GRASS ROOTS RESEARCH

Do savannahs teem with more life than the rain forests? Matthew Brace investigates

In Dr George McGavin's cluttered laboratory in Oxford's University Museum lie 1,000 specimen tubes. They cover his workbench, swarm around his microscope and huddle together in plastic bags under his desk.

Each vial contains hundreds of insects. Such is their number that Dr McGavin, a leading entomologist and the assistant curator of the museum, is expecting to be working overtime ploughing through the lengthy backlog of analysis to be done. This is the latest batch he has carried back to his Oxford lair from an area of seemingly barren, and previously unstudied, East African savannah that is yielding a remarkably abundant insect fauna.

In fact, Dr McGavin and his fellow scientists working in the Mkomazi Game Reserve in north-eastern Tanzania believe they have discovered an ecological treasure chest, an area more diverse than many rain forests, in terms of the numbers of arthropods (animals with segmented bodies and jointed appendages - insects, spiders, centipedes, etc). On an average day, after "mist-blowing" a tree in the reserve with insecticide, Dr McGavin watches as clouds of insects fall from the branches into his collection trays. During years of research in rain forests, deserts and grasslands around the world he has never seen such abundance.

"When I first started working out here I was amazed," he said. "Each morning I'd rush out with my collecting gear knowing I'd be coming back laden with insects at the end of the day."

If Mkomazi is representative of other unstudied areas of savannah in East Africa, these discoveries could go some way to challenging the traditional belief that rain forests are the richest habitats on Earth.

"When people are asked about diversity they always think about rain forests," says Dr McGavin. "Savannahs might have fewer species than rain forests but according to our results in Mkomazi, the numbers of individual insects are immense. In that sense, the savannahs are more diverse [scientifically diverseness is not just richness of species but also abundance]."

His preliminary examinations of 15 acacia trees (from two species, Acacia zanzibarica and Acacia tortilis) show that the average arthropod specimen count is 1,600 per square metre. One small tree can hold more than 70,000 individuals. The major rain forest canopy arthropod studies of the past 12 years have produced much smaller specimen totals. Similar studies have confirmed the relative arthropod diversity of the grasslands. The highest figure in the Toraut lowland forest of Sulawesi, south east Asia, was 461 per square metre; dry evergreen tropical forest in north-east Thailand yielded between 123 and 256 per square metre; and subtropical rain forest near Brisbane, Australia, registered just 34 per square metre. Although it is unlikely Mkomazi will compete with rain forests on overall species richness (the number of different species), the reserve is continuing to reveal a surprisingly large variety of species, some of which are presently undescribed.

While he has identified roughly 500 known species in these two acacia types, Dr McGavin estimates from his early results that one in five of the arthropods from the acacias is unidentified, or, new to science. These new arrivals will help swell the list of the million-plus species of known insects, which make up 56.3 per cent of all living organisms.

And it is not just the plethora of insects that is putting Mkomazi on the diversity map. Ento-mologist and principal scientific officer at the National Resources Institute, Dr Tony Russell- Smith, has logged 575 morpho-species (species yet to be finally catalogued) and estimates up to 20 per cent will turn out to be new to science.

However, it is the number of spider families represented in Mkomazi that has stunned Dr Russell-Smith. "There are between 150 and 200 known spider families, in the tiny area we have been looking at - about 850sq km - we have discovered at least a quarter of that total."

The intense sampling methods used by both Dr McGavin and Dr Russell-Smith could account for some of the high results they have been recording, but neither is in any doubt that in Mkomazi they have found an area of remarkable ecological diversity and value.

Another main faunal survey, of birds, has resulted in one of the highest bird counts ever in East Africa. Dr Peter Lack, head of information systems at the British Trust for Ornithology, has identified roughly 400 species.

"Mkomazi is one of the most diverse areas ever found," says Dr Lack. "And these high bird counts are important from a conservation point of view as they allow us to learn how these systems work and how we can preserve them."

So why such abundance? One main reason is Mkomazi's geography. The reserve lies at the southern end of the great East African savannah - a continuation of the Tsavo system over in Kenya. Within the reserve's 3,600sq km are various different environments - pockets of forest, areas of water, and a range of wooded hills called the Pare mountains. It is this variety that allows Mkomazi to support many species.

Before this project little work had been done in the area, according to Dr McGavin. "We know so little about savannahs but are always striving to study the rain forests, about which we know relatively much more," he said. "I'm not knocking rain forests - they are wonderful things - but the concentration of efforts on them detracts from other areas. Hopefully, the incredible diversity this reserve is revealing will show that savannahs need to be understood so they can be protected if necessary."

He added: "Tropical grasslands cover about one sixth of the world's land. Although we now understand them in terms of the big animals we certainly don't understand them in terms of the small ones. But insects are crucial to this environment. They shape it, they make the savannah system work. Ants, for example, carry off and consume far more animal tissue than the big animals - it's just that you don't see them doing it because they are so small."

The research in Mkomazi is part of an on-going programme co-ordinated jointly by the Royal Geographical Society (with the Institute of British Geographers) and the Tanzanian Department of Wildlife.

The Society got involved when the Tanzanian government (condemned as the reserve has lost much of its game and been ravaged by fires), came to them in 1989 for help in undertaking a major geographical study of the area in order to collect information to prepare a long-term management plan for the future of the reserve.

The programme has remained an Anglo-Tanzanian venture with joint directors Mr Bukari Mbano, director of wildlife in the Tanzanian Ministry of Natural Resources and Environment, and Dr Malcolm Coe, a leading ecologist from St Peter's College, Oxford.

The project won a Darwin initiative grant worth more than pounds l00,000 to fund work to describe, in quantitative terms, the diversity of the terrestrial invertebrate fauna of Mkomazi and to assess the influence upon it of natural and human-induced factors.

And the significance of the work has been recognised by organisations including Friends of Conservation, the British Council, BA, Land Rover, and BP which have contributed towards the pounds 250,000 sponsorship of the project.

The project recently secured a further initiative grant and the possibility of making Mkomazi a long-term field study is being discussed. Such a plan would keep a window constantly open on what is turning out to be one of the most exciting areas for biodiversity research. !