IUBS Unesco ~h

RESPONSES OF SAVANNAS TO STRESS AND DISTURBANCE

A Proposal for a collaborative Programme of Research

edited by P. Frost, E. Medina, J.-C. Menaut, O. Solbrig, M. Swift and B. Walker

Report of a W.orkshop orglanked in QolQbaratian with The .Comrnissiop of ~u~hpean Cammunitieq (CEC),

, . - , *<. . ... 1 : ". . - , :; : A,: - .. .:::- - . . " .

- < ...-.F. , -. &. . . ?' - <* -* 2

RESPONSES OF SAVANNAS TO STRESS AND DISTURBANCE

A Proposa1 f o r a Col laborat ive Programme o f Research

Report o f a Meeting o f an IUBS Working Group on

Decade o f the Tropics Programme/Tropical Savanna Ecosystems

The UNESCO Man and Biosphere (WB) Programme, The Af r ican Biosciences Network (ABN),

The European Economics Commission (EEC) , and The S c i e n t i f i c Council o f Zimbabwe (SCZ)

9 - 13 December, 1985 Harare, Zimbabwe

Edi ted by

Peter F ros t Jean-Cl aude Menaut

B r i an Wal ker Ernesto Medi na

Ot to T. So lbr ig Michael Swi ft

SPECIAL ISSUE - 1 0

BIOLOGY INTERNATIONAL

The In ternat iona l Union o f B i 01 ogical Sciences News Magazine

TABLE OF CONTENTS

SECTION 1: INTRODUCTION AND OMECTIVES 1

OVERVIEW 1

DEFINITIONS Savanna Stress D i sturbance Stab i 1 i t y and Resi 1 i ence

RATIONALE 4

APPROACH AND KEY QUESTIONS 6

SECTION II: THE DETERMINANTS OF SAVANNAS 7

INTRODUCTION 7

SOIL MOISTURE DYNAMICS R a i n f a l l Soi 1 s

SOIL NUTRIENT DYNAMICS S o i l s Organi c matter dynamics

FIRE AND HERBIVORY F i r e Herbi vory

INTERACTIONS E f f e c t s o f vegetat ion on water and n u t r i e n t dynamics E f f e c t s o f animals on water and n u t r i e n t dynamics Soi1 moisture and the tree:grass equ i l i brium Species composition and coexistence Phenol ogy and coexistence P l an t product ion P l a n t qua1 i t y Vegetation dynamics

HUMAN INFLUENCES 36

SECTION III: HYPOTHESES ABOUT THE RESPONSES OF SAVANNAS TO STRESS AND DISTURBANCE

INTRODUCTION

FUNCTIONAL CLASSIFICATION OF SAVANNAS

HYPOTHESES

SECTION I V : PROPOSED PROGRAMME

0 M ECTIVES

PROCEDURE 1. Improving communication between savanna researchers 2. Promotion o f short term co l 1 aborative p ro jec ts 3. In tercont inenta l compari sons 4. Incorporat ion o f research resul t s i n t o management

ORGANIZATION

FUTURE ACTIVITIES

REFERENCES

APPENDIX: ADDRESSES OF WORKING GROUP PARTICIPANTS

RESPONSES OF SAVANNAS TO STRESS AND DISTURBANCE: A PROPOSAL FOR A COLLABORATIVE PROGRAMME OF RESEARCH

SECTION 1: INTRODUCTION AND OBJECTIVES -- - -

OVERV 1 EW

Th is document represents a proposa1 f o r a programme o f c o l l a b o r a t i v e research on t r o p i c a l savanna ecosystems. Our concern stems f rom the c u r r e n t t r end o f degradat ion i n savannas around the world, i n v o l v i n g changes i n composi t ion and p r o d u c t i v i t y t h a t a re adversely a f f e c t i ng the capac i t y o f these systems t o support humans and o the r oryanisms. I n o rder t o a r r e s t o r reverse these changes we need t o improve Our u n d e r s t a n d i n g o f savanna dynamics under p r e v a i l i n g and p r o j e c t e d pa t te rns o f l a n d use. Therefore, the o b j e c t i v e o f t h i s programme i s :

To devel op a p r e d i c ti ve unders tand i ng o f t he - -- ways i n whTch savannas respond & n a t u r a l ---- and man-made s t r e s s e s and d is t ! r rbances. - - -

This can bes t be achieved t h r o u g h a c o m p a r a t i v e , i n t e r c o n t i n e n t a l ana l ys i s o f some se lec ted aspects o f t r o p i c a l savannas i n v o l v i n g a d i v e r s i t y o f research inputs.

Over t h e p a s t few y e a r s t h e r e have been s e v e r a l syn theses o f t h e r e s u l t s o f r e s e a r c h on savannas i n d i f f e r e n t p a r t s o f t h e w o r l d ( H i l l s and Randall, 1968; Bour l i k e and Hadley, 1970; UNESCO, 1979; Wal ker, 1979; Huntley and Wal ker , 1982; Bourl i k e , 1983; Sarmiento, 1984). A l t h o u g h t h e s e a u t h o r s do n o t agree on a l 1 p o i n t s , t h e syn theses p r o v i d e a f o u n d a t i o n on w h i c h t o b u i l d. Some o f t h e disagreements stem from d i f f e r e n t concepts as t o what c o n s t i t u t e s a savanna ,o the rs f r o m t h e d i f f e r e n t a p p r o a c h e s and methods used i n these separate research programmes. An i n t e r n a t i o n a l c o l 1 abo ra t i ve research programme i n v o l v i n g c a r e f u l l y p l anned comparative s tud ies w i t h a common a i m and approach o f f e r s t h e b e s t o p p o r t u n i t y f o r reso l v i ng these conceptual d i f f e r e n c e s and ga i n i ng the i n s i g h t s t h a t a r e so u r g e n t l y needed as a b a s i s f o r t h e r a t i o n a l use and management o f savannas.

Whi le the s c i e n t i f i c r e s u l t s a re impor tan t i n t h e i r own r i g h t , i t i s v i t a l t h a t the m a j o r i n s i g h t s and f i n d i n g s o f the programme should u l t i m a t e l y be a p p l i e d t o t h e p r o b l e m s o f savanna u t i l i z a t i o n and management. It i s n o t enough t o suppose t h a t t he f i n d i n g s pub1 ished i n t h e s c i e n t i f i c l i t e r a t u r e w i l l be p i c k e d up and a p p l i e d b y dec i s ion makers and managers, o r t h a t they w i l l be usefu l t o them i n

t h a t form. A more a c t i v e extension p o l i c y i s t he re fo re envisaged i n which lar id managers and extension o f f i c e r s w i l l be invo lved i n t he programme a t an e a r l y stage, bo th t o he lp evaluate the s i g n i f i c a n c e o f r e s u l t s as t h e y appear, and t o a s s i s t i n d e v e l o p i n g s t r a t e g i e s f o r t r a n s l a t i ng r e s e a r c h r e s u l t s and e inerg i ng concep ts i n t o management dec is ions and ac t i ons i n t he f i e l d .

DEFINITIONS

Savanna

F o r t h e purposes o f t h i s programme, c o r e savannas w i l l be d e f i n e d v e r y b r o a d l y t o i n c l ude a l 1 t h o s e t r o p i c a l and soine n e a r - t r o p i c a l ecosystems cha rac te r i zed by a c o n t i nuous herbaceous cover consi s t i ng m o s t l y o f h e l i o p h i l o u s C4 g rasses and sedges t h a t show c l e a r seasona l i t r e l a t e d t o water s t ress. Woody species (shrubs, trees, palms occur b u t se l dom form a cont inuous cover p a r a l l e l l i n g t h a t o f + t h e g r a s s y 1 ayer. Mary i na1 savanna sys tems i n w h i c h e i t h e r one o f these two vegeta t ion components has an i n s i g n i f i c a n t e f f e c t can be i n c l uded f o r t h e i r c o m p a r a t i v e v a l ue. Savannas encompassed by Our d e f i n i t i o n c o v e r e x t e n s i v e a reas o f South Amer ica, A f r i c a and Aus t ra l i a , and a l so occur i n Centra l America and I n d i a (Figure 1).

F igu re - 1. World d i s t r i b u t i o n o f t r o p i c a l savannas ( f rom Lamotte and Hadley (1984) a f t e r Bour l i k re (1983))

Stress

I n t h i s document we use t h e wo rd " s t r e s s " i n i t s a c c e p t e d p h y s i o l o g i c a 1 sense t o d e s c r i be a c o n s t r a i n i n g env i ronmenta 1 i n f l u e n c e t h a t r e s t r i c t s t h e p r o d u c t i v i t y and e f f i c i e n c y o f an i nd i v i dua l and, by extension, the ecosystem. Such stresses usual ly operate when an environmental variable, such as temperature, l i gh t , water, nu t r i en t s o r defo li a t i on, deviates marked ly f rom i t s norma 1 range o f va lues i n t h e system. S t ress i s seldom accompanied by mor ta l i ty . As s t ress becomes more severe it may cause a disturbance.

D i sturbance

As used here, a 'd is turbance" i s a change i n t h e s t r u c t u r e o f a system which usua l ly a f f e c t s i t s f unc t i on i ng . D is tu rbances a r e o f t e n caused by a p e r i o d i c events, such as f.loods, p a r t i c u l a r l y severe storms, p ro longed d rough t s , e x c e s s i v e h e r b i vo r y , o r a c t i v i t i e s such as bush-c lear ing and cu l t i va t ion . Disturbance may resu lt f r o a one, bu t more o f ten a combination o r sequence o f extreme va lues o f env i ronmenta l va r iab les . These may o r may n o t cause m o r t a l i t y o f i n d i v i d u a l s . O p p o r t u n i t i e s a r e c r e a t e d f o r new ind iv idua ls o f the same o r d i f f e r e n t species t o become established. Increasing ly, d i sturbances are the reçu lt o f human a c t i v i t ies .

Stabi li t y and Resi l ience

A s tab le system i s one i n which those var iables de f in ing the s ta te o f t h e system (eg. spec ies composi t ion, r e l a t i v e abundances, biomass, o r p roduc t i on ) change l i t t l e i n response t o o u t s i d e pressures such as drought, f i r e o r grazing. I f d is tu rbed , t h e y r e t u r n r a p i d l y t o t h e i r o r i g i n a l va lues (Walker, 1980). A s t a b l e system t h e r e f o r e shows l i t t l e v a r i a b i l i t y through t i m e i n t h e amounts o f i t s s t a t e v a r i a b l e s . The d i s t i n c t i o n be tween composi t i o n a l s tab i li t y ( r e l a t i v e constancy o f species composit ion and abundance) and funct iona l s t a b i l i t y ( r e l a t i ve constancy o f the processes m a i n t a i n i n g p r i m a r y p roduc t i on ) needs t o be emphasized since the two can be negat ively cor re la ted (McNaughton, 1977).

A r e s i l i e n t system, on the other hand, i s usual ly no t s tab le and t he va lues o f i t s s t a t e v a r i a b l e s o f t e n change cons iderab l y when subjected t o outside pressure (Walker, 1980). More importantly, the parameters o f the system which in f luence i t s dynamics a lso change, thereby redef i n i ng the boundaries w i t h i n which the system remai ns a t t rac ted t o i t s equi l i b r i um point. As a resu l t , fu tu re disturbances o f the same type are more easi ly accomodated (Walker, e t al., 1981). I n contrast, i n a s tab le system, the parameters o f the system do no t change w i t h d i sturbance.

I n a l 1 o f these de f i n i t i ons the question o f scale arises. Per iodic s t ress, imposed a t r e g u l a r i n t e r v a l s over a long p e r i o d o f ti me, eventually assumes the character is t ics o f a continuous phenomenon, w h i l e a cont inuous s t ress, seen i n t h e very s h o r t t e rm and a t a loca l level, may behave very much as a per iodic phenomenon. Whether o r not a system i s perceived t o be i n equi l ibr iurn of ten depends on t h e s p a t i a l and tempora l sca le o f observat ion. For example, cons iderab le shor t - te rm f l u c t u a t i o n s i n species compos i t ion o r production can be observed a t a loca l sca le, g iv ing the impression o f ins tab i l i t y , b u t when these are viewed over a longer t ime period, as p a r t o f a larger landscape, they simply appear t o be var ia t ions around sone mean value; t he system e x h i b i t s dynamic equi l i b r i um. Simi lar ly , changes which are regarded as i r r eve rs i b l e a t one t ime scale may e x h i b i t slow recovery over longer periods.

The s p a t i a l e x t e n t o f d i f f e r e n t phenomena i s a l so impor tant . A herb ivore 's b i t e i s ve ry loca l i zed , a f i r e may cover hundreds o f hectares, whi l e drought i s experienced a i a regional leve l o r above. Moreover, a t a l o c a l scale, i t i s o n l y poss ib l e t o d e t e c t t h e response o f one o r a few individuals. As the spa t ia l scale expands, so t oo does t h e o r g a n i z a t i o n a l l e v e l a t which a response i s observed. Given t he marked tempora 1 and s p a t i a l heterogenei ty o f savannas, i t w i 11 be p a r t i c u l a r l y i m p o r t a n t t o choose t h e appropriate observationa 1 scale f o r the problems being studied.

RAT1 ONALE

Almost one f i f t h o f the world's population l i v e i n savannas, many o f them i n ru ra 1 soc ie t ies t h a t depend on subsi stence agr icul ture. - Per capi ta food production i s genera 1 ly low, the resu lt o f a var ie ty o f envi ronmenta 1, soc ia l and economic constraints. Rainfa 11 i s high ly seasona 1 and v a r i ab le, l ead i ng o f t e n t o uneven and unpredi c tab l e p a t t e r n s o f p r i m a r y product ion. Many o f the so i 1s a r e r e l a t i v e l y i n f e r t i l e , p a r t i c u l a r l y i n t he h i gh r a i n f a l l zones. Th i s i s re f lected i n low crop y i e l d s and i n the poor n u t r i t i o n a l qua l i t y o f n a t u r a l pastures, e s p e c i a l l y du r i ng t he dry season. Consequently, even i n areas w i t h a h i gh and r e l a t i v e l y cons tan t l e v e l o f p l a n t p roduc t i on n o t a l 1 t h e p l a n t ma t te r can be used p r o d u c t i v e l y by consumers.

Compounding the e f f ec t s o f these factors i s the dual nature o f land use i n many savanna regions. The most productive lands are o f ten se t as ide f o r g row ing cash crops whi l e t h e poorer lands a re used f o r l o c a l food product ion. This, combined w i t h the c o n t i n u i n g r a p i d growth o f the human population i n the tropics, i s placing increasing pressure on marg ina l lands, r esu l t i n g i n many cases i n degraded envi ronments, reduced product iv i t y and lower car ry i ng capaci t ies.

There i s an u rgen t need t h e r e f o r e t o f i n d a l t e r n a t i v e models o f deve lopment des i gned t o i ncrease t h e s a t i s f a c t i o n o f human needs wi thout causing a decl ine i n long-term produc t i v i t y and resi l ience. To be s u c c e s s f u l , t h e s e programmes need t o be based on s c i e n t i f i c a l ly sound techno log ies which take i n t o account t h e r e l e v a n t phys i ca 1, b i o l o g i c a 1 and human elements o f a system. Unfortunately, our know ledge o f these features, pa r t i cu l a r l y t h e i r dynamic aspects, i s s t i 11 q u i t e fragmentary. Th i s p rov i des one o f the main mot ivat ions f o r t h i s programme.

From b o t h a s c i e n t i f i c and land management v iewpo in t , t h e r e a r e a var ie ty o f reasons why t r o p i c a l savannas are an important research subject. Some o f these are:

1. The increas ing ly intense use o f savannas by an expanding human p o p u l a t i o n i s r e s u l t i n g i n adverse changes t o t h e s o i 1 and vegetat ion. When combined w i t h n a t u r a l s t resses such as drought, these changes are leading t o increased erosion and a r i d i f i c a t i o n o f t h e s o i 1, f o l lowed by famine and human misery. Research i n t o t h e causes and consequences o f these changes i n savannas c o u l d h e l p a l l ev ia te some o f the problems i n the future.

2. Within the t ropics, the coexistence and close i n t e rac t i on o f the woody and herbaceous s t r a ta makes savannas unique. Both s t r a ta are o f economic va lue and a b e t t e r unders tanding o f t h e i r coex i s t ence WOU l d c o n t r i bute t o i mproved management.

3. Savannas a r e one o f t h e most seasonal o f t h e wor ld 's m a j o r biomes, experiencing s t rongly cont rast ing c l i m a t i c condi t ions w i t h i n a year, as w e l l as high var iab i l i t y between years. They a lso d isp lay "an appa ren t l y con fus i ng m i x t u r e o f communities o f various sizes, arranged i n d i f f e r e n t p a t t e r n s a t d i f f e r e n t sca les, and each o f which may be changing i n both an order ly and d isorder ly fashion, a t d i f f e r e n t rates" (Walker, 1981: 447). This heterogeneity creates a constant ly vary i ng envi ronment f o r the b i ota and i s probab ly a major f a c t o r i n enab l i ng a r e l a t i v e l y l a rge number o f spec ies and c o n t r a s t i n g l i f e forms t o coex is t . Savannas t h e r e f o r e a r e i d e a l systems i n which t o s tudy how p l a n t s and an imals cope w i t h t h e va r i ous na tu ra 1 s t resses imposed by a v a r i a b l e env i ronment. Th i s WOU l d c o n t r i b u t e t o a b e t t e r unders tandi ng o f t h e e f f e c t s o f man- induced stresses i n these and other ecosystems.

4. F i r e , f r e q u e n t l y assoc ia ted w i t h human a c t i v i t i e s , i s a p rominen t f e a t u r e o f most t r o p i c a l savannas. It a f f e c t s t h e funct ion ing o f these systems i n a va r ie ty o f ways and i n t e rac t s w i t h o t h e r processes, such as herb ivo ry , n u t r i e n t c y c l i n g and p l a n t r e c r u i tment. S ince f i r e can be managed, a b e t t e r unders tanding o f i t s eco log ica l ef fects, and how these can be integrated w i t h other, less managab l e ecosystem processes, WOU l d be extreme ly va luab le.

5. I n a d d i t i o n t o t h e p o s s i b l e con t r ibu t ions t ha t t h i s programme might make t o the be t te r management o f savannas, fu r ther advances i n ecological theory can be expected i n view o f the uniqueness o f the set o f ecologica 1 in te rac t ions t h a t determine the existence o f these sy s tems .

APPROACH AND KEY QUESTIONS

The approach o f t h i s programme i s t h e r e f o r e t o d e v e l o p a c o l l a b o r a t i v e i n t e r n a t i o n a 1 r e s e a r c h e f f o r t wh i ch, t h r o u g h comparative and i n d i v i dua 1 studies, w i 11 improve our understandi ng o f bo th n a t u r a l and managed savannas and t h e i r responses t o d i f f e r e n t kinds o f stress and disturbance. The major problems t o be addressed by the programme are encompassed i n the fo l low ing two key questions :

1. Whatare t h e s t r u c t u r a l and f u n c t i o n a l p r o p e r t i e s o f savannas t h a t render them stable and/or r e s i l i e n t t o seasonal and aseasonal natura 1 stresses and d i sturbance (e.g. f i re, drought)?

2. Are t h e r e c r i t i c a l l i m i t s (i.e. th resho lds ) o f d i s t u rbance beyond which savanna ecosystems do not recover a f t e r the disturbance fac to r i s removed?

The r e s o l u t i o n o f t he problems c a l l s f o r answers t o two f u r t h e r questions:

3. I n which ways (both s t r uc tu ra l l y and functionally), and by how much, do d i f f e r e n t types o f savannas change i n response t o na tu ra l s t resses ( f i r e , below-average r a i n f a l l , f loods, etc.) and n a t u r a l and anthropogenic d i sturbance (e.g. pro longed drought, over-grazing, de-bushing, and CU l t i v a t i o n ) and t h e i r interact ions?

4. What are the mechanisms t h a t determine the manner and the ra tes by which savannas respond t o and recover from a disturbance?

SECTION - II: - THE DETERMINANTS OF SAVANNAS

INTRODUCTION

Recent syntheses on t h e eco logy o f t r o p i c a 1 savannas (Hunt l e y and Wa l ke r , 1982; Bour l i k e , 1983; Sarmiento, 1984; T o t h i Il and Mot t , 1985) revea 1 f o u r p r i n c i pa 1 determinants o f savanna s t ructure and functioning: so i 1 moisture, so i 1 nutr ients, herbivory and f i r e . Soi 1 moisture a v a i l a b i l i t y and so i 1 nu t r i en t status are the key factors, a f f ec t i ng both the balance between grasses and woody p lan ts and the p a t t e r n s o f p r i m a r y p roduc t i on and p l a n t qua l i t y . These i n t u r n i n f l u e n c e t h e k i n d s and e x t e n t o f he rb ivo ry , assoc ia ted an ima l impacts, and t h e f requency and i n t e n s i t y o f f i r e . I n so doing, t h e ava i l a b i l i t y o f m o i s t u r e and n u t r i e n t s and t h e e f f e c t s o f t h e i r i n t e rac t i on are i n d i r e c t ly modi f ied.

A f i f t h determinant o f savanna s t ructure i s human ac t i v i t y . Humans have been associated w i t h Af r ican savannas f o r more than two m i l l i o n years (Harris, 1980). Likewise, i n Asia, there i s evidence o f human occupation going back as f a r as one m i l l i o n years BP. Aus t ra l ia and America were n o t occupied u n t i 1 much l a t e r : about 40000 y r s BP i n A u s t r a l i a , and about 25000 years BP i n South America. These long pe r i ods o f occupa t ion have had a profound e f f e c t on bo th t h e s t ructure o f savannas and t h e i r geographical extent. For example, i t i s b e l i e v e d t h a t t h e savannas o f t he I n d i a n sub-cont inent a r e d e r i v e d p r i m a r i ly f rom woodlands through the a c t i o n o f humans (Singh, 1976; Singh e t al., 1983; Gadgil and Meher-Homji, 1985).

Humans can a f f e c t savanna s t ructure e i t h e r d i rec t l y , as wood-cutters and CU l t i va to rs , o r i n d i r e c t ly, through t h e i r abi li ty t o mani pu l a t e ' f i r e and t o i n f l u e n c e h e r b i v o r e numbers and d i s t r i b u t i o n , bo th by hunting and by in t roduc ing and managing domestic animals. I n more recent t imes humans have developed the capacity t o b r ing about r ap id and considerable change i n savanna s t ructure through mechanical and chemica l means. Overa l l , i n many p a r t s o f t he w o r l d today, human a c t i v i t y c o n s t i t u t e s t h e ma jo r source o f d i s tu rbance t o savanna sys tems . The f o l l o w i n g account represen ts a s e l e c t i v e r ev i ew o f savanna funct ion ing t o provide a background f o r the hypotheses which f o l low. I n t h i s rev iew, we w i sh t o emphasize t he composi te na tu re o f t h e main determinants, each o f which comprise a number o f ra ther weak b u t high ly i n t e rac t i ve forces capable o f producing cumulative strong e f f ec t s (see McNaughton, 1983a, f o r discussion and examples). These e f f e c t s are o f t en contingent on the frequency, i n t ens i t y and t im ing o f t he i n t e r a c t i o n s , on t h e p a s t and p resen t s t a t e s o f t h e system, and on t h e i r i n t e rac t i on w i t h fu tu re events. I n t h i s respect, the

CO-occurrence o f two o r more events, which by themselves may have li t t l e impact, bu t whose i n te rac t i on i s synergi s t ic , may strong ly a f f e c t the system, par t i cu l a r ly when the probabi li ty o f the events o c c u r r i ng i s low. Overa 11, t h e consequences o f these i n t e r a c t i o n s are not s t r i c t l y determin is t ic bu t ra ther the product o f a ser ies o f i n t e r l i nked probabi l i t y funct ions (McNaughton, 1983a).

SOIL MOISTURE DYNAMICS

Soi 1 moi s t u r e regimes i n savannas a re i n f luenced by ( i t h e t o t a 1 amount and seasonal d i s t r i b u t i o n o f annual r a i n f a l l and t h e p r o p o r t i o n o f t h i s t h a t en te r s t h e s o i 1; ( i i 1 the wate r h o l d i n g capac i t y o f t h e s o i 1, which i s l a rge ly a f u n c t i o n o f s o i 1 t e x t u r e and depth, and ( i i i ) t h e amount o f evapotranspi ra t ion, which i s r e l a t e d i n comp l e x ways t o c l imate, so i 1 texture, s o i 1 su r f ace character ist ics, and the type o f vegetation a t a site.

Ra in fa l l

Mean annual r a i n f a 11 i n savannas ranges f rom about 300 mm t o more than 1600 mm and no rma l l y does n o t exceed p o t e n t i a l evapo- t r a n s p i r a t i o n , though i t may do so seasonal ly. Most o f t h e r a i n f a 11s d u r i ng one or, i n equa to r i a 1 regions, two, we 1 1 -de f i ned we t seasons. The d r y season, which l a s t s as long as 9 months i n semi- a r i d areas and as shor t as 3 months i n mois t regions, i s marked by a shortage o f water f o r p l an t growth, a t least i n the iipper 1 - 2 m o f the so i 1 p r o f i le.

The s p a t i a l and tempora l d i s t r i b u t i o n o f r a i n f a l l i s o f t e n h i g h l y var iab le . Rain f a l l s m a i n l y , i n t h e fo rm o f intense, l o c a l i z e d thundershowers, and neighbouring areas may receive d i f f e r e n t amounts o f r a i n on any one occasion. The t i m i n g o f r a i n f a l l w i t h i n a we t season a l s o var ies, g i v i n g r i s e t o d i f f e r e n t p a t t e r n s i n t h e avai l ab i l i t y o f so i 1 moi sture between years. Most. importantly, there a r e o f t e n l a rge year- to-year d i f f e rences i n r a i n f a l l a t a s i t e , par t i cu l a r ly i n the dry areas.

Whi l e the r a i n f a l l i n any one year i s largely unpredictable, there a re suggest ions o f longer-term, quas i -per iod ic f l u c t u a t i o n s i n annua 1 r a i n f a 11 i n some reg ions (southern A f r i ca : Tyson and Dyer, 19781, b u t n o t i n o the rs (Serenget i : Pennycuick and Norton- G r i f f i ths, 1976). Given the large d i fferences i n p r i mary production t h a t occur between years i n apparent response t o f luc tuat ions i n the amount and seasonal d i s t r i b u t i o n o f r a i n f a l l (Poupon and B i 1 le, 1974; Ruther ford, 1981; S i c o t and Grouzis, 1981; Dye and Spear, 19821, the possible existence and o r i g i n o f these r a i n f a l 1 cycles, and t h e i r impl icat ions f o r management, needs more investigation.

S o i l s

The e f f e c t s of va r ia t ion in r a i n f a l l a r e compounded by di f ferences between s o i l s i n permeabil i t y and moi s t u r e r e t e n t i o n p r o p e r t i e s . Savanna s o i l s Vary widely in texture , s t ruc tu re , p r o f i l e and depth, r e f l e c t i n g a t one sca le the i n t e r a c t i o n of geol ogy, geomorphol ogy and c l i m a t e , and a t a n o t h e r , t h e i n f l u e n c e of topography, r e l i c fea tu res of pas t landforms, the kind and ex ten t of vegetation cover, and animal a c t i v i t y (Young, 1976; Montgomery and Askew, 1983). On a r e g i o n a l s c a l e t h e a n c i e n t , h i g h l y weathered s o i l s o f t h e m o i s t savannas , exposed t o moderate t o high r a i n f a l l ( 600 - 1600 m m p.a.), a r e l a t i v e l y s h o r t d ry season ( < 6 months) and high s o i l temperatures ( > 18"C), can be dis t inguished from the younger, l e s s weathered s o i l s of t h e d r i e r savannas which e x p e r i e n c e low and v a r i a b l e r a i n f a l l ( 300 - 700 m m p.a.1, and a 1 ong dry season ( 6 - 10 months).

Mois t savanna s o i l s : These o c c u r mainly on the o l d , e x t e n s i v e ol a t e a u x of t h e Gondwana and mi d - T e r t i a r v ol a n a t i o n s u r f a c e s . a s ;el 1 a s on r e d i s t r i b u t e d sands s i t u a t e d ;n ' the pe r iphery of t h e s e p l a t e a u x and i n upland bas ins . Most o f t h e s o i l s a r e d e r i v e d from f e l s i c o r i n t e r m e d i a t e rocks ( g r a n i t e s , gnei s s e s , phy l l i t e s , s c h i s t s , sands tones ) of t h e u n d e r l y i n g Precambrian c o n t i n e n t a l shields. The s o i l s a r e moderately t o highly weathered and leached, and a re general l y s t ruc tu re l e s s (o f ten massive in highly weathered, c l ayey profi 1 es) t o weakly structured. Well s t ructured soi 1 s occur i n 1 e s s weathered p r o f i l e s , p a r t i c u l a r l y ones d e r i v e d from b a s i c i n t r u s i o n s o r from l i m e s t o n e . K a o l i n i t i c c l a y s and i r o n and aluminium o x i d e s a r e t h e main secondary mi ne ra l S. Expandi ng 2: 1 1 a t t i c e c l ay minera1 s a re uncommon.

The s o i l s a re generally sandy a t the surface, becoming more clayey w i t h depth. T h i s c o n t r a s t i n t e x t u r e between t h e t o p s o i l and t h e subsoil tends t o be gradua1 in ac t ive ly weathering s o i l s but where the s o i l s a r e highly weathered i t i s usually abrupt, giving r i s e t o a duplex profi le. Cateri t e i s commonly present and may form massive sheets. Uniformly textured p r o f i l e s a r e found in sandy s o i l s derived from the weathering of sandstanes and unconsolidated sediments, and in some highly weathered s û i l s of the ra in forest-savanna t r a n s i t i o n zone. The l a t t e r s o i l s contain l a rge amounts of f r e e iron oxides and 1:l l a t t i c e clays. The c lays a r e aggregated i n t o r e l a t i v e l y s t a b l e sand-sized p a r t i c l e s t h a t impart t o the p r o f i l e the consistence and d r a i n a g e c h a r a c t e r i s t i c s o f more loamy s o i l s (Young, 1976; Montgomery and Askew, 1983).

I n f i l t r a t i o n r a t e s a r e g e n e r a l l y h i g h i n sandy s o i 1s and i n c l a y s w i t h a well-developed crack o r crumb structure. However, c lays w i t h an aggregate crumb s t ruc tu re are susceptible t o s t ructura 1 CO 1 lapse when exposed t o r a i n d r o p impac t o r t o o the r forms o f mechanical pressure. These break up t h e aggregates and reduce t h e s i z e o f su r face pores, thereby reduc i ng i n f i 1 t r a t i o n ra tes and increasing r u n o f f (Br idge e t al., 1983). I n f i l t r a t i o n i s a l s o reduced, even i n sandy soi ls, by the presence o f water-repel lan t seals. These can be caused e i t h e r by dormant blue-green algae l i v i n g on t h e s o i 1 surface, o r by t h e accumulat ion o f secondary chemica l compounds leached from the l i t t e r . Since these seals break down w i t h prolonged wetting, t h e i r impact i s greatest ear l y i n the wet season o r a f t e r pro longed dry spe 1 1s.

The upper p a r t o f the so i 1 p r o f i l e i n the high r a i n f a l l savannas i s a t o r c l o s e t o f i e l d c a p a c i t y f o r much o f t h e wet season (Young, 1976). Excess wate r d r a i n s r a p i d l y down t o t he subsoi 1 which can remain m o i s t throughout t h e year, depending on t h e amount o f r a i n f a l l and percolation, p r o f i l e depth, proximity t o groundwater, and the r a t e o f deplet ion o f the accumulated water by plants. Good a e r a t i o n and a low mechanical r e s i s t a n c e t o r o o t p e n e t r a t i o n i n moderate t o wel l -s t ructured soi 1s enables deep-rooted woody p lan ts t o u t i l i z e t h i s moisture, o f t e n w e l l i n t o t he d r y season and long a f t e r t h e wa te r con ten t o f t h e upper 1 - 2 m o f t h e p r o f i l e has f a1 len t o below w i l t i n g point.

The most extreme mo is tu re regimes occur i n s o i 1s w i t h a s t r o n g l y duplex p r o f i l e o r which a r e u n d e r l a i n by an i ndu ra ted i r o n s t o n e ( l a t e r i te ) layer. The in terna 1 drainage o f these soi 1s i s poor and, because o f t h e reduced e f f e c t i v e depth o f t h e p r o f i le, they a l s o have a low wate r s to rage capaci ty . Consequently, t h e s o i l s a re a l t e r n a t e ly wate r logged and d r y (Sarmiento and Monasterio, 1975; Tinley, 1982). However, waterlogging i s not confined t o such soi 1s. S i t e dra inage on t h e a lmos t leve 1 p la teaux i s o f t e n poor and even sandy soi 1s can be saturated i f they occur i n s l i g h t depressions.

Dry savanna s o i 1s: Whereas m o i s t savannas occur on p l a teaux and up lands, the( l rysavannas occur on low a l t i tudes p la ins , i n broad downwarped basins, and i n r i f t - f a u l t e d va l leys , o f t e n a t t h e extremes o f t h e t r op i cs . The predominant ly sandy and s i l t y s o i 1s o r i g inate main ly from unconso li dated sediments, r e d i s t r i buted sands, and from ac id igneous and sedimentary rocks o f Precambrian and l a t e r age. C l a y - r i c h p r o f i l e s a re l ess common but, where present, t hey contain higher proport ions o f expanding 2:l l a t t i c e c lays than f ine- textured so i 1s i n mois t savannas (Young, 1976). Since the c l imate i s dry, t h e s o i 1 p r o f i l e i s se ldom thorough l y wetted. Weatheri ng and leaching ra tes are therefore low. Where h igh ly weathered and leached s o i l s do occur, as i n A u s t r a l i a , these i n d i c a t e t he occurrence o f wetter periods i n the past.

As a r e s u l t o f the louer weathering rates, s o i l s formed i n s i t u f r o i c r ys ta l l i n e rocks, i n freely-drained s i tes, tend t o b e ç h m w and sandy, w i t h l i t t l e o r no p r o f i l e development (Young, 1976). I n f i l t r a t i o n and p e r c o l a t i o n r a t e s a re high. Th is favours t h e conservation o f rainwater since i t lessens the amount o f moisture t h a t can be l o s t through evaporat ion. However, t h i s i s o n l y e f f ec t i ve i n deep sands since most o f the s o i l s are too shallow t o be able t o s tore large amounts o f water.

C layey so i 1s i n dry savannas deve lop i n areas under l a i n by base-ri ch r o c k s ( b a s a l t , sha les , l i m e s t o n e s ) , and i n poo r l y d r a i n e d bottom lands and depressions where c lays and minera 1s (par t i cu l a r ly calcium and magnesium) accumulate. Self-mulching c lays ( v e r t i sols) common ly occur and are pa r t i cu l a r l y widespread i n Austra li a, Ind ia and i n par ts o f Africa. Clays w i t h t ex tu ra l B-horizons, produced by c l a y t r ans loca t i on , a re a l s o common. I n these s o i ls , o f t e n h i g h leve 1s o f exchangeab l e sodium cause the dispersion o f c lays which then accumu l a t e i n t he subsoi 1 where they r e s t r i c t i n t e r n a 1 drainage, aerat ion and r o o t penetration. Subsoil drainage can a lso be i nh ib i t ed by impermeable layers formed by the p rec ip i t a t i on o f c a l c i u m carbonate and s i l i c a t e s . The gradua1 ex tens ion o f these layers towards the surface reduces the e f f ec t i ve depth o f the s o i l and l i m i t s i t s water storage potential.

I n f i l t r a t i o n ra tes and permeabi l i t y are considerably lower i n clayey p r o f i les. This i s o f ten compounded by the presence o f a l ga l crusts o r surface sea 1s resu lti ng from the dispersion o f c lay p a r t i c les by r a i n d r o p impac t (Penning de V r i e s and D j i t b e , 1982). Clayey s o i 1s therefore tend t o be more a r i d than sandy soi 1s under equivalent low r a i n f a 11, desp i t e hav i ng a h i ghe r wate r -ho ld ing capaci ty . I n c o n t r a s t t o sandy s o i l s , much o f t h e wate r e n t e r i n g t he s o i 1 i s r e t a i n e d c l o s e t o t h e s u r f a c e where i t i s s u s c e p t i b l e t o evaporation. Furthermore, a t given water contents, c lays have more negative soi 1 water potent ia ls, so t h a t proport ionately less o f the water i n the so i 1 i s avai lab le t o plants. Soi1 texture therefore has o p p o s i t e e f f e c t s on s o i l m o i s t u r e r e g i m e s and p l a n t w a t e r avai lab i li t y i n moi s t and a r i d regions. Under high r a i n fa 11, c lays are the wet test soi 1s but, where r a i n f a l l i s low, they are generally the most a r i d (Walter, 1971). In sands, the s i tua t ion i s reversed.

The red i s t r i bu t i on and concentration o f water by overland f l o u has important loca l e f f ec t s on so i 1 moisture regimes. I n some s i t e s i n the d r i e r savannas, more than h a l f the incoming r a i n f a l l can be l o s t as sur face runo f f , depending on the amount o f p l a n t cover, s o i l su r face cond i t ions , topography, amount and i n tens i t y o f r a i n f a l l , and antecedant s o i 1 moi s t u r e c o n d i t i o n s (Walker and Cunningham, 1976; Penning de V r i es and D j i t kye , 1982). A sma l l e r percentage o f

t o t a l r a i n f a l l i s l o s t as runo f f i n moist savannas, though absolute amounts may be higher. Much o f t h i s runof f f lows onto adjacent areas where i t i s absorbed. This increases the e f f ec t i ve r a i n f a l l o f these s i t e s and creates refuges f o r drought-sensitive plants i n years o f low r a i n f a l l (Macdonald, 1978). On a more l o c a l scale, t e r m i t e mounds may have the same effect.

The i n f l u e n c e o f topography on s o i 1 p r o p e r t i e s and so i 1 m o i s t u r e regimes i s best r e f lected i n the regu l a r sequence o f d i f fe ren t so i 1 p r o f i les which develop along a gradient from h i 1 l top o r i n t e r f l u v e c res t t o adjacent va 1 ley bottom. These soi 1 catenas are pa r t i cu l a r ly w e l l developed i n ancient, gent ly undu l a t i n g savanna landscapes w i t h d i f f use drainage systems (Mi lne, 1935; Young, 1976). The changes i n soi 1 propert ies are re la ted p r ima r i l y t o differences down the slope i n (il t he e x t e n t o f e ros ion by sur face wash, assoc ia ted w i t h changes i n s lope angle and t he d i s tance o f over land f low; ( i i ) t h e degree o f leaching, downs lope t ranspor t and deposition o f c lays and d i sso l ved so lu tes; and ( i i i ) p r o x i m i t y t o groundwater and i t s seasonal f l uc tua t ions (Young, 1976). The r e s u l t i s the format ion o f we l l - d ra i ned and o f t e n h i g h ly- leached s o i 1s i n t he upper p a r t s o f t h e catena, and base- r i ch g l eys i n t h e poor ly-drained, more r e s t r i c t e d bottomlands. The changes i n soi 1 properties and moisture r e g i mes i n t u r n g i v e r i se t o zona 1 p a t t e r n s i n t he composi t ion, s t r uc tu re , p r o d u c t i v i t y and q u a l i t y o f t he vegetat ion, and i n t h e corresponding p a t t e r n s o f he rb i vo ry (Morison e t al., 1948; Menaut and Cka r , 1979; B e l l , 1981; T in ley, 1982).

SOIL NUTRIENT DYNAMICS

Savanna nu t r i en t dynamics are the outcome over a var ie ty o f t ime and s p a t i a l sca l e s o f i n t e r a c t i o n s be tween c l i m a t e , geology, geomorphology and the biota. Var iat ions i n the physical and chemical

' propert ies o f savanna so i 1s are strongly re la ted t o di f ferences i n bedrock and the degree t o which t h i s has been exposed by weathering and erosion. These p r o p e r t i e s have been subsequent ly modi f i ed by in teract ions between the so i 1, c l imate and geomorpho logy a f f e c t i ng so i1 microclimate, secondary m ine ra l format ion, and t h e eros ion, leach ing and r e l o c a t i o n o f s o i 1 and nu t r i en t s . Soi 1 mo i s tu re and temperature i n t u rn inf luence weathering and leaching ra tes and the leve l o f b i o t i c ac t i v i t y , p a r t i c u l a r l y i n the soi 1.

Soi 1s

Dystrophic savanna soi 1s: So i l s derived from the weathering o f ac id c r y s t a l l i n e rocks o f the Precambrian cont inental shields, as w e l l as those formed from ancient sedi mentary formations and red i s t r i buted sands, genera l ly have a low reserve o f weatherable minera ls . They are most widespread in, bu t are not confined t o areas rece iv ing high

r a i n f a l l and so a r e o f t e n h i g h l y weathered and leached. The predominance o f non-expanding 1:l l a t t i c e c l ays and i r o n and aluminium oxides resu l t s i n low e f f e c t i v e cat ion exchange capaci ty and smal l amounts o f t o t a l exchangeable bases, pa r t i cu l a r l y Ca and Mg (Jones and W i ld, 1975; Lopes and Cox, 1977; M o t t e t al., 1985). The l e v e l s o f ava i l a b l e P a r e a l s o f r e q u e n t l y very low and some s o i l s w i t h an abundance o f sesquiox ides have a h i g h c a p a c i t y f o r f i x i n g phosphorus. Except i n very ac id so i ls, the amount o f organic m a t t e r i s t h e main de te rminan t o f c a t i o n exchange capaci ty . The quanti ty o f soi 1 organic matter i s pos i t i ve ly corre la ted wi t h mean annual r a i n f a l l and t h e l eng th o f t h e wet season but, o v e r a l l , leve ls are generally low (Jones and W i ld, 1975; Lopes and Cox, 1977; Kadeba, 1978; Montgomery and Askew, 1983; Mot t e t al., 1985).

Some h i g h l y weathered s o i ls , p a r t i c u l a r l y i n South America, have high leve ls o f exchangeable aluminium (Lopes and Cox, 1977). These pose prob lems f o r crops and i ntroduced Pasture grasses. Indigenous species, however, appear t o be adapted t o these high levels and some woody species accumu l a t e t h e meta 1 i n t h e i r t i s s u e s (Haridasan, 1982). Outs ide South America, l i t t l e i s known about t h e impac t o f a lumin ium on n u t r i e n t uptake and cyc l ing . Th is i s a t o p i c which needs invest igat ion.

Eu t ro h i c savanna s o i 1s: These s o i 1s occur i n areas u n d e r l a i n by d s m t m c intrusions; i n i n c i sed va l leys where less weathered m a t e r i a l i s be ing exhumed; i n a l l u v i a l depos i t s a long drainage lines, va l ley bottoms o r f loodplains; and i n areas covered by basic volcanic ash. These s o i l s are more prominent i n d r i e r areas where the amount o f r a i n f a l l has been i n s u f f i c i e n t t o promote e i t he r ex tens i ve weather i ng o r l each i ng. Consequently, there are usual ly adequate reserves o f weatherable minerals (Young, 1976). The so i 1s a r e l ess a c i d and may even be moderate ly a l ka l i ne . The o f t e n h i g h p ropo r t i ons o f 2 : l l a t t i c e c lays, p a r t i c u l a r l y i n v e r t i sols, c o n t r i bute t o h i gh ca t ion exchange capaci t ies and higher leve 1s o f exchangeab l e bases. The amount o f so i 1 organic matter i s genera 1 ly low and therefore has l i t t l e e f f e c t on the cat ion exchange capacity o f t h e s o i 1.

Organic matter dynamics

The breakdown o f o rgan ic m a t t e r and t h e subsequent r e l ease o f n u t r i e n t s , as w e l l as t h e r o l e s p layed i n these processes by t h e biota, d i f f e r considerably i n r e l a t i o n t o soi 1 moisture regimes and s o i 1 f e r t i l i t y (Menaut e t al., 1985). I n t h e d r y savannas, p r i m a r y production i s l i m i t e d by the low r a i n f a l l and short wet season. The i npu t o f organic matter t o the so i 1 i s therefore also low and h igh ly seasonal. The qua l i t y o f these inputs i s generally high and much o f t he m a t e r i a l i s r e a d i l y decomposed by t h e so i 1 b io ta . However,

n u t r i e n t r e l ease i s i n t e r m i t t e n t because m ic rob ia 1 a c t i v i t y i s s t r o n g l y l i m i t e d f o r much o f t h e t i m e by t he a r i d i t y o f t he s o i 1. There i s a ne t minera l izat ion o f ni t rogen a t the s t a r t o f the rains, though t h i s i s not sustained because the small organic matter stock i s r a p i d l y decomposed and an i n c r e a s i n g amount o f n i t r o g e n i s immob i l i zed i n t he m i c r o b i a l biomass (Bernhard-Reversat, 1982; Penning de Vries and Dj i t iye, 1982; Menaut e t al., 1985).

In contrast, i n the moist savannas, the higher r a i n f a l l and extended wet season favour high p l an t production and organic matter i npu t t o the soi 1, although t h i s inpu t may be l i m i t e d by frequent f i res. The s o i 1 m i c r o c l i m a t e favours m i c r o b i a l a c t i v i t y and t h e r e i s a cons iderab le p o t e n t i a l f o r m i n e r a l i z a t i o n . However, m i c r o b i a l a c t i v i t y i s l i m i t e d by t h e low q u a l i t y o f t he o rgan ic mat te r , p r i n c i p a l ly t he low l e v e l s o f ass im i l a b l e carbon, h i gh C:N r a t i o s and l i g n i n con ten ts and, i n some cases, h i gh amounts o f condensed tannins and other secondary chemicals. The release o f nu t r ien ts i s s low and r e s u l t s i n a low s tand ing s tock o f ava i l a b l e n u t r i e n t s . Because o f t h e h i g h r a i n f a l l , t he n u t r i e n t s a re suscep t i b l e t o leachi ng.

Microb ia l a c t i v i t y i s st imulated by exudates from p lan t roots and by water soluble carbon compounds which are introduced i n t o the s o i l dur ing i t s passage through the guts o f earthworms (Lavelle e t al., 1983; Menaut e t àl., 1985). These compounds are produced i n inverse p r o p o r t i o n t o t h e i r leve 1s i n t h e s o i 1, resu lti ng i n cons i derab l e b io log ica l regulat ion o f organic matter decomposition. Because the avai lab i li ty o f assimi lab le carbon i s large ly confined t o earthworm cas t s and-the neighbourhood o f r oo t s , t he re lease o f n u t r i e n t s i s h i g h l y l oca l i zed

Savanna p l a n t s ma in ta i n h i g h root:shoot r a t i os . The t u rnove r o f roots, which i s r e l a t e d t o t h e l e v e l o f p l a n t product ion, may be respons ib l e f o r most o f t h e o rgan ic ma t te r i n savanna s o i l s , p a r t i c u l a r l y i n f requent ly burned savannas where f i r e destroys much o f t h e l i t t e r (Sanford, 1982). Since s o i 1 organic m a t t e r i s i m p o r t a n t i n m a i n t a i n i n g s o i 1 s t r u c t u r e and f e r t i l i t y , more i n f o r m a t i o n i s needed on t h e sources, q u a l i t y and r a t e o f decomposition o f organic mater ia l , and the factors a f fec t ing these.

Al though t h e l e v e l s o f p l an t -ava i l a b l e nut r ients i n savanna s o i l s a re r e l a t i v e l y low, t h e g r e a t e r p r o p o r t i o n o f the t o t a l n u t r i e n t poo l l i e s i n the so i 1 and s o i l o rgan ic ma t te r r a t h e r than i n t h e vege ta t i on and l i t t e r (Nye and Green land, 1961; Abbadie, 1983; Sarmiento, 1984; Frost , 1985). I n t he d r i e r savannas, t h i s l a r g e l y r e f lec ts the r e l a t i v e l y low p lan t biomass and the const ra in t imposed by t he seasonal shortage o f wa te r on p l a n t growth and n u t r i e n t uptake. I n t he more m o i s t savannas i t r e f l e c t s t h e s low r a t e o f release o f nu t r ien ts from so i 1 organic matter.

The r a t e a t which n u t r i e n t s c y c l e through savanna vege ta t i on i s r e l a t i v e l y rapid, p a r t i c u l a r ly through the herbaceous layer where n u t r i e n t s tu rnover 2 - 4 t i m e s f a s t e r than through woody p l a n t s (Frost , 1985). The d i f f e r e n c e i n t h e r a t e s o f n u t r i e n t t u rnove r i n woody p lants and grasses extends also t o l i t t e r decomposition. Grass l i t t e r , i n t he absence o f f i r e and under t he same cond i t ions , decomposes 2-7 t imes fas te r than woody lea f l i t t e r , and many t imes f a s t e r than wood i t s e l f ( M o r r i s e t al., 1982; M o t t e t al., 1985). Where annual f i r e s occur, much o f t h e grass and some o f t he woody l i t t e r ge ts b u r n t r a t h e r than decomposed (Hopkins, 1966; Sanford, 1982; F ros t , 1985). Th i s tends t o reduce b u t does n o t e n t i r e l y e l iminate the d i f ference i n ra tes since most o f the grass mater ia l t ha t i s burnt WOU l d norma 1 ly decompose w i t h i n a year, whi l e woody l i t t e r takes much longer.

FIRE AND HERBIVORY

W hereas water and n u t r i e n t ava i l a b i l i t y Vary continuously through t i m e w i t h i n l i m i t s s e t by t h e p reva i l i n g c l imate , s o i 1, and assoc ia ted b i o l o g i c a l processes, f i r e and herb ivory a re events o c c u r r i n g a t p a r t i c u l a r t imes, w i t h s p e c i f i c i n t e n s i t i e s , a t d i s c r e t e i n t e r v a l s . These f ea tu res a re i n t e r r e l a t e d and, taken together, make up the pa r t i cu l a r f i r e o r herbivory regime a t a site. S i nce t h e s e r e g i m e s can be mani pu l a t e d d i r e c t l y , t h e y a r e po ten t i a l l y important management t oo l s . However, t h e consequences are not always easy t o p red ic t because f i r e and herbivory in teract , and t h e outcome i s o f t e n c o n t i n g e n t on t he p a r t i c u l a r t imes, i n t e n s i t i e s and f r e q u e n c y o f t h e i n t e r a c t i o n s . E x t e r n a l , unpredictable factors, such as fu tu re r a i n f a l l o r drought, can a lso intervene and inf luence the outcome.

F i r e s andhe rb i vo res bo th consume and damage p l a n t m a t t e r and a r e able t o k i 11 ind iv idua l plants, p a r t i c u l a r l y seedlings. This reduces the ex i s t i ng density and s t ructure o f a p l an t community and a f f ec t s i t s f u t u r e compos i t ion and dynamics. Al though changes i n p l a n t community composition can r e s u l t from pressures exerted by f i r e o r h e r b i v o r y a lone, m a j o r changes usua 1 l y r e s u l t f r o m t h e i r i n t e r a c t i o n . Both have r e l a t i v e l y s e l e c t i v e e f f ec t s . Herb ivory i s u s u a l l y l i m i t e d t o p a r t i c u l a r p l a n t species and p l a n t par ts , especia 1 ly those o f above-average n u t r i t i o n a 1 qua li ty. Moreover, he rb i vo re impacts tend t o be r e s t r i c t e d i n space b u t a re more u n i f o r m l y d i s t r i b u t e d i n t ime. I n con t ras t , f i r e i s a p e r i o d i c event, o f t e n occu r r i ng over a wide area and a f f e c t i n g bo th l i v i n g and dead mater ia l large ly i r respec t i ve o f t h e i r n u t r i t i o n a l qua lity. However, f i r e i s se lect ive t o the degree t h a t there are considerable d i fferences between species i n t he i r suscepti b i li t i e s and responses t o f i r e (Frost , 1984).

Most savanna f i r e s occur d u r i n g t h e d r y season as su r face f i r e s , bu rn i ng through t he herbaceous vegetat ion. T h e i r i n c i dence and in tens i ty depends on ( i ) the presence o f s u f f i c i e n t fue 1 t o support a f i r e ; ( i i ) the moisture content o f the fuel; and ( i i i ) a source o f i g n i t i o n . Many o f them a r e i g n i t e d by man, though i n t h e e a r l y we t season l i g h t n i n g can be impor tan t . I n m o i s t savannas, f i r e s genera l l y occur every 1 - 3 years, b u t as mean annual r a i n f a l l dec l ines, t h e i n t e r v a l between successive f i r e s inc reases and becomes more variab le.

The inc idence o f f i r e i s l a r g e l y a f u n c t i o n o f the dry-season s tand ing c rop o f grass, i t s e l f a p roduc t bo th o f t h e amount o f r a i n f a l l and p l an t production dur ing the previous wet season and the extent o f herbivory. F i r e i n t e n s i t y i s var iab le and depends la rge ly on t he amount and s t r u c t u r e o f t h e f u e l , i t s degree o f cur ing, and p reva i li ng ambient cond i t i ons . Because o f t h e i r h i ghe r f u e 1 load, f i r e s i n m o i s t savannas a re u s u a l l y more i n t ense than those occurr ing i n the dry savannas.

Annual bu rn i ng has l i t t l e d i r e c t e f f e c t on t he s o i 1. Most e f f e c t s a r e i n d i r e c t , r e s u l t i n g f rom changes t o t he vegetat ion, and a r e con f i ned t o t he su r face s o i 1 (Sanford, 1982). Organic m a t t e r and t o t a 1 n i t r ogen a r e reduced and ava i l a b l e phosphorus i s s l i g h t l y inc reased i n areas exposed t o regu la r , in tense, l a t e d r y season f i r e s . E a r l y o r l ess i n t e n s e burns have much l ess o f an e f f e c t (Harrington and Ross, 1974; Brookman-Ami ssah e t a l., 1980).

Burn ing may speed up t h e r a t e o f n u t r i e n t c y c l i n g by r educ ing l i t t e r , e s p e c i a l l y components such as woody l e a f li t t e r and dead wood which decompose s lowly. Nitrogen, carbon and sulphur are l o s t th rough v o l a t i l i z a t i o n and removal i n smoke and ash. However, t h e o v e r a l l s i g n i f i c a n c e o f these losses has n o t been assessed. By reducing l i t t e r and herbaceous p l a n t cover, f i r e also bares the soi 1 surface, exposing i t t o raindrop impact, wind and Sun. The length o f t ime t h a t the soi 1 remains bare i s var iab le and depends large ly on the t ime o f the f i r e , the r a t e o f regrowth o f the vegetation and the t im ing o f subsequent r a i n f a 11.

F i r e k i 11s seedl ings, sap l i ngs and s m a l l t rees, p a r t i c u l a r l y i n higher r a i n f a l l areas where grass production i s substant ia l and the annua 1 dry-season f i res are intense. Under these conditions, woody p l a n t r e c r u i t m e n t i s g e n e r a l l y e p i sod i c . F i r e a l s o damages aboveground p a r t s o f p l a n t s and r e t a r d s t he growth o f shrubs and saplings, thereby lowering the biomass o f woody p lants (Rutherford, 1981). The e f f e c t s on d e n s i t y a r e v a r i a b l e and depend on t he r a t e s o f m o r t a l i t y and coppicing. Many o f these e f f ec t s are a funct ion o f f i r e in tens i ty . The ho t t es t f i r e s usual ly occur during the l a t e dry-

season and these can reduce a woodland canopy t o coppice. Woodland species, though, are ab le t o regenerate under a reg ime o f l ess intense, ear ly dry-season f i res (Trapne 11, 1959; Brookman-Amissah e t al., 1980; Sanford, 1982; Edroma, 1984).

Protect ion from f i r e s resu l t s i n an increase i n t ree densi ty and a decrease i n grass product ion. I n mesic areas, savanna woodland e v e n t u a l l y develops b u t i n t h e h i g h r a i n f a l l savanna l fo res t t r a n s i t i o n zone, f o res t species gradua 1 ly estab l i s h (Trapne 1 1, 1959; Menaut,1977; Brookman-Amissah etal.,1980; SanJoséand Far inas, 1983). I n the dry savannas, f i r e s seldom occur frequently enough t o l i m i t the density o f woody plants, though when they do occur, o f ten a f t e r pro longed per iods o f above-average r a i n f a 11, mature woody p l a n t s may exper ience cons iderab l e mor ta li ty s ince they a r e more susceptible t o f i r e than moi s t savanna plants.

Young t r e e s grow r a p i d l y i n areas p ro tec ted f rom bo th f i r e and herbivory, more so than i n areas protected from f i r e alone. Growth i n p l a n t s exposed t o b o t h p r e s s u r e s i s s e v e r e l y r e s t r i c t e d (Ha r r i ng ton and Ross, 1974; Belsky, 1984). F i r e and he rb i vo ry t h e r e f o r e i n t e r a c t , w i t h f r equen t f i r e s keeping woody p l a n t s a t a he ight and i n a acceptable s ta te f o r browsers, whi le the e f f e c t o f the browsers i s t o reduce woody p lan t growth and keep p lants w i t h i n the s ize range af fected by f i r e (Trollope, 1974; Pellew, 1983).

Herbi vory

Savannas make up most o f t h e wor ld 's t r o p i c a l g raz ing lands and c u r r e n t ly suppor t a l a r g e b i omass o f domest ic li vestock, main ly c a t t l e , goats and sheep. I n many regions, t he biomass o f these spec ies c u r r e n t l y equals o r exceeds t h a t o f t h e i n d i g e n o u s he rb i vo res which used t o occur (Cumming, 1982; M o t t e t al., 1985). Each o f the main savanna regions i n the past had a d i s t i n c t i v e large h e r b i v o r e fauna, though t h a t o f A f r i c a appears t o have been, and s t i l l i s , the most abundant and diverse.

Fo r b iogeographic reasons, A u s t r a l i a never had any proboscids, a r t i odacty 1s o r p e r i ssodactyls. Instead, these savannas were popu l a t e d main l y by macropods (kangaroos and wa 1 lab ies) , some o f which survive today alongside f e r a 1 and domestic ungulates. South Awerica, i n contrast, had a diverse fauna o f large ungulate and non- ungu l a t e herb ivores throughout t he T e r t i a r y , w i t h most o f t h e spec ies becoming e x t i n c t d u r i n g t he l a t e P le is tocene e i t h e r as a r e s u l t o f concurrent c l i m a t i c change o r the a r r i va1 o f man (Mar t in and Wright, 1967). What p r o p o r t i o n o f these species a c t u a l ly i n h a b i t e d savannas i s n o t known, b u t t h e present-day i ndigenous fauna and t h e c a r r y i n g capac i t y f o r domest ic l i v e s t o c k on n a t u r a l range land i s r e l a t i v e l y low.

O n l y A f r i c a , and t o a l e s s e r e x t e n t I n d i a , c u r r e n t l y have s i g n i f i c a n t popu la t ions o f ind igenous ungulates and o t h e r l a r g e herb ivores. West A f r i c a n savannas suppor t lower numbers and a s m a l l e r biomass o f l a r g e he rb i vo res than do those i n eastern, c e n t r a i and southern A f r i c a (Be l l , 1982; M i l l i g a n e t al., 1982). Th i s p a r t l y r e f l e c t s t h e h i gh hun t i ng pressure and p a r t l y t h e i mpoveri shed s o i ls, low qua li ty vege ta t i on and consequent low carry ing capacity o f the West Af r ican savannas.

S tud ies o f he rb i vo ry i n savannas have focussed l a r g e l y on t h e e f f e c t s o f w i I d l a rge ungulates, ma in l y i n A f r i c a n ecosystems. It must be emphasized however cha t these a re n o t t h e o n l y n o r necessari ly the most important herbivores i n savannas. I n many areas today, domest ic l i ves tock , especia 1 l y c a t t l e , have e f f e c t s t h a t overr ide those o f indigenous species. Moreover, harvester termites, l e a f - c u t t i n g a n t s , l e p i d o p t e r a l a r v a e , and g rasshoppe rs ( p a r t i c u l a r l y locusts) , a l 1 have ma jo r e f f e c t s i n some systems, though i n most cases these have no t been quantified.

There i s cons iderab le v a r i a t i o n among herb ivores i n t h e degree o f se l e c t i v i ty i n t h e i r diets, r e f l ec t i ng a comp lex i nterplay between the k ind o f animal, i t s body s ize and associated energy and nu t r i en t requirements, and t h e g rowth forin, s t ruc tu re , chemi s t r y and phenology o f p o t e n t i a l food p lants . The f u n c t i o n a l responses o f herbivores t o changes i n p l a n t abundance are also important, though these have not been s u f f i c i e n t l y s t ud ied

Ver tebra te and i n v e r t e b r a t e he rb i vo res i n savannas bo th show a marked p re fe rence f o r f eed ing e i t h e r on broadleafed, o f t e n woody p lan ts , o r on grasses and sedges. Mixed feeders a re much l ess common. This spec ia l izat ion r e f l e c t s the very d i f f e r e n t n u t r i t i o n a l and o t h e r f ea tu res o f these p l a n t groups. Crude p r o t e i n l e v e l s i n woody p l a n t s a re g e n e r a l l y h i ghe r than i n grasses, b u t t h e d i g e s t i b i li ty o f p l a n t t i s s u e , i s reduced by h i gh l e v e l s o f s t r u c t u r a 1 carbohydrates and, i n some instances, by secondary chemica 1 compounds such as condensed tannins (Cooper and Owen-Smi th, 1985). Thus browsers tend t o be energy-l imited whereas grazers are often protein- l imited. Moreover, browsers can experience severe food shortages during the dry season when woody plants drop t h e i r leaves.

The e f fec ts o f herbivory depend on ( i l the growth form o f the plant; ( i i ) t h e p l a n t p a r t s removed; ( i i i ) t he i n t e n s i t y , f requency and season o f use; ( i v ) the growth stage o f the plant; (v) so i 1 type and s o i 1 m o i s t u r e c o n d i t i o n s , w h i c h a f f e c t w a t e r and n u t r i e n t ava i lab i l i t y and thereby the p lants capacity t o regrow; and ( v i l the h i s t o r y o f t h e p lan t , p a r t i c u l a r l y t he t i m e s ince a p rev ious occurrence o f d e f o l i a t i o n by o t h e r herb ivores o r f i r e . S ince recovery from defo 1 i a t i on i s no t i nstantaneous, fu ture events such as heavy r a i n f a l l , drought, o r f u r t he r de fo l i a t i on by herbivores and f i r e can a lso inf luence the eventua 1 outcome.

Compensatory growth i s one o f t h e main responses o f p l a n t s t o d e f o l i a t i o n . When d e f o l i a t i o n i s low, p l a n t s j u s t compensate by r e p l a c i n g l o s t t issue. Under moderate l e v e l s o f defo l i a t i o n , and p rov ided t h a t mo i s tu re and n u t r i e n t c o n d i t i o n s f o r growth a r e adequate, some p lants overcompensate f o r amounts removed, leading t o an increase i n aboveground p lan t production (McNaughton, 1985). If a p l a n t i s severely defol iated, the amount o f regrowth i s i n s u f f i c i e n t t o compensate f o r t he amounts o f f o l i a g e removed, i n which case con t inued d e f o l i a t i o n WOU l d eventua 1 ly resu lt i n t h e p l a n t be ing k i l l e d o r so reduced i n s i z e and s t a t u r e t h a t i t i s u l t i m a t e l y over looked by herbivores.

Consumption by herbivores accelerates the processes o f energy f low and n u t r i e n t cyc l ing . D e f o l i a t i o n reduces n u t r i e n t and wate r l i m i t a t i o n s on the remaining t issues and stimulates growth. I n some cases, pho tosyn the t i c and n u t r i e n t uptake r a t e s a r e inc reased (McNaughton, 1979). By reduc ing a p l a n t ' s t r a n s p i r a t i o n a 1 area, de fo l i a t i on also contr ibutes i n i t i a l l y t o a lower leve l o f water-use and, thereby, t o the conservation o f so i 1 moisture. This can lead t o an extension o f the normal growing season (McNaughton, 1985).

Nutr ients t h a t would otherwise be bound up i n t o standing dead p lan t m a t t e r and li t t e r a r e recyc l e d more r a p i d ly. Furthermore, by r e c y c l i n g n u t r i e n t s i n c o n c e n t r a t e d form as dung o r ur ine, herb ivores c o n t r i b u t e t o m a i n t a i n i n g a h i gh ava i l a b i l i t y o f nu t r ien ts f o r plants. Where the po ten t i a l f o r leaching from the soi 1 i s high, t h i s serves t o keep the nu t r ien ts cyc l ing rap id ly through the vegetation and surface so i l , a feature which may be c ruc ia l t o mai n t a i n i ng the longterm nu t r i en t status o f these systems (Botkin e t al., 1981).

Herbivory in te rac ts w i t h f i r e i n both space and time. Many grazers a r e a t t r a c t e d t o r e c e n t l y b u r n t ground t o feed on t he p o s t - f i r e regrowth o f grasses (Frost, 1984). This s t imulat ion o f regrowth a t a t ime o f the year when the p lants are usual l y dormant, together w i t h t h e h i ghe r n u t r i t i o n a l q u a l i ty o f t he regrowth, i s a ma jo r reason f o r t he f r equen t i g n i t i o n o f dry-season f i r e s by p a s t o r a l i s t s i n t r o p i c a l savannas (Medina pers. comm., M i 1 l i g a n and Su le, 1982). Grazers i n t u r n reduce grass biomass and so lower t h e shor t - te rm probab i li ty o f t he area s u s t a i n i ng another burn. Patchy g r a z i ng consequent ly causes patchy f i res and v i sa versa.

For browsers, dry season f i r e s can reduce the a v a i l a b i l i t y o f food and cause the animals t o disperse t o other, unburnt, areas (Bel 1 and Jachmann, 1984). This may increase the browsing pressure on trees i n these areas (Harrington and Ross, 1974). The degree o f concentration o r dispersion, and subsequent e f f ec t s such as overuse and trampling, depend on t h e s i z e o f t h e b u r n r pa tch i n r e l a t i o n t o t h e amount o f unburnt vegetati on, and t he i r r e l a t i v e a t t r a c t i veness t o herbivores.

INTERACTIONS

E f f ec t s o f vegetation on water and n u t r i e n t dynamics

The re la t ionsh ip between so i 1s and vege ta t i on i s i n t e r a c t i v e : t h e na tu re o f t h e s o i 1 g r e a t l y i n f l u e n c e s t h e t ype o f vege ta t i on t h a t occurs a t a s i te , wh i le the vegetation a f f ec t s s o i l proper t ies both d i r ec t l y , through the supply o f organic matter, and i nd i r ec t l y , by a f f ec t i ng so i 1 moi s tu re and temperature r eg i mes, soi 1 chemi stry, and t h e s t a b i li t y o f t h e s o i 1 surface. The e x t e n t o f p l a n t and l i t t e r cover a f f ec t s the f l u x o f water between the atmosphere and the s o i l by in te rcep t ing and r e d i s t r i b u t i n g r a i n f a l l , enhancing i n f i l t r a t i o n r a t e s and l owe r i ng t h e r a t e o f evapo ra t i on o f wa te r f r om t h e s o i 1 surface (Ke l l y and Walker, 1976).

A t the same time, p lants deplete the s o i l moisture s tore by tak ing up and t ransp i r ing water. The r a t e o f water loss depends p a r t l y on s o i l wa te r p o t e n t i a l s , p a r t l y on t h e evapora t i ve demand o f t h e atmosphere, and p a r t l y on p l a n t charac te r i s t i cs such as t o t a l l ea f area, p l a n t wa te r p o t e n t i a l s , t r a n s p i r a t i o n r a t e s and s toma ta l responses m i ncreasi ng water def i c i ts.

There are few data avai l ab le on the t ransp i ra t ion ra tes o f savanna t r e e s and grasses. I n m o i s t South American savannas, where t r e e s have access t o water th roughou t t h e yea r (Sarmiento e t al., 19851, t h e t r e e s have o n l y m a r g i n a l ly lower t r a n s p i r a t i o n r a t e s than grasses (Cioldstein, pers. comm.). Therefore, t h e r e l a t i v e r a t e s a t which t rees and grasses deplete the so i 1 water i n such a system w i 11 depend large ly on the r e l a t i v e l ea f areas o f the two components.

I n t h e d r y savannas, where b o t h t r e e s and grasses exper ience a seasonal wa te r d e f i c i t , t h e r e appears t o be a d i f f e r e n c e i n t h e i r responses t o increasing water s t ress (Walter, 1971; Pendle, 1982). In grasses, t ransp i ra t ion ra tes appear t o be regulated p r imar i l y by atmospheric evaporative demand. By maki ng osmot ic ad jus tments t o l e a f water potent i a 1s through concentrat i ng so lutes i n t he i r leaves (Wi lson e t al., 19801, grasses seem able t o maintain r e l a t i v e l y high ra tes even a t so i 1 moisture leve ls near w i l t i n g point. Whi l e t h i s extends t h e p e r i o d o f p h y s i o l o g i c a l a c t i v i t y o f t h e grasses, i t r e s u l t s i n a r ap id deple t ion o f so i 1 moisture i n the roo t ing zone o f grasses and t h i s eventua 1 ly leads t o l a r g e l ea f -wa te r d e f i c i ts , desiccat ion and, often, lea f death.

I n con t ras t , t h e t r a n s p i r a t i o n r a t e s o f t r e e s i n t h e d r y savannas appear t o be determined more by so i 1 moisture avai l ab i l i t y than by atmospheric demand. As the s o i l d r i es out, the t rees regulate t h e i r

water-use by lowering t ransp i ra t ion ra tes through stomata 1 c losure. o r by reduc ing l e a f area through shedding leaves. Trees t h e r e f o r e deplete the so i 1 moisture s tore less rap id ly and completely than do t h e grasses. The resu lt i s t h a t where r a i n f a l l i s low and grasses predominate, t h e vege ta t i on aggravates the shortage o f wa te r by ex t rac t ing moisture from below the zone i n the soi 1 where i t can be removed by evaporation a lone. Th i s h e i ghtens t he c o n t r a s t between t h e wet and d r y phases o f t he s o i 1. On the o the r hand, where r a i n f a l l i s high and t r ee growth i s favoured, soi 1 moisture tends t o be conserved

Trees a l s o i n f l u e n c e t he s t a t u s and d i s t r i b u t i o n o f n u t r i e n t s i n savanna soi 1s. Woodland regeneration, f o l lowi ng protect ion against sustained heavy browsing and grazing, resu l t s i n an improvement i n so i 1 nu t r i en t status (Hatton and Smart, 1984) . Soi 1 organic matter, pH, a v a i l a b l e phosphorus, and t he l e v e l o f exchangeable c a t i o n s (excep t ing Mn) a l 1 increase. Th i s i s because trees, w i t h t h e i r r e l a t i v e l y deep roo t systems, are able t o ex t rac t nut r ients a t depth i n the so i 1 and so counter the e f f ec t s o f leaching (Kellman, 1979).

The h o r i z o n t a l l y -ex tens ive r o o t systems o f t r ees a 1 low them t o concen t ra te n u t r i e n t s f rom a wide area (Van Donselaar-Ten Bokkel Huinink, 1966; Radwanski and Wickens 1967; Fo lda t s and Rutk is , 1975). Consequently, t h e s o i 1s under t r e e s i n savannas genera l l y have a higher nu t r i en t status which, together w i t h the moister s o i l s and more mesic m ic roc l imate , i n f l u e n c e s t he d i s t r i b u t i o n o f herbaceous plants, favouring more mesic-adapted, palatable grasses. This occurs i n both A f r i ca and Venezuela (Kennard and Walker, 1973; Medina, pers. comm.) b u t n o t i n no r t he rn A u s t r a l i a ( M o t t e t al., 1985) which poses an in te res t ing question why.

E f fec ts o f animals on water and nu t r i en t dynamics

Animals a f f e c t the s t ructure o f the so i 1 and therefore so i 1 moisture dynamics i n va r i ous ways. Soi 1 macrofauna i nco rpo ra te li t t e r and o t h e r o rgan ic m a t e r i a l s i n t o t h e s o i l and c o n t r i b u t e t o t h e syn thes i s o f o rgan ic CO 1 lo ids. These CO 1 l o i d s a re i nvo l ved i n t h e f o r m a t i o n o f s t a b l e s o i l aggregates which i n t u r n enhance t h e aeration, permeabi li ty and water-holdi ng capaci t y o f the soi 1. The network o f underground ga l l e r i es and burrows formed by soi 1 animals presumably has s i m i l a r p o s i t i v e e f f e c t s b u t these have n o t been thorough l y i n v e s t i gated (Woods and Sands, 1977). Earthworms and rermi tes b r ing considerable quanti t i e s o f so i 1 t o the surface i n the form o f f a e c a l casts, g a l l e r i e s , and sur face sheet ings (Wood and Sand, 1977; Bagine, 1984; Lave1 le, 1983). This improves soi 1 surface structure, thereby promoting h i gh water i n f i l t r a t i o n rates.

Much o f t h e m a t e r i a l b rough t t o t h e su r face b y ' t e r m i t e s and earthworms cornes from the subsoil. This counteracts the e f f ec t s o f i l l u v i a t i o n and leaching, and so cont r ibutes t o the maintenance o f u n i f o r m - t e x t u r e d s o i 1 p r o f i l e s (Boyer, 1973; Young, 1976; Josens, 1983; Lave 1 le, 1983). Termi t e s a r e p a r t i c u l a r l y i m p o r t a n t i n t h i s regard as they genera 1 l y se lect c lay-sized p a r t i c les i n preference t o sand when construct ing t h e i r nests and ga l l e r i es (Wood and Sands, 1977). This can have both pos i t i ve and negative e f fec ts on the s o i l moi s ture ba lance, depending on r a i n f a 11.

D i r e c t negative e f f ec t s o f animals on soi 1 moisture regimes r e s u l t from compaction o f the surface s o i l and the breakdown o f aggregates through tramp l ing. Th i s inc reases bu l k s o i 1 dens i ty, reduces i n f i l t r a t i o n r a t e s and lowers t h e wate r -ho ld ing capac i t y o f t h e s o i 1. Run-of f and e ros i on increase. I n d i r e c t nega t i ve e f f e c t s a r e mediated through t h e impac t o f an ima ls on t h e vegetat ion. High dens i t i e s o f l a rge ungu l a t e s and i nvertebrate herbivores such as h a r v e s t e r t e r m i t e s s i g n i f i c a n t l y reduce p l a n t and li t t e r cover (Ke l l y and Walker, 1976; Lepage, 1981). This increases the exposure o f t h e s o i 1 su r f ace t o r a i n d r o p impact , sun and wind, and leads t o more extreme s o i 1 su r f ace temperatures, h i ghe r evapora t ion r a t e s and, u l t i m a t e l y , t o t h e s t r u c t u r a l c o l l a p s e and de f l a t i o n o f t h e so i 1 surface. Impermeable surface sea 1s are o f ten formed, resu l t i n g i n a cyc le o f degradation o f reduced i n f i l t r a t i o n rates, increased r u n o f f , l ower s e e d l i ng es tab l i shment , l ess p l a n t p roduc t i on and f u r t h e r exposure o f t h e s o i 1 su r f ace (Ke 1 ly and Wa lke r , 1976; Macdonald, 1978; Bridges e t al., 1983; Valentin, 1985).

Local d i f ferences i n s o i l type and organic matter production r e s u l t i n cons iderab le s m a l l s ca le v a r i a t i o n s i n s o i 1 n u t r i e n t s ta tus. These d i f f e r e n c e s a r e he igh tened by t h e a c t i v i t i e s o f t e r m i tes. L i t t e r f eed ing t e r m i t e s a r e p a r t i c u l a r ly common i n d y s t r o p h i c savannas where they may consume up to 36% o f annual l i t t e r f a l l , i n c l u d i n g about 60% o f dead wood and grass l i t t e r p roduc t i on (Wood and Sands, 1977; Ohiagu, 1979b; Josens, 1983). Thus c o n s i d e r a b l e amounts o f o r g a n i c m a t t e r and n u t r i e n t s a r e concentrated i n t e rm i t e nests, from where they are s lowly released i n t o the soi 1 as the mounds weather. There have been r e l a t i v e l y few s t u d i e s o f t h e t u rnove r t i m e s o f t e r m i t e mounds. The t u rnove r o f T r i ne rv i termes ermi natus mounds i n a N i ger i an savanna averages 6.3 y e a r s (Ohiagu, h t i n o t h e r spec ies t h e t u rnove r r i m e s a r e much longer. The amount o f o rgan i c m a t t e r and n u t r i e n t s added t o t h e s o i l each yea r i s g e n e r a l l y low (Wood and Sands, 1977). Therefore, wh i le the concentration in, and subsequent slow release o f m a t e r i a l f r om t e r m i t e mounds may conserve o rgan i c m a t t e r and secure n u t r i e n t s a g a i n s t l each ing (Menaut e t al., 19851, i t may equal ly l i m i t t he cyc l i ng o f nu t r i en t s i n savannas.

Soi 1 moisture and the t r ee : grass equi li brium

D i f f e r e n c e s i n s o i l m o i s t u r e r e g i m e s and i n t h e r e l a t i v e a v a i l a b i l i t y o f moiswre and nu t r ien ts t o plants are major fac to rs de te rm in ing t he wide v a r i a t i o n i n t h e s t r u c t u r e o f savanna vegetation. Savannas encompass a range o f physiognomic types, from grassy shrublands a t t he i n t e r f a c e w i t h deserts, through open woodlands and t r e e l e s s edaphic grasslands, t o a lmos t c l osed woodlands w i t h a he l iophyt ic grass understory i n the t r ans i t i on zone t o semi-deciduous and evergreen fo res ts (Huntley and Walker, 1982). The c h a r a c t e r i s t i c f e a t u r e se rv i ng t o l i n k t h i s d i v e r s i t y o f vegetation types i s the norma 1 ly stab l e coexistence o f grasses and t rees, components which i n t he o the r ma jo r biomes tend t o r ep lace one another.

To account f o r the balance between woody plants and grasses i n the dry savannas, Walter (1971) suggested t h a t the s o i l consists o f two func t iona l l y d i s t i n c t layers: a surface layer i n which grasses, w i t h t h e i r shallow compact r o o t systems, r e t a i n and have p r i o r access t o the water enter ing the soi 1; and a lower subsoi 1 layer t o which the deeper-rooted woody p lants have exclusive access. I n terms o f t h i s hypothes i s, t r e e s and grasses can c o e x i s t i n s i t u a t i o n s where t he amount o f water r e g u l a r l y reach ing t h e subsoi 1 i s j u s t enough t o suppor t t r e e s b u t i n s u f f i c i e n t t o enable them t o e s t a b l i s h a cont inuous canopy and shade o u t t h e grasses. The hypothes is has been f o r m a l i z e d by Walker e t al. (1981) and Walker and Noy-Meir (1982).

The l i m i t e d i n f o r m a t i o n ava i l a b l e on t h e r o o t systems o f savanna p lants indicates t ha t there i s considerab l e v e r t i c a l over lap between g rass and t r e e roots , though the maximum dens i t y o f grass r o o t s occurs i n the t op 10 - 20 cm o f t h e s o i l , w h i l e i n t r e e s i t i s g e n e r a l l y below t h i s (Strang, 1969; Rutherford, 1983; Knoop and Walker, 1985). However, bo th have access t o sur face and subsoi 1 water (Russell, 1966; Strang, 1969; Tunsta l l and Walker, 1975; Knoop and Walker, 1985). Desp i te t h i s weakness i n one o f t h e ma jo r assumptions o f the model, i t can s t i 11 be v a l i d p rov ided t h a t the grasses and t r e e s a re each t h e supe r i o r compe t i t o r i n d i f f e r e n t par ts o f the s o i l p r o f i l e o r a t d i f f e r e n t tirnes (Knoop and Walker, 1985).

The h ighe r t r a n s p i r a t i o n r a t e s and more compact r o o t systems o f grasses should g i v e them an advantage i n t he upper l aye rs o f t h e s o i 1. On t h e o the r hand, once t r e e s a re estab l ished, they can outcompete grasses by shadi ng them and, through bei ng longer- li ved, by gradua1 l y a p p r o p r i a t i n g space i n t h e s o i 1 when i t becomes avai lab le. The lower rates o f water-use o f t rees i n the dry savannas a r e a l s o an advantage s i nce t h e ava i l a b l e so i 1 moi s t u r e i n those p a r t s o f t he so i 1 p r o f i l e dominated by t r e e r o o t s w i 11 n o t be depleted so rapidly. This enables trees t o remain act ive f o r longer.

Subsoil moisture levels are determined mainly by the amount o f water draining through from the surface. The r e l a t i v e abundance o f grass and trees w i 11 therefore depend on both the amount o f r a i n f a l l and the water-holding capacity o f the topsoil. I n areas o f low rainfal.1, p a r t i c u l a r l y where t he wate r -ho ld i ng capaci ty o f t he t opso i 1 i s high, most o f t h e incoming wate r remains near t he surface. Th i s a l lows grasses t o reduce the amount o f water eventual l y reaching the subsoi 1, thereby li m i ti ng i ndi r e c t ly the growth o f t rees (Knoop and Walker, 1985). Since t he re i s a shortage o f ava i l a b l e mois ture, ne ighbour ing t r e e s a re p o t e n t i a l compet i tors . Th is i s sometimes r e f lected i n the regu la r i t y o f t h e i r s p a t i a l d i s t r i b u t i o n s (Smi th and Walker, 1983).

Trees are increasingly favoured as more water reaches the subsoil. This general ly occurs on deep, porous sands, stony s lopes, f ractured l a t e r i t i c outcrops and on well-drained so i l s i n regions w i t h high annual r a i n f a l l (Walter, 1971; Walker and Noy-Meir, 1982; Knoop and l ia lker , 1985). As t r e e d e n s i t i e s increase, grass growth dec l i n e s both through shading and through increased cornpetition f o r water and n u t r i e n t s . Tree removal r e s u l t s i n increased grass p roduc t ion (Beale, 1973; Walker e t al., 1972; Dye and Spear, 1982), though t o t a l abovegr'ound p lan t production usual ly declines. I n most cases, grass production i s h i ghest where woody p lants have been comp l e te ly c leared, b u t i n some cases i t reaches a maximum i n l i g h t l y wooded communit ies (Kennard and Walker, 1973). The inc rease i n grass p roduc t i on i s o f t e n accompanied by a change i n grass species compos i t ion and a reduc t i on i n t he abundance o f b e t t e r - q u a l i t y grasses (Dye and Spear, 1982).

Whi l e competi r i o n f o r water may explain the coexistence o f grasses and t r e e s i n t h e d r y savannas, i t does n o t adequate ly account f o r t h e presence o f savanna vege ta t i on under much h igher r a i n f a l l . I n the moist South American savannas, f o r example, the dominant t rees are large ly evergreen, deep-rooted and do no t markedly reduce t h e i r t r a n s p i r a t i o n r a t e s du r i ng t he d r y season (Rawitscher, 1948; Vareschi, 1960; Fo lda t s and Rutk is , 1975; Sarmiento e t al., 1985). Most o f t h e species renew t h e i r leaves i n t he m idd le o f t he d r y season when cond i t i ons seem l e a s t favourab le f o r l e a f expansion (Medina, 1982; Sa rm ien to , 1984). F o r t h e t r e e s t o r e m a i n p h y s i o l o g i c a l l y a c t i v e a t t h i s t ime, they must have access t o adequate moiswre, e i t he r from groundwater sources o r from moisture which has accumulated deep i n the soi 1 p r o f i l e during the previous wet season (Foldats and Rutkis, 1975; Sarmiento, 1984).

A shor tage o f water i s t h e r e f o r e u n l i k e l y t o be t h e main f a c t o r li m i c i ng t r e e d e n s i t i e s i n these savannas. Soi 1 n u t r i e n t leve 1s though a re ex t reme ly low (Lopes and Cox, 1977; Sarmiento, 1984). T h i s i s man i fes ted i n t he s low growth and sc leromorphi sm o f t h e

woody vegetation which, together w i t h the large investment made by these p lants i n t h e i r r oo t systems, l i m i t s the ra te o f development o f a closed t ree canopy. Grasses can therefore coex is t alongside the t r e e s desp i t e be ing dormant f o r most o f t he d r y season. Moreover, because they a r e sha l low- roo ted and r e l a t i v e l y f a s t growing, t h e grasses a r e ab le t o take up n u t r i e n t s r a p i d l y when these a re minera l i zed i n the surface so i 1 dur ing the wet season. This probab ly fu r t he r l i m i t s the avai lab i l i t y o f nu t r ien ts t o trees.

F i r e may also be a fac to r since it l i m i t s the establishment o f t rees and shrubs and retards the development o f a closed canopy, enabling grasses t o c o e x i s t a longs ide t rees. P l a n t s on i n f e r t i l e s o i 1s a r e p a r t i c u l a r l y prone t o f i re . The slowness w i t h which closed t ree and shrub canopies develop enables the grasses, which f ue l the f i res , t o p e r s i s t under t r ees f o r longer (Kellman, 1984). Most o f t h e f i r e s occur d u r i n g t h e d r y season when t h e grasses and o the r herbaceous p lan ts are dormant and therefore r e l a t i v e l y unaffected by f i re .

I n contrast, most woody plants, even those species which tend t o be f i r e - t o l e r a n t as adul ts , a re s e n s i t i v e t o be ing b u r n t du r i ng t h e ear l y stages o f establishment and growth, p a r t i c u l a r l y by i n t e n s e l a t e dry-season f i r e s (S i l v a and Castro, 1985). The biomass o f e s t a b l i s h e d woody p l a n t s i s a l s o reduced as f i r e damages t h e aboveground par ts and retards t he i r growth. F i re, therefore, favours grass p roduc t i on and t h i s , i n turn, f u e l s f u t u r e f i r e s , thereby se t t i ng up a pos i t i ve feedback which maintains both grass and f i r e i n t h e system. Under c o n d i t i o n s f avou r i ng r a p i d and s u b s t a n t i a l grass growth, f i r e s may even be s u f f i c i e n t l y intense t o maintain an open grass land.

Not a l 1 grasslands i n the high r a i n f a l l savannas owe t h e i r o r i g i n t o regu l a r f i r e s . Pure grass lands a l s o occur wherever t h e r e i s poor s i t e dra inage and/or where a sha l low s o i 1 p r o f i l e o v e r l i e s an impermeable c l a y ho r i zon o r l a t e r i t e l aye r (Michelmore, 1939; Sarmiento and Monasterio, 1975; Tinley, 1982). The moisture regime o f these s o i 1s f l u c t u a t e s between an excess o f wa te r d u r i n g t h e r a i n s and exteme water d e f i c i t i n t he d r y season. Th i s does n o t a f f e c t s h a l low-rooted grasses and sedges b u t i s i n i m i c a l t o t h e growth o f the deeper-rooted trees. I n some areas, such as the l lanos o f c e n t r a l Venezuela, t he s c a t t e r e d occurrence o f t r e e s r e f l e c t s l oca l di f ferences i n soi 1 depth and the s t ructure o f the underlying l a t e r i t e layer. The trees occur where the so i1 i s s u f f i c i e n t l y deep and porous so t h a t waterlogging doesn't occur, o r where t h e i r roots are able t o penetrate cracks i n the l a t e r i t e and thereby gain access t o groundwater du r i ng t h e d ry season (Monaster io and Sarmiento, 1968; San Josb and Far inas, 1983). Cornpet i t ion between t r e e s and grasses i s no t a fac to r i n t h i s case.

Changes i n t he abundance o f t r e e s and grasses a t a s i t e i m p l y d i f f e r e n t i a l mor ta l i t y o r s u r v i v a l o f i n d i v i d u a l s o f d i f f e r e n t species. Mater s t ress, f i r e and he rb i vo ry appear t o be t h e main causes o f seed l ing m o r t a l i t y (S i lva, 1973; Penning de V r i es and D j i t k e , 1982; Belsky, 1984). L i t t l e i s known about the processes o f seed ge rm ina t i on and s e e d l i ng establ ishment i n savannas, o r about the r o l e o f competi t ion, pa r t i cu l a r ly from estab li shed plants. Both grasses and t r e e s can i n h i b i t t h e es tab l i shment o f woody p l a n t seedlings, bu t the precise mechanisms are not known (Strang, 1969; Knoop and Walker, 1985). Competition f o r water and nu t r ien ts may be i n d i r e c t l y i nvo lved. Since sma 1 l e r seedl i ngs are general ly the most s u s c e p t i b l e t o wate r s t ress, o r t o be ing b u r n t o r eaten, any l i m i t a t i o n s on g rowth w i 11 inc rease t he r i s k o f m o r t a l i t y . It i s therefore important t o learn what factors favour the recrui tment o f new i n d i v i d u a l s t o a populat ion. Does an inc rease i n p e r c o l a t i o n f avou r an inc rease i n seed p roduc t i on o f a d u l t p lan ts , o r does i t r e s u l t i n improved seed l ing g rowth and s u r v i v a l ? S tud ies o f t h e p o p u l a t i o n b i o l o g y o f se lec ted t r e e and grass species w i l l be essentia 1 i n order t o understand the mechani sms invo lved.

Species composition and coexistence

I n savannas there i s generally an inverse re la t ionsh ip between p l a n t species richness and the moisture and nu t r i en t status o f the soil. Well drained, sandy dystrophic so i 1s can have upto tw ice the number o f spec ies c o e x i s t i n g a t a s i t e compared w i t h s i t e s on eut rophic , poor l y drained c lays (Frost, pers. coma.). The under l y ing mechani sms g i v i n g r i se t o t h i s r e l a t i o n s h i p have n o t been inves t iga ted . Two, n o t mutua 1 ly exc l u s i v e exp l ana t i ons seem p laus ib l e . F i r s t , low n u t r i e n t a v a i l a b i l i t y r e s t r i c t s p l an t growth and reproduction w i t h the r e s u l t t h a t the ra te a t which populations approach compet i t ive equi l i b r i u m i s reduced. Th i s may l i m i t t he dominance o f c e r t a i n spec ies and enable a g r e a t e r number o f p o t e n t i a l compe t i t o r s t o coexi st. I n va r i ab l e envi ronments such as savannas, t h i s coexistence may be prolonged i n d e f i n i t e l y i f t h e compet i ti veness o f d i f f e r e n t spec ies a l t e r n a t e s w i t h t he changes i n t he environment. I n t h i s respect, i t i s worth not ing t h a t whi l e the d i v e r s i t y o f species on c layey s o i 1s i s o f t e n low, t h e changes i n compos i t ion a re more marked, r e f lec t ing the greater amplitude o f change i n condit ions on clays per u n i t change i n r a i n f a l l (O'Connor, 1985).

Secondly, i f nu t r i en t l i m i t a t i o n i s o f prime importance to p lan ts growi ng on dystrophi c so i ls, then se l e c t i on shou l d favour a t t r i butes which enhance a p l a n t ' s c a p a c i t y t o take up and s t o r e nu t r i en t s . These may i n v o l v e s p e c i a l i z a t i o n on a narrow a r r a y o f m i c r o s i t e s w i t h p a r t i c u l a r s o i 1 c h a r a c t e r i s t i c s and n u t r i e n t s ta tus. Where there i s considerable heterogenei t y i n the so i 1, t h i s WOU l d provide more opportuni t ies f o r coexistence and, thereby, a higher diversi ty.

The spat ia 1 heterogenei ty o f savanna soi 1s i s marked. Features such a s - t e r m i t a r i a , l a rge mammal a c t i v i t y s i t e s (dung s i t es , burrows etc.), t he patchy occurrence o f f i r e and grazing, and i n h e r e n t d i scontinui t i e s i n the vegetation (bushc lumps, open and under-tree s i t e s ) , a l 1 c o n t r i b u t e t o d i f fe rences i n s o i 1 c h a r a c t e r i s t i c s . Marked differences i n species composition and abundance can read i l y be observed over shor t distances i n r e l a t i o n t o t h i s heterogeneity, sugges t ing t h a t o p t i m a l s i t e s f o r d i f f e r e n t species are generally l i m i t e d t o one o r a few s i t es along an edaphic gradient. A t the same t ime, s ince es tab l i shment appears t o be a c r i t i c a l stage f o r many plants, some o f these d i f ferences may r e f l e c t processes ope ra t i ng p r ima r i ly a t the seedling stage.

The marked e f f e c t t h a t changes i n s o i l chem is t r y have on grass species composi t i o n i s c l ea r l y indicated by the changes which occur i n response t o f e r t i l i z a t i o n (Mi l l s , 1964; O'Connor, 1985). The e x t e n t o f these changes depends bo th on s o i l type, be ing more pronounced on dys t roph i c sands than on eu t roph i c c lays, and on r a i n f a 1 1, being more apparent i n the wetter savannas, para 1 l e 1 l i n g the observed di f ferences i n species richness.

There a r e a l s o wide d i f f e rences i n t he moi s t u r e requi rements o f savanna plants. This i s r e f lected i n the d i f f e r e n t i a l d i s t r i b u t i o n o f species along so i 1 moisture gradients (S i lva and Sarmiento, 1976; Yeaton e t al., i n prep.); i n d i f f e r e n c e s i n water-use by species g row ing under t h e same c l i m a t i c c o n d i t i o n s (Fo lda ts and Rutkis, 1975; Medina, 1982); and by shor t - te rm changes i n bo th woody and herbaceous community composition occurring during periods o f above- and below-average ra in fa 11 (Poupon, 1979; S i ngh and Krishnamurthy, 1981; O'Connor, 1985). These changes i n spec ies compos i t ion a r e usua 1 ly greater on c layey than on sandy so i ls, r e f lec t ing the more extreme soi 1 moisture regimes o f clays.

The composition o f the herbaceous layer i n dry savannas appears t o be af fected p r imar i ly by year t o year and longer-term var ia t ions i n r a i n f a l l . The e f f e c t s o f g raz ing o r f i r e become more i m p o r t a n t as mean annual r a i n f a l l increases and i t s v a r i a b i li t y dec l ines. For example, i n the Austral ian moi s t t rop ica 1 and subtropica 1 t a l lg rass savannas, introduced ungu lates ( c a t t l e and sheep) and i ncreased f i r e frequencies have apparently caused a change from a Themeda aus t ra l i s dominant understory t o one dominated by Heteropogon contortus (Shaw and Bisset, 1955).

Pheno logy and coexi stence

Seasona 1 v a r i a t i o n s i n water and n u t r i e n t ava i l a b i 1 i ty , t oge the r w i t h coœpe t i t i on f rom o the r p lan ts , and s t r esses imposed by t he annual d r y season, f i r e and herb ivory , c rea te a s p a t i a l l y and temporal ly sh i f t i n g mosaic o f chal lenges and opportuni t ies , which i s

r e f l ec ted i n the wide var ie ty o f phenological behaviour found among savanna p lants . As many as 15 f u n c t i o n a l ly d i s t i n c t pheno log ica l groups have been recognized, based on the seasonali ty/aseasona li ty o f carbon assimi lat ion, continuous versus per iod ic shoot growth, and the ti me o f f lowering (Sarmiento and Monasterio, 1983).

Herbaceous p lants have the w i dest va r ie ty o f pheno logica 1 responses (Menaut and c h a r , 1979; Sarmiento and Monaster io, 1983; Singh e t al., 1985). Most spec ies a re perenn ia l , becoming dormant o r semi- dormant d u r i n g t h e d r y season. Growth d u r i n g t h e we t season i s r a p i d, p a r t i c u l a r ly among grasses. F l owe r i ng and seed s e t usua 1 ly t ake p l a c e l a t e r i n t h e we t season b u t t h e r e a r e wide d i f f e r e n c e s between spec ies i n t h i s r espec t (Menaut and Cka r , 1979; Singh e t al., 19851. Some spec ies f l o u e r i n response t o e a r l y r a i ns , o t h e r s i n response t o l a t e ra ins. Many o f t h e p e r e n n i a l f o r b s f lower towards t h e end o f t he d r y season o r a t beg inn ing o f t h e ra ins . Flowering and seed dispersa1 i n some o f these species i s i n i t i a t e d by f i r e (Coutinho, 1982; Menaut and Cha r , 1979). Most species, though, reproduce on a f i xed schedule, i r respec t i ve o f when the r a i n fa I l s o r f i r e s burn. A few spec ies grow cont inuous ly ; o t h e r s a r e opportuni sts, growi ng whenever condi t ions are favourab le.

A l a r g e number o f annuals occur i n some savannas. These spec ies d i f f e r widely i n t h e i r t im ing and r a t e o f development, and include some spec ies which a r e capable o f comp le t i ng t h e i r l i f e c y c l e s r ap id l y and opportuni s t i c a 1 ly whenever condi t ions are sui tab l e (van Donse laar-Ten Bokke 1 Hui n ink, 1966; Menaut and C k a r , 1979; Sarmiento, 1984). I n a r i d savannas, the di f ferences i n germination t i m e r e f l e c t responses t o two opposing s e l e c t i o n pressures: t h e advantage o f r ap id germination enabling ind iv idua 1s t o e x p l o i t f u 1 l y t he b r i e f f l u s h o f wa te r and n u t r i e n t s a t t h e s t a r t o f t h e s h o r t g row ing season, and t he r i s k t h a t subsequent r a i n s w i 11 f a i 1 and t h a t t h e p l a n t s w i 11 become dess ica ted be fo re they can reproduce (Penning de Vr ies and D j i t k e , 1982). Given the inherent v a r i a b i l i t y o f savanna environments, pa r t i cu l a r ly the stochast ic nature o f the r a i n fa 1 1, the d i ve r s i ty i n pheno logy found among herbaceous savanna p lants i s probab ly a key feature promoting t he i r coexistence.

The phenology o f woody p l a n t s i s much l ess va r iab le . Most o f t h e species, i n A f r i can , Aus t ra l i a n and I n d i a n savannas a r e deciduous (Malaisse, 1974; Menaut, 1983; Sarmiento and Monasterio, 1983; Mot t e t al., 19851. They shed t h e i r leaves du r i ng t h e d r y season, remain lea f less f o r a few weeks t o months, then produce new leaves p r i o r t o o r a t t h e beg inn ing o f t h e f o l l o w i n g we t season. The t i m e o f l e a f f a11 appears t o be r e l a t e d t o water s t ress , t h e t r e e s r e t a i n i n g t h e i r leaves f o r longer i n years o f h igh r a i n f a l l . Some species are on ly b r i e f ly deciduous, the o l d leaves f a l l i n g j u s t before the new ones emerge a t the s t a r t o f the annual rains. Leaf-out i n deciduous species co i nc i des w i t h an i ncrease i n both mean dai ly temperature and photoperiod (Rutherford and Panagos, 1982).

Lea f p roduc t i on i n t h e deciduous spec ies o f t h e m o i s t savannas i s large ly determini s t ic . This and annual shoot growth are apparently der ived from carbohydrate and n u t r i e n t reserves stored i n the plant. The g r e a t e r p a r t o f t h e wet season i s devoted t o pho tosyn tha te p r o d u c t i o n and t h e rep len ishment o f these reserves, as we 11 as t o r a d i a l growth and r o o t extension (Rutherford and Panagos, 1982). I n contrast, i n the d r i e r savannas, shoot growth i s indeterminate and depends on condi t ions dur ing the cur rent growing season.

Evergreen savanna t r e e s and shrubs a r e o n l y common i n t h e h i g h r a i n f a l l savannas o f South America. These species, which a re deep r o o t e d and genera l l y appear t o have access t o an adequate wa te r supply, replace t h e i r fo l i age dur ing the middle o f the dry season. T h i s f a c i l i t a t e s t h e n u t r i e n t economy o f these t r e e s i n t w o ways. F i r s t , i t enables nu t r ien ts t o be withdrawn from senescing leaves and t o be rea 1 l oca ted immed ia te l y t o t h e new ly deve lop ing leaves. Secondly, because the expansion o f the new leaves, which are h igh ly leachable, takes place dur ing the dry season, the r i s k o f n u t r i e n t loss through leaching i s reduced (Sarmiento e t al., 1985).

Flowering and f r u i x i ng are a lso strongly seasona 1. I n Austra l i a n and South American savannas, f lowering occurs mainly dur ing middle o f the dry season and i s f requent ly associated w i t h de fo l i a t i on by f i r e o r drought (Coutinho, 1982; M o t t e t al., 1985). F r u i t s mature r a p i d l y and t h e seeds o f most spec ies a re d ispersed d u r i n g t h e f o 1 l o w i ng wet season (Sarmiento, 1984). I n A f r i c a n savannas, f l ower ing i s concentrated a t the end o f the dry season and beginning o f t h e we t season, o c c u r r i n g a t t h e same t i m e as leaf -out . F r u i t s mature d u r i n g t h e we t season and a r e d ispersed th roughou t t h e f o l lowi ng dry season (Hopkins, 1970; Ma laisse, 1974; Menaut, 1983).

P l a n t production

Organic i na t t e r p roduc t i on and p l a n t q u a l i t y i n savannas depend on the t o t a l amount and seasonal d i s t r i b u t i o n o f r a i n f a l l , and on the a v a i l a b i li t y o f n u t r i e n t s , p a r t i c u l a r l y n i t rogen and phosphorus (San Jose and Medina, 1976; Penning de Vr ies and D j i t b e , 1982; Mot t e t al., 1985). The inf luence o f water avai l a b i l i t y i s most apparent i n t h e d r i e r savannas, as i n d i c a t e d by t h e p o s i t i v e c o r r e l a t i o n s between annua 1 r a i n f a 11 and, f o r examp le, herbaceous aboveground biomass (Wa l t e r 1971; San Joséand Medina, 1976; Ruther ford, 1981; Deshmukh, 1984; Singh e t al., 1985). However, t h e r e l a t i o n s h i p i s on ly very general and i s a f fec ted by fac to rs such as the durat ion o f t h e we t season, s o i 1 t ype and t ex tu re , n u t r i e n t ava i l a b i l i t y , temperature, f i r e and spec ies compos i t i on (San Jose and Medina, 1975; Rutherford, 1981; Penning de Vries and D j i t b e , 1982; Singh e t al., 1985).

The d i f f e r e n t patterns o f p l a n t production on clayey and sandy s o i l s occurr i ng a long r a i n f a l 1 gradients, o r i n response t o f luc tuat ions i n annual r a i n f a 11 a t a s i te, c l ea r l y i 1 lust ra tes the inf luence o f so i 1 texture on water avai l ab i li ty and production. When r a i n f a l l i s low, p roduc t i on on c l a y s may be as low as, o r even l o u e r than t h a t on sandy soi 1s under the same ra i n fa l l , despite the generally higher n u t r i e n t s t a t u s o f c lays. However, p roduc t i on on c l a y s inc reases more r a p i d l y w i t h i n c r e a s i n g r a i n f a l l (Dye and Spear, 1982). The r e s u l t i s tha t , i n r e l a t i o n t o t h e same f l u c t u a t i o n s i n r a i n f a l l , production on clayey so i 1s i s much more var iab le than on sands.

As r a i n f a l l and the length o f the n e t season increase, water becomes less l i m i t i n g and other factors, such a low nu t r i en t ava i l ab i l i t y , begin t o l i m i t production (Mott e t al., 1985). For example, grasses g row ing on i n h e r e n t l y f e r t i l e so i l s , o r on ones which have been a r t i f i c i a l ly f e r t i l ized, are much more productive than those growing on i n f e r t i l e o r u n f e r t i l i zed s i tes (Mi 1 ls, 1968; San José and Garcia Miragaya, 1981; Penning de Vries and D j i t k e , 1982). The e f f ec t s are more pronounced when water i s not l i m i t i n g (Donaldson e t al., 1984). I m p r o v e d s o i 1 f e r t i l i t y i n c r e a s e s t h e use o f w a t e r by t h e vegetation, improves water-use ef f ic ienc ies, and thus contr ibutes t o h i gher production. However, the avai lab l e soi 1 moi sture i s used up more r a p i d l y , thereby shor ten ing t h e p e r i o d f o r a c t i v e growth and i ncreasi ng the ri sk o f physi O logica 1 drought.

The l e v e l o f p roduc t i on i s s t ong l y i n f luenced by t h e r e l a t i v e ava i l a b i l i t y o f n i t r o g e n and phosphorus. Add i t i ons o f n i t r o g e n o r phosphorus alone do not r e s u l t i n s i gn i f i can t increases i n biomass production unless the other nu t r i en t i s already present i n excess. Ni t rogen-f ix ing legumes, f o r examp le, show a pos i t i ve response t o t h e a d d i t i o n o f phosphorus a lone (Penning de V r i es and D j i t b e , 1982). However, when n i t r o g e n and phosphorus a r e added simultaneously there i s a massive i ncrease i n biomass product ion, p a r t i c u l a r l y o f grasses (Norman, 1962; Medina e t al., 1978; San José and Garcia Miragaya, 1981; Penning de Vr ies and D j i t ke , 1982; Mot t e t al., 1985).

Overall, organic matter production i n savannas i s l i m i t e d mainly by n i t r o g e n supply. The grasses o f t h e humid, dys t roph i c savannas o f South America, f o r example, seem t o be adapted t o low phosphorus ava i l a b i l i t y . F e r t i l i z a t i o n exper iments conducted i n A u s t r a l i a (Norman, 19621, South A f r i ca (Weinmann, 1938) and Venezuela (Medina e t al., 1978; San Josb and Garcia-Miragaya, 1981) show t h a t when n i t r ogen i s l i m i ted, f e r t i l i z a t i o n w i t h phosphorus inc reases t h e phosphorus content o f the newly produced aboveground biomass. This i s n o t r e t r a n s l o c a t e d be fo re canopy d r i e s a t t he end o f t h e r a i n y season (Medina, i n press).

Low n u t r i e n t ava i l a b i l i t y i s n o t con f i ned t o t h e w e t t e r savannas. There i s a h i g h p o t e n t i a l f o r n u t r i e n t l i m i t a t i o n i n s i t u a t i o n s where growth must be completed i n a r e l a t i v e shor t period. Rapidly g row ing p l a n t s have a h i g h n u t r i e n t demand and can t e m p o r a r i l y deplete the so i 1 o f avai lab le nutr ients, p a r t i c u l a r l y where n u t r i e n t m inera l i za t ion i s in termi t tent . This i s an important cons t ra in t i n a r i d savannas where the growing season i s shor t and where the stock o f avai l ab le nu t r ien ts i s sustained by m inera l i za t ion o f a l i m i t e d amount o f so i 1 organic matter (Penning de Vries and D j i t b e , 1982). Differences i n the t imes o f peak nu t r i en t ava i lab i l i t y add a fu r the r comp l i ca t i on . I n t h e Sahe 1, f o r examp le, n i t r ogen minera l i z a t i o n peaks a t the s t a r t o f the wet season when the p lants are l i m i t e d by a lack o f phosphorus. Later, as the wet season progresses, n i t rogen i s increasingly immobil ized by the soi 1 b i o ta and t h i s eventual ly l i m i t s p l an t growth (Penning de Vr ies and D j i t b e , 1982).

The increas i ng nu t r i en t const ra in ts faced by savanna p lants through the growing season are c l e a r l y seen i n the seasonal changes i n p l an t qua l i t y (Shaw and Bisset, 1955; Afolayan and Fafunsho, 1978; Mot t e t al., 1981; Penning de V r i e s and D j i t b e , 1982). N i t r ogen and phosphorus con ten t s i n leaves a re h i gh a t t h e s t a r t o f t he wet season and decl ine progressively as the p l an t grows. P lan t qua l i t y i s espec ia l l y low d u r i ng t h e d r y season and and t h i s becomes an i m p o r t a n t c o n ~ t r a i n t on secondary product ion. Average pas tu re q u a l i t y i s o f ten below the minimum maintenance levels f o r domestic l i v e s t o c k and w i l d ungulates. The an ima ls compensate f o r t h i s d e c l i ne t o some extent by foraging se lec t i ve ly on those components i n t h e herb l a y e r which have h i ghe r than average fo rage q u a l i t y (Bremen and de W i t , 1983). Nevertheless, they s t i 11 exper ience a loss i n decline i n body condit ion.

Attempts t o overcome these def ic ienc ies i nc lude the use o f m i nera l supp lements (urea/molasses m ix tu res ) , c r o p r e s i dues, f o d d e r reserves, and the improvment o f pasture qua l i t y by in t roduc ing p l an t spec ies ( p r i n c i p a l l y legumes) which w i 11 p r o v i d e t h e necessary e lements ( T o t h i 11 e t al., 1985). Supplementation general ly resu l t s i n heavier s tock i ng rates, i ncreased natura 1 pasture u t i l i z a t i on and reduced i nc i dence o f f i r e . The i n t r o d u c t i o n o f a l i e n spec ies i n t o n a t u r a l p a s t u r e s r e q u i r e s t h e a d d i t i o n o f n o n - n i t r o g e n o u s f e r t i l i z e r s , u s u a l l y superphosphate. Th i s o f t e n r e s u l t s i n h igher s o i 1 f e r t i li ty , improved fo rage q u a l i ty, and inc reased c a r r y i ng c a p a c i t i e s f o r l i ves tock . I n areas w i t h a long, harsh d r y season, t h i s can cause a breakdown i n the s tab i l i t y o f the herbaceous layer w i t h annual rasses and weeds replacing the nat ive perennia 1s (Mott e t al., 1981 3 . The i n f l u e n c e o f these management a c t i o n s on t h e longterm stab i li ty o f these systems needs t o be investigated.

The e f f e c t o f f i r e on herbaceous production depends both on the t ime o f burni ng and on so i 1 moi s ture avai l ab i li ty. F i r e tends t o i ncrease grass production i n high r a i n f a l l areas w i t h a shor t dry season, b u t i n d r i e r a r e a s p r o d u c t i o n i s g e n e r a l l y reduced r e l a t i v e t o neighbouring unburnt areas (West, 1965; San José and Medina, 1975). The season o f burn has an i m p o r t a n t bear ing on t h e outcome. E a r l y dry season f i res i nduce p lants t o f l u sh a t a ti me w hen so i 1 moi s ture l e v e l s a r e a l r eady dec l in ing . Th i s r eg row th r a p i d l y dep le tes t h e remai n i ng s o i 1 moi s t u r e and t h e ti 1 l e r s do n o t surv ive. The po ten t i a l st imulus t o production i s therefore no t sustai ned. Growth a t the s t a r t o f the fo l low ing wet season i s i n i t i a t e d simultaneously i n bo th b u r n t and unburn t p lants . Burn ing at , o r soon a f t e r t he : s t a r t o f the f i r s t r a i ns general l y has the same outcome except where growth i n unburnt p lants i s l i m i t e d by an accumulation o f standing dead matter. ,.

i n con t ras t , d e f o l i a t i o n by f i r e du r i ng t he l a t t e r p a r t o f t h e d r y season resu l t s i n equ i l i b ra t i on o f p l an t and so i1 water potent ia ls, a l l o w i n g p l a n t s t o grow. Prov ided t h a t t h e so i 1 mo i s tu re s t o r e i s replenished by ea r l y wet season ra ins before it i s again depleted by t h e grasses, t h i s e a r l y s t a r t t o t h e growing season r e s u l t s i n a h i g h e r p roduc t i on by b u r n t p l a n t s (San Jose and Medina, 1975). Depending on the amount o f herbivory, the dry season standing crop o f grass on r e g u l a r l y b u r n t areas can be h i ghe r than on unburn t plots. This increases the probabi l i t y o f f i r e , se t t i ng up a pos i t i ve feedback loop t h a t serves t o maintain a high f i r e frequency and high g rass production.

P lant q u a l i t y

Primary production i n savannas has three possible fates: i t can be consumed by herbivores, i t can be burnt, o r i t can senesce, d i e and decompose. The amount of p l an t matter incorporated d i r e c t l y i n t o the l i t t e r and so i 1 depends on the extent o f both herbivory and f i r e . I n turn, the incidence o f f i r e i s largely a funct ion o f the dry season standing crop o f grass, a product o f the leve l o f production during the preceding wet season and the i n t ens i t y o f herbivory. The major de te rm inan t o f h e r b i v o r y appears t o be p l a n t q u a l i t y (Be l 1, 1982, 1984).

P l an t qual i ty , i n terms o f the su i tab i l i t y o f p l an t matter as food f o r organisms f u r t he r along the food chain, can best be def ined as t h e r a t i o o f a s s i m i l a b l e p l a n t m a t t e r (main ly p ro te in -based compounds and soluble carbohydrates) t o the amounts o f unusab l e and i n h i b i t o r y mate r ia l i n a p l an t (mainly f ib re , l i g n i n and secondary chemica l compounds), r a t h e r than as t h e abso lu te amount o f c rude p ro te i n and soluble carbohydrate present i n a p lan t (Bell , 1981).

The a v a i l a b i l i t y o f s o i l wa te r and n u t r i e n t s such as N and P i n f l uence p l a n t q u a l i ty by a l t e r i n g t h e r e l a t i v e p roduc t i on o f cy top lasm ic and s t r u c t u r a l components.Tota1 biomass and p r o t e i n p roduc t i on b o t h inc rease w i t h i nc reas ing wate r and n u t r i e n t avai l ab i li ty (San José and Medi na, 1976; Medi na e t al., 1978; Be 11, 1981, 1982; Penni ng de V r i es and D j i the, 1982; Breman and de W i t , 1983). However, t h e p roduc t i on o f carbon-based ,compounds such as f i b r e i s l ess a f f e c t e d by a shortage o f n u t r i e n t s than i s p r o t e i n production, so t h a t where nu t r ien ts are l i m i t e d r e l a t i v e t o water, p l a n t q u a l i t y i s low. On t he o the r hand, p r o t e i n p roduc t i on i s s t i m u l a t e d more than f i b r e p roduc t i on by an inc rease i n t h e ava i l a b i l i t y o f nu t r i en t s . Th i s inc rease may be abso lu te (eg. through f e r t i l i za t i on ) o r r e l a t i v e (eg. through a decrease i n water avai l ab i li ty.

P l a n t q u a l i t y , the re fo re , w i 11 be a f f e c t e d by any process which a l t e r s the r e l a t i v e avai lab i l i t y o f water and nu t r ien ts t o plants. Fo r ins tance, low r a i n f a l l , low i n f i l t r a t i o n r a t e s caused by so i 1 capping, low p l a n t cover, o r degradat ion o f t h e s o i l su r f ace by herbivore trampling, a l 1 r e s u l t i n low water avai l ab i l ity. This, i n turn, resu l t s i n t h e p roduc t i on o f a sma 11 amount o f good q u a l i t y p l a n t biomass which may r e s u l t i n a g rea te r p r o p o r t i o n o f t h e v e g e t a t i on b e i ng consumed, less p l a n t cover, more t ramp li ng and i n c r e a s e d s o i 1 deg rada t i on . I n c o n t r a s t , i n c r e a s i ng w a t e r ava i l a b i l i t y r e s u l t s i n t h e p roduc t i on o f more p l a n t biomass o f i n c r e a s i n g l y poor q u a l i t y (Breman and de W i t , 19831, lower an ima l consumption, and an i ncreasi ng proport ion o f the avai lab le nu t r ien ts being concentrated i n the plants. This leads t o a lower avai lab i li ty o f n u t r i e n t s i n t h e s o i l , lower p l a n t q u a l i t y and even l ess consumption.

Vegetati on dynamics

The r e l a t i o n s h i p between wate r and n u t r i e n t a v a i l a b i l i t y a t d i f f e r e n t layers i n the soi l , competit ion between p lants f o r these resources, and t h e r e s u l t i n g pa t t e rns o f p l a n t p roduc t i on and q u a l i t y , a re t h e key t o understanding t h e dynamics o f savanna vege ta t i on i n r e l a t i o n t o he rb i vo ry and f i r e (Be l l , 1981, 1982, 1984). For eastern, cen t ra l and southern Af r ican savannas, four main patterns o f herbivory, f i r e and vegetation dynamics emerge, based on the r e l a t i v e avai l a b i l i t y o f soi 1 moisture and so i 1 nu t r ien ts (Bell, 1984). Wi th t h e excep t ion o f t h e numbers and impac t o f l a rge herbivore populations, which are character i s t i c o f these savannas, the patterns appear t o have counterparts i n other savanna regions.

I n areas o f low wa te r ava i l a b i l i t y and h i g h n u t r i e n t supply, t h e vege ta t i on cons i s t s l a r g e l y o f h i g h q u a l i ty grass lands and open woodlands. Examples include t he Serenget i and Amboseli r eg ions o f East Africa, par ts o f the Deccan o f Ind ia and the Astrebla-dominated

grass lands on t h e Bark ly Tab le lands o f Aus t ra l i a . I n A f r i c a these areas support a high biomass and wide d i ve rs i t y o f indigenous large herbivores i n which smaller-bodied species r e q u i r i n g h i gh qua l i t y d i e t s predominate. Where these species a re now absent, t h e r e i s usua 1 ly an equiva l e n t biomass o f domestic livestock.

A l a rge p r o p o r t i o n o f t he annual above-ground p roduc t i on i s consumed. For example, on t h e Serenget i p l a i ns , an average o f 66%, and up to 94%, o f n e t aboveground r i m a r y p roduc t i on i s consumed annual ly by large herbivores alone f McNaughton, 1985). I n t h i s case, t h e an ima ls a re i nduc ing subs tan t i a 1 compensatory g rowth i n t he p lants thereby c r e a t i ng a h i gh nu t r i en t demand and promoting h i gh n u t r i e n t uptake r a t e s ( ~ ~ N a u g h t o n , 1983b, 1985). Because o f t h i s compensatory growth, the energy f low t o consumers i s add i t i ve and can, i n some cases, be ma in ta ined w i t h l i t t l e o r no r e d u c t i o n i n p l a n t biomass (McNaughton e t al., 1983).

R e g u l a r d e f o l i a t i o n appears t o be one o f t h e m a i n f a c t o r s maintaining the high number o f coexist ing grass species. Changes i n spec ies compos i t ion and a d e c l i n e i n d i v e r s i t y come about when an area i s protected from grazing o r f i re . Para1 l e1 changes, invo lv ing a d i f f e r e n t s u i t e o f species, occur i n areas t h a t a re exposed t o very heavy o r very f requent defo 1 i a t i on (McNaughton, 1979, 1983a). However, t h i s does n o t necessa r i l y i m p l y t h a t t he most d i ve rse community i s the most moderate ly grazed (McNaughton, 1983a1, though t h i s can happen (Singh, 1976). Overa l l , f i r e i s p robab ly less important than grazing, p a r t l y because i t i s a dry season phenomenon occurr ing a t a t ime when the grasses are dormant, and p a r t l y because t he l e v e l s o f consumption a r e u s u a l l y so h i gh t h a t t h e r e i s i n s u f f i c i e n t grass biomass l e f t du r i ng t h e d r y season t o f u e l a f i r e . The i nc i dence o f f i r e i s t h e r e f o r e low, o t h e r than a f t e r exceptional l y wet years. Overa 11, these systems can be characterized as being high l y var iab le bu t r e s i l i e n t (Norton-Gri f f i ths 1979).

A t the other extreme are systems i n which water a v a i l a b i l i t y i s high b u t t he s o i l s a re ex t reme ly nu t r ien t -poor . Such systems a r e dominated by t rees and shrubs and include the miombo woodlands o f the Central Af r ican plateaux, the Cam O cerrados o f Brazi 1, and the monsoonal t a l l g r a s s woodlands O 7- Aus t ra l i a . Because o f t he h i g h r a i n f a l l and prolonged n e t season, water i s seldom l im i t ing , even on s tony o r l a t e r i t i c s o i ls, o the r than on a seasonal basis. The grasses a r e t a l l , f i b rous , and o f low n u t r i t i o n a l q u a l i t y . Grass p roduc t i on i s h i g h e s t i n t he open, between trees. Grasses growing under t r e e s b e n e f i t f rom the more mesic c o n d i t i o n s and h ighe r n u t r i e n t status o f the s o i l s but produce less biomass than grasses growi ng i n the open. This abundance o f low-qua li ty p lan t materia 1, i n A f r i c a a t leas t , suppor ts o n l y a low biomass and d i v e r s i t y o f l a r g e herb ivores. The amount o f p l a n t m a t t e r t h a t they consume annua 1 l y i s genera 1 ly low, though per iodic outbreaks o f lepidoptera larvae, one o f t h e c h a r a c t e r i s t i c f ea tu res o f these systems, can cause severe defol iat ion.

Most p r ima ry p roduc t i on goes i n t o an inc rease i n s tand ing c rop biomass o r i s r ecyc led as l i t t e r f a l l . L i t t e r q u a l i t y i s low and decomposes r a t h e r s l o w l y i n t h e absence o f f i r e . The decomposer biomass, p a r t i c u l a r l y o f t e r m i t e s and, i n t he w e t t e r savannas, earthworms, i s ve ry h i g h and may equal o r exceed t h a t o f p r i m a r y consumers. Dry season f i r e s occur regular ly and are f ue l l ed by of ten large amounts o f dead o r dry mater ia l which has accumulated by the end o f t h e g row ing season. F i r e s a re . s u f f i c i e n t l y i n t ense t o k i 11 seedlings and saplings, thereby e l im ina t ing f i re-sensi t i v e species. However, f i r e i s se ldom a b l e t o c o n t r o l a l 1 woody p l a n t establishment and growth. Once estab li shed, many woody species are ab l e t o s u r v i v e t h e s e f i r e s by r e s p r o u t i ng f rom e x t e n s i v e underground r o o t systems (Frost 1984).

The presence o f a h i g h underground biomass and t h e maintenance o f substant ia l carbohydrate and n u t r i e n t reserves, t oge the r w i t h t h e ab i l i t y t o resprout i f damaged, means t ha t the woody p lants i n these systems are w e l l buffered against per iodic disturbance i n the form o f f i re , drought and the occasional outbreak o f herbivorous insects. The systems are t h e r e f o r e r e l a t i v e l y s tab le . I f sub jec ted t o disturbance, such as the c lear ing o f woodland f o r cu l t i va t ion , they tend t o r eve r t t o woodland once the source o f disturbance has been rel ieved, though t h i s may take some time, p a r t i c u l a r l y i f there has been a loss o f woody p lan t propagules (Robertson, pers. comm.). The r a t e o f recovery depends l a r g e l y on t he e x t e n t o f d e c l i n e i n t h e o rgan ic m a t t e r p o o l and t h e r e f o r e on t he n u t r i e n t s t a t u s o f t h e system f o l lowing clearing.

Two other func t iona l l y d i s t i n c t but less widespread systems occur. The f i r s t consi s t s o f low-medium qua l i t y grasslands w i t h few woody plants. These are found i n areas where both water and nu t r ien ts are l imi ted. They usual ly occur under r e l a t i v e l y high r a i n f a l l on s i t e s w i t h sha l low s o i 1s ( f o r example, some o f t h e up land p la teaux o f C e n t r a l A f r i c a , campo su jo and campo l impo grass lands i n Cen t ra l B r a z i l , and p a r t s o f t h e l l a n o s o f Venezuela), o r under lower r a i n f a l l on nutrient-poor . s a n d s o r examp le, t h e southern Sahel). I n Afr ica, such areas support a low biomass o f f ibre-tolerant, and species- and p lant -par t se lect ive grazers. The densi ty o f herbivores i s generally too low t o reduce the biomass o f grass s u f f i c i e n t l y t o exclude dry season f i res , where these occur. These f i r e s l i m i t the es tab l i shmen t o f woody plants,. b u t have l i t t l e e f f e c t on t h e grasses, most o f which a re dormant a t t h e t ime. Because o f t h e r e l a t i v e l y low p l a n t product ion, s o i 1 o rgan ic m a t t e r l e v e l s a re o f t en low. Any reduct ion i n inpu t resu l t ing from excessive herbivory o r f i r e leads t o a d e c l i n e i n t he amount o f o rgan ic m a t t e r which only recovers slow l y because o f the low leve l o f p l an t production.

Secondly, there are areas i n which ne i ther water nor nu t r ien ts are l i m i t e d These occur i n some R i f t va l ley s i tua t ions i n Africa, and a long t he f o o t h i 11s o f t h e ma jo r escarpment zones. The vege ta t i on consi s ts p r imar i ly of medium-high qua l i t y woodland w i t h palatable grasses. I n A f r i c a , such systems suppor t a h i g h biomass and d i ve rs i ty o f both f i bre-tolerant and select ive grazers and browsers. These a r e p o t e n t i a l l y t h e most uns tab le o f savanna systems (Be l l , 1984). If dis tu rbed , f o r example by the d e s t r u c t i o n o f t h e woody canopy, they change rap id ly t o one o f a number o f a l ternate States which can be maintained by a combination o f f i r e , p lan t cornpetition f o r water, and the type and i n tens i t y o f herbivory.

For example, where woody p lan t density i s reduced, grass production increases, as does the amount o f water used by the grass layer. This may reduce the amount avai lab l e t o trees and so li m i t the i r growth. Subsequent changes depend on t he q u a l i t y o f t he grass a f t e r t r e e c l e a r i ng. I f grass qua l i t y i s low then f i b r e - t o l e r a n t herb ivores w i 1 1 be favoured. However, these species seldom reduce the amount o f grass s u f f i c i e n t l y t o exclude f i r e s . The regenera t ion o f woody species i s i n h i b i t e d and an open grassland maintained.

I f the grass q u a l i t y i s r e l a t i v e l y high, then sma l le r , more s e l e c t i v e g razers w i 11 be favoured. These species can reduce t h e biomass o f grass s u f f i c i e n t l y t o lower bo th t h e frequency and i n tens i t y o f f i res. The system may stabi l i z e a t t h i s point, w i t h a h i g h biomass o f g razers and a h i gh p roduc t ion b u t low biomass o f grass. However, i f the compet i t ive pressures exerted by grasses on woody p lan t seedlings are per iod ica l ly reduced, f o r example by the combined e f f ec t s o f heavy grazing and drought, o r i f the c l imate and soi 1s favour the growth o f trees, then woody p lants wi 11 gradua1 ly re-estab li sh and eventua 1 l y suppress grass production.

Where c o n d i t i o n s do n o t favour the r a p i d growth o f woody p lan ts , browsers a lone can someti mes check t h e i r i ncrease. More usua 1 ly, browsers and f i r e i n t e r a c t . Regular f i r e s ma in ta i n woody p l a n t s w i t h i n the reach o f browsers and i n a n u t r i t i o n a l l y acceptable s ta te (Trol lope 1974). The low biomass o f woody p lants favours the growth o f grasses which i n t u rn provide f u e l f o r l a t e r f i res. Occasionally, f i r e and b rows ing a re unable t o check t he inc rease i n woody p l a n t biomass, f o r example, d u r i n g a s e r i e s o f ve ry d r y years. Woody p lants may then increase t o the po in t where they can suppress grass growth, thereby reducing both f u e l loads and f i r e in tens i t ies .

HUMAN INFLUENCES

Humans have been assoc ia ted w i t h savannas f o r a long t ime, as hunters, pas to ra li s t s and CU l t i v a t o r s . Thei r c u r r e n t i n f l u e n c e i s widespread and involves a range o f land-use pract ices which modify savannas t o v a r y i n g degrees (Table 1). Some o f these a c t i v i t i e s ,

such as wood h a r v e s t i n g and t he c l e a r i n g o f land f o r c u l t i v a t i o n , have d i r e c t e f fects on s o i l s t ructure and f e r t i l i t y , the cyc l ing o f n u t r i e n t s , and on t he compos i t ion o f vege ta t i on (UNESCO, 1979). Other a c r i v i t i e s , such as t he use o f f i r e , t he hun t i ng o f w i l d herbivores, and t h e keeping o f l i ves tock , a f f e c t t h e s o i l and vegetation ind i rect ly . For examp le, i ncreasi ng c a t t l e numbers leads t o greater grass consumption, less p l an t cover, increased trampl ing and compaction o f the so i 1, decreased . i n f i l t ra t ion , greater runof f and e r o s i o n , l e s s p l a n t a v a i l a b l e m o i s t u r e , and i n c r e a s e d a r i d i f i c a t i o n of the soi 1. The reduced grass growth a lso resu l t s i n fewer, o f ten less intense f i r e s and less competit ion f o r woody p lan t seedlings, i n some cases leading t o marked increases i n the density o f these speci es.

Table 1. Types o f land-use i n savannas and t h e degree t o which these G d i f y savanna s t ructure and functioning.

.................................................................... Land use WILDLIFE PASTORALISM WOOD CULTIVATION

Degree CONSERVATION HARVESTING o f AND modi f icat ion UTILIZATION ---------------T----------------------------------------------------

e MINIMUM Protected

c> Nomadi sm Timber Shi f t i n g

4 U

Areas (eg. Transhumance Fue 1 CU l t i v a t i o n

Z Nat i ona 1 .#- ?Y

Parks) - O

MEDIUM Game m

Ranchi ng Timber Conventi ona 1 E Ranching Charcoa 1 agr i cu l tu re V) 4

- dry land aJ L croppi ng U

fi MAXIMUM Domestication Cult ivated A f fo res t - I r r i g a t e d I I pastures: a t ion crop I I z e r o v p r o d u c t i o n graz i ng ....................................................................

Increasing development--------- +

Savannas on t h e d i f f e r e n t con t i nen t s have been exposed t o these impacts f o r very d i f f e r e n t lengths o f t ime and t h i s may be r e f lected i n t h e i r present structure. Moreover, w i t h the sharp increase i n the human popu la t ion , t h e i n t e n s i t y o f many o f these impacts has increased markedly i n recent times. I n the dry savannas especial ly, the combined e f f ec t s o f high animal numbers and reduced p lan t cover,

.S~II IL~~J k o ~ k q ~ a q pue aJkJ ' i u a b ~ i n u aie^ ayz JO s l re iap aq i 6 ~ i p ~ P i S J a p ~ n se ~ e k ~ n ~ 3 se mou S B eaJe ue JO Â ~ o i s k q asn-pue1 J P l n 3 ~ i J P d a q i ~ U ~ M O U ) ( * s u ~ a l l e d pawasqo a q i JO Âueui JO iueui tuJa iap ~ o r e u i e mou aJe suemnq zeql az~u603aJ oz aJoJaJaqi paau aM 'seuue~es Lep-iuasa~d .JO 6 u r u o k i 3 u n ~ PUP a d n z 3 n ~ i s a q i pueisJapun oz 6 u ~ z d u i a i i e UI

*mazsÂs aq i 40 &k~!i3npOJd pue Âzk~ i i ~ a g a q i u i a u i l s a p d~ ieqs e 01 6 u ~ p e a l pue ' u o ~ z e i a 6 a ~ pue Jaiem '1 LOS aqa uo SaJnsSa~d ~ a l e a ~ 6 uaAa 6 u ~ i P a J 3 ' 1 l e j u t e ~ a6e~a~e-r ro laq JO s ~ e a A iuanbasqns JO s i 3 a j ~ a a q i uaiy6iaq asaql -[LOS a q i ub sa6ueq3 a l q LSJaAaJJ 1 Â ~ z u a ~ e d d e 0 1 p a l aAi?i( '~146110Jp 6 u l ~ n p 4 Ji?[ n 3 k i ~ e d

SECTION - III: HYPOTHESES -- ABOUT THE RESPONSES - OF SAVANNAS TO STRESS -- AND DISTURBANCE

INTRODUCTION

The underlying mot ivat ion f o r t h i s programme i s the need t o improve savanna management. Two issues are o f major concern: (1) a l t e ra t i ons t o t h e e c o l o g i c a l s t r u c t u r e o f savanna communit ies, i n v o l v i n g changes i n species composition, r e l a t i v e abundance and re la t ionsh ips between species; and, (2) changes i n f unc t i on i ng , p r i n c i p a l l y dec l i nes i n p r o d u c t i v i t y , r e s u l t i n g f r om changes i n wa te r and n u t r i e n t a v a i l a b i l i ty . To understand t h e i m p l i c a z i o n s o f these changes f o r t h e dynamics o f savannas, and t h e r e f o r e f o r t h e i r management, f o u r key ques t ions have been posed. I n a t t e m p t i n g t o answer these questions, we propose t h a t the main research programme be concen t ra ted on t e s t i n g a number o f r e l a t e d hypotheses. These w i l l serve t o focus t h e research e f f o r t r a t h e r than have i t d i s s i p a t e d ac ross a spectrum o f u n r e l a t e d p r o j e c t s which, whi l e i n t e r e s t i n g i n t h e i r own r i g h t , do n o t s u b s t a n t i a l l y advance our unders tanding o f savanna dynamics i n ways which would lead t o improved management.

The l i s t o f hypotheses i s c l e a r l y incomplete. It i s l i k e l y t h a t some important aspects o f savanna eco logy have not been i nc luded because we cur ren t l y do no t know enough about them. These poor ly understood components and processes w i 11 need t o be i n v e s t i g a t e d w i t h i n t h e broader frarnework o f the programme. The resu l t i ng knowledge can then be used t o r e f i n e t h e e x i s t i n g hypotheses o r t o propose new ones. The r e v i s i o n and f o r m u l a t i o n o f hypotheses i n t h e l i g h t o f new observa t ions and i n s i g h t s w i l l be one o f t h e o b j e c t i v e s o f t h e ser ies o f workshops t o be he ld during the course o f the programme.

FUNCTIONAL. CLASSIFICATION OF SAVANNAS

Savannas encoapass a wide v a r i e t y o f systems w i t h d i f f e r e n t s t r uc tu ra l and func t iona l character ist ics. This makes c l a s s i f i c a t i o n d i f f i c u l t . E x i s t i ng c lass i f i ca t ions , based p r imar i ly on physiognomy, a re u n s a t i s f a c t o r y s i n c e they do n o t convey much about f unc t i on . Moreover, most o f the schemes are regional or, a t best, cont inenta l i n scope. None a r e accepted world-wide. A genera l s t r u c t u r a l - funct iona l c lass i f i c a t i o n o f savannas i s therefore urgent ly needed.

We cons ider t h a t s o i 1 m o i s t u r e and n u t r i e n t ava i l a b i l i t y a r e t h e primary determi nants o f savanna f unc t i oning. Thei r wide va r i a t i on i s probably a major reason f o r the d i ve r s i t y o f savanna types. However, simple ind ices o f moisture and nu t r i en t ava i lab i l i t y based on, f o r example, mean annual r a i n f a l l and s o i 1 type, do n o t adequate ly e x p l a i n a l 1 o f t h i s d i v e r s i ty. B e t t e r measures o f p l a n t - a v a i l a b l e moisture and n u t r i e n t s a re t h e r e f o r e needed be fo re a s t r u c t u r a l -

funct iona l c l a s s i f i c a t i o n o f savannas can be produced. Some o f the fac tors t h a t should be taken i n t o account are: so lar radiat ion; a i r temperature; seasona l i ty, p r e d i c t a b i li t y and v a r i a b i 1 i ty o f r a i n f a l l ; s o i l t e x t u r e ; pH; c a t i o n exchange capac i t y ; n u t r i e n t minera l i z a t i o n rates; topography, and land-use h is tory .

A p o s s i b l e i ndex o f m o i s t u r e ava i l a b i l i t y i s t h e degree t o which evapora t i ve demand i s met by s o i 1 water: t h e r a t i O o f ac tua 1 t o p o t e n t i a l e v a p o t r a n s p i r a t i o n m i g h t be appropr ia te . Such an index would i n t e g r a t e s o l a r r a d i a t i o n , a i r temperature, r a i n f a l l , s o i 1 t ex tu re , topography and seasona l i t y . An index o f s o i l n u t r i e n t avai l ab i li t y cou l d be the m i nera l izab le capaci ty o f the so i 1, w i t h special emphasis on n i t rogen and phosphorus. A l lowance may have t o be made f o r o t h e r s o i 1 f a c t o r s which cou ld a f f e c t t h e vege ta t i on (e.g. exchangeab l e a lumi n i um and sodium percentages, the presence o f heavy metals, and shortages o f micronutrients).

We propose t o develop these two ind ices t o produce a c l a s s i f i c a t i o n o f t h e wo r l d ' s savannas based on an o r d i n a t i o n o f a c t u a l s i t e s i n r e l a t i o n t o these indices. The p red ic t ion t o be tested i s t h a t s i t es w i t h s i m i l a r m o i s t u r e and n u t r i e n t i n d i c e s w i 11 e x h i b i t s i m i l a r s t ruc tu ra l and func t iona l character ist ics.

I n t h e c u r r e n t absence o f a s a t i s f a c t o r y c l a s s i f i c a t i o n , we have used the f o 1 low i ng hypothetica 1 arrangement o f savanna types w i t h i n t h e P l a n t Avai l ab l e M o i s t u r e (PAM) and Avai l a b l e N u t r i e n t s (AN) plane (Figure 2) as a basis f o r formulat ing some o f the hypotheses.

HIGH

LOW

LOW/HIGH I I HIGH/HIGH

( c e r r a d o ) 1

(monsoona i ta1 bgra;s) (wet miombo) ; ( va l ley and escarpment

( 1 lanos I w o d lands (dry miombo)

I I

(p lateau grass lands) : 1 (Serengeti p la ins ) I (Acacia savannas) I (Mopane wood lands :

(Sahel) (Astrebla grasslands) 1 LOW/LOY ! HIGHILOW

LOW HIGH Avai lab l e Nut r ients

Figure 2. Hypothet ical d i s t r i b u t i o n o f savanna types i n r e l a t i o n t o the m a i l determinants o f savannas (modif ied from Bell, 1984)

HYPOTHESES

Each o f the f o l l ow ing hypotheses i s presented i n three parts: ( i l a s ta tement o f t he hypothes is , w i t h b r i e f a m p l i f i c a t i o n where necessary; ( i i ) a statement o f i t s impl icat ions f o r management; and ( i i i ) an ind ica t ion as t o how the hypothesis might be tested.

Hypothesis 1

Given t h e assumption t h a t mo i s tu re and n u t r i e n t ava i l a b i l i t y a re ma jo r de te rminan ts o f savanna func t ion ing , t he expec ta t i on t h a t changes i n PAM and AN cause a p r o p o r t i o n a t e l y g r e a t e r change i n p l a n t p roduc t i on i n those systems where bo th a re most s t r o n g l y l i m i t i n g i s r e l a t i v e l y s t r a i gh t f o rwa rd . I t i s l ess obvious why a s i m i l a r t rend should be expected o f changes i n species composition. 1 t i s based on assumptions t h a t PAM and AN Vary more i n a r id , nutrient-poor systems, and t h a t the species i n these systems d i f f e r w ide l y i n t h e i r c a p a c i t i e s bo th t o su rv i ve pro longed pe r i ods o f water and nu t r i ' en t s t r e s s and t o e x p l o i t occas ional pu lses i n avai l ab i l i ty. Changes i n PAM and AN i n such systems are more l i k e l y t o r e s u l t i n changes i n the dominant species than i n those systems where mo i s tu re and n u t r i e n t s a re more r e a d i l y ava i l a b l e and condit ions f o r growth more favourab le.

Impl icat ion The main imp l i ca t i on f o r management i s t h a t actions which a f f e c t the a v a i l a b i l i t y o f w a t e r and /o r n u t r i e n t s a r e g o i n g t o have proport ionately greater e f f e c t i n those systems i n which PAM and AN are most l im i t ing . These e f fec ts wi 11 r e s u l t not only from actions which increase PAM and AN, such as i r r i g a t i o n and f e r t i l i z a t i o n , bu t a lso from ones which i nd i rec t l y lower them, f o r examp le, harvesting, the use o f f i r e , and overstocking. There are also imp l i ca t ions f o r monitoring. At tent ion needs t o be focussed on the possible e f f ec t s o f management act ions on the physical envi ronment, pa r t i cu l a r ly i n strongly water- and nu t r i en t - l im i t ed systems.

Test The hypothesis can be tested by comparing the responses o f d i f f e r e n t savanna types i n the PAMIAN plane t o changes i n water and nu t r i en t avai lab i li ty (e.g. through i r r i g a t i o n and f e r t i l i za t ion, both s ing ly and i n combination) and monitoring the resu l t ing changes i n species compos i t ion and p roduc t i on i n each case. The hypothes i s p r e d i c t s t h a t t he magnitude of change w i 11 be g r e a t e s t i n those systems i n which both water and nu t r ien ts are most l im i t ing .

Hypothesis 2

The co-occurrence o f two o r more independant events (e drought, above-average raTnfall,frost, f i r e , h e r b i v o r w i 11 have --

- + P synergi s t i c e f f ec t s i n changi ng savanna s t ructure an pro= o r

I n terms o f t h i s hypothesis, marked s h i f t s i n compos i t i on o r abundance are o f ten caused by the co-occurrence, o r c lose sequence, o f events w i t h compoundi ng effects. For example, the establishment o f woody p l a n t s may depend on t h e co-occurrence o f (il a p e r i o d o f e x c e p t i o n a l l y h i g h o r o u t - o f - s e a s o n r a i n f a l l ; ( i i reduced compet i t i o n f rom perenn ia 1 grasses; (i i i ) low herbivore pressure; and ( i v ) a per iod w i thou t f i r e . From a management perspective, some o f these events are managable (e.g f i r e , herbivory); others are not (e.g. above-average r a i n fa 11, droughts, f ros t ) .

Impl icat ions I f t h i s hypothesis i s val id, then management aimed a t inducing, o r avo id ing, ma jo r changes i n community compos i t i on may o n l y be possible when two o r more independent events co i nc i de. Addi t i ona 1 ly, events which co-occur i n f r e q u e n t l y a re li ke l y t o induce g r e a t e r shi f t s i n vegetation composition and production. Understanding the e f f ec t s on key species o f i n te rac t ions between par t i cu l a r c l ima t i c , f i r e and h e r b i v o r y events, and knowing t h e p r o b a b i l i t i e s o f co- occurrence o f these, i s a prerequis i te f o r e f f ec t i ve management.

Test A t e s t o f t h i s hypothesis would be t o impose a va r ie ty o f stresses o r disturbances on a communi ty, s ing ly and i n combination, and then moni t o r t h e subsequent changes i n spec ies composi t ion, r e l a t i v e abundance and production. Since any changes t h a t occur are going t o r e f l e c t d i f f e r e n c e s between spec ies i n estab l ishment , growth, r e c r u i t m e n t and mor ta li ty , a s tudy o f these processes w i 11 be e s s e n t i a l f o r unders tand ing t h e mechanisms o f change. I n any between-si t e comparisons, the h i s tory o f each s i t e w i 11 have t o be taken i n t o account.

Hypothesis 3

& substant ia l change from the reva i l i ng frequency, i n t ens i t y o r sequence o f e v e n m a m a f i b l t -- i n a marked change - structure àn-ucti C.

Impl icat ions R e l a t i v e l y r i g i d f i r e o r h e r b i v o r y reg imes a r e o f t e n a p p l i e d by management. One consequence o f t h i s i s t h a t communi t i e s eventua 1 ly become dominated by those species which are best able t o accomodate t h e p reva i l i n g s t r e s s and d i s t u rbance regimes. Species which a re

be t t e r adapted t o other regimes dec l i ne. Any substantia 1 dev ia t ion from the prevai 1 i ng regime genera 1 l y resu l t s i n sudden changes i n species compos i t ion and product ion, t h e magnitudes o f which a r e r e l a t e d t o t h e degree o f change i n t h e regime. By a a i n t a i n i n g a var iable herbivory o r f i r e regime, and thereby r e g u l a r l y exposing t h e species w i t h i n t h e system t o a w ide r range o f cond i t ions , managers can inc rease the r e s i l i e n c e o f t he system t o these disturbances and minimize the r i s k o f sudden s h i f t s i n composition and production.

Test The hypothesis can be tested experimentally by changing var iab le and f i xed herbi vory o r f i r e reg i mes and monitoring the resu lti ng changes i n species composition and production. The hypoth'esis pred ic ts t h a t t he g r e a t e s t change w i 11 occur i n those cases where a reg ime has been most r i g i d l y a p p l i e d and where t he d i f f e r e n c e between t h e o r i g i na l and the new regime i s greatest.

Hypothesis 4

Impl icat ions This introduces the element o f the t im ing o f events as a fac to r i n savanna dynamics. The e f f e c t s o f s t resses such as f i r e , f r o s t o r herbivory o f ten depend on the phenological s ta te and physiological c o n d i t i o n o f t h e p l a n t s a t t h e t ime. P lan t s which a re dormant a r e u s u a l l y l ess a f f e c t e d than p l a n t s which a r e a c t i v e l y growing. A stress occurring a t one t ime o f the year may stimu l a te reproduction whereas a t another t ime i t may re'tard it. The t im ing o f pa r t i cu l a r management act ions therefore may be as important as t h e i r magnitude and t h i s needs t o be taken i n t o account i n management planning. In t h i s regard i t i s c ruc ia l t o know what are the c r i t i c a l periods i n l i f e cyc les o f key spec ies and hou these m i g h t be a f f e c t e d by d i f f e r e n t management actions.

Test A t e s t o f t h i s hypo thes is would i n v o l v e va ry i ng independent ly t h e t im ing and i n tens i t y o f events such as f i re , herbivory o r a r t i f i c i a l drought, and m o n i t o r i n g t h e e f f e c t s o f t h i s on t h e growth, reproduc t ion and s u r v i v a l o f i n d i v i d u a l s o f t he key species. The hypothes is p r e d i c t s t h a t changes i n species popu la t i ons w i 11 be determined more by the t im ing o f the event, i n r e l a t i o n t o c r i t i c a l periods i n the species' l i f e cycle, than by i t s magnitude.

Hypothesis 5

Increasi ng leve 1s o f herbi vor i n s i ngle-herbivore systems increases the proporti0?lOf - ü n p d s j ë c i e s -- i n the communi tx. Hypothesis 5(a)

An a l t e r n a t i v e t o Hypothes is 5 i s: Chan es i n t h e r o o r t i o n s o f a la tab le and unpalatable s ecies depe&~ine*ix

&=O t he oheno el oav O t he s o e c i e s x h x Z h a e v e o f - -- - I- - - herbi vory.

Impl icat ions The unde r l y i ng p r i n c i p l e o f Hypothesis 5 i s t h a t i n systems dominated by one spec ies o f he rb i vo re (e.g. c a t t l e ) , se l e c t i v e feedi ng by the herbivore changes the competi t i v e balance between the preferred, palatable species and the unpalatable ones i n favour o f t he l a t t e r . Th i s e f f e c t i s thought t o be more pronounced a t h i ghe r g raz ing i n t e n s i t i e s . To coun te r t h i s , a system o f c o n t r o l l e d select ive grazing i s o f ten applied. The system i s based on len ien t use o f preferred species and no de fo l i a t i on o f the unpalatable ones, and involves two main assumptions: ( i l t h a t production o f preferred species i s s t imulated by moderate defol iat ion; and ( i i ) t h a t i n the absence o f d e f o l i a t i o n by g raz ing o r f i r e , unpa la tab le spec ies eventual ly become moribund and dec li ne. An a l t e r n a t i v e management system, non-selective grazi ng, i s based on the opposi t e p r i nc i p le. I t assumes t h a t moderate d e f o l i a t i o n i s more de t r imen ta 1 t o unpalatable species than heavy u t i l i z a t i o n i s t o the palatable ones. Accordingly, h igh grazing pressures are applied f o r b r i e f periods i n o rder t o f o r c e t h e an ima ls t o graze bo th t he p a l a t a b l e and unpalatable species.

The reasoning behind Hypothes is 5(a) i s d i f f e r e n t . For much o f t h e t ime, g raz ing does n o t appear t o have much o f a nega t i ve e f f e c t on grasses; t h e i r p r o d u c t i o n may even be s t imulated. However, de fo l i a t i on during the ear l y stages o f growth i s o f ten deleterious. The d e c l i n e o f a spec ies t h e r e f o r e may be l a r g e l y t h e r e s u l t o f d e f o l i a t i o n d u r i n g t h i s ear ly g rowth phase. R e l i e v i n g t h e g raz ing pressure on p a l a t a b l e spec ies a t t h i s t i m e may be necessary t o mai n t a i n t he i r abundance i n a sward.

I f app rop r i a te g raz ing s t r a t e g i e s a re t o be de f i ned f o r d i f f e r e n t savanna regions, i t i s essent ia l t h a t these d i f f e r e n t hypotheses and t h e i r assumptions be tested across the range o f savannas. Managers need t o know how the t iming, i n t ens i t y and frequency o f de fo l i a t i on a f f e c t t h e g rowth and rep roduc t i on o f key Pasture species ( these i n c l u d e bo th spec ies t h a t decrease and those t h a t increase under grazing), and how t h i s i n tu rn inf luences t h e i r population dynamics and in te rac t ions .

Test The t e s t o f these hypotheses can car r ied ou t i n conjunction w i t h the t e s t o f Hypothesis 4 by va ry i ng independent ly the i n t e n s i t y and t i m i n g o f grazing, and m o n i t o r i n g changes i n t he popu la t ions o f p re fe r red and non-pre fe r red species. Hypothesis 5 p r e d i c t s t h a t pa la tab le species are increasingly adverse ly af fected by an increase i n grazing i n tens i t y and t h a t t h i s favours the unpalatable species, enabling them t o increase i n the sward. I n contrast, Hypothesis 5(a) pred ic ts t h a t p lants defo l ia ted during the ear ly growth per iod w i l l be more suscep t i b l e than those d e f o l i a t e d when t h e p l a n t s a r e dormant o r have completed most o f t h e i r annual growth. Any changes i n species composition which occur w i 11 be caused by di f ferences i n the t ime o f de fo l i a t i on r e l a t i v e t o the t ime o f ear l y p l an t growth.

Hypothesis 6

The e f f e c t o f d is tu rbance on t h e r a t e and e x t e n t o f change i n t h e s p e c i e ç m p s i t i o n o f a s a v a n n a d e n K r i n c i - a T on t h e T i = h i story character i s t i E - a n T a t i o n -57-w 10 ogy -- o f the s p e c x -

The e f fec ts o f disturbance on the species i n a community are usual ly se lec t i ve , some .species be ing more suscep t i b l e t o a p a r t i c u l a r d i sturbance than others. Consequently, species composition tends t o s h i f t i n favour o f those species b e s t adapted t o s u r v i v e t he p a r t i c u l a r d i sturbance, recover f rom i t s e f f e c t s and e x p l o i t t h e post-disturbance environment. A t t r ibu tes which enable a species to recover r a p i d l y a f t e r d i sturbance i nc lude ( i t he presence o f s u b s t a n t i a l belowground reserves; (i i ) t h e capac i t y t o resprout ; (i ii the presence o f r e l a t i v e l y l a r g e reserves o f dormant seeds w i t h va r i ab l e ge rm ina t i on requi rements or, a l t e r n a t i v e l y , t he capac i t y t o d isperse t o and reco lon i ze a s i t e r a p i d l y a f t e r disturbance; and ( v ) t h e a b i l i t y t o e s t a b l i s h under extreme envi ronmenta 1 conditions.

Impl icat ions P r i o r cons ide ra t i on o f t he ways i n which d i f f e r e n t species a re li ke 1y t o respond to par t i cu l a r management actions o r envi ronmenta 1 events i s c r u c i a l t o e f f ec t i ve management. An understanding o f the main f ea tu res o f t h e l i f e h i s t o r y and popu la t i on b i o l o g y o f key species i s therefore necessary. For example, under what condit ions do d i f f e r e n t species e s t a b l i s h f rom seeds, grow t o m a t u r i t y and reproduce? What condit ions cause the w i despread death o f indiv idua 1s o f a species? How are these processes a f f e c t e d by f i r e , grazing, drought and in teract ions w i t h other species? Related species do no t necessari l y respond i n the same way t o the same management action.

Test The hypothes i s can be t e s t e d e i t h e r by imposing on a community a disturbance such as extreme defo l ia t ion, o r removal o f woody plants, o r by s tudy ing n a t u r a l d is turbances such as extreme drought and mon i t o r i ng t h e r e s u l t i n g changes i n abundance and biomass o f pa r t i cu l a r species. The changes need t o be studied a t the population leve l i n order t o understand the mechanisms involved. The species chosen f o r s tudy should encompass species w i t h a wide a r ray o f c o n t r a s t i n g l i f e h i s t o r y a t t r i butes and, where poss ib le , known d i fferences i n response t o the p a r t i CU l a r d i sturbance.

Since t h e r e l a t i o n s h i p s , i f any, between species' l i f e h i s t o r y a t t r i bu tes and kinds o f change i n the abundance and biomass o f these species a re mu l t i v a r i a t e , p o s s i b l e assoc ia t ions can bes t be d i sp layed through correspondence analysis. The hypothesis pred ic ts t ha t the direct ion, magnitude and manner o f change w i 11 be s imi l a r i n species having s imi l a r l i f e - h i s t o r y a t t r ibutes, whereas species w i t h d i f f e ren t a t t r i bu tes w i 11 tend t o respond d i f fe ren t l y .

Hypothesis 7

The res onses o f savanna species t o stress can be pred5ct:d. on the baç is ---- -e- O the'ir-fe h is to ry characi%- - a p o p u l a t i o n i o T o g r

Some o f the a t t r i bu tes which enhance a plant's capacity t o withstand s t r e s s inc lude : ( i ) a h i g h rep roduc t i ve output, p a r t i c u l a r l y one t h a t i s st imulated by stress; ( i i ) a large root:shoot biomass ra t io , which l i m i t s t h e amount o f m a t e r i a l t h a t herb ivores o r f i r e can consume; ( i i i ) unpa la tab i li ty t o herbivores, o r other features which r e s t r i c t the amount o f fo l iage consumed; ( i v ) compensatory growth i n response t o d e f o l i a t i o n ; ( v ) t h e presence o f energy and n u t r i e n t reserves on which t h e p l a n t can draw i n t imes o f s t ress; ( v i l t h e a b i li t y t o r e p r o d u c e v e g e t a t i v e l y ; ( v i i morpho log ica l and phenological p l as t i c i t y , which a l l o w r a p i d adjustments t o s t ress ; and ( v i i i ) physiological quiescence during c l i m a t i c a l ly unfavourable t imes o f the year.

Imp 1 i c a t i ons E f f e c t i v e management depends on b e i n g a b l e t o assess t h e consequences o f d i f f e r e n t a c t i o n s and choose t h a t which comes c l o s e s t t o g i v i n g t h e des i r ed r e s u l t . Since i t i s n o t f e a s i b l e t o determine empi r i c a 1 ly t h e response o f a 11 species t o every management ac t ion , sosie bas i s i s needed f o r be ing ab le t o p r e d i c t t he probable responses o f t h e key species. I f t h i s can be done by using informat ion on the species' li fe-hi story a t t r ibutes, then the po ten t ia l f o r e f f ec t i ve management w i 11 be great ly enhanced.

Test The t e s t o f t h i s hypo thes is i s s i m i l a r t o t h a t o f Hypothesis 6 except t h a t the avai lab le informat ion on the l i f e - h i s t o r y o f the key species i s used t o p red i c t changes i n t h e i r abundance andbiomass be fo re t he exper iment i s c a r r i e d out. ( A t our c u r r e n t l e v e l o f understanding i t i s l i k e l y t h a t t h e p r e d i c t i o n s w i 11 o n l y be qual i tat ive.) The t e s t invo lves making a comparison between these predicted changes and those induced by the par t i cu la r stress. I f the observed changes d i f f e r f rom those p r e d i c t e d then e i t h e r t he l i f e h i s t o r y a t t r i b u t e s o f a species a r e n o t good p r e d i c t o r s o f population change o r our current understanding o f these a t t r i butes and t h e i r e f fec ts i s inadequate f o r making accurate predictions.

A d i r e c t t e s t o f the hypothesi s w i 11 not be possible where there i s i n s u f f i c i e n t l i f e h i s t o r y i n f o r m a t i o n ava i l a b l e a t t h e outset . Instead, the informat ion on l i f e h is to ry a t t r i bu tes w i l l have t o be co l lec ted f o r each species during the course o f the experiment. As i n t he t e s t o f Hypothesis 6, correspondence ana l ys i s can then be used t o d e t e c t any unde r l y i ng assoc ia t i ons between s p e c i f i c a t t r i bu tes and pa r t i cu l a r s t ress- induced changes i n abundance and biomass. The aim would be t o d is t inguish between those a t t r i bu tes which c o r r e l a t e w i t h t h e observed popu la t i on o r biomass changes ( these w i l l be t h e a t t r i b u t e s l i k e l y t o have p r e d i c t i v e value), those which cor re la te w i t h each other and are therefore redundant, and those which do not co r re la te w i t h any o f the changes.

To determine how success fu l l y a c t u a l changes can be p red ic ted , e i t h e r a reserve s e t o f data (i.e. da ta f rom one o r more o f t he r e p l i c a t e d p l o t s , w i t h h e l d f rom t h e i n i t i a l ana lys is ) , o r an i ndependent ly C O 1 l e c t e d data set, can then be ana l ysed t o see t o what e x t e n t t h e i d e n t i f i e d r e l a t i o n s h i p s remain s t a b l e and a re p red ic tab le . However, causa li ty i n these r e l a t i o n s h i p s cannot be i n fe r red from these ana lyses. To understand the mechani sms involved w i 11 r e q u i r e d e t a i l e d s tud ies o f t h e ways i n which s t r e s s a f f e c t s key population processes i n each o f the species.

Hypothesis 8

A decrease i n the e f f ec t i ve r a i n f a l l o f a s i t e leads t o a decline i n

Impl icat ions Management act ions which reduce the inpu t o f water t o the so i1 (e.g. by overstocking t o the po in t where trampl ing and compaction o f the soi 1 resu l t ) , o r which increase the r a t e o f p l an t water-use (e.g. by f e r t i l i z a t i o n ) , may shor ten t he leng th o f t i m e t h a t water i s

a v a i l a b l e t o p lan ts . T h i s w i l l concen t ra te pheno log ica l types i n t ime, r e s u l t i n g i n an even tua l r e d u c t i o n i n species d i v e r s i t y through c o m p e t i t i o n and drought- induced m o r t a l i t y . Management actions which might shorten the per iod o f so i1 water a v a i l a b i l i t y would have t o be avoided.

Test The hypo thes is can be t e s t e d by us ing r a i n - o u t s h e l t e r s and i r r i g a t i o n t o keep constant the t o t a l amount o f water avai lab le t o the p lants whi l e varying the per iod o f avai lab i l i t y . The hypothesis pred ic ts that, under the same t o t a l amount o f p l an t avai lable water, there w i 11 be marked changes i n d ivers i ty , inc lud ing perhaps a loss o f spec ies , i n t h o s e s i t u a t i o n s where t h e p e r i o d o f w a t e r avai lab i li t y i s sharp ly reduced. Species norina 1 ly developi ng i n the middle t o l a te ra iny season w i 11 be most susceptible, whi l e drought- t o l e ran t species w i 1 1 be favoured.

Hypothesis 9

The s t a b i l i t y o f savanna ecos stems, i n terms o f t h e ca a c i t o f - -gr t h e i r c o m D o n e n t s i ë d ë ? t h r o ~ s ~ o~ ~ t u r ance. iI - - s t r o n ï y ;nf l u e n i t t h e d e g r e e o ~ v ~ e n t a l c o n s t r a i n t on 3- a t a b l i s h m e n t an growth.

-

Recovery from stress o r disturbance can be arrested o r red i rected by the subsequent in tervent ion o f drought, f i r e , herbivory, etc. before the process o f recovery has been completed. Where recovery i s slow, the l i ke l ihood o f subsequent events in f luenc ing the eventual outcome i s i ncreased.

Impl icat ions The main i m p l i c a t i o n f o r management i s t h a t where t he process o f recovery from stress o r disturbance i s slow, e i t he r because o f the sever i ty o f the i n i t i a l stress o r disturbance, o r because o f strong env i ronmenta l c o n s t r a i n t s on es tab l i shment and/or growth (e.g. n u t r i e n t - p o o r s o i l s , l o w w a t e r a v a i l a b i l i t y , p o o r seedbed conditions, etc.), subsequent events may occur which w i 11 a f f e c t the even tua l outcome. Where these subsequent events are control lable, managers would need t o ensure t h a t they a l low s u f f i c i e n t t ime f o r t he species t o recover be fo re these o the r pressures a r e imposed. Where subsequent events are uncontrol lable and unpredictable, then t he i n t e n s i t y and t i m i n g o f management ac t i ons would have t o be ad jus ted so t h a t t he p l a n t s cou ld recover i n t he s h o r t e s t t i m e possible.

Test The hypothesis can be tested by comparing the patterns o f vegetation change (spec ies composit ion, r e l a t i v e abundance, p roduc t i on o f se lec ted spec ies) a f t e r d is tu rbance (e.g. a r t i f i c i a 1 drought) on s i t e s which d i f f e r i n t h e i r s u i t a b i li ty f o r p l a n t growth. The hypothes is p r e d i c t s t h a t where c o n d i t i o n s a re unfavourable f o r establishment and growth, the subsequent in tervent ion o f events such as f i r e , herbivory, o r fu r ther drought w i 11 cause greater mor ta l i ty and changes i n species compos i t ion and product ion, than w i 11 t h e same events, occurring a t the same time, on more favourable sites.

Hypothesi s 10

I f d is turbed. savanna ecosvstems tend t o r e t u r n t o t h e i r fo rmer

composition - or production.

Hypothesis 10(a)

An a l t e r n a t i v e . t o Hypothesis 10 i s : S t a b i l i t i s n o t a ma jo r -+ y-- f e a t u r e o f savannas. D i s tu rbance- i nduce changes i n s p e c i es com o s i t ' n and p r o d u c t i v i t y - are accomodated & s t r z t u r a l and &dju~,ments -- with in the community. Autogenic recovery doeS n o t occur. --

Implicat ions The concept o f s t a b l e equi l i b r i a i s c e n t r a l t o much c u r r e n t ecological theory and application. I t s relevance i n the context o f savanna dynamics can be questioned, e s p e c i a l l y i n v iew o f t h e s t o c h a s t i c n a t u r e o f mos t o f t h e d r i v i n g v a r i a b l e s , t h e interact iveness and non- l inear i ty o f many o f the processes, and the resu l t i ng contingency o f many o f t h e i r effects. Yet the concept i s w i de ly app 1 i e d i n savanna management. For instance, a t t emp ts by managers t o reverse undesi rab l e changes i n vegetati on composition o r p roduc t i v i t y caused by, f o r example, overstocking, involve removing the cause o f the disturbance and res t i ng the disturbed area.

The underly i ng assumption i s t ha t species composition and production a t a s i t e tend towards a s ing le , s t a b l e e q u i l i b r i u m which i s determined p r i mari l y by the prevai li ng soi 1 condit ions and c l imate (ma in l y r a i n f a l l ) . Moreover, i f compos i t ion and p roduc t ion a re d i s t u r b e d they w i 11 r e t u r n au toma t i ca l ly t o t h e i r p red is tu rbance s ta te . Hypothesis 10 i m p l i e s t h a t t h e t i m e r e q u i r e d f o r recovery, and hence t h e p e r i o d o f r e s t t h a t i s r e q u i r e d a f t e r disturbance, i ncreases w i t h the i ncreasi ng magnitude o f d i sturbance (as measured by the i n i t i a 1 d i splacement i n species composition o r production).

I n contrast , Hypothesis 10(a) s ta tes t h a t once a savanna has been d i sturbed, species composition and production do not automatica 1 l y return t o t h e i r predisturbance levels but, instead, the changes are accoaodated through s t r u c t u r a l and f u n c t i o n a l rearrangements. Therefore, i n order t o restore the previous relationships, another disturbance, a c t i n g i n the opposi te d i rec t i on , has t o occur. Resolving the quest ion o f how savannas respond t o disturbance has important impl icat ions f o r the kinds o f management actions tha t are requ i r e d t o counter any undesi rab l e changes brought about by disturbance. I n both cases, i t i s c l e a r t h a t regular, f requent monitoring i s needed i n order t o detect and respond rapidly t o any adverse changes which might occur.

Test Both hypotheses can be tested i n the same experiment by disturbing adjacent areas t o d i f f e r e n t degrees (e.g. removing d i f f e r e n t proportions o f woody biomass) and monitoring the resu l t i n g changes i n species composition and production. Hypothesis 10 predicts tha t composi t ion and product ion w i 11 even tua l l y r e t u r n t o t h e i r pre- d is turbance leve ls, w i t h t he t ime taken be i ng propor t iona 1 t o the magnitude o f the disturbance. On the o the r hand, Hypothesis 10(a) predicts t ha t an area, once disturbed, w i 11 not show any consistent tendency w return w the or iginal, predisturbance state unless i t i s subsequently disturbed i n the opposite direction. The system w i 11 e i t h e r remain a t t he p o i n t t o which i t was disturbed, o r undergo fur ther change i n the sane direction.

Hypothesis 11

The r e v e r s i b i l i t o f change i n p l a n t species composit ion and P T o d u d e r ~ l y - r e l a t e h t o -- the degree - o f change -- i n so-7 physico-chemica 1 properties.

An a l t e r n a t i v e t o Hypothesis 11 i s : No i r r e v e r s i b l e chan e i n r ecies con os i t i on wi11 occur wirhout aFncu ryen t and i o d t i q d k @ - , ~ t i c ~ l a ~ a g ~ > - ~ - - s o l 1 p h F i h ~ r o ~ e r t i e s .

The a s s u i p t i o n under ly ing both hypotheses i s t h a t the physico- chemical properties o f the soi1 are a major determinant o f savanna coiposi t i o n and product iv i ty. The d i s t i ngui shi ng feature between the two hypotheses i s the degree t o which plant species composition and production are able t o recover from disturbances t o soi 1 properties. I n t e r m s o f Hypo thes i s 11, any changes t o t h e s o i 1 d u r i n g d i sturbance resu l t s i n a correspondi ng change i n the equi l i b r i u m

Table 2. Some possible e f fec ts o f herbivores on s o i l propert ies and -- processes

--------------------------------------------------------------------- IMPACT EFFECTS CONSEQUENCES ---------------------------------------------------------------------

1. Reductions 1.1 Increase i n area * Increased inso la t ion i n p l an t and o f bare soi 1 * Increased exposure u, li t t e r cover r a i ndrop impact

* Increased so i 1 temperature

* Reduced i n f i l t r a t i o n * Increased run-off * Increased po ten t ia l f o r

water erosion 1.2 Reduction i n * Lower S.O.M.

inpu t o f l i t t e r * Less so i 1 cohesion 1.3 Reduced roo t * Increased po ten t ia l f o r

growth leaching tm groundwater

2. Input o f dung 2.1 Change i n spa t i a l * Increased heterogeneity and u r ine and tempora 1 i n d i s t r i but ion o f so i 1

d i s t r i b u t i o n o f nu t r ien ts nu t r i en t inputs * Increased v o l a t i l i z a t i o n

2.2 Change i n qua l i t y * More rap id O.M. ' o f nu t r i en t inputs turnover

3. Compaction 3.1 Macropore space (increase i n reduced bulk densi ty 3.2 Micropore space o f the so i l ; i ncreased most l i k e l y 3.2 Total pore space when soi 1s reduced are moist) 3.3 Reduced r a i nwater

i n f i l t r a t i o n 3.4 Increased run-of f

* Reduction i n soi 1 water-ho l d i ng capaci ty

* Less p lant-avai lab l e water

* Less favourable soi 1 microclimate f o r bioca

* Poor environment f o r seedling e s t a b l i s h e n t and r o o t growth

* Increased po ten t ia l f o r water eros i on

4. Trampling 4.1 Reduced s ize o f * Increased po ten t ia l f o r (break up o f soi 1 aggregates wind and water erosion soi 1 aggregate * Potent ia l loss o f s t ructure ; surface soi 1 most l i k e l y 4.2 Formation o f soi 1 * Reduced i n f i l t r a t i o n when soi 1s surface sea 1s * Increased run-off are dry) * Loss o f so i1 seed s tore

* Unfavourable envi r omen t f o r seed germination and seedli ng estab li s h e n t

condit ions f o r p l an t growth, and t h i s i n t u r n a f fec ts the degree t o which the vegetation i s able t o re tu rn t o i t s pre-disturbance state. Thus t h e g r e a t e r t he d i sturbance t o t he s o i 1, t he less comp l e t e l y the vegetation recovers t o i t s pre-di sturbance state.

I n contrast, Hypothesis l l ( a ) imp l ies t h a t below some c r i t i c a l p o i n t (which p robab ly d i f f e r s f r o n s i t e t o s i t e , depending on t h e s o i 1 t ype and c l i n a t e ) , any changes i n s o i 1 p r o p e r t i e s do n o t have a l a s t i n g i n f l u e n c e on p l a n t species composi t ion o r product ion. However, i f the soi 1 i s disturbed beyond t h i s point, the changes i n the p l an t community become i rrevers i b le.

Iap l i ca t i ons Large he rb i vo res a re o f t e n i m p l i c a t e d i n changes t o p l a n t s p e c i e s composition and production, bo th d i r e c t l y , as p a r t i a l l y s e l e c t i v e consuiers, and i n d i rec t l y , through t he i r impact on soi 1 propert ies and processes. These i nc lude t rampl i ng and compaction, reduct i ons i n p l a n t and l i t t e r cover, and t h e i n p u t o f n u t r i e n r s i n dung and ur ine, a l 1 o f which have many i n t e r l i n k e d consequences (Table 2). One i m p l i c a t i o n o f these hypotheses i s t h a t t he d i r e c t e f f e c t s o f l a rge herb ivores on long-term vege ta t i on change a re l i k e l y t o be less important these i n d i r e c t ef fects.

T h i s has i m p l i c a t i o n s f o r t he assessment and m o n i t o r i n g o f range condi t ion and c a r r y i ng capaci ty i n savannas. A t present, assessments o f t h e c a r r y i n g c a p a c i t y o f an area a re made a lmos t e n t i r e l y i n terms o f t h e amount o f food ava i l ab l e f o r anima ls , r a t h e r than i n terms o f how much trampl ing and compaction an area can sustain and what amount o f p l a n t cover w i 11 be needed t o p r o t e c t t he s o i 1. I n t h i s r e s p e c t , more c o n s i d e r a t i o n needs t o be g i v e n t o t h e d i f f e r e n c e s which e x i s t between areas i n t h e i r suscep t i b i l i t y t o degradation (e.g. clayey s o i l s are more eas i l y degraded than sandy so i ls, p a r t i c u l a r l y when moist).

Test Since large herbivores are of ten impl icated i n changes t o both so i 1s and vegetation, one t e s t o f Hypotheses 11 and l l ( a ) could involve a m u l t i f a c t o r i a l exper iment i n which ( i s imu la ted grazing; ( i i ) t r a m p l i n g and compaction; and ( i i i) t h e a d d i t i o n o f excreta, a re v a r i e d s i n g l y and i n combinat ion t o c r e a t e a s e r i e s o f changes i n the so i 1 and vegetation a t a site. This would also a l low the e f fec ts o f these d i f f e r e n t disturbances t o be assessed independently o f each other. Some o f t h e p r e d i c t e d changes a re l i s t e d i n Table 2. Key v a r i a b l e s which need m o n i t o r i n g i n c l u d e sur face s o i 1 compact ion and/or bu l k dens i ty, i n f i l t r a t i o n ra te , s o i 1 temperature, s o i 1 organic matter f ract ions, N and P minera l izat ion rates, as ne11 as changes i n p l a n t species composit ion, r o o t growth, and seed li ng establishment. These would be moni w r e d dur i ng the d i sturbance and a f te r the treatments have been withdrawn, t o determine whether there has been any recovery t o pre-treatrnent leve 1s.

The eventua 1 d i f f e rence i n vege ta t i on "composit ion and p roduc t i on between t he p re - t rea tment and post - recovery per iods, and t h e r e l a t i o n s h i p o f these changes t o t h e degree o f change i n s o i 1 p r o p e r t i e s which has occurred over t h e same per iod, p rov ides t he basic t e s t o f the hypotheses. Hypothesis 11 w i l l be inva l idated i f t he d i f f e r e n c e s i n vege ta t ion compos i t ion and p roduc t i on a r e unco r re l a ted w i t h t he degree o f change i n s o i 1 p r o p e r t i e s . Hypothesis l l ( a ) w i l l be i n v a l i d a t e d e i t h e r i f t h e r e i s no re la t ionsh ip between the amounts o f change i n the vegetation and i n soi 1 properties, o r i f the re la t ionship i s pos i t i ve and linear.

Hypothesi s 12

A reduction i n the d i ve rs i t y o f savanna communities resu l ts i n - 7 a d e c l i n e -- i n s o i 1 o rgan ic m a t t e r - and g d i - f f - . e z Tyc 1 i ng processes.

The d isrupt ion o f nu t r i en t cyc l ing ar ises from ( i l the narrowing o f the spectrum o f l i t t e r qual i ty; (i i ) changes i n the patterns o f r o o t growth i n t ime and space; and ( i i i ) a reduction i n synchrony between n u t r i e n t ava i l a b i l i t y and demand. These changes may r e s u l t i n t h e loss o f n u t r i e n t s througR leaching, p a r t i c u l a r l y on we l l - d ra i ned s o i l s i n regions o f high ra i n fa l l .

Imp 1 i ca t ions This hypothesi s has two major impl icat ions f o r the maintenance o f s o i 1 f e r t i l i ty . F i r s t , t he re i s an advantage i n mai n t a i n i n g phenologica 1 and s t ructura 1 d ive rs i t y i n savanna p lan t communities. Secondly, i t i s impo r tan t t o m a i n t a i n those s o i 1 b i o l o g i c a l processes which p reven t o r reverse t h e leach ing o f n u t r i e n t s , especial ly where the soi 1s are well-drained and nutrient-poor. Woody p lants are c ruc ia l i n both instances since they produce substant ia l amounts o f low q u a l i t y l i t t e r , which decomposes s l o w l y and contr ibutes t o a bui ld-up i n surface so i1 organic matter. They also maintain ac t i ve roo t growth and nu t r i en t uptake i n the horizonta 1 and v e r t i c a l planes o f the soi 1 p r o f i le. The removal o f woody plants from a savanna i s therefore l i k e l y t o cause considerable disturbance w i t h long-term e f f ec t s on nu t r i en t status o f the system.

Test The hypothesis can be tested by compari ng nu t r i en t cyc l i n g processes i n savannas w i t h va ry i ng p ropo r t i ons o f t r e e s and grass. The f o l lowi ng processes need invest igat ing: ( i the t im ing and qual i ty o f d i f f e r e n t above- and below-ground organic matter inputs; ( i i 1 the decomposi t i o n rates o f these mater ials and the reçu l t i n g patterns o f n u t r i e n t re lease; ( i i i ) t he e x t e n t o f n u t r i e n t immobi l i z a t i o n i n , and release from, soi 1 organic matter; and ( i v ) the seasonal pat tern o f nu t r i en t uptake by the vegetation. The hypothesis pred ic ts t h a t those communities which are less diverse, p a r t i c u l a r l y i n respect o f