of termite shafts or other holes, presence of rodent runways, presence of rodent droppings, presence of thicket vegetation, and activity of termites. A transect was run over each mound, with a 0.25 M quadrat being randomly placed on the mound top, mound side, mound base, and five meters away, 10 meters away, and >50 meters away from the mound. A l l rodent droppings within the quadrat; were counted. Droppings were noted as being old and partially burnt or green and fresh for the herbivorous species A. niloticus. i i i . The use of mound soil and vegetation.  Soil No animals were followed with the hope of observing geophagia. In an exploratory study on soil utilization in R.N.P. during August 1969, Weir (1974) stated that geophagia was not seen in any species, but indirect evidence of soil excavation and eating by elephants was found. Similar in-direct methods were employed here. During the ground transects (see section III-2—i) mounds were examined for signs of trampling or horning by large mammals. Soil samples were taken from disturbed mounds. It is likely that wallows are the result of animal scrapes originally made to obtain minerals (Weir 1960, Jarman 1972). The methods used to examine the nature and distribution of large wallows in the Ishasha camp area have already been described, including the collecting of soil samples. The species of large mammal using the wallows was recorded. Vegetation. Direct observations of animals eating mound and other vegetation were not made on a quantitative basis, apart from a few on warthog. Indirect methods were employed to compare the grazing pressure on mounds and control areas. First, data on grass biomass were taken from on and off mounds in an area protected from large mammal grazing, and from mound and non-mound control areas. The methods used have already been given in section III— 2—ii. Second a study was made on the heights of Sporobolus  pyramidalis leaf blades occurring on and away from CL kibarensis mounds at Ishasha. The hypothesis was that if this species was under increased 146 grazing pressure on the mounds,then a sample of blades from a mound would have a significantly lower mean length than a sample of blades from a control area nearby. Measurements were taken during November 1972, January 1973, and Feburary 1973. Each time 20 blades were selected at random from a number of tussocks on each of four mounds and four control areas. iv. The use of termitaria as a concentrated source of food. During the ground transects (see section III-2-i) all digs into mounds were noted, and details taken of their shape and size. A more detailed study of all digs and large burrows both in mounds and away from mounds was undertaken at Mweya and also in the Crater Region. Four north/south transects were run on the upper peninsula at Mweya. The first was across the tip of the peninsula and the fourth near to the Parks employees houses. Four transects were also run in the Craters. The transect distance was between 250 M and 750 M. All digs, including large burrows, were recorded within 10 M. either side of the transect line. Their location was noted with respect to termite mounds, subterranean termite nests, and subterranean ants nests. As will be described in the results of the above work, the nature of the digs found indicated that a medium sized animal with broad claws was regularly digging into termitaria at Mweya and in all the other study areas. Staff at the Uganda Institute of Ecology suggested that this animal was Orycteropus afer. To check this and if possible observe the foraging behaviour of this animal, watches were made on the upper peninsula at 147 Mweya at least once every two weeks from November 1972 to February 1973 then at least twice a week from February 1973 until July 1973. Watches were made by the observer (D.A.M.) standing in the back of an open topped Landrover, using a hand held spotlamp with an independent 12 volt supply. A route was followed which criss-crossed the whole of Mweya upper peninsula, from the Parks houses to its tip. The time a watch commenced was recorded, as were the weather conditions. The area was particularly scanned for orycteropus and Manis gigantea. As is described below only orycteropus was found. When this occurred the following procedure was carried out. On encountering an animal its position on the peninsula was recorded, and a marker dropped from the vehicle enabling the position to be checked the following day. The animal was either followed in the Landrover or on foot with the Landrover some distance behind. Observations were made using a pair of 10x50 binoculars. Over a total of twelve observation periods, records were made of the general foraging behaviour, including notes on the attention paid to termite mounds. The method of digging both small digs and large burrows was seen, the specific digs watched being marked and examined the next day. The number of digs made in a given time and the time spent on each dig was noted during one longer observation. Finally, the reactions of orycteropus to other species was recorded. In order to examine the diet of orycteropus, droppings were collected monthly from Mweya peninsula between January 1973 and July 1973. The 148 areas searched were well known and it was possible to accurately date droppings, sometimes to the actual night when they were deposited. Undated droppings were collected from the Craters and Nyamagasani. Analysis of droppings was carried out in the lab on each group of faeces collected. The droppings are unmistakeable oval cakes previously described by Verheyen (1951) and Kingdon (1971). A subsample of two to four of these individual droppings was taken from each group and soaked in a beaker of water overnight. Apart from softening the droppings, this, removed most soluble coloring from them. The colored water was poured off and replaced with fresh water. The droppings were gently broken up with blunt forceps, so freeing pieces of cuticle which floated to the surface. The water plus cuticle was poured through a filter paper in a filter funnel, more water being added as necessary, until all floating pieces were obtained. The mud remaining in the beaker was examined under a binocular microscope with top illumination for seeds and any other non-floating plant or insect remains. The cuticle pieces caught on the filter paper were washed off into a 15 cm petri dish. This was left in a warm oven to evaporate to dryness. The cuticle parts so prepared were easy to observe and manipulate. They were examined under a binocular microscope for the presence or ab-sence of Formicidae, and for individual termite genera only usually distin-guishable by their soldier caste mandibles. The keys of Bouillon and Mathot (1966) and Ruelle (1970) were used, along with a reference collection of termites identified by W.A. Sands of the British Museum (Nat. Hist.). Remains of animals of other insect groups were looked out for, and unidentified cuticle was described as such. The percentage composition of the droppings was then estimated on this cuticle extract, using the number of mandibles of each group as an index of that group's abundance. It was possible to distinguish the following groups by mandibles: individual general of Isopteran soldiers, Macrotermes workers, non-Macrotermes Isopteran workers, and Fbrmicidae (see appendix 5 for details). 150 2. RESULTS: THE USE OF MOUNDS BY VERTEBRATES i . Mounds as fixed elevated points. The species of vertebrates observed using mounds for various activities are given in Table LII. Further details on animals eating mound vegetation and destroying mounds are given in section IV-2-iii. Only the anteater chat (Myrmecocichla formicivora Vieillot) was observed to feed on invertebrates at the mound surface. This species probably also fed on insects inside M. bellicosus mounds, which it was seen to enter on a number of occasions through holes formed by large mammal rubbing. The fish eagle (Cuncuma vocifer Hodgson) was the only species seen feeding on food brought to a mound, in this case Tilapia nilotica (Peters) caught in Lake George (figure 61). A number of species used mounds as lookout points, the animal seen most frequently being the baboon (figure 62). On three occasions a troop was encountered in the area of 0_. kibarensis mounds near Ishasha camp. The first time, a troop of over 24 animals was seen at 5 p.m. foraging on grasses and Securinega virosa (Baill.) fruits, moving slowly back towards the riverine forest where they spent the night. A reaction to our presence was made at a distance of about 50 metres. No fast fleeing occurred, but most animals moved off toward the forest. One young male stayed behind and climbed onto an non-thicketed O. kibarensis mound to watch us. A number of animals joined him and then left him over the next few minutes, eventually leaving three animals on that mound and two adult males on a smaller mound nearby. A young male left the group of three animals, 151 F i g u r e 6 1 . F i s h e a g l e e a t i n g T i l a p i a on a b a r e M . s u b h y a l i n u s mound n e a r L a k e G e o r g e . 152 F i g u r e 62. Baboon use mounds as l o o k o u t p o s t s and r e s t a r e a s a t I s h a s h a . approached cautiously to within 20 metres of us, then sat on another mound. All animals finally left together to move to the forest. Before the last few entered the trees, an adult male climbed a M_. bellicosus mound and watched us, eating blades of grass from the mound at the same time. On the other two occasions again no fleeing was observed, just a more directed movement towards the forest, with less foraging, when we approached to between 50 and 100 metres; each time large males climbed onto mounds to observe us. On two other occasions a troop, probably the same one, was sighted about one kilometer from the mound area and the forest, in open grassland. Both times the whole troop quickly fled into the mound area near the forest before stopping, the reaction occurring when the landrover was over one kilometer away. Details of the occurrence of baboon droppings in the Ishasha study area are given in Tables LIII and LIV. 16.5% of 0_. kibarensis mounds had baboon droppings on them when sampled in November 1972. Significantly more droppings than expected were found between November 1972 and July 1973 on 0_. kibarensis mounds as compared to flat ground or the small Trinervitermes mounds (x = 9.91, df = 2 p<0.01). Results of droppings analysis are shown in Table LV; results of seed germination tests are given in Table LVI. The most common single item taken were fruits of the shrub Secuinega virosa. These were found only to germinate after passage through the gut of baboon. Adult male Kob (Kobus kob) were observed using mounds as lookout posts on two occasions (figure 63) . The indirect observation of a mound trampled by kob but not dug into with the hooves, occurred in the crater region about 100 metres from where two males were seen fighting during the previous week in March 1973. Topi (Damaliscus korrigum) were observed for several hours on each of three days during the 1973 February rut. At no time did animals climb mounds. Animals were not seen on mounds at any other time of the year either. The closely related hartebeest (Alcelaphus buselaphus Palias> was seen on mounds in the flat Serengeti Plains during a visit in December 1972. Marking of mounds other than by defecation was seen in the banded mongoose. On May 19, 1973 a pack of over 25 animals was observed from 16.00 hr. to 16.20 hr. , around a 1 metre high M_. subhyalinus mound on Mweya peninsula. An adult male climbed onto the mound and examined some fresh banded mongoose droppings already present, believed to have been deposited by members of another pack seen in the area the previous day. This male anally marked the mound be dragging itself over the bare top of the mound with its hind legs splayed apart. During the next 10 minutes six other animals climbed the mound and examined the top of it; two rubbed the top with their chins, and another anally marked the mound. Finally an adult male was left; it defecated and then pulled itself over its faeces using the forelimbs, the hindlimbs spread apart. The whole sequence lasted twenty minutes while the other members of this large pack were foraging nearby (figure 64) . 155 F i g u r e 63. A d u l t male kob u s i n g mound as a l o o k o u t p o s t . F i g u r e 64a. Four banded mongooses on M.subhyalinus mound. 157 Figure 64c. Animal on the l e f t i s c h i n marking the mound. 158 F i g u r e 64d. A d u l t male def e c a t e s on the mound bef o r e the pack moves away. 159 Transect results showed that up to 11.3% of 1VL subhyalinus mounds had small carnivore droppings on them (Table LIII). These were usually of banded mongoose, but civet (Viverra civetta) and white-tailed mongoose (Ichneumia albicauda) were also identified. Many birds were seen resting on mounds, and their common presence is evidenced by the many droppings usually present. Records were not kept of the many species seen. No bird of prey was seen to use a mound as a lookout post, a'partf rom a crowned eagle (Stephanoaetus coronatus Sclater) sometimes seen in the tops of O. kibarensis mound thickets prior to exploratory flights or direct attacks on small vertebrates. Red-necked spurfowl (Pternistis cranchii Wagler) commonly roost on mounds in the early evening in the Mweya area. During a thirty minute drive one evening in April 1973 16 birds were seen. 14 were on mounds and the other two were at the base of mounds upon which another bird was standing. No birds were seen roosting elsewhere, nor were any flushed from the grass away from mounds. Many mounds did not have birds on them at this time. Elephant, buffalo, and warthog were seen to rub Macrotermes mounds on a number of occasions. Full details are given in Table LVII. Usually only bare mounds were rubbed, there being a significant correlation (r = +0.932, t = 5.15, df = 4, p<0.01) between the percentage of bare mounds in a study area and the percentage rubbed by large mammals (figure 65). Warthog commonly rubbed mounds near Mweya, often after wallowing, leaving clayey mud on the mound. Hoof prints and damp mud on mounds suggest that a similar practice is carried out by buffalo particularly at o 30. o o 20. 00 Q z ZD O 10. LU CO CQ ZD cn r = 0.93, t=5.15 df=4 , P< 0.01 ' 10 4 0 50 20 30 'BARE MOUNDS (%) Figure 65. Graph of percentage bare Macrotermes i n each study area against percentage of mounds rubbed by large mammals. 161 Ishasha (figure 20). Defecation sometimes accompanied rubbing by both elephant and buffalo. Some mounds were extremely rubbed, suggesting repeated usage, whereas other bare mounds of similar size were not rubbed at all. Samples of ticks taken at the bases of three buffalo rubbed mounds were all identified as Rhipicephalus simus (Koch). i i . The occupation of mound interiors by vertebrates. The species involved. The results of mound and subterranean nest excavations are presented in Tables LVIII to LXII, and figures 66 to 69. All dead Macrotermes mounds with entry points contained evidence of vertebrate presence. M. subhyalinus mounds were a breeding site as well as a refuge site for the murid rodent A_. niloticus and geckoes (Hemidactylus sp_.). Banded mongoose also bred in mound dens, as judged from pack observations. As can be seen from the diagrams of excavated mounds, most rodents were using the smaller galleries inside mounds rather than any large burrows. When a Mus triton was disturbed in a large burrow it ran into the honeycomb of galleries above. No evidence of rodents was found in the large burrow under mound NK3 at Nyamagasani. Termites were still active in this mound, and had sealed the rest of the mound off from the burrow, filling in all small galleries. A large tick, Ornithodoros moubata porcinus (Agassiz) , was found in the burrow, suggesting use by a medium sized mammal. Because of the complexity of the internal mound space it was found that a number of species could be present in one mound at the same time. For example in mound D3 at Mweya six species were present at the time of digging; 2 species of murid rodent, banded mongoose, a bat, a frog, and a snake. Another species, (X afer, had recently used the large burrows under the mound as well. M. subhyalinus mound ,2F was particularly interesting. This dead mound had been invaded by Odontotermes termites, building a nest with three small ( 5 cm) ventillation shafts. An Arvicanthis niloticus nest was found at the base of these shafts where they joined, positioned with a vertical termite gallery immediately below it. The termites had walled off their nest from the rodent nest. Observations made during ground transects identified a few additional species using M_. subhyalinus mound interiors. The use of large burrows in mounds, and elsewhere, is dealt with below. A pair of adult blue bodied agama lizards (Agama atriocollis Blake) were seen to enter small, approximately 5 cm diameter, holes produced by weathering in the top of a 1 M high mound. Civet (Viverra civetta Schreber) droppings were found inside a hollow dead mound with a 20 cm diameter entry hole; probably originally dug by banded mongoose as judged by many old droppings on the mound. On two occasions a side-striped skink (Mabuya striata Rafinesque) entered 20 cm diameter holes in the sides of dead mounds in a low lying area of Mweya peninsula liable to seasonal flooding. In both cases these holes had formed by weathering around stems of Capparis tormentosa shrub growing out of the mound. Finally, snake eggs were twice found associated with termite mounds. On the first occasion a group of six eggs had been laid a few 163 F i g u r e 66. Excavated M.subhyalinus mounds. Numbers r e f e r t o s i t e s of v e r t e b r a t e s i g n , as giv e n i n t a b l e L V I I I . a)Mound ND; b) mound D 12. 164 Figure 6 6 . cont. c) Mound D 3 . 165 centimeters below the surface of a tall bare mound in the Imperata grassland of the crater region. The second time, three eggs were collected from just below the surface of a pile of loose soil which was the result of an orycteropus mound dig at Nyamagasani. No eggs of either set were successfully incubated in the lab and so the species involved are unknown. Rodents built nests in live as well as dead M_. bellicosus mounds. The unidentified carnivore droppings were probably of banded mongoose as judged by brown hairs found both in the droppings and inside the mound. Other small carnivores seen in the area were the white-tailed mongoose (Ichneumia albicauda Cuvier) , the marsh mongoose (A til ax paludinosus Cuvier) , and the common genet (Genetta genetta L) . Mounds 19(4) and 19(7) clearly showed how mounds could be the centres of activity for a small mammal. These two mounds were examined soon after a burn (figure 67). They were approximately 12 meters apart and connected by well worn runways which radiated out into the surrounding grassland and then faded. When 19(4) was examined one of the pair of A. niloticus found escaped and ran to 19 (7) , following the runway, and disappeared down hole 7. A number of ticks of the species Rhipcephalus simus were collected from the A_. niloticus nest in mound 19(4) and in the large nest in mound II. O. fulleri shafts were used for breeding by A^ niloticus and Lemniscomys  striatus, and as a refuge by an unidentified rodent. In all cases the nests were positioned with a termite shaft below them, as described above for M. subhyalinus mound 2F. In no nest had enlargement of the termite F i g u r e 67. Excavated M . b e l l i c o s u s mounds, a) Mound I I . Numbers r e f e r to s i t e s of v e r t e b r a t e s i g n , as - given i n t a b l e LIX. live Pseudocanthotermes nest (c) F i g u r e 67. cont. b) Mound 14; c) mound 110. Figure 67. cont. d) Mounds 19(4),=MI, and 19 (7),=M3. Plan view plus s e c t i o n s . F i g u r e 68. Excavated Odontotermes n e s t s , a) nest C7-9; b) nest CIO-9. Numbers r e f e r to s i t e s of v e r t e b r a t e s i g n , as gi v e n i n t a b l e LXI. F i g u r e 68. cont. c) Nest CIO-4; d) nest C9-6. 171 hoiel (plan) i i 25cm i 1 Figure 69. Excavated 0.kiba r e n s i s mound IS-20. See t a b l e L X I I . 172 galleries taken place. Worn pathways similar to the one over the turret of O. fulleri nest C9-6 were seen occasionally during small mammal trapping in the Crater Region, usually accompanied by a worn runway with herbi-vorous rodent droppings. This appears to be typical of prolonged occupation of a shaft by _A_. niloticus. The results of O. kibarensis mound excavations are shown in Table LXII. niloticus was again found to be the main user of termite shafts. Droppings were common on runways, especially near a hole entrance. The often crumbling mound soil showed signs of additional rodent burrowing, in contrast to the situation in the solid structured Macrotermes mounds. Both live and dead mounds were occupied by rodents, but a rodent nest was positioned away from that part of a mound inhabited by termites. The use of mounds as den sites by a small carnivore. The details of twenty-three banded mongoose dens located near Mweya are given in Table LXIII. Fifteen were located in dead M_. subhyalinus mounds and one in an Odontotermes shaft under a thicket. Six of these fifteen were actually in large orycteropus burrows underneath mounds. Of the remaining six not connected with termitaria, four were on steep slopes in erosion gulleys, one was under a dense thicket with no burrows found, and one was under a man-made concrete slab. Small carnivore burrowing was seen in twelve of the dens, but only in one away from mounds and three in mounds was this the only agency for entrance hole formation. The mean number of entrances was 3.8, with a range of one to seven. The mean size of entrance was 22 cm in diameter, with a range of < 10 cm to 90 cm. 173 The use of large burrows in mounds and elsewhere by medium-sized mammals The results of monitoring 14 burrows in Mweya peninsula over an eight month period are shown in Tables LXIV and LXV. There was no tendency for any species to use mound burrows in preference to other sites, either over the period as a whole or during any one month. Mound holes were used 7.8% of the time, with a range of 0% to 18.8% for individual months. The same figures for non-mound holes were a mean of 6.1% and a range of 0% to 12.5%. Warthog was the most common user of burrows. Also identified were banded mongoose, hyaena, and orycteropus. Solitary small carnivores entering burrows were not distinguished from each other; regularly seen at night in the area were civet, white-tailed mongoose, and genet. A genet at Ishasha was seen to run into a large burrow when alarmed. The only other species not mentioned above which was also seen in a large burrow was the gaboon viper (Bitis gabonica Gray). The four burrows excavated in addition to those under mound D3 are drawn in figure 70. A l l the tunnels had tapering ends, and in burrow 10 broad scratch marks, similar to those of orycteropus, were present at the end. Termitaria and small mammal distribution. Crater Region Table LXV gives all snap-trapping data for the Crater Region, including the number of Odontotermes shafts on each grid. The correlation coefficients for the number of rodents caught and the density of shafts are 174 F i g u r e 70. Four l a r g e burrows excavated a t Mweya. a) Burrow I. 175 ,Q5M (c) Figure 70. cont. b) Burrow 10. c) Burrow H. .Q5M . gure 70. cont. d) Burrow T. shown in Table LXVI and figure 71. No one species, or group of species, showed a density which was significantly correlated with the density of shafts. Nor was any species absent from an area lacking shafts. Details of the weights and length measurements of murid rodents caught are given in figures 72 to 78. The monthly capture results are given graphically in figures 79; and 80, distinguishing between adults and juveniles. Juveniles were classified as those animals with a head plus body length less than 2/3 the maximum recorded for the species. To examine whether species were trapped more frequently near shafts than expected from a random distribution of captures on the grids, the results were analysed as follows. For each of the twelve grids where Odontotermes shafts occurred, the forty trap points were ranked with respect to their distance from the nearest shaft. If the rodents were distributed randomly over a grid then a cumulative frequency plot of the number of captures summed for the ranked traps, would not be significantly different from a straight line starting at 1/Nxl00% cumulative captures (where N = number of traps) for the trap nearest a shaft, and ending with N/NxlOO, i.e., 100% cumulative captures, with the last trap furthest from a shaft. If animals were being caught nearer shafts than expected the % cumulative captures/ ranked trap plot would lie always above the expected figure. If animals were trapped less often than expected near shafts, then the plot would lie below the expected one. This analysis was made for each grid with shafts and also for combined data, ranking all 480 trap points. On each plot the significance of the difference between observed and expected results was 178 •DDNTL7TERK€5 SHAFTS/SMALL MAMvlALS 30. 20-1S.I 10-5-1 ^ — 1 • 1 1 1 1 1 1 1 I 0- ID- 30 • 30- 40- SO- GO. 70- 80- 90- 100. ODDNlUILMvCB SHAFTS Figure 71. Graph of t o t a l small mammals captured i n Crater Region grids against number of Odontotermes shafts per g r i d . 100, x male o female X 0 O) r -X CD LU 50. O X X o o X KK X X 0 OK x * o x© o 40 60 80 100 120 HEAD PLUS BODY LENGTH (mm) 140 F i g u r e 72. Graph of weight a g a i n s t length f o r M . n a t a l e n s i s . o 00 r-t 120. U) IE CD LU 80. 40. 40 x male o female X * 0 K 60 80 100 o o0> X o 120 140 HEAD PLUS BODY LENGTH (mm) F i g u r e 73. Graph of weight a g a i n s t length f o r A . n i l o t i c u s . 00 80 60 U) H 40 I E CD U 20J 40 x male o female X X 80 120 HEAD PLUS BODY LENGTH (mm) F i g u r e 74. Graph of weight a g a i n s t length f o r L . s t r i a t u s . CN CO 150, 100 O) f-1 o LU 50. x male o female 80 120 N 160 HEAD PLUS BODY LENGTH (mm) F i g u r e 75. Graph of weight a g a i n s t l e n g t h f o r L . s i k a p u s i . 183 20 10 4 0 O) 6 0 .40 20 x O L±J ^ 12 8 4 6 0 X o 0 X oX Figure 76, M . t r i t o n 60 8 0 100 X 0 8 0 100 Figure 77, Z .hildegardeae 120 Figure 78, M.minutoides 30 50 70 9 0 HEAD PLUS BODY LENGTH (mm) Figure 76,77,&78. Graphs of weight against length f o r murid rodent s p e c i e s . 184 2 0 10 CO I— X O < cr O C\] cr LU QL Q LU cr ZD t < 1 2 CO 8 u Q O o r 4 A Figure 79, 'Mound' g r i d s . A - adult J - juvenile Figure 80, 'Non-mound'grids NOV D E C JAN F E B M A R APR MAY J U N J U L Figure 79 an:d 80. Results of snap t r a p p i n g i n the C r a t e r Region. 185 found by using the Kolmogorov/Smirnov one sample test described by Siegel (1956). For the individual monthly results, the only significant differences (p<0.05) found were for the two diurnal species Lemniscomys striatus and Arvicanthis niloticus, being caught nearer shafts than expected during the month of March (Table LXVII). When considered overall no significant differences occurred, either for individual species, or when nocturnal and diurnal species were considered separately. The only species which was always caught nearer shafts than expected was_L. striatus (figures 81 to 85). One similarity is seen in all these cumulative percentage capture plots. This is that more rodents than expected were caught in traps where the trap/shaft distance was less than the trap/shaft distance for the trap point which coincided with the maximum positive difference in cumulative percentage of captures (Table LXVIII). Considering all species, significantly more animals than expected (x 2 = 9.31, df = 1, p< 0.01) were trapped < 1.5 meter from shafts, and apart from L_. striatus, animals appear to be distri-buted randomly with respect to shafts at greater distances. Snap-trapping after the Crater Region burn resulted in captures only around bushes, no animals were caught in the open, not even around Odonototermes shafts, (Table LXIX) . The live-trapping results after the burn are shown in Table LXX. The largest difference in total captures between the shaft and non-shaft grid occurred in the first trapping period, two weeks after the burn. The small number of animals on the non-shaft grid was caused by an absence of L_. striatus. Direct observation of released CO • v O observed 100 200 300 400 RANKED TRAP POINTS F i g u r e 81. Graph of percentage cumulative captures ( a l l s p e c i e s ) a g a i n s t t r a p p o i n t s ranked with r e s p e c t to distance.to the ne a r e s t Odontotermes s h a f t . CO z o §1.0-OC L L _ observed __ expected 4- max. deviation 100 200 300 RANKED TRAP POINTS 400 F i g u r e 82. Graph of percentage cumulative captures ( M.n a t a l e n s i s ) a g a i n s t t r a p p o i n t s ranked with r e s p e c t to d i s t a n c e . t o the ne a r e s t Odontotermes s h a f t . CO CO rH _ observed expected I max. deviation 100 200 RANKED TRAP POINTS 300 400 F i g u r e 83. Graph of percentage cumulative captures ( L. s t r i a t u s ) a g a i n s t t r a p p o i n t s ranked with respect to d i s t a n c e . t o the nearest Odontotermes s h a f t . 00 F 1.0 i t .8. oo L±J cr 3 .6. Q_ < u .4. LU > ZD Z> U observed — expected 1 max. deviation 100 200 300 RANKED TRAP POINTS 400 F i g u r e 84. Graph of percentage cumulative captures ( n o c t u r n a l s p e c i e s ) a g a i n s t t r a p p o i n t s ranked with r e s p e c t to d i s t a n c e t o the n e a r e s t ©dontotermes s h a f t . 190 CUMULATIVE CAPTURES (FRACTION) p-t-i CD OO U l 3 0) cn CD l-i OJ 0) &) ^ H- *v CD t? cn cn rt rt o hh o rt & i-< T3 o 0) CD 1-1 rt O O CD rt O CD rt H P) li- iQ CD en CD (/) i-i O W 0) C tr B 0) Hi CD t—• r+ rt s: < rt CDtr O n CD cn rt •a £ CD H o CD rt cn rt O Cb H- c cn rt 0) 0) H " O CD cn T3 rt CD O n rt CD cn CD 191 animals confirmed that both species captured at this time, _L_. striatus and M. natalensis, were regularly using shafts as refuge sites. By April the grass was once again too high for observations to be made. Of eleven releases on the shaft grid which were successfully followed for more than ten meters, ten went down shafts. Of five such releases on the non-shaft grid, four went down shafts, each > 30 meters beyond the edge of the grid. One Mastomys juvenile male was seen to use the same hole on consecutive days, and one Mastomys adult male disappeared down the same hole on four occasions over a two month period. Different individuals of both species were found to sometimes use the same hole on consecutive days. The results for stomach analyses for rodents caught in the Craters are given in Table LXXI. Odontotermes termites were particularly taken by Lophuromys sikapusi, and were also the dominant item in the single Zelotomys stomach examined. Overall Pseudocanthotermes termites were the most often selected termite species. Ishasha The Ishasha snap-trapping results are given in figure 86. The main species caught was _A_. niloticus; also captured were M_. natalensis, M. minutoides , and Crocidura sp.. Considering just A_. niloticus, only one animal was caught in a non-mound grid, that being an adult female at the beginning of the first 1973 dry season. Total captures as well as A^ niloticus captures, show definite peaks at the ends of the two wet seasons. The capture of the only three juveniles (<30 g.) took place in November, May, and June, suggesting breeding at the end of the second 1972 rainy season, and at the 192 1 1 fi r - ^ r ^ * - r-^= r - i - = , a -^ , S O N D J F M A M J J A 1972 1973 Figure 86. Results of monthly snap-trapping at Ishasha, on mound and c o n t r o l g r i d s . 193 end of the first 1973 rainy season. The two peaks in number observed may also be at least partially ex-plained by emigration and then immigration of A. niloticus in the snap-trapping area. This is suggested by the live trapping results shown in figure 87. Although A_. niloticus droppings were present on mound grid LI in January 1973, no animals of this species were caught until June 1973. Of the eleven captures that month only three were juveniles suggesting a large movement of animals. Captures were made in the non-mound grid at this time as well. None of the animals marked in June were recaptured in July, suggesting a movement through this area between the forest and the seasonal floodland snap-trapping area in the preceding month. The live trapping also showed that M_. natalensis makes use of mound holes as regular home sites. An adult pair was trapped in both March and April on mound LI; on each release they went down mound holes. Finally in this section, the results of rodent dropping surveys are given in Tables LXXII and LXXIII. Droppings were virtually confined to mounds. Significantly more droppings (x^ = 11.3, df = 1, p<0.001) were found on holed mounds than non-holed ones, suggesting that the mound in-terior is the important feature rather than the usually present thicket vegetation. M. minutoides x mound o control 0 X o x ® ^^L- ,—sS r-a , 194 M. natalensis _0 ,—0 , — Q , 2 . — Q _ A. niloticus vSi , & , Q , — & , , — 2 , J F M A M J J A F i g u r e 87. R e s u l t s o f l i v e - t r a p p i n g on a mound g r i d and a c o n t r o l g r i d a t I s h a s h a . 195 i i i . The use of mound soil and vegetation  Soil No animals were observed eating soil in R.N.P. apart from a male warthog on Mweya peninsula. This animal was seen digging into a low M. subhyalinus mound, apparently taking in soil. Two IVL subhyalinus mounds in the Crater Region had been trampled by buffalo, and another showed evidence of extensive horning by buffalo. In all cases the dead mineral-rich inner galleries had been exposed. Six bare or sparsely grassed M. subhyalinus mounds in two local areas near the Kazinga channel, were found completely demolished by hippopotamus. One was a live mound, with dead termites exposed in the still damp inner galleries. Analysis of soil from one demolished mound at Mweya gave similar figures to other mound surface samples taken at Mweya, with increases over the surrounding surface soil in calcium, sodium, magnesium, and nitrate nitrogen (Table XXXI). At Ishasha the interiors of two I V L bellicosus mound shells had been used as wallows by warthog. In addition a low excavation had been made near the base of one by an unidentified animal (figure 88). The soil from this scrape showed a similar mineral content to nearby control samples, but was of finer texture (Table XXXI). Details of the distribution of large wallows in the CL kibarensis mound area near Ishasha camp have been given in section III— 3—ii. Buffalo were the only species seen wallowing in these. Wallow mud was not analysed but samples from both the mounds and control areas contained larger amounts of all minerals tested for when compared to other Ishasha F i g u r e 88. M . b e l l i c o s u s mound s h e l l and adjacent s o i l ( foreground ) used as a wallow by a medium s i z e d mammal. surface soil, especially sodium, (Table XXXI). Vegetation On a number of occasions buffalo were seen in apparently awkward positions eating thicket and other vegetation from mounds in the Mweya area; the front legs high up on the side of a mound. In the rainy season deep hippopotamus prints were seen on a number of mounds covered with Cynodon dactyIon. This grass is the preferred grazing species of hippo-potamus , (Field 1970) . On April 14, 1973, two male warthog were seen eating together on a grass covered M. subhyalinus mound on Mweya peninsula. After three minutes they moved off. Without grazing in between they walked straight to another grass covered mound approximately 150 meters away. They ate here for four minutes before moving away. The dominant grass on each mound was Cynodon dacty Ion. The results of the Ishasha study on comparative blade length of Sporobolus pyramidalis on mounds and away from mounds, are given in Table LXXIV. For each sampling period the mean blade length was signi-ficantly less (p<0.02) on mounds than off. A l l leaves measured showed signs of grazing, suggesting increased pressure on mounds, iv. The use of termitaria as a concentrated source of food A l l the diggings into live termite mounds were classified as having been made by one species. It is described below how direct observation confirmed that the animal responsible was orycteropus. It has been shown above that small carnivores such as banded mongoose, can dig into dead 198 mounds to make den entrances. It has also been shown that these carnivores do eat both Macrotermes and non-Macrotermes termites. No evidence was found, however, during any of the digging surveys or ground transects, to suggest that these animals dug into live mounds or subterranean nests for food. Evidence was not found for the presence of the Giant Pangolin. (Manis gigantea) in R.N.P. The digs of this species resemble those of orycteropus, but are clearly distinguishable because the claw marks of the pangolin are much narrower (Kingdon per. comm.) This pangolin species is recorded as being present in R.N.P. and other areas of Uganda (Kingdon 1971). Orycteropus afer Three basic types of dig were distinguished for orycteropus, all being attributable to this species because of the unique broad, almost spade like, marks left by the claws. The first type was the "food dig" (figure 89) . This was v-shaped when viewed from above, with claw marks at the narrow end. It was usually 20-30 cm wide, 30-40 cm long, and under 40 cm deep. This type was found singly or in groups, entering termite mounds or away from mounds. The second type of dig was the "defecation scrape" (figure 90) . This was formed by an animal first digging a small "food dig" only 5-10 cm deep, then defecating in it, and covering it with soil from an area approxi-mately 40 cm x 40 cm square. Only on two occasions were droppings collected away from such a scrape. Both of these were in the Crater Region where the droppings were deposited in the spoil from a freshly dug large burrow. 199 F i g u r e 89. M.subhyalinus mound showing orycteropus food d i g s i n s i d e . 200 F i g u r e 90. a) Orycteropus d e f e c a t i o n scrape i n the f i e l d , b)Scrape w i t h s o i l to show the many o v a l droppings. as found removed The third type of dig was the large burrow . Diagrams of the five burrows which were excavated have already been given (figures 66 and 70). Of eighteen burrows examined or excavated on Mweya peninsula, 13 had a single entrance, 2 had two entrances, 2 had three entrances, and 1 had five entrances. It is not certain that orycteropus was the only animal to dig all of these burrows. During the study period only orycteropus was observed to dig and extend burrows. Tracks left in fresh spoil indicated that warthog rebored a number of burrows, but did not start any or extend any. All burrows dug showed tapering ends to the tunnels, suggesting that the animal digging used its forefeet; warthog use their tusks and flat snout only (Bradley 1971). Burrow number 10 showed marks at the end similar to those left by the claws of orycteropus in food digs. Different marks were seen on the side walls of this burrow and a number of other burrows. These were 2 cm deep parallel grooves, 2 cm apart, running at an angle from the floor of the burrow to the roof, and directed upward and inward. These were in the opposite direction to orycteropus strokes seen in the larger food digs, and were not arranged in groups of two or three as seen with orycteropus. It seems likely that these were made by warthog raking the tunnel wall with their tusks. The large burrows used by this species ranged from a single short tunnel to a branching network of tunnels. This may be related to the function of the burrow as either a nocturnal refuge site for orycteropus or a day-time den. Two of the five burrows excavated had been widened at the ends of tunnels, enabling an orycteropus 202 or warthog to turn around. Two burrows, "T-burrow" and "Den 3", had additional widened areas along their length. Den 3 has already been des-cribed as showing evidence of orycteropus occupation, possibly as a day-time den. It consisted of 13 meters of tunnels under a large M. subhyalinus mound, with three entrances and two large tunnel cave-ins. The tunnel walls were very smooth suggesting frequent use by a medium-sized mammal. No warthog tracks were found, only the mark left by an orycteropus tail. The other burrows were generally less than two meters long and un-branched. Of the three single entrance tunnels excavated out, only one, "T-burrow", split into two tunnels. The largest group of possibly connecting tunnels was found at Ishasha, on high ground to the south-east of the low seasonal flood plain, adjacent to transect IJ (figure 91). Eight tunnels were present within an approximately 400 M area. It is believed that these were originally dug by orycteropus, judging from the presence of food digs in the area. At the time of observation the only occupants were hyaena, genet, lizards and a bat. The distribution of orycteropus digs The results covering the distribution of digs associated with mounds, as found from the ground transects, are given in Table LXXV. In the north of the Park digs were found in all study areas. It was found that the density of dug mounds per hectare was significantly correlated (r=+0.51, t=3.52, df=35, P<0.01) to the density of Macrotermes mounds, with the highest number of dug mounds and the highest percentage of dug mounds, occurring at Nyamagasani. w entrance (not to scale) i I i i \ i \ i 5M F i g u r e 91. R e l a t i v e p o s i t i o n of e i g h t l a r g e burrows near t r a n s e c t 13 at Ishasha. 204 In the Crater Region orycteropus dug mounds or dug Odontotermes nests, were only present in the outer area. No dug mounds were found in the central crater region, or on the higher rocky crater rims in the outer area. Both 0_. fulleri shafted nests and Macrotermes mounds were dug, but only large burrows were made into the Odontotermes nests, completely destroying them. No typical food digs were made into these subterranean nests. At Ishasha, mounds with orycteropus food digs were only found on transect IE, and mounds with large burrows were only found on transects IE and IL. Both of these transects were in flat bushed areas, not liable to seasonal flooding. On transect IE 12.5% of mounds had food digs, which means that 2.5 dug mounds per hectare were locally present. Away from transects dug and tunnelled mounds were only found additionally near transect IJ, on high ground south-east of the Ishasha plains. No food digs or large burrows were present in 0_. kibarensis mounds in seasonal floodland at Ishasha. The results of the orycteropus dig transects are shown in Table LXXVI for Mweya and the Craters. At Mweya a trend is seen for reduced orycteropus activity towards the houses on Mweya peninsula. This trend occurs in both food dig density and defecation scrape density, but not in large burrow density. The mean number of food digs per group was 4.6, with a range of 1-43. Overall 0.2 dig groups per hectare were into subterranean ants nests; 2.0 dig groups per hectare were into subterranean termite nests (only 205 Macrotermes and Odontotermes dug nests were identified); 2.6 dig groups per hectare were into or around M. subhyalinus mounds; and finally 5.8 dig groups per hectare were away from mounds but not connected with an obvious subterranean nest. This means that at Mweya 20.9% of dig groups were into M_. subhyalinus mounds. In addition 37.5% of large burrows were under Macrotermes mounds, presumably originally dug for food. In the outer Crater Region, the mean number of food digs per group was 3.2, with a range of 1 to 9. 2.1 groups per hectare were into under-ground nests, excluding Odontotermes shafted nests; 1.3 groups per hectare were under Macrotermes mounds; 2.3 groups per hectare were away from Macrotermes mounds but not into an obvious subterranean nest (on one occasion a Pseudocanthotermes mound had been dug into); no groups of digs were found into shafted Odontotermes nests. This means that 23.7% of groups of digs were into Macrotermes mounds in the Craters. There were over twice as many large burrows into Odontotermes nests as into Macrotermes mounds, giving 1.4 burrowed nests per hectare, and 0.5 burrowed mounds. It is therefore seen that orycteropus usually made a number of food digs into Macrotermes mounds rather than destroying the mound with a large burrow into the centre. The latter method was the sole one used for Odontotermes subterranean nests, however. On a number of occasions it was observed that a mound was visited on consecutive nights. In the periods between visits the holes were built up by termites with moist earth; this meant that an area was created affording easier access into the mound on the following night. Orycteropus foraging behaviour and use of large burrows from direct  observation A summary of the results from 12 watches when an animal was found, is given in Table LXXVII. Only one animal was ever seen at any one time, giving a total of 4 hr. 20.5 minutes observation time. The longest continuous watch was 35 minutes long, and the longest total observation during one evening, was 58 minutes. No animal was encountered before 22.35 hr. , even though watches were often made from 21.00 hr. onwards. Individuals were observed on bright moon-lit nights as well as ones without a moon, but no animal was seen in the rain. The general method of foraging was as follows. The animal walked slowly or quite briskly, with its nose near to the ground. It followed a zig-zag path which resulted in a strip of ground about 30 meters wide being inspected. The ears usually pointed forward and slightly to one side. Frequent short stops were made at the end of which the animal would either move on or commence to dig rapidly with its forefeet. On the watch of 23-5-73 an animal was observed foraging at running pace. It made sharp zig-zag turns, the head raised. Having thus sharply turned it proceeded directly to a spot some meters away and commenced digging. Since the head was raised it seems likely that hearing as well as smell played a part in termite location. No particular attention was paid to termite mounds, and only on one occasion (22-7-73) was an animal seen to dig into the base of a mound. When digging orycteropus rested back on its haunches and tail, 207 leaving the forelimbs free for digging (figure 92). An animal was seen to dig a burrow on 24-2-73. This individual had entered a short burrow and proceeded to dig itself in a further meter in five minutes. Soil was pushed back with the hind feet and also the tail, which contracted to make horizontal waves, and so force the soil out. The time pattern of digging whilst foraging is illustrated by the notes taken during the watch of 19-7-73 (Table LXXVIII) . Three distinct types of digging stop were distinguished. If the animal stopped and dug for less than 10 seconds and then moved on, the dig was termed unsuccessful. Visits to a freshly dug area during the day showed numerous shallow 5 cm scrapes, believed to be the result of these presumably exploratory digs. The second type of dig was the food dig described above. This dig was usually completed and left in under 30 seconds, although a stay of 2 minutes was recorded. Thirdly an animal was seen to make a group of digs in one local area. We have seen from the digging transects that up to 43 food digs were found in these groups. On 19-7-73, 2 unsuccessful dig stops, 11 single dig stops, and 3 group dig stops, were made in a 35 minute period. A visit to the area the next day showed numerous small scrapes to be present, presumably unsuccessful digs not recorded because they were made too quickly, the animal hardly stopping. Nightly distance travelled by orycteropus Details of the distance travelled during all observation periods are given in Table LXXVII. The hourly distance travelled was calculated from these data, and ranged from 406 M to 9,600 M. To obtain the best estimate oo o F i g u r e 92. O r y c t e r o p u s a f e r d i g g i n g n e a r a M . s u b h y a l i n u s mound on Mweya P e n i n s u l a . 209 of the nightly distance travelled, a ten hour foraging period was assumed, and data only used from the five longer watches (> 30 minutes) , when the animal was believed to have been foraging undisturbed. This gave a mean distance of 9.59 km travelled each night, with a range from 4.12 km to 14.74 km. Reaction to predators Most watches were terminated by the animal reacting to our presence and moving to a large burrow. The burrow was immediately entered if the animal had been pursued closely. On 24-2-73 an individual was pursued closely during the final part of the watch as it bounded away at an estimated 10-15 m.p.h. It soon stopped at the entrance to a large burrow in a termite mound, briefly examined i t , and then entered. Soil was then pushed back for a few seconds. The animal then reappeared, hindquarters first, looked at us, and then raced off again. After a further 100 meters it somer-saulted over the top of an open burrow entrance which was facing away from its line of flight, landed on its back with a loud thud. The burrow was immediately entered, and the animal dug itself in as described above. The exit from a burrow was observed on the 28-6-73. After 23 minutes of observation the animal had run down a large burrow at the top of the kyambura slopes (figure 93). After 30 minutes it reappeared head-first at the entrance, where it sat for one minute, completely still except for movement of the ears. It then suddenly ran off, only to abruptly stop 20 meters from the hole. For 15 seconds it again waited, completely still except for the ears. Finally it moved away in a normal foraging manner. o H F i g u r e 93. Map showing the routes followed by 0..afer d u r i n g 12 s u c c e s s f u l watches on Mweya p e n i n s u l a . 211 The diet of orycteropus The results of droppings analyses are given in Table LXXIX. Macrotermes termites, Odontotermes termites, and Formicidae, formed the bulk of the diet in all three areas sampled. At Mweya Macrotermes soldiers were present in 12 of the 15 groups of droppings sampled, Odontotermes soldiers were present in 8, and Formicidae were present in 11. Overall, on a percentage composition basis, at Mweya, Macrotermes workers formed 27.6% of the diet, Formicidae formed 57.5%, and other Isopteran workers 12.5%. In the Craters Odontotermes termites were taken more frequently than at Mweya. Overall, non-Macrotermes termite workers (probably mainly Odontotermes as judged from soldier mandibles) made up 39.3% of the diet, formicidae made up 36.0%, and Macrotermes workers 19.0%. At Nyamagasani the few data taken showed that Macrotermes workers were the main item eaten, making 73.4% of the diet. The only other species of termite found in orycteropus droppings were Trinervitermes sp. on four occasions, Pseudocanthotermes sp. on two, and Basidentitermes on one. At Mweya a trend was found for less ants to be eaten in the wet months, when Macrotermes workers formed most of the diet. In figure 94 the per-centage of ants in droppings is plotted with respect to the date of deposition; the monthly rainfall is also given. In figure 95 an index of rainfall is plotted against the percentage of ants in the eight groups of droppings where the date dropped was precisely known. The index of rainfall is the number of 1 mm. graph paper squares under the rainfall histogram in the two weeks t-i fl> o\° ANTS IN DIET o —1 1 u_ RAINFALL (mm) 6 4^  o O O O o co O > tt > m rH CN 200 x a z 100. z < cr 0 r= - .805 , t= 5.076 P< 0.001 100 50 % ANTS IN DIET F i g u r e 95. A graph of r a i n f a l l index ( see t e x t ) a g a i n s t percentage of ants i n O.afer d i e t . preceding the date the faeces were dropped. The percentage of ants taken is significantly inversely related to this rainfall index (r=-0.805, t=5.08, df=14, p< 0.001.). 3. DISCUSSION: THE USE OF MOUNDS BY VERTEBRATES i . Mounds as fixed elevated points. A number of vertebrate species within R.N.P. use mounds for resting, as lookout posts, as a place to eat prey brought from elsewhere, and for certain other activities where the importance of the mound is that it is a raised platform. Hediger et al (1948) recorded that reptiles and birds used mounds as fixed points to return to for rest, and as lookout posts. He also noted that baboon and topi used mounds for observation posts, but that waterbuck (Kobus defassa) never did. Weir (1974) reported that topi used mounds as lookout points in the Ishasha area of R.N.P. I did not see this. Hartebeest were seen on mounds in the Serengeti Plains of Tanzania and in the open grassland of Kabalega National Park, Uganda. No medium-sized mammals used mounds for lookout sites in bushed grassland. Mounds can therefore be said to increase the 'comfort level' of the environment for prey species by providing elevated platforms from which animals can scan the surrounding area. Mounds appear to mainly serve this function in open grassland, where a small increase in height results in a greatly extended range of vision. No predator, large or small, was seen to climb a mound in order to visually search the surrounding area for prey. The decreased fleeing distance of Ishasha baboons in a mound area as opposed to open grassland was an example of this increase in 'comfort level', although the close proximity of the forest to the mounds would also have contributed (Rowell 1966) . A direct example of the importance of mounds as elevated points was given by Vos (1969) who recorded that lechwe (Ontotragus leche smithemani) used them as islands in their otherwise flooded habitat; as did predators that preyed on lechwe. Other examples of the importance of mounds in floodland are discussed below in the section on mound interiors. Baboons were shown to defecate more frequently on mounds than else-where in the Ishasha camp area. In other parts of the Ishasha region baboon droppings were noted on the raised soil pushed up by a large track levelling vehicle, indicating that the important feature of mounds was that they were raised areas; often the only raised areas in otherwise flat grassland. Hediger (1948) reported that large male baboons often defecated on termite mounds. Jackson and Gartlan (1965) noted that in a long grass area vervet monkeys (Cercopithecus aethiops) would usually climb onto large boulders or termite mounds to defecate and urinate. These observations again suggest that mounds increase the 'comfort level' of the environment for these primate species by allowing good all-around vision. The only time a number of baboons were observed defecating away from a mound was on a large bare area of soil near the Ishasha river: a position also allowing unobstructed vision. Vervet monkeys were only seen to defecate when in trees and large bushes, which comprised a large amount of their habitat in the north of R.N.P. Such preferential defecation by primates can have effects on mound vegetation. Burtt (1942) and Jackson and Gartlan (1965) suggested that if preferential defecation occurs it could significantly affect the habitat by aiding in forest regeneration. Jackson and Gartlan (1965) found that the base of rocks and termite mounds were the most frequent sites for thicket re colonization. The viable seeds of many species of tree and shrub occur in primate droppings. These seeds may show a higher percentage germin-ation after passing through the gut of an animal; a fact noted before in other species (Lamprey 1963). A number of ways have already been discussed showing how mounds can offer a superior location for thicket growth over flat grassland (section III-4-ii). These included protection from flooding and possibly fire, and increased minerals in mound soil. A concentration of droppings on mounds would further increase minerals on and around them, so adding to their suitability for thicket growth. The faeces of large mammals such as elephant, buffalo, and hippopotamus were often seen on the tops and sides of termite mounds wherever these species occurred, but no quantitative measurements were taken. If preferential grazing of mound vegetation occurred (see section IV-3-iii) then an increase in faecal deposits by grazing and browsing mammals may be expected, further enriching the local mound area. Many mounds in R.N.P. had small carnivore droppings conspicuously deposited on them. The droppings were of solitary species as well as the gregarious banded mongoose. It is likely that all these deposits are important in marking behaviour, as was shown for the banded mongoose. Rood (pers. comm.) stated that banded mongoose will mark any raised object; rocks, termite mounds, tin cans from a garbage dump. He has also shown that there was considerable overlap in home ranges of packs in the Mweya area, and that they are not strictly territorial. The larger pack always won out 218 when two packs came into conflict, no matter where the meeting took place. Display on mounds was observed in two bird species. One, the red-necked spurfowl, used M_. subhyalinus mounds as the regular roosting site. Since no spurfowl was seen to roost on any other site it is possible that mound density could affect the distribution and density of this species. It has been shown that mound populations can vary significantly in density even between adjacent areas, but that within an aggregation overdispersion often occurs. Such patterns of mound distribution could act to space out the individuals of a species, either through their use as permanent territorial markers, or as the preferred centre of the individual's breeding area, such as in the lek system of kob. In the latter case mound dispersion could affect the density of of the breeding population. It is likely that the kob-trampled mound des-cribed in the Crater Region, did in fact represent the centre of a male kob's territory as described by Beuchner (1961) , but no concentration of such areas into a definite lek system was seen. The rubbing of mounds by large mammals may have significance as a marking process (Hediger 1948). Many bare mounds are not rubbed whereas others in the same area show signs of frequent usage. Hediger (1948) re-corded that elephants return again and again to rub the same mound or tree. He also noted many ticks at these sites, as was recorded in this study, the removal of which may represent the main purpose of rubbing. Sikes (1968) proposed that in areas where tree cover has for any reason been drastically reduced, an elephant's skin will become abnormally dry. Wallowing followed by dust baths are used to attempt to relieve this condition but as the mud 219 cracks the irritation returns and the animals constantly rub against the few remaining trees, effectively ring barking and killing them. Without con-sidering the effects of elephants browsing trees, it is possible that a high density of large termite mounds may be slowing tree destruction caused by rubbing in areas such as R.N.P. which have high large mammal populations, i i . The occupation of mound interiors by vertebrates Small and medium sized vertebrates enter all four termitaria types in R.N.P., for breeding as well as places of refuge. A l l dead mounds with entry points showed signs of vertebrate occupation, often by more than one species. Some live mounds and subterranean nests with open shafts, showed no signs of occupation. Excavation results did show that all these live open types of nest could tolerate vertebrate presence. This is particularly surprising for the well-named M_. bellicosus whose soldiers immediately attack a would-be intruder who disturbs a mound in any way. Other workers have noted animals seeking refuge or breeding in mounds. Hartmann (1969) reported that monitor lizards often lay their eggs in termite mounds, and Cowles (1930) made a detailed study of the use of Trinervitermes trinervoides mounds as incubation sites for eggs of this species in Natal. De Bont (1964) noted a number of species of bird nesting in mounds in West Africa. Hediger (1948) stated that birds and reptiles made use of mound interiors in Central Africa. Dorst and Dandelot (1970) state that three small carnivores apart from banded mongoose, use mounds as preferential den sites. These were Pousagues mongoose (Dologale  dybowskii Pousagues) , Dwarf mongoose (Helogale parvula Sunderval) , and 220 the dark mongoose (Crossarchus obscurus F. Cuvier) . As far as is known all three are similar to the banded mongoose, being of small size, diurnal, and gregarious, which is in contrast with the other nocturnal solitary species. The preference for these sites by such animals may be a combination of their need for a large space for the whole pack to den in, along with small entry points for protection. Goodland (1965) saw armadillos, fox lizards, and rattlesnakes use holes at the base of large termitaria in South America. Mitchell (1965) found a gecko living in mounds and feeding on the termites in Australia, together with a small python which preyed on the gecko. Lee and Wood (1971) reported that several species of Australian parrot excavate, and nest in, holes in mounds. Macrotermes mounds in particular provide a well fortified complex space, often additionally protected by thicket vegetation. The mounds are usually harder than the surrounding soil, which means that of all the various hole forming processes identified, direct action by the small species occupy-ing the interiors are of the least significance. Even for a good digger like the banded mongoose, only in 4 out of the 15 mound dens did they completely dig the entrances. Animals occupying mounds particularly use galleries or burrows which are nearest to the minimum diameter possible for their pass-age. This means that a rodent and a small carnivore can occupy one mound at the same time, the rodent being effectively sealed off from a potential predator. 221 The mound microclimate is probably an important factor favouring mound usage, in addition to protection from predation and large mammal trampling. Southern and Hook (1963) observed that small mammal nest sites were usually situated in a particularly well buffered environment. One example being under large rocks which would provide shade during the day, and then at night slowly radiate the heat which had been accumu-lated during the day. It has been demonstrated that mound soil can act in the same way. Snakes laying their eggs in the exposed mound surface is likely related to a good incubation temperature rather than protection. In R.N.P. all small mammal nests in termitaria were well protected from flooding, either by their position high up in a mound, or by having vertical termite galleries directly below them. This aspect of mounds was directly responsible for niloticus being able to inhabit seasonal floodland at Ishasha. A similar situation was described by Genelly (1965) for the mole rat (Cryptomys hottentotus) in Rhodesia. The tunnels of this species were concentrated in large termite hills, where food storage areas, resting areas, and feeding areas were situated. An absence of mounds restricted these centres of activity to high ground. Good drainage may also be a major factor in den selection by small carnivores. Sixteen out of the twenty-three banded mongoose dens examined were on slopes, while all those on lowland liable to flooding were associated with termite nests. The protection from fire afforded by mounds is discussed below with reference to small mammal distribution. Most areas of R.N.P. burn at least 222 once a year. In contrast to reports by Griffith (1938) no mounds were seen to protect thicket vegetation from burning, other than that vegetation which would have survived equally well away from a mound. Mound interiors therefore,rather than mound vegetation is the way by which mounds provide increased protection from fire. The use of large burrows in mounds and elsewhere by medium-sized mammals. Medium-sized mammals do not frequent large burrows in mounds more often than large burrows elsewhere in R.N.P. Hediger (1948) believed that orycteropus only used burrows associated with termite mounds as a day time sleeping site. He suggested that since this animal usually sealed the burrow system before retiring for the day, the mounds served an important function by aerating the tunnels. The only set of burrows in R.N.P. identified as a probable day-time refuge for orycteropus was under mound D3 at Mweya, where termite galleries did extend from the tunnel roof to the exterior at the top of the mound. The only difference recorded in temperature or relative humidity, between mound and non-mound large burrows, was the possibility of a higher relative humidity in burrows under termite occupied mounds. This could cause the smaller vertebrates to favor mound burrows, especially in the dry season. The details of burrow structure and size are discussed below in the section on orycteropus. Termitaria and small mammal distribution. Crater Region None of the ten species of small mammals trapped in the Crater Region were limited in distribution by an absence or low density of Odontotermes shafts. Neither was the density of any species significantly correlated with shaft density, apart from M. natalensis. For this species, however, the significance rested on the position of a single point, and so the result is not in fact believed to be significant. Two main points came from the analysis of capture results with respect to trap position. First, one diurnal species only, 1^ . striatus, was always trapped nearer shafts than expected from a random distribution. Second, when all species were considered, significantly more animals than expected were trapped less than 1.5 M from shafts. These patterns suggest that a proportion of animals of all species were entering shafts, but that only L. striatus may have been foraging so as to remain near shafts. Therefore shafts may be an important refuge site for this species. The monthly capture results indicated significant increased activity of L. striatus and_A. niloticus near shafts during March 1973. The restriction of an association between rodents and shafts to this period may be indicative of an increased use of shafts for breeding at that time. The finding of L. striatus juveniles entering the trappable population only in June and July, and the occurrence of a female L^ striatus giving birth to a litter of four in a live trap on April 20, 1973, show that breeding was beginning again in March after the January dry period. Like many species of tropical rodent, striatus is known to restrict its breeding largely to the rainy periods- (Delany 1964, Delany and Neal 1969). A. niloticus usually breeds all year (Delany and Neal 1969), but if peaks occur they do so at the end of the rains (Neal 1967) . Rodents, termitaria, and fire, in the Crater Region. Most of this region is burnt at least once a year. Much of it has no bush cover and Odontotermes shafts represent the main potential refuge site for small mammals, the comparatively few other holes found being attributable to small mammal burrowing, dung beetle activity, and orycteropus digging. The shafts were present in a patchy distribution over much of the region, but were often regularly dispersed within aggregations. It was considered likely that the frequency of use of shafts as refuge sites would increase after a burn, when the often thick surface covering of dead grass, as well as living grass, had been removed. Neal (1970) found evidence of diurnal rodent species in R.N.P. becoming more nocturnal after a burn. This suggested that diurnal species may use shafts more than nocturnal ones, being more affected in their habits by the burn. Snap-trapping after the January burn showed a restriction of animal activity to bush cover, with no usage of shafts recorded. The shafts were, however, within 100 M of bushes, suggesting preference for the latter. During subsequent live-trapping in February, animals of the two species caught, striatus and M. natalensis, usually disappeared down shafts if successfully followed after release. Neal (1970) and Cheeseman (pers. comm. both noted this for all species they live-trapped in this area after burns. 225 A lack of striatus in the unshafted grid during February, but their presence in March and Apr i l , suggests that this diurnal species may be restricted to shafted areas after a burn, until the Imperata grass provides a certain amount of cover. By March 23rd, only two months after the burn, I. cylindrica leaves were over 1 meter high and flowering had already ceased. Neal (1970) showed that although some species populations in the craters were particularly reduced by burns, few animals were killed as a direct result of burning. No dead animals were found after the January burn either. It seems likely that animals seeking refuge in shafts would survive a burn. This statement is supported by the work of Lawrence (1966) in a temperate region. He found that small mammals probably experienced no i l l effects from heat when in burrows greater than 8 cm deep. Of more importance was sufficient air circulation to prevent suffocation. In an experimental situation he found that all animals in burrows with a single entrance were suffocated, while those in burrows with two or more openings survived. The space below the often numerous connected termite shafts are therefore particularly well suited as refuge sites during a burn. Lawrence (1966) considered that postburn predation could appreciably reduce small mammal numbers. In agreement with Neal (1970), no increase in raptores or small carnivores was observed in the Craters after the burn; an increase expected if increased predation occurred. Neal did account for some post burn density changes by identifying both immigration and emmi-gration in burnt areas. Delany (1964) suggested that in tropical regions, aridity and a reduced food supply in burnt areas might account for any reduction in numbers observed. The relative importance of these two factors in the Craters is not known. From the live-trapping results it is suggested that Odontotermes distribution can affect the distribution of small mammals after a burn, particularly diurnal species, and may therefore dictate whether a species population emigrates from a burnt area. Mounds, flooding, and rodents. Trapping at Ishasha showed that 0_;_ kibarensis mounds allowed a population of A. niloticus to survive in seasonal floodland, with animal activity restricted to mounds and adjacent areas. Dropping counts confirmed this and identified mounds with open termite shafts as preferred sites, suggesting the mound interior to be of importance. Excavations revealed that _A. niloticus was indeed nesting in these holes. It is not clear whether the lack of A_. niloticus in March and April 1973 was because of genuine low numbers and a small sample grid, or whether emmigration and then immigration occurred as suggested by the live trapping results. There was much surface water lying in November 1972 when animals were caught, so it seems unlikely that animals regularly leave the area in the wet season. Two factors may have been responsible for an initial emmigration from the area. First, an area adjacent to the mound trapping area burnt during February 1973, the area where the droppings counts were made. This may have caused a movement of animals out of the trapping area in advance of the fire, even though the fire was stopped by a track 800 meters before the trapping area. The number of 227 droppings found in the burnt area showed that niloticus had been present over most of that mound area before the burn. No fresh droppings were found before July 1973, however. It seems likely, therefore, that animals left the area sometime before the burn, possibly immediately before i t , and then returned five months later, when the dominant grass Sporobolus  pyramidalis was flowering up to 50 cm with 25 cm leaves. This again fits in with the live trapping results suggesting a movement of A. niloticus from high ground riverine bush to low grassland in June. A second factor, possibly connected with an initial emmigration from the snap-trapping area, was that Dory line ants were using one in three mounds on the March mound grid, and two in three mounds on the April mound grid. These were the only months when Doryline ants were seen in the area, and could have caused a local lack of Arvicanthis in these two grids. The ability of Doryline ants to disturb other animals including vertebrates is well documented, (Wheeler 1912). i i i . The use of mound soil and vegetation.  Soil Warthog were described above as taking in mound soil. Field (1970) recorded geophagia in warthog on seven occasions during a one year study in R.N.P. Buffalo and hippopotamus were identified as destroying termite mounds and exposing the mineral rich inner galleries. The extent to which this soil was eaten is not known. The case of a buffalo horning a mound may have been an example of aggressive behaviour often seen in males. 228 Edmond-Blanc (1947) described how a male Cambodian wild ox (Bibos sauvelli) horned every termite nest that it passed; it also horned the soil. These activities caused the shredding of the horn sheath, leaving a very pointed tip. The animal did not stop at the mounds and so no consumption of nest material occurred. Weir (1974) noted that the base of a M. subhyalinus mound in the north of R.N.P. had been kicked out by elephant. The main mineral increase in the soil from that mound was for calcium. Weir (1974) also described ex-cavation by elephant of a calcium rich cliff near the shore of lake Amin in R.N.P. The soils of the north of the Park are generally rich in minerals, having developed from volcanic deposits. It seems likely that large mammals derive enough minerals for their physiological needs from plant material and drinking water. Cattle and game animals do use termite mounds as mineral licks in other parts of Africa (Hesse 1955, Weir 1969, Jarman 1972). Water soluble sodium has been suggested to be the primary attractant for elephant in mineral licks (Weir 1969, 1972), but other minerals will be associated with sodium. Jarman (1972) suggested that sodium ions would also attract ruminants to mineral licks and particular drinking sites. In R.N.P. the highest percentage increase of any one exchangeable base in termite mound soil over control soil, was for sodium. Therefore, although mineral licks of any kind are not common in R.N.P., mounds are identified as good sources of minerals, especially sodium. 229 In the south of the Park the soils are less mineral-rich in general, having been derived from Rift-Valley sediments. As stated by Weir (1960) and Jarman (1972) the formation of zoogenous pools must be considered in conjunction with wallowing and mineral licks. Scrapes dug when dry as salt licks can develop into wallows and drinking sites when wet. The area near Ishasha Camp where CL kibarensis mounds and large wallows occurred, was an example of a mineral rich hydromorphic soil developed under seasonal waterlogging (Dept. L. and S. 1962). The wallows were therefore probably originally dug as mineral licks. The significant association between the position of wallows and mounds suggests that large mammal activity was initially preferentially directed to areas near mounds. The reasons for this, and the associations of wallows and mounds elsewhere in Africa, have already been discussed (see section III-4-i). Vegetation A high percentage of Macrotermes mounds may be vegetated. Both the absolute density, and the percentage of mounds with a certain type of vege-tation cover, can vary significantly between even adjacent areas. Edaphic, topographic, and climatic factors have been identified as influencing mound ' vegetation cover in R.N.P.; either directly, or indirectly by influencing the size and shape of mounds built by termites. In general mound grasses showed a higher crude protein content and a higher mineral content than control samples. It is likely that all grazing mammals eat mound vegetation in R.N.P. as part of their diet. The significance of mound vegetation in their diet as a 230 whole will depend on two factors; first, the absolute surface area of mound grasses present, which depends on the density of mounds and the processes outlined above; second, it depends on whether selective grazing of mound vegetation occurs. The surface area taken up by mounds can be large (see section III-4-i) . At Ishasha 0_. kibarensis mounds added 20.68 per hectare to the land surface, whilst occupying approximately 6% of the surface area. Meyer (1960) recorded Macrotermes mounds occupying 30% of the ground surface in one area. It is expected that the amounts of mound vegetation present in areas such as these would form a significant proportion of the diet of even non-selective grazers. It is suggested that selective grazing of mound vegetation does occur in R.N.P. Warthog were seen to graze mound grasses preferentially; buffalo have been seen in apparently awkward positions in order to eat mound vegetation. The heights of S. pyramidalis leaves at Ishasha camp suggest that this species was being preferentially grazed on kibarensis mounds. The biomass results for grasses on and off M_. subhyalinus mounds in the Mweya area indicate that mound grasses were being preferentially taken in that area. Weir (1974) suggested that in one area near Mweya where water soluble minerals were abundant, they may form a local mosaic related to grazing intensity. Field (1968) has demonstrated that in R.N.P. hippopotamus and buffalo selectively graze the more nutritious species as they appear through-out the year. Hippopotamus showed a consistent preference for Cynodon  dactylon, which maintained a high average crude protein content through 231 the year, despite a high stem to leaf ratio. Cynodon dactylon survived this intense grazing by growing close to the ground. Lock (1970a) recorded that in R.N.P. Cynodon dactylon was particularly well grazed in this small form, and that the small form was mainly seen on termite mounds, a condition noted in this study. These observations suggest that hippopotamus and perhaps other mammals were preferentially grazing Cynodon dactylon on termite mounds in the Mweya area, and that this increased grazing pressure may have been responsible for the low growth form of this species on mounds. Lamprey (1963) recorded that in the Tangire Game Reserve, the main habitat for Cynodon dactylon was on Macrotermes mounds, and that this species of grass was the main dry season food for Impala (Aepyceros melampus Lichtenstein) . These relations formed part of a diagram to illustrate an inter-dependence in plant and animal species in the reserve. Field (1968) found that buffalo selectively ate plants with a high protein content. Buffalo were one of the most frequent grazers in the 0_. kibarensis mound area near Ishasha camp. Field (1968) also demonstrated that buffalo showed a high preference for S. pyramidalis in both wet and dry seasons. Where this species was heavily grazed it was maintained in an actively growing condition throughout the year. Where it was undergrazed it became dormant in the dry season. The grass on mounds at Ishasha camp remained greener than that off mounds during the dry season. Selective grazing by buffalo on S. pyramidalis may at least partially explain the lusher condition of mound grass in the dry season (see also section IV-3-iii). Taken along with the data on S^ pyramidalis leaf blade height, these observations strongly suggest 232 that^)_. kibarensis mound grasses were being preferentially grazed. Vos (1969) reported that termite mound vegetation was often more heavily grazed/browsed than the surrounding vegetation, but no data or references were given to support the statement. Conclusion In order to summarize the results discussed above on usage of mound soil and vegetation, along with the processes previously discussed (see section III-4-ii) whereby mounds affect mineral movement, a diagram has been made (figure 96). This gives all the pathways of mineral flow which have been identified as centered around Macrotermes mounds. iv. The use of termitaria as a concentrated source of food. Orycteropus afer was the only species identified as digging into live termitaria in R.N.P. to harvest this vast wealth of insect food. It is possible that the giant pangolin was present with a restricted distribution, and was just not encountered. Other animals which feed on insects at the ground surface can benefit from orycteropus digs. Kingdon (1971) observed baboon picking up live termites from a fresh orycteropus dig; he also suggested that the aardwolf (Proteles cristatus Sparrman) may follow orycteropus and bene-fit from the digs. Evidence of orycteropus was found in all areas of R.N.P. except those liable to seasonal flooding and those with shallow soils. This is in agreement with the distribution in Botswana described by Smithers (1971). In addition no digs were found in the dense Maramagambo forest on the few exploratory visits made. The species is known to frequent such habitats (Pages 1971). cn cn CN pref. grazing pnpf defecation im mv mound vegetation hb herbivore dung t via termites st subterranean termites w weathering I leaching drp deep-rooted plants F i g u r e 96. Pathways of m i n e r a l movement i n f l u e n c e d by vegetated t e r m i t e mounds, 234 The correlation found between the use made of an area, as indicated by the number of digs, and the density of Macrotermes mounds may indicate that mounds affect the distribution of this species; or as suggested by Kingdon (1971) , mounds may just be an indicator of a high total termite and ant population from which most of the diet is selected. Droppings analyses did show that in R.N.P. Macrotermes termites can form over 70% of the diet in an area of high mound density. The change in composition of diet through the year, with few termites being taken in the dryer months, is probably related to the quiescent nature of termites at that time (Bodot 1967) , rather than a change in the animal's digging abilities which remain good even in the driest ground. Three stomachs examined by Smithers (1971) in Botswana all contained a dominance of ants, with traces of termites in two, and traces of hemipteran larvae in one. Smithers (1971) also described the contents of 8 stomachs collected in Rhodesia. Formicidae occurred in 7, forming 100% in 4; Isoptera occurred in 4, in two cases forming 100%. The time of year when these animals were captured was not given. In one stomach 25% of the contents consisted of melon pips, probably originating from melons grown as cattle food, and presumably taken in for moisture. The significance of the unidentified red seeds found in many of the droppings collected at Mweya is not known. Orycteropus paid no particular attention to mounds whilst foraging, which is in contrast to reports which stated that this species forages by moving from one mound to another (Dorst and Dandelot 1970). Day time surveys showed that mounds were often extensively dug, but it is likely that it was not the fact that the sites were mounds which caused them to be attacked. It was probably the senses of hearing and smell, identified as being important for finding food away from mounds, which also brought the animal to mounds. Kingdon (1971) recorded vigorous sniffs made with the nose pressed flat to the ground and suggested that the fleshy tentacles which occur around the nostrils may be highly sensitive to chemical stimuli or perhaps to minute vibrations set up the insects' reaction to the animal's sniffing. Verheyen (1951) on the other hand emphasized the probable im-portance of hearing in food location. He believed that hearing allowed orycteropus to find termites under leaf litter, as well as groups of allates leaving mounds, and large parties on the move on the ground surface. The sense of sight is very poor in orycteropus, an animal fully illuminated by a powerful spotlamp foraged to within 2 meters of me of four occasions. If, however a metallic or other material sound was made it would bound away at much greater distances. Kingdon (1971) recorded that when an animal was released in Naivasha Kenya, it seemed blind in the sun and walked into objects in its path. When mounds were dug, the method was to either make small food digs around the base and sides, or just a single large tunnel into the centre. The latter method usually killed the mound, while the former allowed repeated visits to the same mound, so maintaining a "living sore" on it. This agrees with the observations of Kingdon (1971) in Uganda, but disagrees with Bigourdan (1950) who stated that the same mound was not visited on consecutive nights but at intervals of 5 to 8 days. The use 236 of only large burrows to dig into subterranean Odontotermes nests, and therefore always completely destroy them, may be related to the smaller size of these nests, and a reduced ability to withstand a repeated harvesting method of foraging. The mean distance of 9.6 Km travelled per night for an animal at Mweya was similar to previous estimates made from tracks. Bigourdan (1950) suggested that 2-4 Km was travelled in one night, and Kingdon (1971) gave a maximum distance of 10 Km. Verheyen (1951) estimated a nightly range from 10 to 30 Km. This large distance covered, coupled with the high frequency of digging necessary to supply food to such a large animal, results in an area becoming covered with signs of its presence. As described above, the large burrows in mounds and elsewhere remain for a long time, providing refuge sites for many species in addition to orycteropus. 237 V. CONCLUSIONS: THE USE OF MOUNDS BY VERTEBRATES 1. MOUNDS AS FIXED ELEVATED POINTS In section III it was concluded that the high density of mounds in R.N.P. and their uniformity of spacing, was likely to result in many vertebrate/mound interactions. This aspect of mound presence, mounds as fixed points; has been shown to be utilized by many species of medium sized vertebrates; mounds being used as marking posts, as lookout posts, and for other activities where the importance of the mound is that it is a raised platform. Mound activities involving a sentinel component occur particularly in grassland areas where no other vantage points are available. It seems unlikely that this mound characteristic is affecting the distri-bution of any species in R.N.P., although this could occur for those animals using them as territorial markers. More probably mounds increase the 'comfort level' of the environment for many species, as was demonstrated for the Ishasha baboons. 2. THE OCCUPATION OF MOUND INTERIORS In section III it was stated that small and medium sized vertebrates were expected to favour the complex space within live and dead mounds as den sites and refuge sites; physical protection and a well buffered microclimate being the important factors. Results have shown that these areas are generally used when available, usually by a number of species at any one time, in-cluding predators and their prey, or in succession. No data were obtained on the preference for these sites by small vertebrates, but the fact that all dead mounds with entrance points had signs of their occupation suggests that 2 3 8 it is high. Medium sized mammals showed no preference for burrows in mounds over burrows elsewhere, probably because they themselves were able to modify conditions inside the large volume of a burrow to a greater extent than any modifications caused by a dead mound. Such maintenance of an equitable microclimate in a large burrow has been described by Geigy (1955) for warthog. As expected, mound interiors were found to offer particular protection during times of fire and flooding. Under these conditions termitaria affect the distribution of certain small mammal species in R.N.P. In other areas and at other times mound interiors are usually occupied when present, but their availability probably does not affect the distribution of any species. 3. THE USE OF MOUND SOIL The results presented in section III led to the conclusions that large mammals would be expected to create wallows in the fine textured wash-off area around mounds, and that if the surrounding surface soil was particularly leached, geophagia would occur at the mineral enriched mounds. The lack of signs of geophagia at mounds or elsewhere especially in the north of the Park, suggests that the mineral content of vegetation and drinking water is usually sufficient to meet animal requirements. In the south of the Park where wallows are most abundant, there is a significant correlation between their position and the position of large valley-bottom mounds. Since the whole mound/wallow area was composed of very fine soil it is suggested that this association is connected not with the finer soil texture near mounds attracting large mammals, but with increased water run off, or increased mineral content of soil around mounds, or a combination of both. 4. THE USE OF MOUND VEGETATION It was suggested that the restriction of thicket vegetation to mounds may cause a local increase in small mammal activity because of increased cover. This possibility was investigated at Ishasha. It was shown for small mammals that accessable mound interiors were the features determining centres of activity rather than thicket vegetation. However, since virtually all mounds with termite holes were also thicketed, both features may be important. The main mound vegetation/vertebrate interaction investigated was the possibility of animals selectively grazing from mounds. Mounds do increase the range of nutritive value of grasses available to herbivores in R.N.P. It is believed that the amounts of grass involved could be significant to even non-selective grazers. Therefore, because it was shown in sections III and IV that preferential grazing did occur, the biomass of vegetation taken is likely to be important to the herbivores. Mound vegetation may complete the nutrient requirements of animals throughout the year, or have particular use as a dry season forage as was found by Lamprey (1963) for Impala. 5. THE USE OF TERMITARIA AS A CONCENTRATED SOURCE OF FOOD Termite mounds contain a very large amount of secondary production. A single mound may contain over two million individuals (Luscher 1955) , and the biomass of termites per hectare may equal or exceed biomasses recorded for East African ungulates (Lee and Wood 1971). Termites feeding on herbaceous plants and grasses may compete for food with herbivores. Therefore, it is surprising that in R.N.P. only O. afer is regularly digging into mounds for food. This species does not preferentially forage at mounds, but digs where its senses of smell and hearing locate termites, mounds being included along with subterranean nests. The extensiveness of diggings suggests that this animal could be a controlling influence on termite populations. Kingdon (1971) noted that in areas of South Africa where afer and other insectivorous mammals have been exterminated, pastures have suffered enormous depredation from termites of the genera Hodotermes and Trinervitermes. In R.N.P. the main species taken was the wood feeding Macrotermes subhyalinus, therefore no such obvious beneficial effect of termite predation is expected. The observed increase in_0. afer digs in areas of high Macrotermes mound density may mean that mounds are affecting the distribution of this animal, or mounds may just be an indicator of a general high termite and ant population, from which most of the diet is selected. 6. CONCLUSION The role of termitaria in the lives of vertebrates varies greatly. Features of importance range from those which represent a discrete site used for a specific activity, such as a mound interior as a den, to those which add to the quantity or quality of a resource also found elsewhere, such as vegetation for food. The small mammal work was an example of how, by a detailed study of species populations, it is possible to identify under what conditions specific termitaria characteristics affect an animal's distribution 241 and abundance. This enables the varied environmental contributions of termitaria to be considered along with all other habitat requirements and prevailing environmental conditions, and then related to the total ecology of the animal. It is possible that this ubiquitous insect increases the carrying capacity of many savanna areas of Africa through the diverse mound processes des-cribed above. Processes which are a combination of passive phenomena resulting from mound structure, and the actual activity of termites. This is in contrast with the frequent casting of termites as villains, denuding range-land vegetation and so adversely affecting livestock. It should be remembered that little termite food consists of living vegetation, and such adverse effects will only occur after extensive overgrazing and trampling. In a diverse natural community it seems likely that the beneficial effects of mound building termite populations on the environment of vertebrate species will outweigh any detrimental effects produced by their foraging activity. The possible importance of termitaria in the environment of vertebrate species has been largely uninvestigated. Their wide distribution, especially in Africa, their often high density, their regularity of spacing, and their many aspects of difference from their surroundings, suggest that a closer consideration is warranted. TABLES I to LXXIX Table I . Between area comparison of M. subhyalinus mound d e n s i t i e s . Study Area Number of Transects Mean Number of Mounds Ha~l Nyamagasani 10 12.9 Crater Region 10 8.1 Mweya 10 15.0 Kasenyi 3 8.0 Kamulikwezi 4 6.0 F = 5.86, p < 0.01 Table I I . Dispers ion of M. subhyalinus mounds within study areas. Area N 1 D 2 R £ 4 R5 Comments C 6 S i g . 7 Nyamagasani 78 0 .00130 18.99 13.71 1. 385 Regular 6.51 p < .001 Crater Region 35 0 .00080 24.97 17.68 1. 421 Regular 0.90 Not s i g . Kamulikwezi 20 0 .00073 20.25 18.50 1. 095 Random 0.81 Not s i g . Kasenyi 11 0 .00070 39.55 37.80 1. 046 Random 0.59 Not s i g . Mweya 63 0 .00162 20.57 12.42 1. 660 Regular 9.95 p < .001 1. Number of nearest neighbour p a i r s . 2. Mean density of mounds M~2. 3. Mean observed nearest neighbour dis tance. 4. Mean expected nearest neighbour distance. 5. Dispers ion c o e f f i c i e n t . * 6. Standard var i a te of the normal curve . * 7. S igni f icance of the dif ference between the type of d i spers ion found (see "comments") and a random d i sper s ion . * See appendix 1. Table I I I . Dispersion of M. subhyalinus mounds within two Crater Region aggregations.* Transect N D R A R E R Comments C S ig . CB - F l a t Region 8 0.0008 30.63 17.67 1 .733 Regular 3.96 0.001 DG - F l a t Region 8 0.0010 23.75 15.81 1 .493 Regular 2.70 0.01 * Notation as for Table I I . Table IV. Dispersion of M. subhyalinus mounds from a e r i a l photographic t r ansec t s . * Sample Area •&• Number Transect Height (ft) Area Covered (Ha) Number of Mounds Number Ha~l R A * E R Comments C S ig . Kasenyi 1 500 2.74 9 3.28 25.10 27.53 0.912 Clumped 0. 41 Not s i g . n 2 500 2.70 5 1.85 69.20 36.80 1.880 Regular 3. 35 p <0.001 II 3 500 2.61 6 2. 30 45.75 32.9 7 1. 388 Regular 1. 81 Not s i g . II 4 500 2.51 2 0.78 175.15 56.60 3.090 Regular 5. 44 p <0.0001 II 5 500 2.73 5 1.83 50.24 36.96 1.360 Regular 1. 52 Not s i g . Nyamagasani 1 600 3.40 23 6.76 27.90 19.23 1.45 Regular 4. 14 p <0.0001 II 2 600 3.53 16 4.53 30.50 23.49 1.30 Regular 2. 14 p < 0.05 Rutanda 1 600 3.63 10 2.75 34. 33 30.15 1.14 Regular 0. 83 Not s i g . II 2 600 2.43 5 2.06 35.49 34. 84 1.02 Random 0. 08 Not s i g . Notation as for Table I I . T a b l e V . I s h a s h a g r o u n d t r a n s e c t r e s u l t s . T r a n s e c t T r a n s e c t Number Number Number N o n - Number N o n - S o i l * G e n e r a l D e s c r i p t i o n Number L e n g t h M a c r o t e r m e s M a c r o t e r m e s M a c r o t e r m e s M a c r o t e r m e s T e x t u r e o f T r a n s e c t A r e a (M) Mounds Mounds Ha"J- Mounds Mounds H a - 1 R a t i n g IA IB IC ID I E IG IH I I I J IK I L 500 250 250 500 200 500 500' 500 150 500 500 3 10 8 1 11 4 1 0 8 3 20 0 0 20 1 11 4 3 . 3 0 8 34 7 10 26 13 27 22 10 11 34 14 20 65 33 27 22 10 37 3 2 - 5 2 - 5 2 - 5 F l a t d e n s e b u s h -n o t s e a s o n a l l y f l o o d e d Open g r a s s l a n d -n o t s e a s o n a l l y f l o o d e d Open g r a s s l a n d -s e a s o n a l l y f l o o d e d S e a s o n a l l y f l o o d e d -t h i c k e t s o n K i b a r e n s i s mounds F l a t b u s h e d a r e a -n o t s e a s o n a l l y f l o o d e d Open g r a s s l a n d - on s l o p e -s e a s o n a l l y f l o o d e d i n p a r t On s l o p e f l o o d e d n o t s e a s o n a l l y B u s h e d - on s l o p e - n o t s e a s o n a l l y f l o o d e d B u s h e d - on s l o p e - n o t s e a s o n a l l y f l o o d e d F l a t d e n s e b u s h -n o t s e a s o n a l l y f l o o d e d Open b u s h f l o o d e d n o t s e a s o n a l l y * See s e c t i o n I I I - 3 - i i ** No d a t a Table V I . Variance to mean analyses for mound dens i t ie s i n transects only where M. be l l i co sus occurred. Sample N D <5Z <5VD Comments fx S igni f icance Ishasha - 8 8.791 57.7 6.56 Aggregated 45.94 p < 0.001 between transects Ishasha - 25 1.2 2 2.25 1.88 Aggregated 45.00 p <0.01 between 50M units of 3 transects 1. Mean number of mounds per transect (= per hectare) . 2. Mean number of mounds per 50 meter u n i t . Table V I I . Disperison of mounds at Ishasha from two a e r i a l photgraphic t ransects . F i lm Location Area N D R A ^ R Comments C S igni f icance Number Covered (Ha) a) M. be l l i co sus 2- 26-36 Ishasha - 3. 69 20 0.00062 18.85 20 .04 0. 931 Aggregated 0 .51 Not s i g . dry thicket 2- 1-10 Ishasha - 5. 33 45 0.00107 11.75 15 .29 0. 768 Aggregated 2 .97 P < 0.005 grassland b) Non-Macrotermes 2- 26-36 dry thicket 3. 69 40 0.00136 14.17 13 .58 1. 043 Random 0 .53 Not s i g . 2- 1-10 grassland 5. 33 58 0.00833 6.88 5 .47 1. 256 Regular 0 .49 Not s i g . 1. Notation as for Table I I . 250 Table V I I I . Comparison of M. b e l l i c o s u s and non-Macrotermes mound dispers ions from two a e r i a l photographic t ransects .1 a) Photographic transect no. 2-26-36 Base Mound -p o W Xi Q) x! U tn (0 -H (D <D S3 S3 Macrotermes Non-Macroterme s Tota l Macrotermes 6 2 (7 .25) 3 17 (15.75) 23 Non-Macrotermes 17 33 (34.25) 50 Tota l 23 50 73 2 = 0.46, p > 0.1 b) Photographic transect no. 2-1-10 Base Mound u •p o w Xi Qi A U tn (d -H <U <D Macrotermes Non-Macrotermes Tota l Macrotermes Non-Macrotermes Tota l 3 (5.56) 54 (51.44) 57 46 (43.44) 399 (401.56) 445 49 453 502 X 2 = 1.38, p> 0.1 1. See appendix 1. 2. Observed number of nearest neighbours. 3. Expected number of nearest neighbours. Table IX. 0. Kibarensis mound dispers ion from three a e r i a l photographic t ransects .1 F i lm Height Location Area N D RA R Comments C S ig . Number (ft) Covered (Ha) 2-12-26 600 Ishasha 5.41 20 0.00051 28.95 22.04 1.313 Regular 2.67 p<0.01 (L.M. 1) 4-11-25 1200 Ishasha 10.03 51 0.00051 31.39 22.18 1.415 Regular 5.67 p<0.0001 (L.M. 1) 4-26-36 1200 Ishasha 9.17 58 0.00077 26.33 18.01 1.462 Regular 6.73 p<0.0001 (L.M. 1) 1. Notation as for Table I I . CN IT) CN Table X. 0. f u l l e r i nest densi t ies from ground transects run af ter a burn i n the Crater Region. Transect Number CE CF CG CH CI Number of Mounds Ha -1 14 (9) 1 0 (5) 5 (10) 0 (8) 0 (12) Descr ipt ion of Area Crater bottom -no bare rock Crater side -bare rock High ground -no bare rock High ground -rocky High ground -rocky 1. M. subhyalinus mound densi t ies for comparison. Table XI. M. subhyalinus d i s t r i b u t i o n i n the Crater Region. Crater Sides Crater Bottoms Transect Per Hectare Density Transect Per Hectare Density CB 8 CA 9 CD 4 CC 8 CF 5 CE 9 CJ 8 CH 8 x = 6.25 CI CG 12 10 x = 9.33 t = 2.75, p < 0.05 Table XII . 0. f u l l e r i nest d i s p e r s i o n s from Crater Region t r a p p i n g g r i d s . AREA N D Ra Re R Comments C S i g n i f i c a n c e A l l c r a t e r grids 50 0.0012 18.88 14.43 1.310 . Regular 4.16 P<0.001 Highest density g r i d Cl. 13 0.0039 12. 70 8.01 1.590 Regular 4.04 P<0.001 1. Notation as f o r t a b l e I I . Table X I I I . Odontotermes & Macrotermes r e l a t i v e d i s -p e r s i o n p a t t e r n s i n the C r a t e r Region using data from small mammal tr a p p i n g g r i d maps.1 a) A l l g r i d s . Base Mound Odontotermes Macrotermes T o t a l nearest Odontotermes 4 0 2 ( 3 7 . 1 ) 3 20 (22.9) 60 neigh-bour Macrotermes 20(22.9) 17 (14.1) 37 T o t a l 60 37 97 X 2 = 1.57, df = 1, P>0 .1 b) S i x g r i d s i n f l a t c r a t e r bottom Base Mound Odontotermes Macrotermes T o t a l nearest Odontotermes 12 (13.2) 11 (9.8) 23 neigh-bour Macrotermes 11(9.8) 6 (7.2) 17 T o t a l 23 17 40 X 2 = 0.65, df = 1, P>0 .1 1. See appendix 1. 2. Observed number of nearest neighbours. 3. Expected number of nearest neighbours. Table XIV. Comparison of a e r i a l & ground transect results for per hectare Macrotermes mound densities. Area A e r i a l v i s u a l transects -Aerial photographic transects Ground Transects A l l Mounds Mounds >1 High Unthicketed Mounds Unthicket-ed >lm High Mounds Nyamagasani 1.65 (1.68) 3 5.64 12.9 4.0 11.6 3.6 Mweya'*" + 2 + 15.0 4.4 6.4 1.9 Kazinga 1.10 (1.10) + + + + + Craters 0.37 (0.45) + 8.1 2.2 6.6 1.8 Kasenyi 0. . 72 (1. ,02) 2.01 8.0 2.8 4.0 1.4 Kamulikwezi 0. .82 (0. . 83) + 6.0 3.5 3.7 2.2 Rutanda 1. .28 (1. .45) 2.41 + + . .+ + 1. 2. 3. Adjacent areas considered equivalent. No data. Adjusted data; see text. Table XV. Comparison of a e r i a l v i s u a l density-estimates (adjusted) with ground transect r e s u l t s for four M. subhya-linus mound classes. Mound Class from Ground Transects Correlation C o e f f i c i e n t t Significance A l l Mounds •- +0 .61 1.35 not s i g . Mounds >lm High and unthicketed +0. 73 1.87 not s i g . Mounds >lm High +0. 70 1. 72 not s i g . Unthicketed Mounds +0.68 1.59 not s i g . Table XVI. Mound dimensions. Study Area Mound Type N Mean Height (M) Mean Diameter (M) Nyamagasani M. subhyalinus 127 0.77±0.22 1 2. 34 ±0.54 C r a t e r s M. subhyalinus 56 0.81±0.09 2. la 47±0.24 Mweya M. subhyalinus 110 0.75±0.08 2. 45+0.17 Kasenyi M. subhyalinus 24 1.02±0.27 2. 61±0.70 Kamulikwezi M. subhyalinus 23 1.20±0.34 2. 21±0.52 Ishasha M. b e l l i c o s u s 41 1.02±0.15 2. 12±0.27 Ishasha 0. k i b a r e n s i s 25 1.52±0.23 10. 80±1.15 1. 95% conf i d e n c e l i m i t s . T a b l e X V I I . Comparison o f M. s u b h y a l i n u s mound volumes 1 and n e a r e s t n e i g h b o u r d i s t a n c e s . Mean _2 Mound -R A r e a N Volume , , C o r r e l a t i o n t S i g n i f i c a n c e (M3) I™' C o e f f i c i e n t A l l A r eas 99 4. .23 22 . .1 +0. . 23 2. . 43 P:<0 . 02 Nyamagasani 32 2. .28 18. .9 +0. .53 3. .42 P<0 . 01 Mweya 24 2 . 34 15. .9 +0. .41 2. .11 P<0 . ,05 C r a t e r s 8 1. .92 23. .8 -0. .25 0. .63 n o t s i g . K a m u l i k w e z i 24 3. .96 21. .8 +0. » .11 0. .53 n o t s i g . K a s e n y i 11 7. .25 40. .9 +0. .09 0. .27 no t s i g . 1. See appendix 3. 2. Mean n e a r e s t n e i g h b o u r d i s t a n c e . Table XVIII. Mound surface area and the increase i n surface area caused by mounds. Mound type Study Area Mean Number Ha"l Mean Mean . Mean R 1 Hi O SL 1 (M) (M) (.M) Mean In-Mound Mound crease i n Surface Basal Area per Area Area Mound (M2) (M2) (M2) Mean per hectare i n -crease i n area caused by Mounds (M2) % Area covered by Mounds M. subhyalinus Nyamagasani 12 .9 1 .17 0 .•77 32°2l' 1.41 5 .19 4 30 0 .89 11 .47 0 .55 M. subhyalinus Mweya 15 0 1 .23 0 75 31°28' 1.44 5 53 4 71 0 .82 12 .23 0 71 M. subhyalinus Craters 8 1 1 24 0 81 33°15' •1.48 5 73 4 79 0 .94 7 .60 0 39 M. subhyalinus Kasenyi 8 0 1 31 1 02 38° l ' 1.66 6 79 5 35 1 45 11 51 0 43 M. subhyalinus Kamulikwezi 6 0 1 105 1 20 47°2l' 1.63 5 66 3 84 1 82 10 94 0 23 M. b e l l i c o s u s Ishasha 8 8 1 06 1 02 43°54' 1.47 4 90 3 53 1 37 12 05 0 44 q. kibarensis Ishasha Camp 5. 9 5 40 1 52 15°44 ' 5.60 95 09 91. 62 3 47 20 68 5 97 1. Surface area approximated to that of a cone: where area=ixR.SL. Table XIX. The a s s o c i a t i o n between 0. k i b a r e n s i s mounds and wallows, from two a e r i a l photographic mosaics. a) F i l m number 4-26-36. Base  Mound Wallow T o t a l nearest ~ neighbour Mound 13 (18.4) 21(15.6) 34 Wallow 27(21.6) 13(18.4) 40 T o t a l 40 34 74 X2=6.34, d f = l , P<0.02 b) F i l m number 4-11-25. Base  Mound Wallow T o t a l nearest neighbour Mound 27(29.8) 22(13.4) 49 Wallow 24(15.2) 1(6.9) 25 T o t a l 51 23 84 X 2- 11.19, df = 1, P<0.001 1. See appendix 1. 2. Observed number of nearest neighbours. 3. Expected number of nearest neighbours. Table XX. The nature and d i s t r i b u t i o n of holes i n the four t e r m i t a r i a types studied. Mound Type Study Area N Mean % With % With % With % With % With % With 1 Number Cave-ins Aardvark Small Holes Large Mammal Any Entry Ha"l Burrows Burrows Due To Rubbed Holes Point Weathering M. Subhyalinus Nyamagasani 129 12, .9 9. .3 8.5 7.7 4. .0 0.0 19.4 M. subyalinus Mweya 118 15. .0 7. .6 5.9" 5.9 3, .4 0.0 21.2 M. subhyalinus Craters 65 8. .1 4. .6 4.6 15.4 1. .5 0.0 20.0 M. subhyalinus Kamulikwezi 24 6. ,0 4. .2 16.6 12.5 8, .3 0.0 37.5 M. subhyalinus Kasenyi 24 8. .0 0. .0 12.5 0.0 0, .0 0.0 12.5 M. b e l l i c o s u s Ishasha 46 6. .4 8. .7 13.0 4.5 30. .4 10.9 56.5 0. kibar e n s i s Ishasha 32 5. .9 0. ,0 0.0 40.6 2 0. .0 0.0 59.4 0. f u l l e r i Craters 136 3. .8 0. .0 8.7 3.8 2 0. .0 0.0 9.0 1. Excluding termite shafts. 2. Those shafts showing sign of small mammal usage. Table XX. Continued. Mound Type Study Area % With Mean Number Range i n Mean Shaft Range i n Termite Shafts per Shaft Size (cm) Shaft Size Shafts Mound When Number (cm) Present M. subhyalinus Nyamagasani 0 - - -M. subhyalinus Mweya 0 - - - -M. subhyalinus Craters 0 - - -M. subhyalinus Kamulikwezi 0 - - - -M. subhyalinus Kasenyi 0 - - - -M. Be l l i c o s u s Ishasha 34. .8 3.3 1-10 7.4 4-25 0. kibarensis Ishasha 59. 4 4.9 1-14 5.6 4-7 0. f u l l e r i Craters 100. 0 6.1 1-22 5.8 2-12 Table XXI. D e t a i l s of a l l holes i n one hect-are of c r a t e r bottom examined a f t e r the January 1973 burn. Number of Hole Type Holes Odontotermes s h a f t s 105 Large Burrows (Aardvark) 2 Small Burrows (small mammal and Dung Beetle) 6 Large mammal cave-ins i n T e r m i t a r i a 1 Table XXII. Between area comparison of the den s i t y of M. subhyalinus mounds wi t h cave-ins. Study Area Number of Transects Mean Density of mounds wi t h cave-ins Ha7"1 Range i n den s i t y of mounds wi t h cave-ins Ha -1 Nyamagasani 10 1.2 0 - 3 Mweya 10 1.0 0 - 3 Crater s 9 0.4 0 - 2 Kamulikwezi 4 0.3 0 - 1 Kasenyi 3 0 0 F = 2.05, P>0.1 Table X X I I I . Between area comparison of the dens i t y of M. subhyalinus mounds with l a r g e burrows. Study Area Number of Transects Mean Density of mounds wi t h la r g e burrows Ha" 1 Range i n dens i t y of mounds w i t h l a r g e burrows Ha--1 Nyamagasani 10 1.1 0 - 3 Mweya 10 0.8 0 - 3 Crater s 9 0.4 0 - 2 Kamulikwezi 4 1.0 0 - 2 Kasenyi 3 0.7 0 - 2 F= 0.75, P>0.1 Table XXIV. Between area comparison of the density of M. subhyalinus mounds with small burrows. Study Area Number of Transects Mean Density of Mounds with Small Burrows H a - 1 Range i n Density of Mounds with Small Burrows H a - 1 Nyamagasani 10 0.8 0-4 Mweya 10 0.8 0-3 Craters 9 1.9 0-8 Kamulikwezi 4 0.8 0-1 Kasenyi 3 0.3 0-1 F = 0.68, p > 0.1 Table XXV. Between area comparison of the density of M. subhyalinus mounds with holes from weathering. Study Area Number of Mean Density of Range in Density Transects Mounds with of Mounds with Weathered Holes Ha-1 Weathered Holes Ha~l Nyamagasani 10 0.5 0-1 Mweya 10 0.5 0-2 Craters 9 0.2 0-2 Kamulikwezi 4 0.5 0-2 Kasenyi 3 0 0 F = 0.60, p > 0.1 00 CN Table XXVI. Between area mounds with comparison of the any entry point. density of M. subhyalinus Study Area Number of Transects Mean Density of Mounds with Any Entry Point H a - 1 Range Mounds Entry i n Density of with Any Point H a - 1 Nyamagasani 10 2.5 1-4 Mweya 10 3.1 0-6 Craters 9 2.3 0-8 Kamulikwezi 4 2.3 1-5 Kasenyi 3 1.0 0-2 F = 0.88, p > 0.1 Table XXVII. Within area comparisons of mound dimensions, a) M. subhyalinus N Height (M) S i g n i f i c a n c e 1 Diameter (M) S i g n i f i c a n c e 1 NYAMAGASANI Mean dimensions of a l l mounds 127 0, .77 + .22* 2 .34 + .54 Mean dimensions of mounds + holes 25 0, .71 + .18 t = .87 2 .36 + .41 Mean dimensions of mounds without holes 102 0, .80 + .25 not s i g . 2 .37 + .87 not s i g . CRATERS Mean dimensions of a l l mounds 56 0. .81 + .09 2 .47 + .24 Mean dimensions of mounds + holes 12 0. .57 + .24 t = 2.43 2 .55 + .57 t = . 7 Mean dimensions of mounds without holes 44 0. .86. + .11 s i g , . p <0.02 2 .40 + .26 not s i g . MWEYA Mean dimensions of a l l mounds 110 0. .75 + .08 2 .45 + .17 Mean dimensions of mounds + holes 23 0. .65 + .16 t = 1.3 2 .57 ± .45 t =» 1.35 Mean dimensions of mounds without holes 87 0. .78 + .09 not s i g . 2. . 34 ± .19 not s i g . KAMULIKWEZI Mean dimensions of a l l mounds 23 1. .20 + .34 2 .21 ± .52 Mean dimensions of mounds + holes 4 1. . 35 + 1.45 t = .74 2, .18 ± 2.04 t = .47 Mean dimensions of mounds without holes 19 1. ,04 + .40 not s i g . 1. .99 ± .49 not s i g . KASENYI Mean dimensions of a l l mounds 24 1. ,02 + .27 2, .61 ± .70 Mean dimensions of mounds + holes 3 1. ,38 + 2.15 t = 1.03 4, .33 + .02 t = 2.04 Mean dimensions of mounds without holes 21 0. 97 ± .28 not s i g . 2, .37 ± .32 s i g . , p <0.1 1. S i g n i f i c a n c e of dimension d i f f e r e n c e s between holed and non-holed mounds. * 95% confidence l i m i t s . Table XXVII. Continued. b) M. b e l l i c o s u s Height (M) S i g n i f i c a n c e 1 Diameter (M) S i g n i f i c a n c e 1 ISHASHA Mean dimensions of a l l mounds 41 1, .02 + .15* Mean dimensions of mounds + termite shafts 16 1, .20 + .27 Mean dimensions of mounds - termite shafts 25 0. .90 + .20 Mean dimensions of bare mounds + non-termite holes 14 1. .05 + .28 Mean dimensions of bare mounds -non-termite holes 7 1. .15 + .68 1.9 .1 not s i g . 2.12 ± 2.29 + 1.85 + 2.03 ± 1.87 + .27 .29 .21 •27 .76 t = 2.61 s i g . p <0.02 not s i g . c) O. kibarensis Height (M) S i g n i f i c a n c e Diameter (M) S i g n i f i c a n c e ISHASHA Mean dimensions of a l l mounds 25 1.52 ± .23 Mean dimensions of mounds + termite holes 18 1.43 ± .19 t = .62 Mean dimensions of mounds - termite holes 7 1.33 ± .31 not s i g . 10.80 ± 1.15 10.99 ± 1.27 9.89 ± .95 t = 1.1 not s i g . IV S i g n i f i c a n c e of dimension differences between holed and non-holed mounds. * 95% confidence l i m i t s . Table XXVIII. Volumes of mounds with non-termite made entry holes. Termite Species Study Area Number of Mounds Mean Volume of Mean Volume of Ha-1 with Any a Mound with Any Holed Mounds Entry Point Entry Point (M3) H a - 1 (M3) M. subhyalinus Nyamagasani 2. M. subhyalinus Mweya 3. M. subhyalinus Craters 2. M. subhyalinus Kamulikwezi 2. M. subhyalinus Kasenyi 1. M. bel l i c o s u s Ishasha 3. 0. Kibarensis Ishasha 3. 5 1.55 3.88 1 1.69 5.23 3 1.46 3.35 3 2.52 5.80 0 10.17 10.17 6 1.80 6.48 5 67.87 237.57 CN CN Table XXIX. Volumes within shafts of 0. f u l l e r i nests . Mean Mean Mean Mean Mean Mean Mean Maximum Maximum Shaft Shaft Shaft Number Number Number Volume Number Volume Diameter Depth Volume of Nests of Shafts of Shafts of Shafts of Shafts of Shafts (cm) (cm) (M3) H a - 1 per Nest H a - 1 H a - 1 (M3) Ha-1 Ha~l (M ) 5.8 45.2 0.0012 3.8 6.1 23.2 0.031 700 1 0.840 1. From trapping gr ids ; see Table LXV. 273 Table XXX. Densit y and Volume of f i v e l a r g e bur-rows i n the n o r t h o f R.N.P. a) Volume Burrow Number T o t a l Tunnel Length (M) Mean Tunnel Radius (M) Number of E n t r a n -ces Volume of Tunnels (M 3) Volume per En-trance (M 3) 10 1.95 0. 32 1 0. 31 0. 31 T 3.25 0.29 1 0.43 0.43 D3 13.25 0.42 3 3.70 1.23 H 2.10 0.36 1 0.43 0.43 1 2.50 0.54 2 1.13 0.57 b) Density Number Mean num- Mean num- Mean Ha - 1Range of ber of bur- ber of de n s i t y i n Mean 500M rows i n burrows of number volume Transects Mounds elsewhere burrows Ha -1 Ha~-1 Study Area Mweya Cra t e r s 5 5 0.6 1.2 0.9 3.4 1.5 1-2.5 0.89 3.8 1-15.0 2.24 H a l f o f Mweya upper-p e n i n s u l a 14- 0.6 0-8.0 0.34 1. Number i n 37.0 hectare area 2. Range between map g r i d squares; (1 g r i d square= 1 h e c t a r e ) . fD >< cr CD O O 3 CD H (D tt c in >< £ 0) o -• 3 o in o 3 01 c n W Ml cr 0 o n < CD CD O O 3 01 3 01 £ C I-K H-hh X 0> £ 0 CD (D N O 3 0 O tt CD 01 01 cr M O o CD O O 3 rt H O CD 0> 3 O C 3 tt 0) tr o < CD — rt M CD 0 3 o 3 3 " o H C - 3 tt Z 01 i< Cu 3 01 01 C iQ H Oi Hi 0] ru Oi O 3 CD 0 C H O CD O m cn 01 3 number o f samples s o i l t exture conduct-i v i t y (mmhosi N i t r a t e Nitrogen (ppm) Phosphor-ous (ppm) Calcium (Meq.%) Magnesium (Meq. %) Potassium (Meq.%) Sodium (Meq.%) T o t a l exchangable| c a t i o n s (Meq.%) Ol NI Ln O to LJ it* H* ~J (-* CO PH rt O !» cr o < ro cr o < CD rt O fl> cr o < 01 rt C 3 c cr~ (/) O fl> O 3 rf C ~ 3 o C O rl 3 C z ro s: o n *• 3 cr O < ro o o 3 " rt cn H o O I 03 o rt 0 0 3 — Bi cr o < ro 3 35 O H-C -O 3 -0 0, o a ro cn rt n o >< tn o e 1 o (II number of samples s o i l t exture conduct-i v i t y (mmhos) N i t r a t e N i t r o g e n (ppm) Phosphor-ous (ppm) Calcium (Meq.%) Magnesium (Meq.%) Potassium (Meq.%) Sodium (Meq.%) T o t a l exchangable c a t i o n s (Meq.%) PH o M 0 3 3 rt 0 — . H, ~ c l- 0 t- 3 o t—* O O w 0 w 3 3 •• — rf 0 M 01 0> 3" cr o Cn < 3" CD 0; IO M -tr o> n (I 3 3 n 0 o CTi cn 0 0 0 0 o C 3 3 3 3 e 3 rt rt rt 0 H Q . M. — M — H 0 0 ui 0 !-• 0 3 CD H- h—' o M o t-< o 3 01 0 0) O 01 n oi 0 rt 3 3 E " Ml CD rt — rt — rt H. 0 0 0 t-*H cn H- 01 Oi 0 CU 0> 01 3" 3 K cr cr cr 01 •5 0 o 0 01 i-" . — < < < 3" CD a CD CD CD a CD •• a 0. .—. 3^  ui cn o ui ui M H " M O O O H* Ul H» -J O . U O (Tt cn Ul LJ vo ID number of samples soil texture conduct-ivity (mmhos) Nitrate Nitrogen (ppm) Phosphor-ous (ppm) Calcium (Meq.%) Magnesium (Meq.%) Potassium (Meq.%) Sodium (Meq.%) Total exchangable cations (Meq.%) PH 277 Table XXXII. Comparison of surface s o i l from M. subhyalinus mounds w i t h t h a t of adjacent c o n t r o l areas using the Student's t - t e s t f o r p a i r e d v a r i a t e s . S o i l Property Number of Sample P a i r s Mean Mound Value Mean Co n t r o l Value S i g n i f i c a n c e of Mound/Control D i f f e r e n c e C o n d u c t i v i t y (mmhos) 4 N i t r a t e Nitrogen (ppm) 5 Phosphorus (ppm) 5 Calcium (Meq.%) 4 Magnesium (Meq.%) 4 Potassium (Meq.%) 4 Sodium (Meq.%) 4 T o t a l c a t i o n s (Meq.%) 4 PH 5 0.59 8.94 87.0 20.02 7.63 2.47 0.46 30.59 7.6 0 . 35 3.00 114. 7 14.92 5.03 2.35 0.08 22.37 7.1 P<0.01 P>0.05 P<0.05 P<0.05 P<0.01 P>0 . 05 P<0.05 P<0.01 P>0 . 05 Table XXXIII. Comparison of mineral weights i n M. subhyalinus mounds wi t h mineral weights i n t o p s o i l , 1. Nyamagasani. Mean mound b a s a l 1.16M n i A M Mean mound height Mean above ground mound volume U • / flJYL 1.5645M3 129.OKg 36 4.lKg 1.33% Mean above ground 3 Mean above ground mound weiqht H a - 1  40, Mound weight Ha _ 1/mound & t o p s o i l X 100 weight H a - 1 S o i l Type Ca Weights of Mine r a l s Per K Mg Hectare Na N i t r a t e Nitrogen Mounds ( A l l ) 201.82 42.48 49.14 3.52 0. 71 Mounds (Bare) 12.52 2.64 3.06 0.22 0.04 T o p s o i l (0-15cm) 12822.0 4192.6 2601.0 72.0 15.75 T o p s o i l and Mounds ( A l l ) 1302 3.82 4235.08 2650.14 75.52 4.51 Bare Mounds „ Mounds and X 1 0 0 °- 1 0 0.06 0.12 0.29 0.25 T o p s o i l Mounds ( A l l ) Mounds and X 100 1.55 1.00 1.90 4.66 4.31 T o p s o i l Table XXXIV. Comparison of mineral weights i n M. subhyalinus mounds wit h mineral weights i n t o p s o i l , 2. Mweya. Mean mound b a s a l radius 1.25M Mean mound height 0 . 72M 3 Mean above ground mound volume 1.7730M Mean above ground mound weight 3546.OKg Mean above ground mound weight Ha ^ 5 3190.OKg Mound weight Ha ^/mound & t o p s o i l weight X 100 ... 1.74% Ha" I Weights of Minerals per Hectare (kg) S o i l Type Ca K Mg Na N i t r a t e Nitrogen Mounds ( A l l ) 190.68 43.90 48.54 6.90 0.74 Mounds (Bare) 20.98 4.84 5.34 0.75 0.08 T o p s o i l (0-15cm) 5289.0 3645.0 1686.0 69.0 31.5 T o p s o i l & Mounds ( A l l ) 5479.69 3688.90 1734.54 75.90 32.24 Bare Mounds 1 Q 0 Q 3 g Q > 1 3 Q > 3 1 1 > 0 Q Q 2 5 Mounas and T o p s o i l !! O U n^ S A l ^ X 100 3.48 1.19 2.80 9.09 2.29 Mounds and T o p s o i l Table XXXV. Comparison of mineral weights i n M. subhyalinus mounds with mineral weights i n t o p s o i l , 3. Crater Region. Mean mound basa l radius 1.2 3M Mean mound height 0.78M 3 Mean above ground mound volume 1.8600M Mean above ground mound weight 3720.OKg Mean above ground mound weight Ha 30,132. OKg Mound weight Ha ''"/mound & t o p s o i l weight X 100 . . . . 0.99% Ha" 1 Weights of Mineral Per Hectare (Kg) Ni t ra te S o i l Type Ca K Mg Na Nitrogen Mounds (All) 146.40 33.06 36.18 1.50 0.42 Mounds (Bare) 11.28 2.56 2.80 0.12 0.03 Topso i l (0-15 cm) 11076.0 2940.0 2544.0 60.0 1.5 Topso i l and Mounds (Al l ) 11222.40 2973.06 2580.18 61.50 1.92 Mounds and X 1 0 0 ° - 1 0 ° - 0 9 0.11 0.20 1.56 Topso i l Mounds (All ) v n . . . _ Mounds and X 1 0 0 1 ' 3 0 1-40 2.44 21.88 Topso i l Table XXXVI. Comparison of mineral weights i n M. b e l l i c o s u s mounds at Ishasha with mineral weights i n t o p s o i l . Mean mound bas a l radius 1.13M Mean mound height 0.9 3M Mean above ground mound volume 1.8550M Mean above ground mound weight 3710.OKg Mean above ground mound weight Ha 1 2 3,706.9Kg Mound weight Ha "'•/mound & t o p s o i l weight X 100 .... 0. 78% Ha" 1 Weights of Minerals Per Hectare (Kg) N i t r a t e ••! S o i l Type Ca K Mg Na Nitrogen Mounds ( A l l ) 75.60 15.72 10.70 2.92 0.64 Mounds (Bare) 34.46 7.14 8.52 1.32 0.29 T o p s o i l (0-15cm) 3516.0 643.6 882.0 111.0 1.5 T o p s o i l and Mounds ( A l l ) 3591.6 659.32 892.70 113.92 2.14 Bare Mounds Q > 9 6 Mounds and T o p s o i l Mounds ( A l l ) x 1 0 Q 2 1 Q 2 3 Q 1 2 Q 2 5 6 29.91 Mounds and T o p s o i l CM 00 CN Table XXXVII. Comparison of mineral weights i n 0. k i b a r e n s i s mounds at Ishasha with m i n e r a l weights i n t o p s o i l . Mean mound basa l radius 5.6 3M Mean mound height 1.34M Mean above ground mound volume 88.968M Mean above ground mound weight 177,936.OKg Mean above ground mound weight Ha ^ 1,062,277.92Kg Weights of Minerals per Hectare (kg) S o i l Type Ca K Mg Na N i t r a t e Nitrogen Mounds ( A l l ) 10,245.0 1139.4 532.6 255.0 12.75 Mounds (Bare) 0 0 0 >• 0 0 T o p s o i l (0-15cm) 5,124.0 396.0 1500.0 675.0 0.75 T o p s o i l and Mounds ( A l l ) 15,369.0 1535.4 2032.6 930.0 13.5 Mounds ( A l l ) 1 0 Q 66.66 74.21 26.20 27.42 94.44 Mounds and T o p s o i l Table XXXVIII. Percentage of M. subhyalinus mounds i n four vegetation cover c l a s s e s . Study Area Number of Mounds Examined Mean Number Ha-1 Percen-tage Bare Percen-tage & Basal Grass Percen-tage & Grass Cover Percen-tage Thicket-ed Nyamagasani 129 12.9 6.2 32.6 51.1 10.1 Mweya 118 15.0 11.0 12.7 18. 7 57.6 Craters 65 8.1 7.7 13.8 60.0 18.5 Kamulikwezi 24 6.0 48.8 4.2 19.5 37.5 Kasenyi 24 8.0 41.1 0 8.3 50.0 Table XXXIX Percentage vegetation of M. b e l l i c o s u s mounds cover c l a s s e s . i n four Ishasha 46 6.4 45.6 0 30.5 23.9 Table XL. Percentage of 0. k i b a r e n s i s v e g e t a t i o n cover c l a s s e s . mounds i n four Ishasha 32 5.97 0 0 15.6 84.4 Table XLI. Vegetation de ta i l s of 0. kibarensis mounds from three a e r i a l photographic transects . F i lm Number Number Number of Percentage Percentage Number of Percentage Number of of Mounds and of Mounds of Thickets Thickets of Thickets Mounds Thickets Thickets Thicketed on Mounds on Stream on Mounds Slopes Leaving Out Slope Data 4-11-25 51 27 22 43.1 81.4 _1 81. 4 1 4-26-36 58 40 21 36.2 52.5 15 84.0 2-12-26 20 39 14 70.0 35.9 17 63.6 1. Photographic transect 4-11-25 d id not cover any stream slope area. Table XLII. Between area comparison o f M. subhyalinus v e g e t a t i v e cover. 1. Number of bare mounds per h e c t a r e . Study Area Number of T r a n s e c t s Mean De n s i t y Per Hectare Range i n Den s i t y per Hectare Nyamagasani 10 0.8 0.0 - 4.0 Mweya 10 1.6 0.0 - 4.0 C r a t e r s 9 0.6 0.0 - 2.0 Kamulikwezi 4 2.8 1.0 - 9.0 Kasenyi 3 3.3 3.0 - 4.0 F = 2.22 , P<0.1 Table X L I I I . Between area comparison of M. subhyalinus v e g e t a t i v e cover. 2. Number of mounds wi t h b a s a l grass per he c t a r e . Study Area Number of T r a n s e c t s Mean Densit y per Hectare Range i n Den s i t y per Hectare Nyamagasani 10 4.2 0.0 - 9.0 Mweya 10 2.1 0.0 - 6.7 C r a t e r s 9 1.7 0.0 - 6.0 Kamulikwezi 4 0.3 0.0 - 1.0 Kasenyi 3 0 -F = 1.15, P>0.1 Table XLIV. Between area comparison of M. subhyalinus v e g e t a t i v e cover. 3. Number of grass covered mounds per h e c t a r e . Study Area Number of T r a n s e c t s Mean Densit y Per Hectare Range i n Density per Hectare Nyamagasani 10 7.0 3 - 1 3 Mweya 10 1.7 0 - 6 C r a t e r s 9 4.0 0 - 1 0 Kamulikwezi 4 0.25 0 - 1 Kasenyi 3 0 -F = 7.86, P<0 .001 Table XLV. Between area comparison of M. subhyalinus v e g e t a t i v e cover. 4. Number o f t h i c k -e ted mounds per h e c t a r e . 'J Study Area Number of T r a n s e c t s Mean Densit y Per Hectare Range i n Den s i t y Per Hectare Nyamagasani 10 1.0 0 - 4 Mweya 10 8.3 2 - 1 4 C r a t e r s 9 1.8 0 - 4 Kamulikwezi 4 2.5 1 - 4 Kasenyi 3 4.3 2 - 6 F = 14.21, P<0.0001 Table XLVI. Within area mound dimensions, with a comparison between bare and vegetated mounds. a) M. subhyalinus N Height (M) S i g n i f i c a n c e Diameter (M) S i g n i f i c a n c e NYAMAGASANI Mean dimensions Of a l l mounds 127 0. 77 + .22 1 2.34 + .54 Mean dimensions of vegetated mounds 119 0. 77 + .23 t = 0.08 2.37 + .56 t = 1.25 Mean dimensions of bare mounds 8 0. 77 + .15 not s i g . 2 1.89 + .73 not s i g . CRATERS Mean dimensions of a l l mounds 56 0. 81 + .09 2.47 + .24 Mean dimensions of vegetated mounds 51 0. 80 + .10 t = 1.42 2.55 + .25 t = 2.03 Mean dimensions of bare mounds 5 0. 86 + .41 hot s i g . 1.62 ± .97 s i g . p <0. .05 MWEYA Mean dimensions of a l l mounds 110 0. 75 + .08 2.45 + .17 Mean dimensions of vegetated mounds 95 0. 79 + .08 t = 1.16 2.56 + .18 t = 2.00 Mean dimensions of bare mounds 15 0. 52 ± .21 hot s i g . 1.78 + .52 s i g . p <0. .05 KAMULIKWEZI Mean dimensions of a l l mounds 23 1. 20 + . 34 2.21 + .52 Mean dimensions of vegetated mounds 12 1. 69 ± .39 t = 2.95 2.88 + .83 t = 2.79 Mean dimensions of bare mounds 11 0. 92 + .43 s i g . . p <0. 01 1.64 + .23 s i g . p <0. .02 KASENYI Mean dimensions of a l l mounds 24 1. 02 ± .27 2.61 + .70 Mean dimensions of vegetated mounds 12 1. 44 + .35 t = 4.18 3.88 + .90 t = 5.82 Mean dimensions of bare mounds 11 0. 60 .27 s i g . p <0. 001 1.35 + .32 s i g . p <0. .001 1. 95% confidence l i m i t s . 2. p >0.05 Table XLVI. Continued. b) M. b e l l i c o s u s Height (M) Sig n i f i c a n c e Diameter (M) Si g n i f i c a n c e ISHASHA Mean dimensions of a l l mounds 41 1.02 ± .15 Mean dimensions of vegetated mounds 21 .91 ± .02 t. = .8 Mean dimensions of bare mounds 20 1.02 ± .23 not s i g . ' 2.12 ± .27 1.89 ± .49 2.2 ± .63 t = .9 not s i g . c) 0. kibarensis Height (M) Sig n i f i c a n c e Diameter (M) Si g n i f i c a n c e ISHASHA Mean dimensions of a l l mounds 25 1.52 ± .23 Mean dimensions of grassed mounds 5 1.15 ± .61 not Mean dimensions of thicketed mounds 20 1.61 ± .26 s i g . 10.80 +1.15 10.20 ± .03 10.97 ± 1.34 not s i g . 1. 95% confidence l i m i t s . 2. p >0.05. Table XLVII. Within area var i a t ion i n mound s ize and vegetation cover at Mweya. Seasonally Flooded High F l a t Areas Steep Slope Low Lying Areas Not Seasonally Flooded Mean Height (M) 0.96 ± 0. 18 1 (26) 2 0.64 ± 0 .11 (44) 0.78 ± 0. 14 (27) Mean Diameter (M) 3.24 ± 1. 83 (26) 2.47 ± 0 .29 (44) 2.74 ± 0. 36 (27) Mean Number Mounds Ha -1 12 .4 (26) 20.8 (44) 13 .5 (27) Mean Number Thickets Ha -1 24 .9 (26) 14.0 (28) 20 .0 (20) Mean Number Mounds with Thickets 10 .8 (26) 8.6 (44) 12 .0 (27) Percentage of Mounds with Thickets • 88 .1 (26) 40.6 (44) 88 .3 (27) Percentage of Thickets on Mounds 54 .1 (26) 57.4 (28) 50 .0 (20) 1. 95% confidence l i m i t s . 2. Number of mounds or thickets sampled. Table XLVIII . S igni f icance of differences between the means of mound and th icket charac te r i s t i c s between three topographic areas at Mweya. Mound Condit ion Areas Compared S igni f icance Mean Height Mounds (Low Lying X (Steep Slope (Low Lying X High Ground) X High Ground) Steep Slope) t = not not 3.28 s i g . 1 s i g . P <0. 01 Mean Diameter Mounds (Low Lying X (Steep Slope (Low Lying X High Ground) X High Ground) Steep Slope) not not not s i g . s i g . s i g . Mean Number of Mounds per Hectare (Low Lying X (Steep Slope (Low Lying X High Ground) X High Ground) Steep Slope) t = t = not 3.159 3.08 s i g . P P <0. <0. 05 05 Mean Number of Thickets per Hectare (Low Lying X (Steep Slope (Low Lying X High Ground) X High Ground) Steep Slope) not not not s i g . s i g . s i g . % Mounds + Thickets (Low Lying X (Steep Slope (Low Lying X High Ground) X High Ground) Steep Slope) t = t = 6.50 7.71 P P <0. <0. 01 01 % Thickets + Mounds (Low Lying X (Steep Slope (Low Lying X High Ground) X High Ground) Steep Slope) not not not s i g . s i g . s i g . 1. p >0.05 Table XLIX. Detai led vegetation descr ipt ion of ten thicketed M. subhyalinus mounds near Mweya. Mound Number 1 2 3 4 5 6 7 8 9 10 Height (M) 1.58 1.08 0.66 1.50 0.50 1.33 1.33 1.00 0.8 1.10 Mean Diameter (M) 4.05 3.15 2.50 3.50 0.50 3.30 5.00 3.25 3.00 3.20 Fresh Working / / / / / Mound Dead Shrubs/Trees • Euphorbia candelabrun / / / / / • Tarenna qraveolens / / / / • / / Turraea sp. / / Erythrococca qangensis / / / / Hoslundia opposita / / Capparis tormentosa / / / / / / / / Cassia spp. • Herb/Grass Layer / / Commelina gangetica / / Mariscus dubius / / Dyschoriste radicans / / / / / Tribulus t e r r e s t r i s • Wissadula sp. / Talinum sp. ^Sporobolus pyramidalis / / / / / / / / 2Cenchus c i l i a r i s / / ^Chlor i s guyana / ^Cynodon dactylon / 1. Presence of character . 2. Grasses. CN CN Table L . Grass biomass on four M. subhyal inus mounds and four contro l areas at Mweya. Mound Location Wet Weight (g) Dry Weight (g) % Water Species of Grass Present Enclosure Mound 321.4 163.5 49 A : CG : : CP : CC. Enclosure Control 169.2 105.0 38 A : CG : : CP : CC. Enclosure (Burnt) Mound 98.3 52.0 47 A : CG : : SP. Enclosure (Burnt) Control 53.1 21.8 59 CC : CG : E . Outside Mound 1 27.9 15. 3 45 CD : CC : SP. Outside Control 1 40. 7 22.0 46 SP : CG ;: E : CC : B. Outside Mound 2 35.8 23.2 35 CD : SP : B : CG. Outside Control 2 41.9 28.6 32 SP : CG : CP. A - A r i s t i d a sp. CP - C h l o r i s pycrothr ix B - Bothr iocloa sp. CD - Cynodon dactylon CC - Cenchus c i l i a r i s E - Eragros t i s sp. C G ~ Chlor i s guyana SP - Sporobolus' pyramidalis Table L I . Comparison of analysis means from i n d i v i d u a l mound (A) and contro l (B) samples. a) M. subhyalinus (Jan. 1973) Grass Species A/B Number Mean % Dry S ig . Mean % S ig . Mean % Dry S ig . of Weight S i l i c a - Weight Samples Protein free Ash Potassium Sporobocus py. A 7 4.37 + 1.41* P< 2. 71 + 0. 26 P< 0. 77 + 0. 11 P< II B 7 2.66 + 1.03 0.05 2. 06 + 0. 37 0.01 0. 52 + 0. 11 0.01 Cenchus o i l . A 6 4.48 + 3.88 not 4. 54 + 1. 59 not 1. 40 + 0. 60 not II B 6 4. 73 + 2.58 s i g . 1 2. 98 + 1. 15 s i g . 1. 13 + 1. 95 s i g . Bothr ioc loa sp. A 3 6 .135 + 11.76 not 4. 44 + 1. 42 not 0. 92 + 0. 35 not II B 3 4.58 + 5.16 s i g . . 3. 3 + 2 . 2 s i g . 0. 82 + 0. 12 s i g . Ch lor i s guyana A 2 4.86 + 8.96 not 4. 77 + 6. 78 not 1. 3 + 2. 93 not II B 4 3. 70 + 2.09 s i g . • 5. 72 + 3. 14 s i g . 1. 02 + 0. 15 s i g . b) 0. Kibarensis (Nov. 1972) Sporobolus py. A 5 6.00 + 7.66* not 3. 32 + 0. 68 P< 1. 20 + 0. 30 P< II B- 5 3.85 + 2.72 s i g . 2. 35 + 0. 05 0.01 0. 87 + 0. 15 0.05 1. * 2. p >0.05. 95% confidence l i m i t s . From Student's t - t e s t . Table L I . Continued. a) M. subhyalinus (Jan. 1973) Grass Species A/B Number Mean % Dry S ig . Mean % Dry S i g . Mean % Dry S ig . of Weight Weight Weight Samples Calcium Magnesium Sodium Sporobocus py. A 7 0. 40 + 0 .03* not 0. 14 + 0. 06 not 0. 015 + 0 .006 P< B 7 0. 34 + 0 .03 s i g . x 0. 12 + 0. 13 s i g . 0. 009 + 0 .005 0.05 Cenchus c i l . A 6 0. 38 + 0 .39 not 0. 23 + 0. 15 not 0. 024 + 0 .003 not n B 6 0. 37 + 0 .08 s i g . 0. 14 + 0. 06 s i g . 0. 019 + 0 .011 s i g . Bothr ioc loa sp. A 3 0. 43 + 0 .03 P< 0. 31 + 0. 10 not 0. 021 + 0 .011 not II B 3 0. 36 + 0 .03 0.01 0. 11 + 0. 08 s i g . 0. 011 + 0 .008 s i g . Ch lor i s guyana A 2 0. 62 + 0 .04 P< 0. 22 + 0. 76 not 0. 018 + 0 .057 not B 4 0. 44 + 0 .10 0.05 0. 14 + 0. 28 s i g . 0. 021 + 0 .014 s i g . b) 0. k ibarensis (Nov. 1972) Sporobolus py. A 5 0. 39 + 0 .02* P^ 0. 09 + 0. 02 P< 0. 025 + 0 .025 not B 5 0. 35 + 0 .04 0.05 0. 07 + • 0. 01 0.05 o . 025 + 0 .005 s i g . 1. p 0.05. * 95% confidence l i m i t s . 2. From Student's t - t e s t . Table L I I . Animal a c t i v i t y at termite mounds i n R.N.P. Dir e c t Observation I n d i r e c t Animal Species rji c 01 K to 0 •u 0> •P M-l XI C <u a It) c 1 H 0 0 0 K 0 3 0 c •r t XI o -a •rt o O M tn +l Cn 01 U U c •p JJ •H >i 4J ro tn C ra c r.t c m 3 3 ra •P ro •rt S C •rt T) 4J -rt TJ r.t •H S o o ra r-4 U •rt Tj •d c oi TJ C 0) A; 01 c a ~ XI C a) 3 o> 0! 3 >: 01 0 0 <« •r-t to 0) H XI 3 m o ai a) o c oi o o o © u Frl ro 3 O fc< s > a. a H Utfc. 4J a D Q H « S c c " O 0 0 IH Cn C -rt Cn •rt 0 G 3 4-> C 4J •rt O O •rt 4J O Cn rtg 3 ^ 3 •3 C a. n Q, 0 •0 H •rt -a e xi +J e x; C 4J c c rO 4-> to ra 4J 3 Ul H 3 u -H ai U -H O 01 0 0 B 3 s a CC s A f r i c a n elephant (Loxodonta africana) B u f f a l o (Syncerus c a f f e r ) Hippopotamus (Hippopotamus amphibius) Lion (Panthera leo) Uganda kob (Kobus kobus) Domestic goats Warthog (Phacochoerus aethiopicus) Baboon (Papio anubis) Vervet monkey (Cercopithecus aethiops) Banded mongoose (Hungos mungo) F i s h eagle (Cuncuma v o c i f e r ) Crowned eagle (Stephanoaetus coronatus) Maribou stork (Leptoptilos coumemf erus) Wattled plover T f t f r i l y x senegallus) Anteater chat (Myrmecocichla formicivora Red-necked spurfowl ( P t e r n i s t i s c r a n c h i l ) / / •* / / /* / • / /* • / • • * >five observations. Table L I I I . D e f e c a t i o n by medium s i z e d mammals on t e r m i t e mounds i n R.N.P. Mound Type N 1 % with Small Carnivore % with Baboon Droppings & L o c a t i o n Scats (Ishasha only) M. subhyalinus 80 6.3 Nyamagasani M. subhyalinus 53 11.3 Mweya M. subhyalinus 29 0 C r a t e r s M. subhyalinus 24 4.2 Kamulikwezi M. subhyalinus 24 0 Kasenyi M. b e l l i c o s u s 37 10.8 0 Ishasha 0. k i b a r e n s i s 18 0 16.5 Ishasha 1. Sample s i z e . Table LTV. Baboon faeces counts i n 40 meter X 40 meter area a t Ishasha c o n t a i n i n g four 0. k i b a r e n s i s mounds and two s m a l l bare T r i n e r v i t e r m e s mounds. Month On On On Ground Away T o t a l Number 0. k i b a r e n s i s T r i n e r v i t e r m e s from Mound of Droppings Mounds Mounds Nov. '72 2 0 0 2 Jan. '73 3 0 0 3 Feb. '73 1 0 0 1 March '73 0 1 0 1 A p r i l '73 0 0 0 0 May '73 1 0 0 a> June '73 1 0 0 I J u l y '73 0 0 1 I T o t a l 8 1 1 10 X = 9.91, d f = 2, p < 0.01 (on mounds versus not on mounds) \ cn \ cn \ CO \ \ r—' r-1 M o M (-• M M O D 0 P> CTi CTi cn co \ \ \ \ r—1 rt \ \ \ \ \ \-> r-1 —1 (-' fD - J o \ \ NJ NJ fD co co co co CO —1 CO —1 co O. rt fD I I I I I I I I I I I I I I I I I I I I I I I ^ I I I I I I * I I I I I I I I I Maerua e d u l i s Capparis tormentosa seed Securinega v i r o s a seed H o s l u n d i a o p p o s i t a seed Grass seed Veg. f i b r e s U n i d e n t i f i e d shrub/ t r e e seed 1 U n i d e n t i f i e d shrub/ t r e e seed 2 Non-Macrotermes. t e r m i t e s Non-Macrotermes a l l a t e s Formicidae A n n e l i d a U n i d e n t i f i e d c u t i c l e Table LVI. Germination t r i a l s for shrub seeds from baboon faeces and shrub seeds from ripe f r u i t at Ishasha. Sample Species No. Seeds/ No. Seeds 1 % Difference Fruit s Tested Germinated Germination Significances for %'s Control Hoslundia opposita 40 10 seeds inside intact f r u i t s 7 17.5 " L P> 0.1 Control Hoslundia opposita 39 10 seeds from f r u i t s 10 25. 6 -1 Control Control Tarrena graveolens Tarrena graveolens seeds from 10 intact f r u i t s 10 opened f r u i t s 0 0 0 1 0 l - 0.1 Control Securinega vinosa 20 intact f r u i t s 0 0 1 Control Securinega vinosa 86 20 seeds from f r u i t s 0 0 2 X = 54.67, Baboon Dung Securinega. vinosa 40 seeds 11 27.5 . p :< 0.001 Baboon Dung Capparis sp. 20 seeds 15 75.0 1. After ten days: 11th June 1973 - 21st June 1973. Table LVII. The rubbing of mounds by large mcutmals. Mound Type N* % Mounds Rubbed % Bare Mounds Rubbed % Vegetated Mounds and Location by Large Mammals Rubbed M. subhyalinus 129 0 0 0 Nyamagasani M. subhyalinus 118 1.7 1.7 0 Mweya M.. subhyalinus 65 '6.2 6.2 0 Craters M. subhyalinus 24 29.2 25.0 4.2 Kamulikwezi M. subhyalinus 24 16.7 16.7 0 Kasenyi M. bellicosus 46 17.4 17.4 0 Ishasha * Sample size Table LVIII. Vertebrate species occupying the interior of holed M. subhyalinus mounds. J * ™ 3 M 5 u n c ^ D e i d Vertebrate Species Actual Evidence for Occupation Site of Evidence in Mound Original Source of Entry Number or Alxve Found Into Mound N.D. Dead D3 Dead D12 Dead 2F Alive 1. Banded Mongoose Droppings + canine tooth 2. Insectivorous Droppings + grass nest Murid Rodent 3. Gecko 15-20 broken eggs, one containing dried embryo 1. Arvicanthis 2 nests + droppings niloticus 2. Banded Mongoose Droppings 3. Mus triton Live animal 4. Rana sp. Live animal 5. Unidentified Live animal 50 cm green snake 6. ' Unidentified bat Live animal 7. OrycteroDus afer Tail impression in soi l of burrow 1. Banded Mongoose Droppings 2. Lophuromys  sikapusi 1. Arvicanthis niloticus One large group of droppings Adult male + adult female with nest On compacted floor of mound Banded Mongoose burrowing interior Termite gallery. 30 cm Banded Mongoose burrovring below compacted floor Small termite gallery below Banded Mongoose burrowing compacted floor Galleries at top of mound On mound and inside Aard-vark burrows under mound In large burrow, then ran into small galleries above Large burrow 50 an into central compac-ted galleries Large burrow Large burrow Weathering Aardvark burrows Aardvark burrow Aardvark burrow Weathering Aardvark burrow Aardvark burrow Top of mound + hollow interior Termite gallery below mound Odontotermes shaft Cave-in + Banded Mongoose burrowing Rodent burrowing + cave-in Odontotermes shaft NK3 Alive None None None Aardvark burrow CN O ro Table LIX. Vertebrate species occupying the i n t e r i o r of M. b e l l i c o s u s mounds. Mound Mound Dead or A l i v e Vertebrate Species Found Actual Evidence f o r Occupation S i t e of Evidence i n Mound O r i g i n a l Source of Entry i n t o Mound II Dead 12 13 14 19 (4) 19 (7) 110 K l A l i v e A l i v e Dead Dead Dead Dead A l i v e 1. Small Carnivore Droppings 2. Small Mammal 3 large nests 3. Crocidura sp. None None Live young i n nest None None Small Vertebrate Droppings 2. Arvicanthis  n i l o t i c u s 3. Small Carnivore 1. Arvicanthis  n i l o t i c u s 1. Arvicanthis.  ni l o t i c u s " 1. Mus minutoides 2. Arvi c a n t h i s  n i l o t i c u s 1. Mastomys natalensis S k u l l + droppings + nest Droppings Adult p a i r + nest + droppings Live animal + droppings Live animal i n nest Live animal i n nest Adult specimen On top of collapsed g a l l e r i e s On top of collapsed g a l l e r i e s Termite g a l l e r y under mound None None Termite g a l l e r y Termite g a l l e r y On collapsed g a l l e r i e s Halfway down basal termite shaft i n side g a l l e r y Hollow i n t e r i o r G a l l e r y o f f basal termite shaft C e l l a r region Basal termite shaft 10cm-20cm basal termite shafts 10cm-20cm basal termite shafts 10cm-20cm basal termite shafts 20cm basal termite shaft Six 4cm-10cm basal shafts Termite shaft/rodent burrow Rodent burrow/weathering Termite shaft/weathering Termite shaft/weathering Weathering/rodent burrow Termite shaft Termite shaft Basal termite shaft Table LX. Summary of r e s u l t s from 0.. f u l l e r i nest e x c a v a t i o n s . Number of l i v e nests excavated 10 Number of l i v e nests excavated t h a t showed v e r t e b r a t e s i g n 1 Number of dead nests excavated 3 Number of dead nests excavated t h a t showed v e r t e b r a t e s i g n 3 Table LXI. D e t a i l s of excavation r e s u l t s from four O. f u l l e r i nests with signs of vertebrate occupation. Termite Mound Dead Vertebrate Species Actual Evidence for S i t e of Evidence i n Termite O r i g i n a l Source of Nest or A l i v e Found Occupation Nest Entry Number C7-9 Dead U n i d e n t i f i e d insectivorous rodent Droppings Hollow i n t e r i o r at base of shaft Termite shaft C10-4 Dead 1. Lemniscomys s t r i a t u s Omnivorous rodent droppings + grass nest with h a i r In large termite shaft Termite shaft C10-9 C9-6 A l i v e Dead 1. Arvicanthis n i l o t i c u s A r v i c a n t h i s n i l o t i c u s Herbivorous rodent droppings + shredded grass nest Herbivorous rodent droppings + shredded grass nest + l i v e adult On runway near shaft entrance Termite shaft and i n nest In g a l l e r y at side of v e r t i c a l shaft On runway near shaft entrance Termite shaft and i n nest At base of 80 cm v e r t i c a l shaft Trapped by entrance Table LXII. Results of 0. kibarensis mound excavations. Mound Number Mound Dead or A l i v e Vertebrate Species Found Evidence Found for Occupation S i t e of Evidence i n Termite Mound O r i g i n a l Source of Entry 111-4 p 1. Arvi c a n t h i s n i l o t i c u s Nest + hairs + droppings 30 cm beneath top of mound i n burrow Termite shaft + rodent burrow IS-20 A l i v e 1. Arvi c a n t h i s n i l o t i c u s Nest Droppings 35 cm i n side of mound In nest + on runways on mound Termite shaft + rodent burrow IL-2 A l i v e None None None 3 X 4 cm diameter termite shafts IT Dead 1. Mastomys natalensis' 2. Arvi c a n t h i s . n i l o t i c u s Live adult Droppings Burrow entrance In burrow + on runways Termite shaft + rodent burrow Termite shaft + rodent burrow Table LXIII. Details of 23 banded mongoose dens on Mweya peninsula. Den Den Location Topography Number Range in Size Probable Origin of Number of Den Site of of Entrances Den Holes i Entrances 1 Large burrows under + near M. subhyalinus mound Top of slope 4 • 30 — 35 cm Aardvark. burrows 2 5 burrows + 2 cave-ins near two thicketed M. subhyalinus mounds Flat high ground 7 30 — 90 cm Aardvark burrows 3 Large burrows under large M. subhyalinus mound Lowland l iable to seasonal flooding 5 28 — 36 cm Aardvark burrows 4 Large burrows under 2 M. subhyalinus mounds in bushed area Lowland l iable to seasonal flooding 3 30 — 35 cm Aardvark burrows 5 Large burrows under M. subhyalinus mound in bushed area Top of slope 3 30 - 35 cm Aardvark burrows 6 Burrows around M. subhyalinus mound Top of slope 15 10 - 35 cm Aardvark burrows + mongoose tunnelling 7 8 Erosion gulley Erosion gulley On steep slope On steep slope 7 ? 7 •> — 9 Erosion gulley On steep slope 7 ? -10 Hollow interior of thicketed M. subhyalinus mound On steep slope 2 20 — 25 cm Small carnivore burrowing 11 Caved-in subterranean M. subhyalinus On slope 4 < 10 cm Cave-in + small nest (dead) carnivore burrowing 12 Caved-in subterranean M. subhyalinus nest On slope 3 15 20 cm Cave-in + small carnivore burrowing 13 Under concrete sewage cover Flat high ground Many < 20 cm wide Man-made 14 Caved-in worn M. subhyalinus mound On slope 2 10 cm Small carnivore burrowing Table LXIII. D e t a i l s of 23 banded mongoose dens on Mweya peninsula , cont. Den Den Location Topography Number Range i n Size Probable O r i g i n of Number of Den S i t e of of Entrances Den Holes Entrances 15 Thicket On slope ? 7 ? 16 Burrow On slope 1 10 cm Small carnivore burrowing 17 Erosion gulley Top of steep slope ? ? -18 Densely thicketed termite mound On slope 2 .. 20 cm Small carnivore burrowing 19 Thicketed termite mound On slope 1 17 cm Small carnivore burrowing 20 M. subhyalinus mound ins i d e large Lowland l i a b l e to 1 15 cm Small carnivore burrowing t h i c k e t flooding 21 Underground nest cave-in, caused by Lowland l i a b i e to 1 20 cm Cave-in + small f a l l i n g E. candelabrum tree (in flooding carnivore burrowing thicket) 22 Odontotermes shafts under large dense Lowland l i a b i e to 4 < 10 cm Odontotermes shafts + th i c k e t flooding small carnivore burrowing 23 Hollow i n t e r i o r of worn M. subhyalinus On slope 6 10 - 24 cm Weathering + small mound carnivore burrowing 308 Table LXIVa. Burrow use by s i x c l a s s e s of medium-sized mammals on Mweya P e n i n s u l a : Nov. •72 June '73 • i -d £ CD o SH cu u 0 r H CO u to -p r H o 3 -p •H m 0 W xi G s tn CO tn 0 Ui G . M-l - H Tj 4-1 2 cu O O S 0) 0 >i 5-1 2 o u CO SH O tn 5-1 G u + o u £ to > -ri rn o CU & 0) -P CD O -P -H cu G xi -P O r H X! CO -H •H G T! cu -P o G tO g >iC e u H 5-1 G to 5H >1 M •P 3 to o O rrj to >i (0 5H G O 3 Q 2 53 CO. Ui u m O D EH Mound Burrow 169 6 0 1 1 5 2 22 37 Non-Mound Burrow 293 8 1 0 1 1 3 0 2 0 3 5 Table LXIVb. Monthly burrow usage expressed as percent of burrow-days or burrow-n i g h t s monitored. Mound Burrows Non-Mound Burrows Day Night Day + Night Day Night Day + Night Nov. '72 6.25 13.00 9.38 6.25 18.75 12.50 Dec. '72 4.16 3 8 3.32 6.25 2.50 15.00 8.75 Jan. '73 0 0 0 0 5.00 2.50 Feb. '73 0 16.67 8.34 0 23.30 11.67 Mar. '73 0 16.67 8.34 0 4.00 2.00 Apr . 17 3 25.0 12.50 18.75 2.50 5. 00 3 .75 May '73 0 14.58 7.29 0 10.00 5.00 Jun. '73 0 8.33 4.16 0 5.00 2.50 Means1 4.43 11.25 7.84 1.41 12.10 6.09 1. The mean percent usage of mound burrows was not s i g n i f i c a n t l y (p > 0.05) d i f f e r e n t from t h a t of non-mound burrows f o r days, n i g h t s or t o t a l s . Table LXV. Results of small mammal removal trapping from "shafted" and "non-shafted" gr ids i n the Crater Region, expressed as the number of animals caught per 120 t rap-n ight s . Grid Number and Date tn - P rd xi cn w CD S m u O CD -p u o CD - P xi fl e o 2 O rH CD rH - P rd • fd X ! in cn CD • e Cn •H cn CD cn in H tn cn •H •O S-l cn • rd A o tn •H rd • w IH o CD +> -H c 0 o Cn rd 0 rH rd - P o 0- -p CD in rH n in rd •H o -p fd -a rH 3 5H rH - P 5-1 rH -H fl rH rd •H Xi CD rd rd -p •H M •H -H -H 5-1 •H fl cn c - P in X ! CD CD o - P - P 0 H (d rd rd 5-1 53 6 a <: a a N EH EH U C2/Nov. ' 722 0 3 2 0 0 1 0 0 0 0 0 0 C l / N o v . ' 72 2 0 8 3 2 1 2 0 0 0 0 0 Of! C4/Dec. 1 72 1 19 1 0 4 0 0 3 3 0 0 0 0 0 C3/Dec. ' 72 2 -i 10 3 2 0 0 0 0 0 0 0 0 1 C6/ Jan . ' 72 1 4 7 4 2 0 0 0 0 0 0 0 2 0 C5/ Jan . 1 73 2 28 5 4 0 0 0 1 0 0 0 0 0 C7/Feb. 1 7 3 2 98 17 13 3 1 0 0 0 0 0 0 0 C8FEebl 7 3^ 13 15 13 2 0 0 0 0 0 0 0 0 CIO/Mar. ' 7 3 1 23 25 11 3 6 0 2 1 0 2 0 0 C9/Mar. 73 2 4 8 3 2 3 0 0 0 0 0 0 0 C l l / A p r • 73 1 39 19 5 8 0 2 0 0 1 3 0 0 C12/Apr. ' 73 2 8 12 8 4 0 0 0 0 0 0 0 0 C14/May 73l 23 17 4 7 0 2 1 0 1 2 0 0 C13/May 73 2 0 10 3 3 1 0 0 1 0 2 0 0 C16/Jun ' 7 3 1 33 18 6 4 4 0 2 0 1 1 0 0 C15/Jun . *73 2 0 13 5 1 0 0 4 0 .3 0 0 0 C18/Jul . ' 7 3 1 33 2 0 1 0 0 0 0 0 1 0 0 C17/Jul . ' 73 2 0 7 2 3 0 1 1 0 0 0 0 0 1. "shafted" g r i d 2. "non-shafted" g r i d Table LXVI. C o r r e l a t i o n between the numbers of sm a l l mammals caught and the numbers of Odontotermes s h a f t s o c c u r r i n g on 16 1 t r a p p i n g g r i d s i n the C r a t e r Region. Small Mammal C o r r e l a t i o n S i g n i f i c a n c e Group C o e f f i c i e n t A l l s m a l l mammals +0 .33 p > 0 . 0 5 D i u r n a l +0 .22 p > 0.05 Murid Rodents N o c t u r n a l +0 .38 p > 0.05 Murid Rodents M. n a t a l e n s i s + 0.47 p < 0.05 L. s t r i a t u s + 0.21 p > 0.05 1. J u l y g r i d s omitted s i n c e they were the June g r i d s r e t r a p p e d ; see t e x t . CN H cn T a b l e L X V I I . S m a l l mammal t r a p p i n g : m o n t h l y 1 c o m p a r i s o n s o f c u m u l a t i v e p e r c e n t c a p t u r e s on t r a p s r a n k e d w i t h r e s p e c t t o d i s t a n c e f r o m Odontotermes s h a f t s , and e x p e c t e d c u m u l a t i v e p e r c e n t c a p t u r e s a s s u m i n g e q u a l c h a n c e o f c a p t u r e a t e a c h t r a p p o i n t . G r i d Number and d a t e 4J x: rH rH 3 rH ro ro rH e u 0 co. e ui crj in Vi rH rH m ro f i rH ro e •u e 3 B o ro 2 S EH S 01 •H 0) 0) m 3 3 c 0 4J <u •H c ro rH •U 0 •rl ro 0 4J u •P rH •r| +> ro •H IH (fl c c 4J r3 s < s ai ro . 0) 01 ai T3 01 •rl ^ 01 •r| ro • 01 3 0 CH a 3 a 4J 01 0) rH ro a TJ rH M c rH ro •rl •H •H •rl u U 01 8 XI o HI 4J JJ • • • ro ro E B EH C3/Dec.'72 3 D 3=-.45,p> .2 . N=0 N<3 N=0 N=0 N=0 N=0 N=0 N=0 N=0 N<3 C4/Dec.'72 10 D=-.28,p<. 2 N=0 D=-.28,p>. 2 N=0 D=- 50,p> 2 D=+.24,p> 2 N=0 N=0 N=0 N=0 . N=0 C5/Jan.'73 5 D=-.4 3,p>. 2 N=0 D=-.4 3,p>. 2 N=0 N=0 N<3 N=0 N=0 N=0 N=0 N=0 C6/Jan.'73 4 D=+.18,p>. 2 N=0 D=-.32,p>. 2 N=0 N=0 N=0 N=0 N=0 N=0 N<3 N=0 C7/Feb.'73 17 D=+.10,p>. 2 D=-.35 n-» 7 , D=-.14,p>. 2 N<3 N=0 N=0 N=0 N=0 N=0 N=0 N=0 C8/Feb.'73 15 D=+.10,p>. 2 p> . L N<3 D=+.12,p>. 2 N=0 N=0 ' N=0 N=0 N=0 N=0 N=0 N=0 C9/Mar.'73 8 D=+.15,p<. 2 N<3 D=-.35,p>. 2 D=+.78, N=0 N=0 N=0 N=0 N=0 N=0 N=0 p<0.05 CIO/Mar.'73 25 D=-.13,p>. 2 D=+.8, D=-.22,p>. 2 D=-.5,E <.l N=0 N<3 N<3 N=0 N<3 N=0 N=0 p<0.05 N <3 C l l / A p r . ' 7 3 19 D= + . 13, p> . 2 p>.2 D=+.4,p<.2 N=0 N=0 N=0 N<3 D=-.45,p>.2 N=0 N=0 C12/Apr.'73 12 D=+.22,p> 2 p >.2 D=+.18, p 2 N=0 N=0 N=0 N=0 N=0 N=0 N=0 N=0 C14/May"7 3 17 D=+.13,p>. 2 D=-.48 , D = + « 5 3, p<. 1 N=0 N<3 N<3 N=0 N<3 N<3 N=0 N=0 C16/Jun.'73 18 D=+.08,p>. 2 P < • £• D=+.58 p < . l , D=-.39,p<. 2 D=-.38, p> . 2 N=0 N<3 N=0 N<3 N<3 N=0 N=0 T o t a l G r i d s 153 D=+.05 D=+.18 D=-. 7 D=+.15 D= =-.09 D=+.14 N<3 D=+.65 D=+.24 N<3 N<3 p>.2 p>.2 p> . 2 p>.2 P>.2 _ p>.2 p < . l p>.2 1. 2. 3. O n l y f o r g r i d s w i t h s h a f t s . U s i n g t h e Ko l m o g o r o v - S m i r n o v one sample t e s t ( S i e g e l 1956) D = maximum d i f f e r e n c e (above (+) o r below (-)) between o b s e r v e d and e x p e c t e d c u m u l a t i v e c a p t u r e s . Table LXVIII. S i g n i f i c a n c e of p o s i t i v e d ifference between observed and expected number of small mammals caught i n traps nearer to shafts than the distance c o i n c i d i n g with the maximum p o s i t i v e d i f f e r e n c e i n cumulative percent small mammal captures. Number of "D" Trap/Shaft Number Expected Number Expected X 2 Degrees S i g n i f i c a n c e Animals Caught Distance c o i n c i - Animals Number Caught Number of ding with maxi- Caught to be > "D" to be Freedom mum increase of < "D" Caught < "D" from Caught > "D" cumulative per- from Shafts cent captures Shafts from from over that expec- Shafts Shafts ted (m) L. s t r i a t u s 33 6.0 11 7.6 22 25.0 1.92 1 p > 0.05 M. nat a l e n s i s 76 1.5 8 3.8 68 71.9 7.82 1 p < 0.01 A l l small mammals 153 !- 5 14 6.4 139 146.6 9.31 1 p < 0.01 "3< H ro ' Table LXIX. Snap-trapping resul t s i n the Crater Region immediately after the January 1973 burn. Location Number Captures on Four Days Fol lowing Burn of Traps Day 1 Day 2 Day 3 Day 4 Under Bushes 30 4 3 6 2 Near Six Odontotermes shafts 6 0 0 0 0 Controls (open burnt grassland) 36 0 0 0 0 Table LXX. Number of rodents captured during 36 trap-days and nights each month, l i v e - t r a p p i n g i n the Crater Region a f t e r the January 19 73 burn. Month Number of Weeks a f t e r January Burn Area 1 With Odontotermes Shafts L. s t r i a t u s M. natalensis T o t a l Area 2 Lacking Odontotermes Shafts S i g n i f i c a n c e L. s t r i a t u s M. n a t a l e n s i s T o t a l Totals Feb.'73 Mar. 173 Apr.'73 2 7 11 7 8 20 9 15 3 16 23 23 0 15 9 7 11 8 7 3.52 p < 0.1 26 0.18 p > 0.1 17 0.90 p > 0.1 Table LXXI. Small mammal stomach analyses. Food Item o a ra rrj •rt 0 CM n TJ in t l 4J <u •rt 0) Ml-rt II JJ rH II 4J am ra ra ra Sl-U z EH > Z EH Macrotermes s o l d i e r termite 0 1 Odontotermes s o l d i e r termite 0 2 Pseudocanthotermes s o l d i e r termite 3*2 7*2 Trinervitermes s o l d i e r termite 0 0 Macrotermes worker termite 1 0 Non-Macrotermes worker termite 7*2 9*3 Formicidae 1 3 U n i d e n t i f i e d c u t i c l e 5*1 14* 6 Vegetation 0 0 Vertebrate remains 0 1 P a r a s i t i c nematode 7 3 8*a 0 2*1 11 8*5 6*1 I * 1 0 0 1 i * l i o b i 2 i i o o o o o o o o o 2*1 0 0 0 2 1 2*2 0 1 4 1 4 1 0 0 0 0 0 0 0 0 0 1 0 0 0 *( ) = number of times item was dominant. Table LXXII. 0. k i b a r e n s i s mound rodent droppings t r a n s e c t s : A p r i l 1973. Mound Number 1 2 3 4 5 Height (M) 1.75 1.75 1.33 1.50 1.66 Mean Diameter (M) 15 11 16 10 15 Number of Holes i n Mound 6 3 3 5 4 Presence of Rodent "Runways" x x x / / Presence of Thicket Vegetation / / / / / Termites A c t i v e / / x x / Number of Droppings^ on Mound Top 9 4 8 3 10 2 Number of Droppings on Mound Side 1 2 - - -Number of Droppings on Mound Base 0 0*,1* 12 20*,2* 1 Number of Droppings 5M Away 0 1*,2* 0 0*,0* 7 Number of Droppings 10M Away 0 1*,0* 0 0 4 Number of Droppings >50M Away 0*,0* 0*,0* 0*,0* 0*,0* 0*,0* 1. A l l droppings recorded i n randomly p l a c e d 0.25M^ Quadrat. R e p l i c a t e s . 318 Table L X X I I I . 0. k i b a r e n s i s mound rodent droppings t r a n s e c t s : comparison between holed and non-holed mounds -March 1973. Number of Mounds Number of Mounds with Holes without Holes T o t a l Number of Mounds wit h Rodent Droppings Number of Mounds without Rodent Droppings T o t a l 12 x (8.3^) 1 (4.7) 13 2 (5.8) 7 (3.3) 14 22 1. 2. Observed Expected X 2 = 11.29, df = 1, p < 0.001 Table LXXIV. Heights of Sporobolus pyramidalis l ea f blades on 0. k ibarens i s mounds and contro l areas at Ishasha. Sample Date Number of Number of Mean Height (cm) S igni f icance of Mounds or Grass Mo un d/C on t ro1 Control Blades Differences . Areas Measured Mound Nov.'72 4 22 33.5 ± 4 . 8 * t = 3.65 Control Nov. 1 72 4 22 44. 7 ± 5.8 p < 0.001 Mound Jan. '73 4 22 27.8 ± 6.8 t = 2.42 Control Jan. 1 73 4 22 36.1 ± 3.9 p < 0.05 Mound Feb. * 73 4 22 22.6 ± 2.7 t = 6.13 Control Feb. '73 4 22 36.7 ± 4.1 p < 0.001 * 95% confidence l i m i t o CN cn Table LXXV. D i s t r i b u t i o n of 0. af e r digs and large burrows i n t e r m i t a r i a from ground transects. Termite Nest Type Area Number of Termite Nests Examined Number of Transects Examined Percentage of Mounds + Aardvark Digs Percentage of Mounds + Large Burrows Mean Number of Mounds + Digs Ha" 1 Range i n Number + Digs Ha" 1 M. subhyalinus M. b e l l i c o s u s 0. f u l l e r i 0. k i b a r e n s i s Nyamagasani 129 Mweya 118 Craters 65 Kamulikwezi 24 Kasenyi 24 Ishasha 46 Craters (CE, outer region) 14 Craters (CG, c e n t r a l region) 5 Ishasha 32 10 10 9 4 . 3 11 56.2 19.5 4.4 12.5 16.6 2.2 8.5 5.9 4.6 16.6 12.5 13.0 14.3 0 0 7.1 3.4 1.1 0.8 1.3 0.1 2 - 1 5 0 - 1 6 0 - 4 0 - 2 1 - 2 0 - 2.5 H CN ro T a b l e LXXVI. R e s u l t s o f o r y c t e r o p u s d i g t r a n s e c t s . a) Mweya T r a n s e c t T r a n s e c t T o t a l Food T o t a l Food D e f e c a t i o n Food D i g Food D i g F o o d D i g % F o o d D i g Number o f Number o f Number A r e a (Ha) D i g s H a _ l D i g Groups D i g s H a ~ l Groups Groups H a - 1 Groups Groups i n T u n n e l s T u n n e l s H a - 1 H a - 1 i n i n t o S u b t e r - i n t o F o r m i - Mounds i n t o E l s e w h e r e Mounds r a n e a n T e r - c i d a e N e s t s Mounds Ha-1 m i t e N e s t s Ha-1 MAI 1.50 110.7 13.3 4.7 4.0 3.3 . 0.67 40.0 0.7 0 MA2 1.30 52.3 13.1 3.1 3.8 0.8 0 29.4 1.3 0 MA3 1.44 17.4 9.7 0.7 1.4 1.4 0 14.3 1.4 0.7 MA4 0.80 13.8 6.3 0 0 2.5 0 0 0 2.5 Means of Above 1.26 48.6 10.6 2.4 2.6 2.0 0.2 20.9 0.9 0.8 CN CN cn Table LXXVI. Continued. b) Crater Region Transect Transect T o t a l T o t a l T o t a l Food Dig Food Dig Food Dig % Food Dig % Food Dig Number Number Number Area Number Food Defeca- Groups Groups Groups Groups of -.' Dig t i o n H a - 1 i n t o H a - 1 i n t o H a - 1 i n t o i n t o Odonto- M. sub- Odonto- Mounds Food Digs Ha-1 Groups Digs H a - 1 H a - 1 termes hyalinus termes Nests Mounds Subter-ranean . Termite Nests Groups i n t o Odonto- termes Nests of of Tunnels Tunnels i n t o into Mounds Ha-1 Odonto-termes Nests Ha-1 Number Tunnels Elsewhere CA1 1.0 ' 3.0 3. .0 1.0 0 1, .0 0 33.3 0 0 0 1.0 CA2 1.0 12.0 5. .0 0 0 1, .0 2;0 20.0 0 0 2.0 1.0 CA3 0.5 40.0 8, .0 4.0 0 2. .0 4.0 25.0 0 2 2.0 0 CA4 0.9 18.9 6. . 7 1.1 0 1. .1 2:2 16.7 0 0 1.1 16.7 Means of Above 0.9 18.5 5. ,5 1.7 0 1. .3 2;1 23.7 0 0.5 1.4 4.7 Table LXXVII. Deta i l s of 12 successful orycteropus watches. • rH G X ! G rd 0 o m -H g 1 HH -H G -p •H Xi 0 -P 0 fd G Q rd •H X ! -H •O X ! > . . -P G •P T i -P CD -P H . -H CD CD rd Cn CD H CD CO > Cn CD G G ti -P g Cn G g CD g u SH G cn •H 0 G fd •H CD CD -H -H •H -H CD CD X I g 0 0 Q EH eq ^ EH fa EH fa CO J O a u T! rH CD CD fd o G fd CD -p > cn rd •H H g •H G <: >i's CD CD o G fd (U K •P > cn rd U M CD a o 5H CD X ! g G O •H •P rd CO > t) >H O CD -H CO SH d X. CD Q E H X ) ^ - ' Q E H O , — S O C H fd g •r| G <: & o to O 20- 2- 73 23. 15. 210 24. 05 4. 5 3/4, c lear 195 9600 1 Two obs. large pe r iods / lo s t near burrows each time 22- 2-•73 23. 30 180 24. 30 5. 0 3/4, cloudy 225 2700 One obs. p e r i o d / l o s t i n bush 23- 2- 73 24. 30 180 1. 30 1. 0 No moon, cloudy 120 7200 One obs. bush period/burrow i n 24- 2- 73 23. 30 160 i l . 00 30. 0 No moon, cloudy 560 13432 One obs. large period/see t ex t , burrow 27- 2- 73 22. 30 60 22. 35 2. 0 No moon, cloudy 60 1800 One obs. gu l ley per iod/eros ion 23- 5- 73 21. 30 165 23. 10 20. 0 Moon, cloudy 190 570 One obs. text period/burrow, see 28- 6- 73 22. 00 165 22. 57 23+35 No moon, c lear 40+860 14743 Two obs. period/burrow i n 4064 bush, see text 30- 6- 73 22. 00 140 22. 55 30. 0 No moon, c lear 575 Two obs. periods/bush 19- 7- 73 22. 00 120 22. 50 35. 0 No moon, c lear 490 875 One obs. text period/burrow, see 22- 7- 73 23. 00 150 24. 15 10. 0 No moon, c lear 60 360 One obs. slope period/Lake Edward 25- 7- 73 23. 30 150 23. 50 35. 0 No moon, c lear 240 4412 One obs. period/burrow 27- 7- 73 23. 30 180 24. 30 30. 0 No moon, c lear 345 "790 One obs. s lope period/Lake Edward 1. Based on f i r s t observation only . 2. Based on f i r s t 25 minutes before f l ee ing occurred. 3. Based on second longer observation period only . 4. Based on t o t a l distance covered over 1 h r . 25 minute per iod . Table LXXVIII. Deta i l s of 16 stops which an animal made during 35 minutes of observation on 19th July 1973. Dig Type Number of Digs Made Maximum Time Spent at One Dig Type Locat ion of Digs Mound Bush Base Open Unsuccessful 2 Single 11 Groups 3 10 seconds 2 minutes 6 minutes 0 0 0 0 0 2 2 11 1 Totals 16 14 Table LXXIX. Analysis of orycteropus droppings. Presence/Absence 1 Dated Droppings from Mweya \ \ \ ro r o c- « ^ r> p p r- \ — i n m x o > \ ^ W \ \ i o ro r o * ) ' . ' ? TJ< i-i —. I I I i n i n i o o i Macrotermes / / / / / / / V Odontotermes / / / / / Trinervitermes / j / Pseudocanthote'rmes / Other Formicidae / / / / / / Other U n i d e n t i f i e d C u t i c l e / / Red Seeds 2 / / / / V / cn r -cn cn r -\ \ \ vx> \£> o> o \ rH CM cn CNCO o\ 1 O CN CM CN / / / / / % Composition Macrotermes S o l d i e r s Odontotermes S o l d i e r s Other Isoptera Soldiers Macrotermes Workers Other Isoptera Workers Formicidae N (no. mandibles sampled) Mean 0.9 0.6 0.4 1.2 0.5 0.4 0 1.1 1.1 0 0.5 , 0 0 13.6 0 1.35 0. 0 0 1.2 0 0.9 2.5 2.8 0.6 3.4 1.0 1.7 0 0 0 0.94 0 0 0 0.8* 0.25 ** 0 0.6** 0 0 0.3** 0 0 0 0 0 0.13 74.8 57.1 3.4 45.9 19.5 5.6 19.9 63.2 87.2 1.3 1.4 6.0 11.8 8.5 8. .8 27.60 0 5.1 5.1 19.7 3.4 1.7 21.1 27.7 4.7 93.7 2.9 2.5 0 0 0 12.50 24.3 37.2 91.1 31.1 76.3 91.5 55.9 5.3 6.4 1.3 94.2 89.7 88.2 78.0 91. . 2 57.45 107 301 527 244 410 468 161 361 468 149 416 116 170 59 6S ! 1. S o l d i e r s only, except Macrotermes workers-. 2. See text. * Pseudocanthotermes sp. ** Trinervitermes sp. Table LXXIX. Continued. Undated, C r a t e r Regi on Undated, Nyamaga s a n i in CA VO t^* m r—i rH 1 i i i 1 rH Presence/Absence 1 in <! CTl < 00 <! in 00 rt! CN u CQ Z Macrotermes / / / / / / Odontotermes / / A y / T r i n e r v i t e r m e s / / Pseudocanthotermes Other / / Formicidae / / / / Other U n i d e n t i f i e d C u t i c l e / / Red Seeds2 % Composition Mean Mean Macrotermes S o l d i e r s 0 0.7 5.5 0.5 1.4 1. 62 1. 8 6.9 1.2 3. 3 Odontotermes S o l d i e r s 2.3 6.0 6.8 1.5 0.4 3. 4 0 0 0 0 Other I s o p t e r a S o l d i e r s 2.3 : k* o 0 0 . 4 * / . 4 + 0. 62 0 0 0 0 Macrotermes Workers 9.1 3.4 23.5 2.0 56.8 18. 96 31. 33 92.1 96.9 73. 4 Other I s o p t e r a Workers 42.4 84.5 52.7 5.7 11.2 39. 3 2. 8 0 0 0. 93 Formicidae 43.2 5.4 11.6 90.3 29.5 36. 0 64. 1 1.0 1.9 22. 33 N (no. mandibles sampled) 44 149 311 392 285 39 '6 202 161 1. S o l d i e r s only, except Macrotermes workers 2. See t e x t . * Pseudocanthotermes sp. ** T r i n e r v i t e r m e s sp. t ' Basidentitermes sp. LITERATURE CITED Agricultural Research Council. 1965. The Nutrient Requirements of Farm Livestock. No. 2, Ruminants: Technical Reviews. London: Her Majesty's Stationary Office. Anon. 1971. Uganda National Parks Handbook. Longman, p. 160. A.O.A.C. 1970. Official methods of analysis of the Association of Official Analytical Chemists. 11th ed. Washington, D.C. Araujo, R.L. 1970. Termites of the Neotropical Region. In "Biology of Termites", ed. Krishna & Weesner. pp. 527-575. Aumann, G.D. 1965. Microtine abundance and soil sodium levels. J. Mammal 46(4): 594-604. Aumann, G.D. and J.T. Emlen. 1965. Relation of population density to sodium availability and sodium selection by microtine rodents. Nature, 208: 198-199. Bentley, J.C. 1969. Dependence of desert rodents on winter annuals and precipitation. Ecology 50(4): 721-724. Bequaert, B.J. 1922. The Predacious Enemies of Ants. Bull. Amer. Mus. Nat. Hist. 45: 271-329. Best, T.L. 1973. Ecological separation of three genera of pocket gophers (Geomyidae). Ecology 54(6): 1311-1319. Bett, I.M. and G.P. Fraser. 1958. Rapid micro method for determining serum calcium. Biochem. J. p. 13. Bett, I.M. and G.P. Fraser. 1959. Micro-determination of serum calcium. Clin. Chim. Acta. 4: 346-356. Bigourdan, J. 1950. Sur quelques caracteres et habitudes de l'orycterope. 1st. Int. W. Afr. Cont. Dakar: 186-188. Blair-West et al. 1968. Physiological, morphological and behavioural adaptation to a sodium deficient environment by wild native Australian and introduced species of mammals. Nature, Lond. 217: 922-928. Boer, P.J. den. 1968. Spreading of risk and stabilization of animal numbers Acta, biotheor. 18: 165-194. 328 Bodot, P. 1967a. Cycles saisonniers d'activitie collectives des Termites des savanes de Basse Cote-D'lvoire. Insectes Soc. 14(4): 359-388. Bodot, P. 1967b. Etude ecologique des termites des savanes de Basse Cote-D'lvoire. Insectes Soc. 14(3): 229-258. Bouillon, A. 1970. Termites of the Ethiopian Region. In "Biology of Termites" ed. by Krishna and Weesna. Bouillon, A. et S. Kidieri. 1964. Repartition des termitieres de Bellicositermes bellicosus rex Grasse et Noriot, dans L'Ubangi, d'apres les photos aeriennes, et correlations ecologique qu'elle oevele, note preliminaire. In "Etudes sur les Termites Africains", ed. Bouillon: 373-376. Bouillon, A. and G. Mathot. 1965. Quel est ce termite Africain? ed. de l'universite Leopoldville. p . 115. Boyer, P. 1958. Influence des remariements par le termite et de l'erosion sur revolution pedogenetique de la termitie^re epigee de Bellicositermes rex. C r . hebd. Seanc. Acad. Sci. Paris, 247: 749-751. Bradley, R.M. 1971. Warthog (Phacochoerus aethiopicus Pallas) burrows in Nairobi National Park. E. Afr. Wildl. J. 9: 149-152. Bredon, R.M. and J. Wilson. 1963. The chemical composition and nutritive value of grasses from semi arid areas of Karamoja as related to ecology and soil types. E. Afr. Agric. J. 29: 134-142. Brown, J.H., Lieberman, G.A. and Denyler, W.F. 1972. Woodrats and Cholla: Dependence of a small mammal population on the density of cacti. Ecology 53 (2): 310-313. Buechner, H.K. 1961. Territorial behaviour in Uganda Kob. Science, 133: 698-699. Burtt, B.D. 1942. Some East African Vegetation Communities. J. Ecol. 30: 65-146. Clark, P.J. and F.C. Evans. 1954. Distance to nearest neighbour as a measure of spatial relationships in populations. Ecology 35: 445-453. Coaton, W.G.H. 1962. Nesting habits and mounds of the termites of Northern Rhodesia. Afr. Wild. Life, 16: 61-70. Cook, S.F. Jr. 1959. The effects of fire on a population of small rodents. Ecology, 40(1): 102-108. Cowles, R.B. 1930. The life history of Var anus niloticus (Lin) as observed in Natal, South Africa. J. Ent. Zool. 22: 1-31. Dale, F.H. 1939. Variability and environmental responses of the kangaroo rat, Dipodomys heermanni saxatilis. Amer. Mid. Nat. 22: 703-731. De Bont, A.F. 1964. Termites et densite d'oiseaux. In "Etudes sur les termites Africains", ed. Bouillon, A. Unexo. pub. 274-283. Delany, M.J. 1964. A study of the ecology and breeding of small mammals in Uganda. Proc. Zool. Soc. Lond. 142: 347-369. Delany, M.J. and B.R. Neal. 1969. Breeding seasons in rodents in Uganda. J. Reprod. Fert., Suppl. 6: 229-235. Department of Lands and Surveys. 1962. Uganda Atlas. Entebbe: Government Printer. Derscheid, J.M. 1925. Un Orycterope de l'Aruwimi. Revue Zoologique  Africaine. Vol. 13(L): Bull, du Cercle Zoo. Congo:!ais: 13. Dorst, J. and P. Dandelot. 1970. A field guide to the larger mammals of Africa. Collins, p. 287. Du Toit et al. 1940. A study of the mineral content and feeding value of natural pastures in the Union of South Africa (Final Report). Onderstepoort J. Vet. Sci. 14: 123-329. Edington, J.M. and M.A. Edington. 1972. Spatial patterns and habitat partition in the breeding birds of an upland wood. J. Anim. Ecol. 331-357. Edmond-Blanc, F. 1947. A contribution to the knowledge of the Cambodian wild ox or kouproh. J. Mammal. 28(3): 245-248. Emerson, A.E. 1928. Termites of the Belgian. Congo and the Cameroon. Bull. Amer. Mus. Nat. Hist. 57: 401-574. Emerson, A.E. 1956. Regenerative behaviour and social homeostasis of termites. Ecology, 37(2): 248-258. Field, C.R. 1968. A comparative study of the food habits of some wild ungulates in the Queen Elizabeth National Park, Uganda. Preliminary report. Symp. Zool. Soc. Lond. No. 21: 135-151. Field, C.R. 1970. Observations on the food habits of tame warthog and antelope in Uganda. E. Afr. Wildl. J. 6: 1-17. 330 Field, C R . 1972. The food habits of wild ungulates in Uganda by analyses of stomach contents. E. Afr. Wildl .J. 10: 17-42. Foran, I. 1961. The Elusive Ant-Bear. Wildl. (Nairobi) 3(3): Fossati, L. 1937. Abitudini dell'Orirteropo d'Eritrea. notizie e varieta. Natura 28(1) Milan: 16-23. Frechkop, S. 1938. Exploration du Pare National de la Kagera. Inst. des pares. Nat, du Congo Beige. Fascicule 1. Mammiferes. p. 56. Geigy, R. 1949. La capture d'un cochon de terre au Tanganyika pour le Jardin Zoologique de Bale. Zoologischen Garten Basel, 77, Jahrcshericht des verwaltungsrates an die Aktoinare. Genelly, R.E. 1965. Ecology of the common mole-rat (Cryptomys holtertotus) in Rhodesia. J. Mammal. 46: 647-665. Ghilarov, M.S. 1962. Termites of the U.S.S.R., their distribution and importance. In "Termites in the Humid Tropics", New Dehli Symposium. UNESCO, Paris, pp. 131-135. Glover, P.E. et al. 1964. The chemical composition of Kenya browse and pasture herbage. E. Afr. Wildl. J. 2: 86-121. Glover, P.E. , E.C. Trump, L.E.D. Wateridge. 1964. Termitaria and vegetation patterns on the Loita Plains of Kenya. J. Ecol. 52: 367-377. Glover, P.E. 1949. Some aspects of the ecology of the Abercorn district of Northern Rhodesia. Unpublished M.S., in Glover et al 1964. Golley, F.B. and H.K. Buechner. 1968. IBP Handbook No. 7. A Practical Guide to the Study of the Productivity of Large herbivores. Blackwell Sci. Pub. Goodland, R.J.A. 1965. On Termitaria in a savanna ecosystem. Can. J.  Zool. , 43: 641-650. Greaves, T. 1962. Studies of foraging galleries and the invasion of living trees by Coptotermes acinaciformis and C_. brunneus (Isoptera). Aust. J.Zool. 10: 630-651. Griffith, G. 1938. A note on termite hills. E. Afric. Agric. For. J. 4: 70-71. Grinnell, J. 1928. Presence and absence of animals. Univ. Calif. Chronicle. Oct. 1928: 429-450. 331 Hardy, R. 1945. The influence of types of soil upon the local distribution of some mammals in Southwestern Utah. Ecol. Monogr. 15: 71-108. Harris, W.V. 1956. Termite mound building. Insectes Soc. 3: 261-265. Harris, W.V. 1966. The role of termites in tropical forestry. Insectes Soc. 13(4): 255-265. Harris, W.V. 1971. Termites their recognition and control. Longman, p. 186. Harrop, J.F. 1960. The soils of the Western Province of Uganda. Uganda  Dept. Agric. Memoirs., Series 1, No. 6,. p. 105. Hartmann, F. 1969. From another age. Africana, 3(10): 17-18. Hediger, H. and J. Verschuren. 1948. Exploration des Pares Nationeaux du Congo Beige. Fasc. I, Observations sur la Psychologie Animale: 31-72. Hesse, P.R. 1955. A chemical and physical study of the soils of termite mounds in East Africa. J. Ecol. 43(2): 449-461. Hesse, P.R. 1957. Fungus combs in termite mounds. E.Afr. Agric. J. 23: 104-105. Holdaway, F.G. and F.J. Gay. 1948. Temperature studies of the habitat of Entermes exitosus with special reference to the temperature within the mound. Aust. J. Suert. Res. B l : 464-493. Howard, D.A., M.L. Burdin, and G.H. Lampkin. 1962. Variation in the mineral and crude protein content of pastures at Muguga in the Kenya Highlands. J. Agric. Sci. (Camb.), 59: 252-256. Howell, A.B. 1924. Theories of distribution - acritique. Ecology 5(1): 51-53. Howse, P.E. 1970. Termites: A study of social behaviour. Hutchinson University Lib. Jackson, G. and J.S. Gartlan. 1965. The flora and fauna of Lolui Island, Lake Victoria. A study of vegetation, men and monkeys. J. Ecol. 53(3): 573-597. Jarman, P.J. 1972. The use of drinking sites, wallows and salt licks by herbivores in the flooded Middle Zambezi Valley. E. Afr. Wildl. J. 10(3): 193-210. Jeffrey, A.C. 1973. Seasonal changes in the food habits, reproduction and body composition of two equatorial rodents. M.Sc. thesis. Makerere University. J e f f r e y , A.C. 1973. Seasonal changes i n the food h a b i t s , r e p r o d u c t i o n and body composition of two e q u a t o r i a l rodents. M.Sc. t h e s i s Makerere U n i v e r s i t y . Juko, C D . and R.M.Bredon. 1965. The chemical composition of leaves and whole p l a n t s as i n d i c a t o r s of the range of n u t r i e n t s a v a i l a b l e f o r s e l e c t i v e g r a z i n g by c a t t l e . Trop. A g r i c u l t u r e T r i n . 38: 179-187. Kemp, P.B. 1955. The t e r m i t e s of North-Eastern Tanganyika; t h e i r d i s t r i b u t i o n and b i o l o g y . B u l l . Ent. Res. 46: 113-135. Kingdon, J . 1971. Ea s t A f r i c a n mammals. An a t l a s of e v o l u t i o n . V o l . 1 . Academic P r e s s . Kozlova, A.V. 1951. Accumulation of n i t r a t e s i n t e r m i t e mounds i n Turkmenia. Pochnovedenie 1951 (10): 626-631. Lamprey, H.F. 1963. E c o l o g i c a l s e p a r a t i o n of the l a r g e mammal s p e c i e s i n the Tang i r e Game Reserve, Tanganyika. E. A f r . W i l d l . J . (1): 63-92. Langdale-Brown, H.A. e t a l . 1964. The v e g e t a t i o n of Uganda and i t s b e a r i n g on land use. Uganda government p u b l i c a t i o n : p.159. Lawrence, G.E. 19 66. Ecology of v e r t e b r a t e animals i n r e l a t i o n t o c h a p p a r r a l f i r e i n the S i e r r a Nevada f o o t h i l l s . Ecology 47. Lee, K.E. and T.G.Wood. 1971. Termites and s o i l s . Academic P r e s s . 251p. Lock, J.M. i970a. The grasses of Queen E l i z a b e t h N a t i o n a l Park. Uganda J o u r n a l , 34 (l):49-63. Lock, J.M. 1970b. Unpublished v e g e t a t i o n map of Queen E l i z a b e t h N a t i o n a l Park. Lock, J.M. 1970c. Unpublished l i s t of t e r m i t e s p e c i e s c o l l e c t e d i n Queen E l i z a b e t h N a t i o n a l Park. Long, M.I.E., B. M a r s h a l l and D.D. Thornton. 1969. N u t r i t i v e v a l u e of grasses i n Ankole and the Queen E l i z a b e t h N a t i o n a l Park Uganda. I I - Crude p r o t e i n , crude f i b r e and s o i l n i t r o g e n . Trop. A g r i c u l t u r e T r i n . 46: 31-42. Luscher, M. 1956. Die Lufterneueerung im nest der t e r m i t e Macrotermes n a t a l e n s i s ( H a v i l a n d ) . I n s e c t e s Soc. 3 : 273-276. Luscher, M. 1961. A i r - c o n d i t i o n e d t e r m i t e n e s t s . S c i e n t . Am. 205 (1): 138-145. 333 MacArthur, R . H . 1964. Environmental factors a f fec t ing b i r d species d i v e r s i t y . Amer. Nat. 98: 387-397. Macfadyen, W.A. 1950. Vegetation patterns i n the semi-desert p la ins of B r i t i s h Somali land. Geogr l . J . 116: 199-211. Meause, A . D . J . 1959. Aarvarken en aardkomkommer. Levende Nat. 62 (1) : 1-7. Meyer, J . A . 1960. Resultats agronomiques d'un essa i de nivellement des termit ieres re'alise' dans l a Cuvette Centrale Congolaise . B u l l . A g r i c . Congo Beige 51: 1047-1059. Mi lne , A . 1961. D e f i n i t i o n of competition among animals. 15th. Symp. Soc. Exp. B i o l : 40-61. M i l n e , G. 1947. A s o i l reconnaissance journey through parts of Tanganyika t e r r i t o r y . J . E c o l . 35: 192-264. M i t c h e l l , F . J . 1965. A u s t r a l i a n geckoes assigned to the genus Gehyra Gray. Senckenberg. b i o l . 46: 287-319. Morr i son, C . G . T . and A.cSCHoyle and J . F . Hope-Simpson. 1948. T r o p i c a l so i l -vege ta t ion catenas and mosaics. J . E c o l . 36 (1): 1-84. Murray, J . M . 1938. An inves t i ga t ion of the i n t e r r e l a t i o n s h i p s of the vegetat ion, s o i l s and termites . S. A f r . J . S c i . 35: 288-297. Naik, A . H . 1965. Minera l status of grasses and s o i l s of part of Mbulu d i s t r i c t of Tanzania i n r e l a t i o n to animal hea l th . E . A f r . A g r i c . J . 31: 175-182. Neal , B.R. 1970. The habi tat d i s t r i b u t i o n and a c t i v i t y of a rodent population i n Western Uganda, with p a r t i c u l a r reference to the ef fects of burning. Rev. Z o o l . Bot. A f r . 81 (1-2): 29-49. Neal , B.R. and A.G.Cook. 1969. An analys i s of the s e l ec t ion of small a f r i can mammals by two break-back t raps . J . Z o o l . Lond. 158: 335-340. Nea l , E . 1970. The banded mongoose. E . A f r . . W i l d l . J . 8: 63-71. Neal , W.M. 1941. Present knowledge of the n u t r i t i o n a l value of grassland herbage. J . Amer. Soc. Agron. 33: 666-670. N e l , J . J . C . 1968. Aggressive behaviour of the harvester termites Hodotermes mossambicus (Hagen) and Trinervitermes  t r inervo ides (S jo s t . ) . Insectes Soc. 15: 145-156. Nelson, B.W. and <EQE.Sommers. 1973. Determination of t o t a l nitrogen i n plant m a t e r i a l . Agron . J . 65:109-112. N o i r o t , C.H. 1970. The nests of t e r m i t e s . In "Biolog y of t e r m i t e s " ed. by Kr i s h n a and Weesner, Academic P r e s s . N u t t i n g , W.L. 1970. F l i g h t and colony f o u n d a t i o n . In "Bi o l o g y of t e r m i t e s " ed. by K r i s h n a and Weesner, Academic P r e s s . Nye, P.H. 1955. Some s o i l - f o r m i n g processes i n the humid t r o p i c s . IV. The a c t i o n o f the s o i l fauna. J . S o i l . S c i . 6: 73-83. Pages, E. 1970. Sur l ' e c o l o g i e e t l e s a d a p t a t i o n s de l ' o r y c t e r o p e e t des p a n g o l i n s sympatrique du Gabon. B i o l . Gabonlca (1): 27-92. Pianka, E.R. 1966. Convexity, d e s e r t l i z a r d s , and s p a t i a l h e t e r o g e n e i t y . Ecology 47 (6): 1055-1059. P i e l o u , E.C. 1961. S e g r e g a t i o n and symmetry i n two-species p o p u l a t i o n s as s t u d i e d by nea r e s t neighbour r e l a t i o n -s h i p s . J . E c o l . 40: 255-269. Robinson, J.B.D. 1958. Some chemical c h a r a c t e r i s t i c s of te r m i t e s o i l s i n Kenya c o f f e e f i e l d s . J . S o i l . S c i . 9: 58-65. Rood, J.P. 1974. Banded mongoose males guard young. Nature Lond. 248: 176. Rosenzweig, M.L. and J . Winakur. 1969. P o p u l a t i o n ecology of d e s e r t rodent communities: h a b i t a t s and e n v i r o n -mental complexity. Ecology 50 (4): 558-572. Ross, B.A. , J.R.Tester, and W.J.Breckenridge. 1968. Ecology of mima-type mounds i n Nothwestern Minnesota. Ecology 49 (1): 172-177. Rowell, T.E. 1966. F o r e s t l i v i n g baboons i n Uganda. J . Z o o l . Lond. 149: 344-364. R u e l l e , J.E. 1964. L ' A r c h i t e c t u r e du n i d de Macrotermes  n a t a l e n s i s e t son sens f o n c t i o n e l . In "Etudes sur l e s t e r m i t e s a f r i c a i n s " ed. by B o u i l l o n : 327-362. R u e l l e , J.E. 1970. A r e v i s i o n of the t e r m i t e s of the genus Macrotermes from the E t h i o p i a n r e g i o n ( I s o p t e r a : T e r m i t i d a e ) . B u l l . B r i t . Mus. (Nat.Hist.) Ent . 2 4 (9). Sands, W.A. 1965. Termite d i s t r i b u t i o n i n man-modified habtats i n West A f r i c a , w i t h s p e c i a l r e f e r e n c e t o sp e c i e s s e g r e g a t i o n i n the genus T r i n e r v i t e r m e s . J . Anim. E c o l . 34: 557-571. S i e g e l , S. 1956. Nonparametric s t a t i s t i c s f o r the B e h a v i o u r a l S c i e n c e s . Mcgraw H i l l , New York. S i k e s , S.K. 1968. Observations on the ecology of a r t e r i a l d i s e a s e i n the A f r i c a n Elephant (Loxodonta afr1cana) i n Kenya' and Uganda. Symp. Z o o l . Soc. Lond. "21: ~ 251-273. Smithers, R.H.N. 1971. The mammals of Botswana. Museum memoir 4. S a l i s b u r y . Southern, H.N. and O.Hook. 1963. Notes on the breeding of s m a l l mammals i n Uganda and Kenya. Proc. Z o o l . Soc. Lond.r-140 : 503-515. Stoops,G. 1964. A p p l i c a t i o n of some p e d o l o g i c a l methods to the a n a l y s i s of t e r m i t e mounds. In"Etudes sur l e s t e r m i t e s a f r i c a i n s " , ed. by B o u i l l o n : 379-398. Thomas, A.S. 1941. The v e g e t a t i o n of the Sese I s l a n d s , Uganda. J . E c o l . . 29: 330-353. Thornton, D.D., M.I.E.Long, and B . M a r s h a l l . 1968. N u t r i t i v e v a l u e of grasses i n Ankole and Queen E l i z a b e t h N a t i o n a l Park, Uganda. I . - C o l l e c t i o n s i t e s , s o i l s and g r a s s e s . Trop. A g r i c u l t u r e T r i n . 45: 257-268. T r o l l , C. 1936. Termitensavannen. In "Landekundliche F o r s c h r i f t f u r Norbert Krebs" : 275-312. Uganda department of lands and surveys (U.D.L.S.). 1965. Maps of R.N.P. : s e r i e s Y732. U.D.L.S. 1969. Maps of R.N.P. : s e r i e s Y732. Ur b a i n , A. 1954. B i o l o g i e de 1'orycterope. Ann. Mus. Congo  Tervuren, in-4°, Z o o l . , 1: 101-105. U.S. Dept. of A g r i c u l t u r e and C o r n e l l U n i v e r s i t y A g r i c . Exp. S t a t i o n . 1965. The e f f e c t of s o i l s and f e r t -i l i z e r s on the n u t r i t i o n a l q u a l i t y of p l a n t s . A g r i c . I n f o . Bui1. 299. Verheyen, R. '1951. C o n t r i b u t i o n a 1'etude e t h o l o g i q u e des mammiferes du Pare Nat. de L'Upemba. I n s t . Parc Nat. Congo Beige: 94-98. Vos, A de. 1969. E c o l o g i c a l c o n d i t i o n s a f f e c t i n g the p r o d u c t i o n of w i l d h e r b i v o r o u s mammals on g r a s s l a n d s . Adv. E c o l . Res. 6: 137-183. Waloff, N. and R . E . B l a c k i t h . 1962. The growth and d i s t r i b u t i o n of mounds of L a s i u s f l a v u s F a b r i c u s (Hym. Formicidae) i n Silwood Park, "Berkshire. J . Anim. E c o l . 31: 421-437. Watson, J.P. 1962. The s o i l below a t e r m i t e mound. J . S o i l . S c i . 13: 46-51. Watson, J.P. 1967. A t e r m i t e mound i n an i r o n age b u r i a l ground i n R h o d e s i a . J . E c o l . 55: 663-669. Watson, J.P. 1969. Water movement i n two t e r m i t e mounds i n R h o d e s i a . J . E c o l . 57: 441-451. W e i r , J.S. 1969. C h e m i c a l p r o p e r t i e s and o c c u r r e n c e on K a l a h a r i sand o f s a l t l i c k s c r e a t e d by e l e p h a n t . J . Z o o l . Lond. 158: 293-310. Weir, J.S. 1973a. E x p l o i t a t i o n o f water s o l u b l e s o i l sodium by e l e p h a n t s i n M u r c h i s o n F a l l s N a t i o n a l P a r k , Uganda. E. A f r . W i l d l . J . 11: 1-7. Wei r , J.S. 1973bi A i r f l o w , e v a p o r a t i o n and m i n e r a l a c c u m u l a t i o n i n mounds o f Macrotermes subhya1inus (Rambur). J . Anim. E c o l . 42 ( 3 ) : 509-520. Wheeler, W.M. 1913. A n t s . Columbia U n i v . P r e s s , p.663. W i l d , H. 1952. The v e g e t a t i o n o f S o u t h e r n Rhodesian tegmifeafiba. R h o d e s i a A g r i c . J . 49 : 280-292. W i l l i a m s , M.A.J. 1968. T e r m i t e s and s o i l developement near B r o c k s Creek, N o r t h e r n T e r r i t o r y . A u s t . J . S c i . 31': 153-154. W i l l i a m s , R.M.C. 1959. Co l o n y development i n Cubitermes  Ugandensis F u l l e r ( I s o p t e r a : T e r m i t i d a e ) . I n s e c t e s "Soc. 6: 291-304. Wood, T.G. and K.E.Lee. 1971. Abundance o f mounds and c o m p e t i t i o n among c o l o n i e s o f some a u s t r a l i a n t e r m i t e s p e c i e s . P e d o b i o l o g i a 11: 341-366. Yakushev, V.M. 1968. I n f l u e n c e o f t e r m i t e a c t i v i t y on t h e development o f l a t e r i t e s o i l . S o v i e t S o i l S c i . 1968 ( 1 ) : 109-111. 337 APPENDIX 1. Nearest Neighbour Calculations Dispersion within a single population. The calculations made follow Clark and Evans (1954). N: the number of nearest neighbour distance (NND) measurements taken. r: the distance from the center of one mound to the centre of its nearest neighbour, (NND). D: the mean density of mounds. R A N : the mean of the series of distances to nearest neighbour. R A= — — : the mean distance to nearest: neighbour expected in an infinitely large 2/B~ random distribution of density 'D'. R^ A : this dispersion coefficient is a measure of the degree to which the R E observed distribution departs from random expectation with respect to the distance to nearest neighbour. In a random distribution R=l. Under maximum aggregation R=0, since all of the mounds occupy the same position and the distance to the nearest neighbour is therefore 0. Under conditions of maximum spacing, mounds will be distributed in an even, hexagonal pattern, and every one will be equidistant from six others. In such a situation R=2.1491 (see Clark and Evans (1954), appendix). Test for significance. <$R = 0 i 2 6 ^ 6 : the standard error of R £, (see Clark and Evans (1954) , J ND appendix). C = R A " \ R E : t n e standard variate of the normal curve. 6R E c=1.96, represents the 5% level of significance -(tor a; two tailed test). Other values are obtained from a table of the normal distribution. 338 Relationships between two populations. When a group of loci are considered consisting of two different types, such as mounds (M) and wallows (W) , there are four possible nearest-neighbour relationships. 1. A mound having a mound as its nearest neighbour. 2. A wallow having a wallow as its nearest neighbour. 3. A wallow having a mound as its nearest neighbour. 4. A mound having a wallow as its nearest neighbour. The numbers of each type can be compared in a 2x2 contingency table .(Pielou 1961) as follows: -MOUND BASE WALLOW TOTAL NEAREST MOUND: MM MW NM1 NEIGHBOUR WALLOW: . WM WW NW2 TOTAL: NM NW N A X 2 test on the observed numbers in the four classes and the expected numbers will show the significance of any differences, indicating clumping within mounds or within wallows, or between mounds and wallows. 339 APPENDIX 2: A Comparison of Mound Density Figures obtained from Ground  Transects and Aerial Visual Transects. The results of aerial visual transects, aerial photographic transects and ground transects, for areas in the north of the Park, have been presented in Table XIV. As well as total mound densities from ground transects, three classes of mound have been selected based on results presented in section III-3—ii; these are mounds >1 meter high; mounds not thicketed; and mounds not thicketed and also >1 meter high. Each of these in turn was compared with their respective adjusted aerial visual density estimates in order to assess which mounds were particularly being missed on the aerial surveys. As can be seen in Table XIV up to 95% of mounds recorded on ground surveys were missed on aerial visual surveys. These comparisons are shown in Table XV, and correlation coefficients given; because the sample size was the same in each case the coefficients can be directly compared. Notice that the highest correlation was for the relation between the adjusted aerial visual results and mounds >1 M high which were also unthicketed. Also either of these classes separately gave a higher correlation than when all mounds were considered. The highest percentage of the density recorded in ground transects was for the Kamulikwezi area, where the highest percentage of mounds over 1 meter occurred. The second highest percentage of the ground transect density recorded by a visual aerial transect, was at Nyamagasani. This area did not have a high percentage of mounds over 1 meter, but it did have the lowest percentage of thicketed mounds. The lowest recorded percentage of a ground 340 transect result was taken on the aerial transects in the Crater Region; this area contained the lowest percentage of mounds > 1 meter, however it did have the second highest percentage of unthicketed mounds. These results show that many Macrotermes mounds were not recorded on aerial visual transects if they were of smaller size or thicketed. Even bare mounds were missed if they were small, probably because they were confused with small bare non-Macrotermes mounds. The use of aerial visual methods may be useful in obtaining comparative density figures for uniform mound types. In other situations aerial photographic methods should be employed in close conjunction with ground transects. APPENDIX 3: Mound Volume Estimation. The shape of Macrotermes mounds was assumed to be intermediate between that of a cone and half a spheroid. 1. Vol. of cone = 1/12 d 2h 2. Vol. of half a spheroid = 1/6 d 2h 3. Intermediate formula used: V = 1/8 dj^h-, where d^ and d£ are two measurements of basal diameter, and h is mound height. The shape of large O. kibarensis mounds was assumed to approximate to half a spheroid. APPENDIX 4: Density of Termite Mound Soil Density here refers to the total volume of epigeal structure in relation to its weight. This will be less than the actual bulk density of mound soil since the mound contains numerous galleries. Little data are available on this density measurement. Hesse (1955) gives the above ground volumes and weights of four Macrotermes mounds, which produce a mean density figure of 2.1 g/cc. Williams (1968) used a figure of 2.5 g/cc, but allowed separately for gallery spaces. A figure of 2.0 g/cc was used in the present calculations. 342 Appendix 5: The soldier and worker caste mandibles found in droppings of Orycteropus afer 343 a) Pseudocanthotermes spiniger, major soldier. b)_P_. spiniger, minor soldier . c) Odontotermes fulleri, soldier. d) _0_. kibarensis , soldier. e) Ch stercorivorus, soldier. f) Trinervitermes gratiosus, major soldier nasus. 344 a a) M. subhyalinus, minor worker, b) Non-Macrotermes sp. (O. stercorivorus) , worker. Arrows denote major difference between a and b. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058259/manifest

Comment

Related Items