Search

Vast areas of tropical evergreen forests have been selectively logged in the past, and many areas continue to be logged. The impacts of such logging on amphibians are poorly understood. I examined the response of anuran communities to historical selective logging in a wet evergreen forest in south India. Anuran assemblages in unlogged forest were compared with assemblages in selectively logged forest. Forty 10 m 9 10 m quadrats in forest, riparian zones, and streams of unlogged and selectively logged forests were searched at night for anurans. Species richness did not appear to be affected by logging. However, anuran density varied significantly and was 42 percent lower in selectively logged forests compared to unlogged forests. Anuran densities also varied significantly across microhabitats, with highest densities in streams of both selectively logged and unlogged forests. Patterns of niche overlap varied with selective logging as niche breadth either expanded, contracted, or remained neutral for different species. Ordination analysis explained 95 percent of the variation in species assemblage across selectively logged and unlogged forests. The assemblage in selectively logged forest was nested within unlogged forest. Among the habitat characteristics, litter thickness and water depth had the highest influence on the assemblage. This was followed by litter/water temperature, air temperature, and lastly relative humidity. It appears that species richness and composition of anurans in selectively logged forests is converging with unlogged forests, but the effects of historical logging seem to persist on anuran densities and their niche characteristics even ca 40 yr since logging ceased.

Movement patterns of animals can vary dramatically as a function of their reproductive cycle or social structure; however, little is known about how changes in the social structure of dispersers affect patterns of seed dispersal. We examined the movement patterns of the forest-dwelling and cooperatively breeding Puff-throated Bulbul (Alophoixus pallidus) in relation to different stages of their reproductive cycle, time of day, and group size, to determine potential impacts on the shape and scale of dispersal curves generated using a combination of gut passage time and displacement distance data. There were significant differences in dispersal distances depending on group size, season (breeding, non-breeding), incubation (vs. other times of the year), and time of the day. The estimated median seed dispersal distance was 28 m. The median dispersal distances produced by birds in larger groups were longer than those of smaller groups (29 m vs. 25 m). During the breeding season, median dispersal distances were longer than during the non-breeding season (31 m vs. 25 m), but the median dispersal distances were significantly shorter during incubation than during outside incubation (24 m vs. 28 m). The median dispersal distance produced in the early morning (30 m) was also longer than that of other times of the day (23 m late morning, 28 m early afternoon, and 26 m late afternoon). This study suggests that various aspects of an animal's behavior are likely to have significant effects on seed shadows and that this may vary significantly even among individuals of the same species. บทคดยอ รปแบบการเคลอนทของสตวสามารถเปลยนแปลงไปตามชวงเวลาของการสบพนธหรอโครงสรางทางสงคม แตอยางไรกตามเรายงมความรนอยมากเกยวกบการเปลยนแปลงโครงสรางทางสงคมของสตวผกระจายเมลดตอรปแบบการกระจายเมลด เราทำการศกษารปแบบการเคลอนทของนกปรอดโองเมองเหนอ ซงเปนนกปรอดปาและมระบบการสบพนธแบบมนกผชวย โดยศกษาความสมพนธของการเคลอนทตอชวงระยะในการสบพนธ ชวงเวลาของวน และขนาดของกลม ซงอาจจะสงผลตอรปแบบการกระจายเมลดทสรางขนจากการรวมขอมลระหวางระยะเวลาทนกใชในการขบถายเมลดและระยะทางการเคลอนท จากการศกษาพบวาระยะทางการกระจายเมลดมความแตกตางอยางมนยสำคญขนอยกบ ขนาดของกลม ฤดกาล (ฤดผสมพนธ-นอกฤดผสมพนธ) ระยะการกกไข (เปรยบเทยบกบระยะเวลาอนๆของป) และชวงเวลาของวน นกปรอดโองเมองเหนอมคากงกลางระยะทางการกระจายเมลดอยท 28 เมตร นกทมขนาดกลมใหญกวาจะมคากงกลางระยะทางการกระจายเมลดมากกวานกทมขนาดกลมเลกกวา (29 เมตร ตอ 25 เมตร) และในฤดผสมพนธนกปรอดโองเมองเหนอจะมคากงกลางระยะทางการกระจายเมลดมากกวาในชวงนอกฤดผสมพนธ (31 เมตร ตอ 25 เมตร) แตเมอพจารณาในชวงเวลาการกกไขคากงกลางระยะทางการกระจายเมลดมระยะทางสนกวาชวงทไมมการกกไขอยางมนยสำคญ (24 เมตร ตอ 28 เมตร) คากงกลางของระยะทางการกระจายเมลดทเกดในชวงเชาตรมคาระยะทางมากทสด (30 เมตร) เมอเทยบกบชวงเวลาอนๆของวน (ชวงสาย 23 เมตร ชวงบาย 28 เมตร และ ชวงเยน 26 เมตร) การศกษานแสดงใหเหนวาพฤตกรรมสตวมผลตอรปแบบการกระจายเมลดอยางมาก แมกระทงความแตกตางของพฤตกรรมของสตวแตละตวทเปนชนดเดยวกนไดสงผลตอรปแบบการกระจายเมลดอยางมนยสำคญ

Making generalizations about the impact of commercial selective logging on biodiversity has so far remained elusive. Species responses to logging depend on a number of factors, many of which have not been studied in detail. These factors may include the natural forest conditions (forest types) under which logging impacts are investigated; but this question has so far remained unexamined. In a large-scale replicate study we aimed at clarifying the relationship between logging and forest types on leaf litter frogs. We contrast three distinct and naturally occurring forest types, including wet evergreen, moist evergreen and semi-deciduous forests. Selectively logged sites were compared with primary forest sites for each forest type. We found that the response of frog communities to logging varies in different forest types. In the wet evergreen forest, richness was higher in logged forest than primary forest, while diversity measures were not different between logged and primary forest habitats. In the moist evergreen, richness and diversity were higher in selectively logged areas compared with primary forest habitats. In the semi-deciduous, logged forests were characterized by drastic loss of forest specialists, reduced richness, and diversity. These results indicate that the net effect of logging varies with respect to forest type. Forest types that are characterized by adverse climatic conditions (i.e., low rainfall and protracted dry seasons) are more likely to produce negative effects on leaf litter anuran communities. For comparisons of the impact of logging on species to be effective, future research must endeavor to include details of forest type.

Clouded Leopard, Leopard, and Tiger are threatened felids in Southeast Asia, but little is known about the factors in! uencing their distributions. Using logistic regression, we assessed how habitat variables, prey detection patterns, and presence of intraguild predators affect the occurrence of these felids across 13 protected areas within Thailand. Our analysis is based on data from 1108 camera-trap locations (47,613 trap-nights). Clouded Leopard and Leopard are associated with habitat where Red Muntjac and Eurasian Wild Pig were most likely to be present. Tiger are associated with habitat with a higher likelihood for the presence of Gaur, Eurasian Wild Pig, and Sambar. Clouded Leopard and Tiger were both weakly associated with areas with mature evergreen forest. Besides availability of prey, associations with potential competitors also appear to in! uence the distribution of these felids, although the strength of these effects requires further investigation. Occurrence rates for Clouded Leopard were no different in protected areas with Leopard versus without Leopards. Leopard had similar occurrence rates regardless of the presence of Tiger, but Leopards were less likely to be detected at the same camera-trap points with the larger felid. Our results suggest that the two most commonly photographed prey species in the study areas serve as key prey species, Eurasian Wild Pig for all three carnivores and Red Muntjac for Leopard and Clouded Leopard.

Although it is known that forest loss and degradation negatively impact most forest-dwelling primates, such relationships are difficult to quantify because many primates are difficult to survey over large areas. Furthermore, recovery times are also difficult to assess due to a lack of long-term data. Here, we determined how forest characteristics and habitat disturbance correlate with the abundance of pileated gibbons, Hylobates pileatus. We studied a population in Khao Ang Rue Nai Wildlife Sanctuary in southeastern Thailand, assessed its density using an auditory method combined with distance sampling at 24 randomly placed sample sites. In addition, we determined how simple forest structural characteristics and habitat disturbance correlate with the gibbon abundance. Average gibbon density per site was 1.02 ± 0.16 (SE) groups/km 2 (range 0—2.74). Bivariate analyses indicated that densities depended on food tree biomass, level of disturbance, evergreen forest cover, time since protection, and distance to the sanctuary boundary. Multiple regression analysis suggested evergreen forest cover and distance to boundary were the most influential factors. Because evergreen forest cover, time since protection, and habitat disturbance are correlated, these results suggest a direct dependence of gibbon densities on mature, undisturbed evergreen forest. While gibbons can persist in disturbed areas if the forest is protected, it appears that recovery to previous densities may take decades. We suggest that this is due to the slow pace of forest regeneration and/or poor recovery potential of gibbons.

Montane evergreen forest in SE Asia is structurally and floristically different from lowland habitats. The response of large mammals to this variation is largely unexplored. We used sign transects to compare community structure of large mammals in montane (>1100 m), and lowland (

Diet and habitat use of the closely related and size-overlapping sympatric Cynopterus brachyotis and C. sphinx were established in lowland dry evergreen forest, Thailand, between March 1998 and March 2000. Feces from netted bats were analyzed, and the recapture rate determined. Although both species share a set of food plants, and fruits from early successional forest contribute about half of their diet, C. brachyotis, the smaller of the two species, ate a significantly greater proportion of fruits from early successional forest than C. sphinx. The latter ate a significantly greater proportion of fruit species in larger size classes. More C. brachyotis were captured in early successional forest in almost every month, while C. sphinx is more common in old-growth forest. However, the capture rate of C. sphinx increased in early successional forest in the mid-dry season when its preferred fruits become available. The recapture rate of C. brachyotis in early successional forest was significantly higher than that of C. sphinx, and the reverse situation was observed in old-growth forest. Male C. sphinx had a significantly higher recapture rate in early successional forest than females. Fruit size and habitat use are the major determinants of resource partitioning between these size-overlapping congeners.

We compared bird diversity and frequency in selection logged and unlogged forest to determine the effects of recent selection logging on avian biodiversity in a subtropical, moist evergreen forest. We used a combination of mist netting and fixed-radius point counts to assess bird communities in February and March 1993 in northwestern Belize. Vegetation structure and composition was similar in logged and unlogged forest. The 66 most common species occurred with statistically similar frequency in logged and unlogged forest although 13 species were two times more frequent in intact forest. Numbers of total bird species were similar between logging gaps and the logged forest matrix, and between the logged forest matrix and unlogged forests. A comparison of numbers of species in 26 guilds based on migration strategy, diet, foraging substrate, and height strata also showed them to be similar regardless of logging history. Our results differed from previous studies that reported lower bird species richness and abundance of individual species in logged tropical forests than in unlogged forest. The differences might be explained by the lower logging intensity and/or greater levels of natural disturbance in our study area compared to previous studies.

Wongabel, a northeastern Queensland tropical, wet, evergreen forest, contains 22 species of Scarabaeinae dung beetles. Five of these species were observed to perch commonly on leaves at night. Length of the beetle and the height of its perch were recorded for each of 56 1 specimens. Unlike the New World tropical dung beetle perchers, no clear evidence was found that small species perched closer to the ground than larger species. The evidence gathered, at least for the four most common perchers, supports the hypothesis that perching is one type of foraging strategy. The similarities and differences between the Australian and New World perchers are discussed.

I assessed damage caused by a cyclone in November 1978 to a dry evergreen forest in Sri Lanka. Damage included defoliation, breakage of twigs, branches and trunks, tree fails and post-cyclone tree mortality. Within the 3 km2 study area there was a trend for increasing damage with forest height. Tree species in the upper forest layers had significantly more falls and post-cyclone mortality than trees in the more sheltered subcanopy and shrub layers. Mortality was significantly greater among trees which lost 40 percent or more of their branches and trunks than among trees with lesser crown damage. Defoliation and twig loss were extreme in the discontinuous upper layers and probably contributed to the greater tree mortality evident there. Overbrowsing by folivorous primates after the cydone may have contributed to the demise of some preferred feeding trees. Total tree loss was 46 percent from the upper forest layers, 29 percent from the subcanopy, or 40 percent from all tree layers. Five upper layer tree species were subject to tree losses of 80 to 100 percent representing a virtual elimination of 22 percent of species from the upper layers or 12 percent from the forest. Dry evergreen forest formerly covered 80 percent of the island's land area, and often has been described as "old secondary climax" in recognition of past disturbance. Extrapolations from meterological data indicate that 33 to 44 percent of the range of dry evergreen forest may be subject to cyclone damage per century. The species composition of dry evergreen forest on a local and wider geographical scale typically is fairly uniform in the subcanopy layer, but variable in the upper layers. I suggest that recurrent cyclone damage may be an important factor contributing to succession in dry evergreen forest and to the variation in species composition of the upper layers. TROPICAL CYCLONES OR HURRICANES are known to cause extensive damage to human life and property in the coastal regions of the Bay of Bengal in the Indian Ocean. Although natural plant and animal populations are also subject to cyclone damage, published reports addressing this phenomenon are rare and concern areas outside the Indian region (e.g., Webb 1958, Whitmore 1974). A recent cyclone which swept through the dry evergreen forest of Sri Lanka provided an opportunity to study the effects on phenology, forest structure and floristic composition. Quantitative data concerning these features were collected before the cyclone, allowing their comparison before and after the cyclone. The objectives of this paper are threefold: first, to assess the nature and extent of cyclone damage to the forest; second, to examine the causes of tree mortality in the 42 months following the cyclone; and third, to consider the effect of recurrent cyclones on the geographic variation in species composition of the dry evergreen forest of Sri Lanka. The long-term recovery of the forest from cyclone damage will not be examined in this report. Data concerning phenology are considered briefly, however, in order to show the immediate effects of the cyclone. NOMENCLATURE AND NATURAL HISTORY The forest types of Sri Lanka have been classified according to several schemes (e.g., Chapman 1947; Holmes 1956; Koelmeyer 1957, 1958; Gaussen et al. 1964; Fernando 1968). Mueller-Dombois (1968) offers an evaluation of these schemes and additional refinements have been made recently by Greller and Balasubramaniam (1980) and Greller et al. (1980). The forest type considered here is commonly known in Sri Lanka as the "dry-zone forest." It has also been referred to as "Tropical Dry Evergreen Forest" (Champion 1935, Chapman 1947, Holmes 1956), "Tropical Dry Mixed Evergreen Forest" (de Rosayro 1950; Koelmeyer 1957, 1958; Andrews 1961), "Semideciduous Forest" (Gaussen et al. 1964, 1965), "South Tropical Moist Deciduous Forest" (Chapman 1947, Koelmeyer 1957), "Tropical Lowland Seasonal Rain Forest" (Perera 1975) and "Semi-evergreen Forest" (Dittus 1977). The last label had been selected to conform to Walter's (1971) worldwide classification of tropical vegetation types, although Walter does not specifically refer to the dry-zone forest of Sri Lanka. In the literature the dry-zone forest is most frequently referred to as "Dry Evergreen Forest" and therefore I will follow this tradition here. Taxonomic nomenclature follows Abeywick-rama (1959) and Dassanayake and Fosberg (1981, 1983). I Received 2 August 1983, revised 11 May 1984, accepted 14 June 1984. 2 Field address: Smithsonian Institution Primate Project, 4/4 Galkanda Road, Anniewatte, Kandy, Sri Lanka. BIOTROPICA 17(1): 1-14 1985 1 This content downloaded from 207.46.13.127 on Wed, 12 Oct 2016 04:32:28 UTC All use subject to http://about.jstor.org/terms

This report compares flowering seasons in perennial woody plants at four altitudinal levels in the Harar Highlands of eastern Ethiopia. In this area a majority of woody species flower in the dry season only in the montane evergreen forest formation above about 2200 meters altitude. In the other three deciduous formations at lower altitudes flowering is primarily in the wet season. In the drier Acacia (microphyll) formations there is a burst of flowering at the beginning of the rainy season. THE AREA OF the Harar Highlands in eastern Ethiopia above 1000 m altitude and north of latitude 80 30' North experiences a single wet season with two maxima, in April and August. The rains that fall between March and June are called the "little rains" by local people. They may be heavy but vary greatly from year to year and are dependable for agriculture only above about 2200 m elevation. The months of July, August, and September are the period of the "big rains," and these permit the maturation of long-season (90 day) crops at altitudes of about 1700 m and higher. The long dry season, from early October to March or April, may total as little as 50 mm of rain at 2000 m elevation. The dry season in this area of Ethiopia is made more severe by the nature of the prevailing winds. From late September or early October to March the winds are usually quite strong out of the north and northeast. They are part of a large monsoon system bordering the Indian Ocean and sweep across parts of Arabia before moving southward over Ethiopia and Somalia. These winds are thus very dry. The ameliorating effects of higher elevations are pronounced and produce several easily distinguished vegetation types or life zones. The vegetation types in this area range from subdesert grasslands at 700 to 1000 m altitude to elfin forest and subalpine formations above 3000 m. The boundaries of these

During a 20-month study that spanned a severe drought and forest fire in a Bornean rain forest, a group of long-tailed macaques adapted successfully in the short term to extreme habitat damage by shifting to less preferred food items and exploiting an insect outbreak while fruits and flowers were scarce. Charred and desiccated fruits were major food sources just before the insect outbreak. The group dispersed widely to forage and increased its terrestrial travel. Riverine refuging became more erratic late in the drought and after the fire. The continued presence of sizable populations of this and other diurnal primate species, relative to those of some other frugivores with fixed home ranges and more restricted diets, suggests that dietary breadth is an important factor mediating frugivore responses to habitat degradation. FROM JUNE 1982 THROUGH MAY 1983, the El Ninio Southern Oscillation caused drought and forest fires in East Kalimantan, Indonesia (eastern Borneo), severely damaging approximately 3 million ha of evergreen forest, including large tracts of swamp forest, peat forest, and both selectively logged and undisturbed lowland rain forest (Anonymous 1984, Leighton & Wirawan, in press). My field study of the behavioral ecology of long-tailed macaque monkeys (Macacafascicularis) in the Kutai National Park, conducted from January 1982 through September 1983, spanned the entire drought, fire, and a 5month post-fire period, enabling me to record the responses of one group of macaques to drastic ecological change. In this paper I describe some of the monkeys' feeding and ranging responses. Leighton and Wirawan (in press) provide a general account of the effects of drought and fire on Kutai flora and fauna.

Sweep samples of insects at 200, 1600, 3550, and 3600 m elevation in secondary vegetation in the Venezuelan Andes near Merida show that the greatest number of insect species and dry weight occurs at the intermediate elevation, that the species richness of Diptera and parasitic Hymenoptera is not as proportionately reduced by elevation as is the species richness of other insect groups, that at higher elevations there are a reduced number of species and an increasingly unequal distribution of the individuals among the species, and that large insects are much less abundant at high than low elevations. These results confirm trends first indicated in an earlier elevational transect of Costa Rican insects. WE REPORT the results of sweep sampling the insect community at 3600 m, 3550 m, 1600 m, and 200 m elevation in the Venezuelan Andes near Merida, and at 3380 m in Costa Rica. The study is a continuation of the examination of large scale heterogeneity in tropical insect community structure that was initiated by Janzen (1973a,b). Here, we ask how sweep samples of low secondary vegetation differ at elevations characterized by paramo, mid-elevation evergreen forest which is generally converted to coffee and sugar cane plantations in Venezuela, and lowland ("tierra caliente") evergreen forest, which is generally converted to cattle pasture and sugar cane in Venezuela.

The study of the seeds rejected during a year by a colony of Artibeus jamaicensis Leach in a cave located in Los Tuxttlas Region of Veracruz, Mexico, gave us information concerning the fruit species on which these bats feed, their degree of utilization and bats' preference for some fruits. Information was also obtained on some peculiarities in the reproductive biology of plants, for example, the length of fruiting seasons and the possible role played by bats in the interchange of seeds between primary and secondary growth vegetation in a tropical humid evergreen forest area. FRUGIVOROUS BATS play an important role in the dispersal of several tropical plant species whose

Semi-evergreen forests cover the dry zone plains of Sri Lanka and constitute four-fifths of the island's vegetation. In a sample area of 3 kM2, in the Polonnaruwa Sanctuary, 63 tree species were found; 46 were characteristic of other dry zone forests, and 17 occurred only under special edaphic and biotic conditions. As in other dry zone forests, Drypetes sepiara (Euphorbiaceae) prevailed with a relative density of 21.3 percent. The Shannon index of diversity was 4.23 bits per individual, of which 79.4 percent was attributable to evenness; most species had few individuals. Dominance was shared between species typical of the subcanopy and canopy. Measures of diversity and of dominance between species placed the semi-evergreen community in Polonnaruwa as a type intermediate between tropical rainforest and deciduous monsoon forest. Measures of diameters (DBH) and estimates of the height of trees indicated that all species with typically very large trees had few individuals that were distributed more or less evenly through all the size classes. Typically smaller species had large numerical representation. These facts are discussed in light of dominance relationships and regenerative patterns in the community, and are related to possible evolutionary trends. The distribution of most species was clumped, but that of certain rare species was random. Clumping at 2,500 m2 plot size usually meant clumping at smaller-sized plots. Clumping on a large scale reflected local differences in species dominance. Five shrub associations were distinguished by density and constitute dominant and co-dominant species. Differences were related to the amount of light penetrating through the tree canopy and to edaphic factors. Glycosmis pentaphylla (Rutaceae) was the dominant shrub species in the climax association which flourished under a fairly closed tree canopy and lacked an herbaceous layer. ALTHOUGH EVERGREEN RAINFORESTS and monsoon forests of tropical Asia have received considerable attention from eaologists (e.g., Richards 1962, Ashton 1964, Ogawa et al. 1961, 1965), the semi-evergreen forests have been less intensely studied. In Sri Lanka the forests have been surveyed primarily from an economic perspective (e.g., Andrews 1961). This paper aims first, quantitatively to describe some structural features of a community of semi-evergreen forest in Sri Lanka; second, to examine the dominance relationships between species in this community. These relationships are briefly related to possible successional patterns in the dry zone of Sri Lanka, and to possible evolutionary trends among plants in general. The semi-evergreen forest has been described by Walter (1971: 209) as one in which the "upper tree layer is defoliated for some of the year [during the droughti,' while the lower tree layer retains its foliage." According to certain structural features, annual rainfall and drought duration, it is a type intermediate between tropical evergreen rainforest and deciduous monsoon forest. The tropical island of Sri Lanka has a gradation of forests from lowland and montane rainforests to lowland scrub forests and savanna (fig. 1). The majority of the forests may be roughly classed as semievergreen, however. These cover the most extensive physiographic feature of Sri Lanka, the lowland plains in the dry zone. Several dassificatory schemes have been proposed for the vegetation of Sri Lanka (e.g., Koelmeyer 1958, Gaussen et al. 1964, Fernando 1968). Mueller-Dombois (1968) evaluates these schemes and offers an up-to-date reclassification according to long-term climatic information. The community of forest described herein is located within the Polonnaruwa Sanctuary in the northwestern dry zone at 070 56' N and 81? 00' E. Polonnaruwa was the capital of Sri Lanka for varying periods between the 8th and 13th centuries. It was completely abandoned by the end of the 13th century, and natural vegetation reclaimed the land for the next 700 years. The area was resettled only at the beginning of the 20th century. Most of the natural vegetation has been maintained within the Sanctuary, although the shrub layer has been removed in areas immediately surrounding archaeological sites. The cutting of trees is prohibited, but poaching, particularly of young trees of construction-pole size, is not completely prevented. Parts of the study area are completely devoid of young trees and shrubs. These areas have been largely omitted in the present sample. The study site encloses approximately 3 km2, is bounded by water and cultivation, and is continuous with more extensive expanses of forest by a narrow

Studies of vegetation on 10 islands in the Galipagos demonstrate that the pattern of distribution of various vegetation types, ranging from desert scrub to evergreen forest, is closely associated with the pattern of variation in the local climatic conditions. The variation in average leaf size of the plant communities and of different layers in the communities is on the one hand associated with variations in altitude and exposure, and on the other hand with variations in the structure of the communities. The average leaf size of a plant community is considered to be a useful diagnostic feature for classifying the vegetation types. Structural features of the plant communities, for example, stratification and coverage, influence the leaf size of lower layers within the same community: the variation in leaf size of the shrub and of the herb layer is associated with the variation in coverage of the layer above. Variation in leaf size in the Acalypha complex and in Croton scouleri is suggested to be related to the structural variations in the communities: the structure of the plant communities may impose a selection pressure on the species of Acalypha and Croton, whereby they tend to develop the leaf size optimally fitted for the requirements of their particular environment.

A cyclone in November 1978 caused extensive damage to a site of natural dry evergreen forest in Sri Lanka. The cyclone destroyed more than 50 percent of the woody vegetation that had produced most of the food for two species of leaf-eating monkeys or langurs. Apparently, this caused an imbalance between these langur populations and their natural food supply and resulted in overbrowsing on those feeding trees which were not destroyed by the cyclone. Preferentially browsed tree species that were relatively rare and/or small in size died at significantly greater rates due to overbrowsing than those which were buffered against overbrowsing by virtue of being large in tree size and/or relatively abundant in the forest. The virtual disappearance of three overbrowsed tree species from the forest suggests that langurs may contribute to the change in floristic diversity in cyclone-disturbed areas. However, such an effect of langur folivory is thought to be short-lived and specific to this kind of rare disastrous environmental situation. ON NOVEMBER 23, 1978, A CYCLONE SWEPT across the island of Sri Lanka in the Indian Ocean causing extensive damage. Studies concerning the forest and its primate inhabitants were conducted before and after the cyclone at a site, Polonnaruwa, which lay directly in its path. This provided a unique opportunity to examine the effects of such a natural disaster on the relationship between the primates and their natural forest habitat. Earlier (Dittus 1985) I described the cyclone damage to the forest at Polonnaruwa and considered the influence of recurrent cyclones on the dry evergreen forest of Sri Lanka in general. At Polonnaruwa one immediate effect of the cyclone was the destruction of much of the food supply for two folivorous monkeys, the grey langur Presbytis entellus and the purple faced langur Presbytis senex (Figs. 1 and 2). My aim in this paper is to examine the effect that browsing by these langurs had on the survivorship of several of their feeding trees within 42 months after the cyclone. The focus is on the trees rather than on the langurs during what probably represents a transient phase of imbalance between the langur populations and their feeding trees. The examination is of interest on two counts. First, it clarifies the nature of the relationship between the langur populations and their feeding trees at a time of sudden drastic environmental change. Second, as overbrowsing by langurs contributed to the cyclone damage and consequent death of several of their feeding trees, it defines langur browsing as an additional agent influencing the floristic diversity of dry evergreen forest of Sri Lanka in areas that have recently undergone cyclone damage. The meteorological details of the cyclone, its path, and the kind and degree of damage it inflicted on the forest at Polonnaruwa were described earlier (Dittus 1985). Cyclone damage involved the uprooting of trees (tree falls), bre kage of trunks and large branches (crown damage), and extreme defoliation and loss of twigs particularly in t e upper layers of the forest. In addition, many trees that survived the immediate effects of the cyclone died within 42 months thereafter. Some of this post-cyclone tree mortality could be attributed to extensive crown damage; thus trees which lost 40 percent or more of their branches and trunks died at significantly greater rates after the cyclone than trees with lesser crown damage. But, not all post-cyclone tree mortality could be thus explained.

Confidence intervals for the estimates of leaf biomass and twig/wood biomass in a tropical rain forest were calculated using data from a complete biomass inventory. The influence of sample size and calculation method (mean tree technique or regression technique) on the accuracy and precision of the biomass estimates is discerned. A stratified random sample of 10-15 percent of all trees, selected on the basis of D2H, will yield confidence intervals (P = 0.05) of 7 5-125 percent and 90-110 percent for the estimate of leaf biomass or twig/wood biomass per unit area, respectively. FOR THE STUDY OF THE FUNCTIONAL ASPECTS OF FORESTS (primary productivity, nutrient cycling, energy flow), the quantification of aboveground organic matter (biomass) is of paramount importance. On a more practical level, the assessment of the effect of human interferences, such as harvesting, on a forest also requires the quantification of biomass. Two methods are commonly used to estimate the tree biomass of forest stands: the "mean tree technique" calculates forest biomass from trees with average dimensions, and the "regression technique" estimates forest biomass on the basis of regressions between tree weight and tree dimensions. Although both methods have been widely applied (e.g., IUFRO Biomass Studies 1974), the accuracy and precision of most biomass estimates remains unknown, and only a few authors present error terms for such estimates (Madgwick 1970, Tanner 1980, Hase and Foelster 1982, Lescure et al. 1983). The biomass of forests rich in species and/or with a complex structure, like many tropical forests, should be quantified with particular attention to error terms. The purpose of this paper is to present information about the accuracy and precision with which the aboveground biomass of such a highly heterogeneous forest stand can be calculated from a sample. We emphasize that this study estimates fresh biomass only (i.e., green weight), and the associated error terms should not be ignored at subsequent stages of the inventory of forests (dry weight biomass, productivity, bioelement stores, and so forth).