Grazing lawns are a distinct grassland community type, characterised by short-stature and with their persistence and spread promoted by grazing. In Africa, they reveal a long co-evolutionary history of grasses and large mammal grazers. The attractiveness to grazers of a low-biomass sward lies in the relatively high quality of forage, largely due to the low proportion of stem material in the sward; this encourages repeat grazing that concomitantly suppresses tall-grass growth forms that would otherwise outcompete lawn species for light. Regular grazing that prevents shading and maintains sward quality is thus the cornerstone of grazing lawn dynamics. The strong interplay between abiotic conditions and disturbance factors, which are central to grazing lawn existence, can also cause these systems to be highly dynamic. Here we identify differences in growth form among grazing lawn grass species, and assess how compositional differences among lawn types, as well as environmental variables, influence their maintenance requirements (i.e. grazing frequency) and vulnerability to degradation. We also make a clear distinction between the processes of lawn establishment and lawn maintenance. Rainfall, soil nutrient status, grazer community composition and fire regime have strong and interactive influences on both processes. However, factors that concentrate grazing pressure (e.g. nutrient hotspots and sodic sites) have more bearing on where lawns establish. Similarly, we discuss the relevance of enhanced rates of nitrogen cycling and of sodium levels to lawn maintenance. Grazer community composition and density has considerable significance to grazing lawn dynamics; not all grazers are adapted to foraging on short-grass swards, and differences in body size and relative mouth dimensions determine which species are able to convert tall-grass swards into grazing lawns under different conditions. Hence, we evaluate the roles of different grazers in lawn dynamics, as well as the benefits that grazer populations derive from having access to grazing lawns. The effects of grazing lawns can extend well beyond their borders, due to their influence on grazer densities, behaviour and movements as well as fire spread, intensity and frequency. Variation in the area and proportion of a landscape that is grazing lawn can thus have a profound impact on system dynamics. We provide a conceptual model that summarises grazing lawn dynamics, and identify a rainfall range where we predict grazing lawns to be most prevalent. We also examine the biodiversity associated with grazing lawn systems, and consider their functional contribution to the conservation of this biodiversity. Finally, we assess the utility of grazing lawns as a resource in a rangeland context.

Summary 1.We expect tree species that regenerate primarily by sprouting to produce fewer seedlings than co-occurring species that regenerate mainly from seedlings, because of the trade-off between allocating resources either to ensuring vegetative reproduction (e.g. protective bark/latent buds) or to sexual reproduction (e.g. seeds). 2. Furthermore, resprouting species, because of their multi-stemmed nature, should be at a relative disadvantage, and therefore relatively infrequent, in tall forests. This is because a resprouting individual allocates resources to a number of basal branches/stems and buds rather than maximizing vertical extension of a single leader, as is the case in a seeder. Also, many tall stems arising from the same multi-stemmed base, as is the case in resprouters, will be relatively poorly supported in comparison to the single stem of a reseeder. 3. To test these two ideas we surveyed a number of plots in a range of South African forests and thicket communities. We noted the numbers of seedlings and resprouts for each species and determined a mean for each site. 4. Short forests and thickets were dominated by multi-stemmed species and there was an almost total absence of seedlings. In contrast, tall forests were dominated by singlestemmed reseeding species and were accompanied by seedlings. Key-words: Forest dynamics, persistence, recruitment, regenerationFunctional Ecology (1997) 11, [101][102][103][104][105] will first develop an argument that resprouters should be rare in tall forests and second that resprouters will only rarely produce seedlings.In tall forests where the ability to grow tall is critical for potential canopy species, a resprouter is likely to be at a disadvantage for both resource allocation and structural reasons. Resprouters divert resources to developing, maintaining and protecting a bud-bank instead of devoting it to vertical growth. Because they are often multi-stemmed they also share their resources amongst several stems. Reseeders, in contrast, can divert resources to maximizing rates of vertical extension of a single stem. Based on this resource allocation argument we suggest that reseeders will be able to grow taller and/or faster than resprouters. Again, because of their multi-stemmed nature, each stem of a resprouter will be more poorly structurally supported than will the single stem of a reseeder. This too will disadvantage resprouters in tall forests. Our first prediction then is that resprouters and multi-stemmedness should be more common in short forests than tall forests.Resprouters should produce fever viable seeds and seedlings than reseeders. First, this is because parent resprouters must devote resources to maintaining and protecting latent buds. Second, this is because seedlings of resprouters will be less competitive because they too must allocate resources to developing and protecting buds rather than growing tall. These differences between resprouters and reseeders are well known for shrublands (e.g. Hansen, Pate & Hansen 1991) but have not been we...

ISBN 978 1 86814 479 2All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the express permission, in writing, of both the author and the publisher.Cover photograph by Donald Cook at stock.xchng Cover design, layout and design by Acumen Publishing Solutions, JohannesburgPrinted and bound by Creda Communications, Cape Town Foreword SOUTH AFRICA and its people are blessed with diverse and thriving wildlife. We are also a developing economy with a growing population. From these facts emerges the particular situation of having most of our protected areas surrounded by land that has been transformed, to a greater or lesser extent, by human development. Large mammals, such as elephants, no longer roam the entire landscape, and their populations are no longer completely governed by the laws of nature. Protecting elephants and the ecological systems in which they exist in a practical and sustainable way that balances the needs of humans, elephants and the environment is a challenge to which I am committed.This Assessment was undertaken to reduce the degree of scientific uncertainty associated with decisions that must be made very soon and in the medium-to-long term. It helps to evaluate the costs and benefits associated with each choice, both in economic and ecological terms, and clarifies the legal framework within which they must be made. Collectively the chapters in this report reveal the many successes our country's experts, in collaboration with their peers in neighbouring countries and abroad, have achieved in understanding elephants and their needs, in fields as diverse as veterinary science, ecology, animal behaviour, population and resource modelling.Importantly, the Assessment exposes important gaps in our understanding and thus outlines necessary future avenues of research. This Assessment represents a key milestone in an ongoing Elephant Research Programme.Science does not provide all the information required to resolve the difficult issues raised by the management of elephant in a changing and humandominated world. Many of the required decisions have a strong element of human values implicit in them. How do South Africans wish to treat the other species with which they share our land? Extensive consultation and careful consideration of the values expressed by a wide range of stakeholders is also an essential part of the process of managing elephant in a democratic country.I am grateful to the many experts and interested persons who invested their time, experience and intellect to deliver this Assessment. I look forward to their continued engagement on the issue of elephant management, which is of great interest to many. Marthinus van Schalkwyk Minister of Environmental Affairs and Tourism, 2008 ContentS LiSt oF FigureS LiSt oF boxeS About the AuthorS And ContributorS Brandon Anthony is an assistant professor at the Department of EnvironmentalScienc...

Baobab size class distributions were surveyed in the Limpopo National Park (LNP), Mozambique, and the Kruger National Park (KNP), South Africa. There are very few elephants in the LNP and the baobab population there had a reverse J‐shaped size class distribution with many small baobabs. In contrast, the elephant‐impacted baobab population of KNP displayed a mono‐modal size‐class distribution, with a lack in recruitment. Within KNP, elephant impact (percentage bark stripped up to the height of 3 m) decreased with increasing rockiness and slope steepness. We interpret this to suggest that steep rocky slopes are inaccessible to elephants and therefore these sites may act as a refuge for baobabs. In such inaccessible areas, the baobab population has a similar size‐class distribution to that of the populations in the LNP. However, these baobab refugia are restricted in the northern KNP landscape and are therefore probably not large enough to sustain a viable baobab population.

No abstract