Summary1. Current emphasis on species conservation requires the development of speci®c distribution models. Several modelling methods are available, but their performance has seldom been compared. We therefore used discriminant analysis, logistic regression and arti®cial neural networks with environmental data to predict the presence or absence of six river birds along 180 Himalayan streams. We applied each method to calibration sites and independent test sites. With logistic regression, we compared performance in predicting presence±absence using map-derived predictors (river slope and altitude) as opposed to detailed data from a standardized river habitat survey (RHS). 2. Using the entire calibration data, overall success at predicting presence or absence was only slightly greater using arti®cial neural networks (89±100%) than either logistic regression (75±92%) or discriminant analysis (81±95%), and on this criterion all methods gave good performance. 3. When applied to independent test data, overall prediction success averaged 71± 80%, with logistic regression marginally but signi®cantly out-performing the other methods. Encouragingly for researchers with limited data, model performance in jack-knife tests faithfully represented performance in more rigorous validations where calibration (n = 119) and test sites (n = 61) were in separate geographical regions. 4. All three methods predicted true absences (83±92% success) better than true presences (31±44%). Results from logistic regression were the most variable across species, but positive prediction declined with increasing species rarity in each method. 5. Applications with logistic regression illustrated that signi®cant habitat predictors varied between data sets within species. Hypotheses about causal eects by habitat structure on distribution were thus dicult to erect or test. Logistic regression also showed that detailed data from the river habitat survey substantially improved positive prediction by comparison with prediction using slope or altitude alone. 6. We conclude that discriminant analysis, logistic regression and arti®cial neural networks dier only marginally in performance when predicting species distributions. Model choice should therefore depend on the nature of the data, on the needs of any particular analysis, and on whether assumptions for each method are satis®ed. All three methods share drawbacks due to systematic eects by species rarity on performance measures. They also share limitations due to the correlative nature of survey data often used for model development at the spatial scales required in macro-ecology and conservation biology. Tests with independent data, *Correspondence: Dr Steve Ormerod (e-mail: ormerod@cardi.ac.uk). {Present address: Laboratoire de Biologie des Populations d'Altitude, CNRS UMR 5553, Universite Joseph Fourier, BP53-38041 Grenoble Cedex 9, France. Journal of Applied Ecology 1999, 36, 734±747# 1999 British Ecological Society using a wider range of performance measures than those used traditionally, will be...

Summary1. Piecemeal changes in land use might have cumulative eects on regional biodiversity. However, this hypothesis is dicult to test experimentally at the scales involved, so alternative approaches are required. Here, we illustrate some of the strengths and weaknesses of surveys for evaluating the eects of land use on rivers and river birds over a large area of the Himalayan mountains. 2. We surveyed 180 streams and their catchments in north-west India and Nepal in 1994±96. We then used analysis of covariance (ANCOVA), multiple linear regression and multiple logistic regression to assess how stream habitat structure, stream chemistry, aquatic invertebrate abundance and the occurrence of river birds were aected by land use after accounting for altitudinal pattern. 3. Streams draining terraced catchments diered signi®cantly in habitat structure from other streams. They had more physical modi®cations, wider channels, fewer cascades, ®ner substrata and simpler riparian vegetation with fewer trees. We detected no clear eects of land use on stream chemistry, but terracing was accompanied by signi®cantly increased abundances of benthic dipterans, ephemeropterans and total aquatic invertebrates. 4. River bird occurrence was best explained by altitude, and secondarily by habitat structure. Some of the habitat features in¯uenced by terracing signi®cantly aected birds both positively (grey wagtail Motacilla cinerea) and negatively (little forktail Enicurus scouleri, river chat Chaimarrornis leucocephalus, brown dipper Cinclus pallasii, plumbeous redstart Rhyacornis fuliginosus). However, only in the grey wagtail did the presence of terracing per se aect occurrence unequivocally; eects on other species were either small or confounded by altitude. 5. We cannot refute the hypothesis that catchment land use aects Himalayan river ecology, but our data on the regional consequences for river birds were equivocal. We suggest that large-scale surveys, although providing one of the few pragmatic methods of assessing large anthropogenic eects on ecosystems, will need careful design to factor out potential confounds if they are to be used to test hypotheses robustly. They should also be supported where possible with process studies, intervention studies and model applications to independent data.

1. Despite the growing view that biodiversity provides a unifying theme in river ecology, global perspectives on richness in riverine landscapes are limited. As a result, there is little theory or quantitative data on features that might have influenced global patterns in riverine richness, nor are there clear indications of which riverine landscapes are important to conservation at the global scale. As conspicuous elements of the vertebrate fauna of riverine landscapes, we mapped the global distributions of all of the world's specialist riverine birds and assessed their richness in relation to latitude, altitude, primary productivity and geomorphological complexity (surface configuration). 2. Specialist riverine birds, typical of high‐energy riverine landscapes and dependent wholly or partly on production from river ecosystems, occur in 16 families. They are represented by an estimated 60 species divided equally between the passerines and non‐passerines. Major radiation has occurred among different families on different continents, indicating that birds have evolved several times into the niches provided by riverine landscapes. 3. Continental richness varies from four species in Europe to 28 in Asia, with richness on the latter continent disproportionately larger than would be expected from a random distribution with respect to land area. Richness is greatest in mountainous regions at latitudes of 20–40°N in the riverine landscapes of the Himalayan mountains, where 13 species overlap in range. 4. Family, genus and species richness in specialist riverine birds all increase significantly with productivity and surface configuration (i.e. relief). However, family richness was the best single predictor of the numbers of species or genera. In keeping with the effect of surface configuration, river‐bird richness peaks globally at 1300–1400 m altitude, and most species occur typically on small, fast rivers where they feed predominantly on invertebrates. Increased lengths of such streams in areas of high relief and rainfall might have been responsible for species–area effects. 5. We propose the hypothesis that the diversity in channel forms and habitats in riverine landscapes, in addition to high temperature and primary productivity, have been prerequisites to the development of global patterns in the richness of specialist riverine organisms. We advocate tests of this hypothesis in other taxonomic groups. We draw attention, however, to the challenges of categorically defining riverine organisms in such tests because (i) rivers grade into many other habitat types across several different ecotones and (ii) `terrestrialisation' processes in riverine landscapes means that they offer habitat for organisms whose evolutionary origins are not exclusively riverine.

1. The monsoon causes major flood events in some Himalayan streams, but their seasonal predictability might reduce the resulting disturbance. We assessed seasonal change in the benthos of 16 streams in central Nepal over a gradient of declining rainfall and increasing altitude from 600 to 3800 m. All sites were surveyed on four occasions, two in winter (November) and two pre‐monsoon (June), with additional sampling during the monsoon (August) at four low altitude sites. Invertebrate abundance, taxon richness and persistence were assessed at all sites, and density and meso‐habitat distribution at the four low altitude sites only. 
2. Strong seasonal variation among invertebrates was confined primarily to streams at low altitude (600–800 m) where monsoon rainfall was greatest and catchments were dominated by terraced agriculture. At these sites, a significant reduction in benthic density (on average by 77%) and taxon richness (by 20%) occurred between the winter and pre‐monsoon periods, so that invertebrate numbers were already low before the monsoon. A further significant decline occurred in all meso‐habitats during the monsoon, but the change in density was small in absolute terms. 
3. Persistence in rank abundance was equally low at all sites, but turnover in composition was significantly lower at sites in semi‐natural forest than in catchments managed for terracing or alpine pasture. 
4. These data provide no evidence that monsoonal floods represent major disturbance, instead supporting the view that the ecological response might reflect an adjustment to predictable flow pattern. However, catchment land use in the Himalaya appears to be a significant source of ecosystem instability, and confounds the simple interpretation of monsoon effects.

The Vietnamese Mekong Delta has undergone much anthropogenic change in the last 100 years, and few areas of natural or semi-natural habitat remain. Despite the likely consequent loss of biodiversity, little documentation exists about the region's avifauna at any time in its history. Here, we present a review of the avifauna of the Vietnamese part of the Mekong delta based on various surveys and other fieldwork carried out since 1988, and an assessment of the few earlier data that were available. A total of 247 species has been recorded from the Vietnamese delta since 1988. Most pre-war data are now considered to be of uncertain validity. Some 50% of the species recorded since 1988 are dependent on wetlands. Of these, 20 are listed as globally threatened or Near Threatened. The delta supports, or is likely to support, internationally important numbers of 21 species of waterbird, including two threatened and three Near Threatened species. Annotations are provided for records of globally or regionally threatened species; for those species for which the Mekong Delta holds internationally important numbers; and those for which records described herein constitute range extensions. Information on the avifauna of the Mekong Delta presented here suggests it is of regional and global importance for the maintenance of wetland biodiversity.