Studying species turnover along gradients is a key topic in tropical ecology. Crucial drivers, among others, are fog deposition and soil properties. In northern Peru, a fog-dependent vegetation formation develops on mountains along the hyper-arid coast. Despite their uniqueness, these fog oases are largely uninvestigated. This study addresses the influence of environmental factors on the vegetation of these unique fog oases. Accordingly, vegetation and soil properties were recorded on 66 4 × 4-m plots along an altitudinal gradient ranging from 200 to 950 m asl. Ordination and modelling techniques were used to study altitudinal vegetation belts and floristic composition. Four vegetation belts were identified: a low-elevation Tillandsia belt, a herbaceous belt, a bromeliad belt showing highest species richness and an uppermost succulent belt. Different altitudinal levels might reflect water availability, which is highest below the temperature inversion at around 700 m asl. Altitude alone explained 96% of the floristic composition. Soil texture and salinity accounted for 88%. This is in contrast with more humid tropical ecosystems where soil nutrients appear to be more important. Concluding, this study advances the understanding of tropical gradients in fog-dependent and ENSO-affected ecosystems.

Volcán Barva, Costa Rica, has on its northern slope an unbroken sequence of rain forest on volcanic parent materials from near sea level at La Selva Field Station up to its summit at 2906 m. It provides a good area to study forest changes with altitude and their causes. In the present paper we describe the forests as a background for soil and litterfall studies from 1 ha plots at each of the following altitudes: 100 m, 500 m, 1000 m, 1500 m, 2000 m and 2600 m. The canopy heights (with height of the highest emergent in parentheses) ranged from 35–40 m (45 m) at 100 m to 20–23 m (32 m) at 2600 m; basal area was least (22.7 m2) at 100 m and highest (51.2 m2) at 2600 m; the tree (≥10 cm dbh) density ranged from 391 ha–1at 500 m to 617 ha–1 at 2600 m. Most trees were identified and on samples of them we recorded presence of buttresses, lianes, skiophytic climbers, vascular epiphytes and bryophytes; and drew profile diagrams. In the classification of Whitmore (1984) the two lower plots are evergreen lowland rain forests; the other four are lower montane rain forest. Species richness was highest in the plot at 500 m, with at least 135 species of tree, and least at 2600 m, with at least 35 species. The Volcán Barva forest altitudinal sequence is briefly compared with those elsewhere.

The numbers and biomass of litter and soil invertebrate fauna were investigated in six plots at altitudes of 280 m, 330 m, 480 m, 610 m, 790 m and 870 m on Gunung Silam, Sabah, East Malaysia. There were relatively high numbers and biomass in the lower plots, where the Oligochaeta were a high proportion of the total invertebrate biomass. The biomass of other invertebrate groups was low in the soil. There was a marked effect of altitude, particularly for the Oligochaeta above 610 m. The low biomass of the Oligochaeta in the higher plots suggested that the importance of this group may be diminished at relatively low altitudes on small mountains. However, several other soil faunal groups showed no evidence of an altitudinal effect, and it is suggested that the stunted forests near the summits of small mountains such as Gunung Silam may have a different litter and soil fauna from forests of a similar physiognomy at high elevations on large mountains.

The altitudinal zonation of forests on Mount Kerinci, Sumatra (3800 m) is described and compared with that in the temperate region of east Asia. Nine plots were selected between 1750 m in altitude and the upper limit of vegetation at 3250 m, at intervals of about 200 m in altitude. The plots are distinguished according to their main dominants, and the population structure of the dominant species is examined. The lower forests have species showing the whole range of size classes as well as solitary giants as dominants, but the upper forests lack these giants and are floristically poorer. Between 1750 m (the start of well preserved natural vegetation) and 2950 m (the forest limit) three forest zones are distinguished, and between 2950 and 3250 m a scrub zone. Upper forest zones tend to be dominated by species of the same genus or family which form important understory components of the zone below. Based on floristic comparisons with mountains of higher latitudes (i.e. Himalayas and Japan), the two lower forest zones (up to 2400 m) represent a subtropical zone, and the upper forest zone a warm-temperate zone. Climatic conditions at the forest limit on Mount Kerinci are similar to those at the latitudinal limit of warm-temperate evergreen trees; in the Himalayas the forest limit represents the latitudinal limit of the cool-temperate, and in Japan of the subarctic Altitudinal zonation patterns change with latitude, reaching their most complex on subtropical mountains where the two floristic realms, the Boreal and Palaeotropical, meet. A scheme for the pattern of vegetation zonation in east Asia is proposed.

Bukit Belalong, Brunei, is a small mountain (913 m) of uniform shale lithology with continuous primary rain forest from its base to its summit. Three 0.25 ha plots were established at each of three altitudes (200, 500 and 850 m) to investigate the altitudinal zonation of the soils and the vegetation. One soil profile from each altitude is described and chemical analyses of the soils indicate that they are similar at all altitudes. In each plot all trees ≥10 cm dbh were measured and identified as far as possible. Maximum tree height was greatest at 200 m (60 m) and least at 850 m (33 m). Tree density (number of individuals ha−1) increased with altitude. The forest at 850 m was the most species- and family-rich, but since the ratio of species to individuals did not vary significantly among altitudes, the higher species richness is attributed partly to the larger number of trees sampled. The Dipterocarpaceae was the most important family in terms of basal area at all altitudes, but its proportion of basal area was much smaller at 850 m than at the other altitudes. The Dipterocarpaceae was the most diverse family at 200 m and 500 m, but at 850 m its diversity was exceeded by the Lauraceae, Myrtaceae and Euphorbiaceae and equalled by the Clusiaceae. The forests at 200 m and 500 m are classified as lowland rain forest and that at 850 m is classified as a lower montane forest.

Montane tropical and subtropical rain forests are complex ecosystems, characterized by marked rainfall and temperature gradients with altitude, which in turn control the vegetation altitudinal zones (Hueck 1978). Montane forests are often referred to as cloud forests in recognition of the important influence of a dense and frequent cloud cover that conditions forest structure and functioning (Bautista-Cruz & del Castillo 2005, Holder 2004).

Although some studies in tropical forests suggest the existence of elevational discontinuities, zones or critical altitudes in floristic composition (briefly reviewed by Lovett 1996) two data sets from the Usambara mountains of Tanzania independently show a continuous elevational change in moist forest large tree communities (Hamilton et al. 1989, Lovett 1996). However, one data set is from the West Usambara mountains (Lovett 1996) and the other is from the East Usambara (Hamilton et al. 1989). The junction of the two data sets at around 1000–1200 m in elevation is regarded as a critical altitude (Lovett 1996), and so neither conclusively prove continuity over the entire elevational range of the forests. The two data sets were collected using the same methodology, the twenty-tree variable-area plotless technique (Hall 1991), and together constitute a transect covering nearly 2 km in elevation. They are combined here to test for continuity in change in forest large-tree communities from lowland to upper montane forest types.