using four heat-flux plates and four temperature probes. Evapotranspiration and CO 2 flux were measured using the eddy-covariance method with an Applied Technologies sonic anemometer and LI-COR 6262 infrared gas analyser mounted on 2-m towers 25 . The mean and standard error for energy flux, gross primary production and evapotranspiration at sites 3 and 4 were calculated on the basis of 30-min averages. CO 2 fluxes at sites 11, 17 and 21 were determined using eddycovariance methods and 2.5-m towers 26 . Mean values and standard errors at these sites were calculated using the daily mean CO 2 fluxes. The daily methane fluxes were integrated over the thaw period to obtain annual emission. Winter methane fluxes were assumed to be zero. CH 4 flux was measured during the thaw season, Jun-Aug, at 27 MNT and MAT sites along the Dalton Highway in 1996 using a static chamber method 30 . Air samples were taken over periods of 30-45 min and were analysed on a gas chromatograph equipped with a flame ionization detector.

Uganda is one of the most biologically diverse countries in Africa, with much of its biodiversity represented in a system of 10 national parks, 10 wildlife reserves, and 710 forest reserves, covering 33,000 km2 (14%) of the country's area. We focus on the role of the forest reserves in biodiversity conservation and describe a procedure we developed to design a national system of forest nature reserves. In the late 1980s a policy was instituted to dedicate half the area of forest reserves to sustainable timber production and the other half to environmental protection (with 20% as nature reserves). To select suitable sites, a 5‐year, US$1‐million program of biodiversity and resource assessment was undertaken, focusing on five biological indicator species groups and covering all the major forest reserves. Based on data generated by the field studies, we ranked each forest in terms of various criteria—(species richness, rarity, value for nonconsumptive uses, timber production, and importance to local communities)—and used an iterative site selection procedure to choose the most suitable combination of forests for nature reserve establishment. Our procedure maximized complementarity in representing species and habitats in reserves across the whole protected‐area system. We initially selected sites using purely biological criteria but later modified our procedure to ensure that opportunity costs and potential land‐use conflicts were minimized. Our preferred network of sites included 14 forests that, in combination with the existing national parks, would account for 96% of species represented in the country's protected areas. These 14 forests were classified as “prime” and “core” sites and were selected for the establishment of large nature reserves (averaging 100 km2). The addition of 25 smaller “secondary” forest nature reserves (averaging 32 km2) would protect more than 99% of the indicator species.

S -wThe floristic and structural changes resulting from the long-term exclusion of large herbivores from an experimental area set up in 1971 have been analysed, comparing it to two plots in the surrounding grazed and trampled area.The vegetation of the area is grassland with thicket clumps, with Sporobolus pyramidalis P. Beauv. dominating the grassland and Capparis tomentosa Lam. dominating the thicket layer. The survey showed that long-term exclusion of herbivores results in: (i) higher density and cover in the grass and thicket layer, (ii) lower biodiversity in the grass layer and in the isolated shrubs, (iii) higher root biomass, probably due to the absence of trampling. Despite the difference in area, no difference was noted between the biodiversity of the thicket clumps of the ungrazed area and the grazed and trampled plots. Ri%UIll6On a analysk les changements floristiques et structurels resultant de l'exclusion a long terme des grands herbivores d'une zone experimentale ktablic en 197 1, en la comparant a deux zones situkes dans la region voisine, pbturee et piktinke. La vegetation de la region est une prairie avec des massifs de buissons, avec Sporoboluspyramidalis P. Beauv. dominant la prairie et Capparis tomentosa Lam. dans les buissons. L'Ctude a montre que l'exclusion 9 long terme des herbivores aboutit a: (i) une densite et une couverture plus grandes de l'herbe et des buissons, (ii) une biodiversitk plus faible dans la couche herbeuse et dans les fourrks isoles et (iii) une plus forte biomasse des racines, due probablement a l'absence de piktinement. MalgrC la region beaucoup plus ClevCe, on n'a note aucune difference de biodiversite dans les agglomkrats de buissons dans la zone non pgturee.

No abstract

BioOne Complete (complete.BioOne.org) is a full-text database of 200 subscribed and open-access titles in the biological, ecological, and environmental sciences published by nonprofit societies, associations, museums, institutions, and presses.