Alien plants can disrupt interactions between native plants and frugivores, alter frugivory and seed dispersal, and reshape plant community structure and dynamics. Their effects can range from impeditive, outcompeting native species for frugivores and seed dispersal, to facilitative, increasing consumption (neighborhood effects) and dispersal (contagious dispersal) of coincidentally fruiting native species. We investigated such effects in the case of Coffea canephora (robusta coffee): a fleshy-fruited mammal-dispersed alien species in India’s Western Ghats mountains. We compared seasonal trail camera-based encounter rates (Relative Abundance Index: RAI) of wild civets and macaques (known frugivores of coffee), seasonal civet scat encounter rates and composition, and juvenile tree communities between a 40-year-old discontinued shade-coffee agroforest (DC) with remnant coffee bushes and a mature rainforest (RF). Frugivore RAIs and scat encounter rates were similar or lower in DC compared to RF and did not differ between the coffee fruiting and non-fruiting seasons, offering no evidence that coffee attracts frugivores into DC. However, civet scats in DC were dominated by coffee seeds while also comprising 1.5-2.0 times more native seeds and species than RF scats (which predominantly comprised invertebrates) during the coffee fruiting season. Correspondingly, mammal-dispersed species including coffee and a few coincidentally-fruiting native species dominated juvenile tree communities in DC. Our results suggest that altered mammal frugivory can facilitate the persistence and spread of alien coffee and select native species, highlighting potential roles for coffee removal and enrichment with underrepresented species and dispersal guilds in ecological restoration strategies in forests recovering following discontinuation of coffee cultivation.

Forest restoration success depends crucially on the reinitiation of ecological processes such as seed arrival that drive natural regeneration. We know little about whether, by increasing and diversifying local seed sources to alleviate seed limitation, and attracting animal frugivores to alleviate dispersal limitation, restoration could shift seed arrival rates and species composition in the direction of mature rainforests, particularly for large-seeded species in degraded fragments. We tested this by examining seed arrival over a year in 189 1-m2 seed fall traps within 63 seed trapping stations across sites ecologically restored (through maximum diversity mixed-native species planting 13-21 years ago) and those left unrestored within degraded rainforests fragments, and in mature 'benchmark' rainforests, in India's Western Ghats. We found that overall, large-seeded species arrival rates in restored sites was comparable with benchmark, and was higher than unrestored sites. Seeds arriving through dispersal (from fruiting trees absent in the overstory) also showed a recovery towards benchmark levels, but were dominated by alien species abundant in the matrix habitat. Restoration also had no effect on species richness and composition of all arriving seeds, with restored sites closely resembling unrestored sites and remaining distinct from benchmark sites. Further, restoration did not alleviate the effect of fragment isolation on large seed arrival rates, which declined with distance from benchmark forests similarly in restored and unrestored sites. Our results suggest that restoration can increase local availability and dispersal of large seeds, and highlights the value of diverse native species restoration plantings within degraded forests and promoting their use instead of alien tree species in the surrounding matrix. However, our findings also show that recovery of natural regeneration can remain slow or uncertain in restored forest fragments, and underscore the need for sustained monitoring and evaluation of restoration over multiple decades.

Restoration of degraded tropical forests is often impeded by invasive species. For regenerating native seedlings, the presence of invasives in the neighbourhood can alter insect herbivory patterns, ultimately shaping restoration trajectories; however, such indirect effects are rarely examined. Here, we investigated the effect of robusta coffee ( Coffea canephora ) - a shade-tolerant invasive species under closed-canopy secondary forests in the Western Ghats – on the incidence of herbivory (proportion of leaves with any sign of damage) and the extent of leaf damage (percentage leaf area consumed) in seedlings of 10 rainforest species in plots from which coffee plants were either weeded out or left intact. We further examined whether local neighbourhood densities of coffee and other saplings, and species’ leaf traits, explained herbivory patterns. Removal of invasive coffee did not influence herbivory incidence or the extent of damage across our focal species. However, the incidence of herbivory declined with increasing neighbourhood plant density, suggesting that neighbourhood plants provide a resource dilution effect. Both incidence and extent of herbivore damage were strongly species-specific and partly explained by leaf traits: greater leaf carbon content was correlated with lower herbivory incidence. Contrary to expectations, plants with resource-acquisitive traits (high leaf nitrogen and specific leaf area) experienced lower incidence of herbivory and extent of damage. Our findings suggest that in this system, the indirect effects via herbivores are perhaps not as important in influencing restoration as the more direct effects of invasive species, such as competition for resources. ### Competing Interest Statement The authors have declared no competing interest. National Centre for Biological Sciences, https://ror.org/03gf8rp76, Project Identification No.RTI 4006 AMM Murugappa Chettiar Research Centre Hindustan Unilever Ltd. Department of Science and Technology, Ministry of Science and Technology, India R. M. Tulpule Charitable Trust Rohini Nilekani Philanthropies

1) DESCRIPTION The dataset contains annual woody stems (shrubs and trees) census data collected from two long-term ecological monitoring plots spanning one hectare each in the Anamalai Hills of the Southern Western Ghats, India. These two plots represent one situated in a mature forest located within relatively undisturbed rainforest of the Anamalai Tiger Reserve (ATR) and one in secondary forest on the Valparai Plateau, respectively. Both plots have been censused and measured from 2017 to 2022 following the standardized protocol (RAINFOR-GEM, Marthews et al. 2014). 2) CONTACTS CONTACT #11. Name: Akhil Murali2. Work Address: Nature Conservation Foundation, 1311, 12th A Main, Vijayanagar 1st Stage, Mysuru 570017, Karnataka, India3. Work Phone: +91 82812 974414. Email address: akhil@ncf-india.org5. ORCID: 0000-0001-6149-6458 CONTACT #21. Name: Srinivasan Kasinathan2. Work Address: Nature Conservation Foundation, 1311, 12th A Main, Vijayanagar 1st Stage, Mysuru 570017, Karnataka, India3. Work Phone: +91 821 25156014. Email address: srini@ncf-india.org5. ORCID: 0000-0001-7323-6653 CONTACT #31. Name: Kshama Bhat2. Work Address: Nature Conservation Foundation, 1311, 12th A Main, Vijayanagar 1st Stage, Mysuru 570017, Karnataka, India3. Work Phone: +91 821 25156014. Email address: kshama@ncf-india.org5. ORCID: 000-0002-6190-2687 CONTACT #4 1. Name: Jayashree Ratnam 2. Work Address: National Centre for Biological Sciences, TIFR, Bellary Road, Bengaluru 560065, Karnataka, India3. Work Phone: +91 80 23666001 4. Email address: jratnam@ncbs.res.in 5. ORCID: 0000-0002-6568-8374 CONTACT #51. Name: Mahesh Sankaran 2. Work Address: National Centre for Biological Sciences, TIFR, Bellary Road, Bengaluru 560065, Karnataka, India3. Work Phone: +91 80 236660014. Email address: mahesh@ncbs.res.in 5. ORCID: 0000-0002-1661-6542 CONTACT #61. Name: Divya Mudappa2. Work Address: Nature Conservation Foundation, 1311, 12th A Main, Vijayanagar 1st Stage, Mysuru 570017, Karnataka, India3. Work Phone: +91 821 25156014. Email address: divya@ncf-india.org5. ORCID: 0000-0001-9708-4826 CONTACT #71. Name: T. R. Shankar Raman2. Work Address: Nature Conservation Foundation, 1311, 12th A Main, Vijayanagar 1st Stage, Mysuru 570017, Karnataka, India3. Work Phone: +91 821 25156014. Email address: trsr@ncf-india.org5. ORCID: 0000-0002-1347-3953 CONTACT #81. Name: Anand M Osuri 2. Work Address: Nature Conservation Foundation, 1311, 12th A Main, Vijayanagar 1st Stage, Mysuru 570017, Karnataka, India3. Work Phone: +91 821 25156014. Email address: aosuri@ncf-india.org 5. ORCID: 0000-0001-9909-5633 3) GEOGRAPHIC COVERAGE and SITE DESCRIPTION a) Site type: : Tropical Forestb) Geography: : Anamalai Tiger Reserve, Southern Western Ghats.c) Habit: : Mid elevation Wet evergreen Forestd) Site History: : i) MANAMBOLI- The Mature Forest plot (10.357748° N, 76.889747° E; 825 m asl) is situated within a relatively undisturbed 200-hectare mid-elevation tropical wet evergreen rainforest tract at the core of the Anamalai Tiger Reserve (ATR). This area has been protected from logging and other significant disturbances since its establishment as a protected area in 1979. ii) CANDURA- The Secondary Forest plot (10.30855411° N, 76.83391853° E; 875 m asl) is situated within a 124-hectare rainforest remnant on the Valparai Plateau: the Candura rainforest remnant. The Candura site experienced episodic selective logging in the 1990s and early 2000s, with the last logging episode occurring in 2004. In the early 2000s, the understorey of the remnant was cleared for Vanilla (Vanilla planifolia) cultivation in the central and southern parts (abandoned in 2007), robusta coffee (Coffea canephora) in the northwestern corner (abandoned in the early 2000s), and pepper in 21 hectares in the northeastern part (established in 2015, abandoned in 2021). Climate: Humid tropical with about 2400 mm rainfall annually, falling mainly during the southwest monsoon. 4) TEMPORAL COVERAGE a) Begins: 2017-11-30 (Year, Month, Day)b) Ends: 2022-11-12 (Year, Month, Day) 5) SAMPLING DESIGN AND METHODS a) Plot Design: Each 1 ha plot of 100 m × 100 m, sub-divided into 100 continuous sub-plots of 10 m × 10 m, was surveyed and mapped to maximum accuracy using a theodolite in the field, with grid corners permanently staked. b) Data collection period and frequency: After the plot establishment in NOvember -- December 2017, the plots were recensused each year (around November). c) Research Methods: All woody plant individuals with girth at breast height (GBH, at 1.3 m) ≥10 cm were tagged with numbered aluminum tags and spatially mapped. Plant species were identified using standard floral keys. Stem GBH was measured for all single stemmed individuals. For trees with buttresses, the GBH point of measurement (POM) was taken at 50 cm above the buttresses or at the height where the stem is regular. New saplings that recruited into the ≥10 cm GBH class were identified, mapped, tagged, and added to the monitoring. Stems that appeared to be dead were recorded at each monitoring and those that showed no signs of recovery in subsequent visits were recorded as mortality. 6) FILES INCLUDED The dataset includes the following 9 files, whose details and contents are explained below. (Wherever used in the various files, NA implies not available.) 01_README.txtMetadata (this file) including information on the dataset explaining associated files and their contents. 02_Candura_annual_census.csv This contains the Annual census data with the following column headings: site: Site name (Can = Candura)cno: Census Number (1 = 2017, 2 = 2018..., 6 = 2022)ymd: Date in DD-MM-YYYY format (Day Month Year)gno: Grid Numbertno: Unique tag number for the plantpno: Pole Number (unique alphabetic code for each stem of multi-stemmed individuals)sps: Species name as codeslx: X coordinate of tree in the 10 m × 10 m subplot (in metres)ly: Y coordinate of tree in the 10 m × 10 m subplot (in metres)ht1: Point of measurement at 1.3 m above the ground or 50 cm above the top of the highest buttress or stilt root (POM1)c1: Alive status of the stem at the POM1 (coded according Marthews et al. 2014, page: 97)g1: Stem girth at POM1 (in centimetre)ht2: 20 cm above the ht1 or point of measurement 2 (POM2) recording girth at which the dendroband is attachedc2: Alive status of the stem at the POM2 (coded acording Marthews et al. 2014, page: 97)g2: Girth at POM2 (in centimetre)dyn: whether the dendroband is attached to the tree or not (y-Yes, n-No)da: alive status of stem (d-dead,a-alive)remarks: remarks or notes 03_Manamboly_annual_census.csvThis contains Annual census data with the following column headings: site: Site name (Man = Manamboli)cno: Census Number (1 = 2017, 2 = 2018..., 6 = 2022)ymd: Date in DD-MM-YYYY format (Day Month Year)gno: Grid Numbertno: Unique tag number for the plantpno: Pole Number (unique alphabetic code for each stem of multi-stemmed individuals)sps: Species name as codeslx: X coordinate of tree in the 10 m × 10 m subplot (in metres)ly: Y coordinate of tree in the 10 m × 10 m subplot (in metres)ht1: Point of measurement at 1.3 m above the ground or 50 cm above the top of the highest buttress or stilt root (POM1)c1: Alive status of the stem at the POM1 (coded according Marthews et al. 2014, page: 97)g1: Stem girth at POM1 (in centimetre)ht2: 20 cm above the ht1 or point of measurement 2 (POM2) recording girth at which the dendroband is attachedc2: Alive status of the stem at the POM2 (coded acording Marthews et al. 2014, page: 97)g2: Girth at POM2 (in centimetre)dyn: whether the dendroband is attached to the tree or not (y-Yes, n-No)da: alive status of stem (d-dead,a-alive)remarks: remarks or notes 04_Candura_vernier.csvThis file has the girth measurement of trees with lianas where digital vernier calipers were used to measure stem diameter since it was not possible to measure stem girth using measuring tape.site: Site name (Can = Candura)cno: Census Numberymd: Date in DD-MM-YYYY format (Day Month Year)gno: Grid Numbertno: Unique tag number for the plantpno: Pole Number (unique alphabetic code for each stem of multi-stemmed individuals)sps: Species name as codesvern1_d1 First measure of diameter at POM1 (in millimetre) vern2_d1 Second measure of diameter at POM1 (in millimetre) vern3_d1 Third measure of diameter at POM1 (in millimetre) calc_g1: Girth at POM1 (in centimetre; calculated using the averaged value as diameter from the three measurements)vern1_d2 First measure of diameter at POM2 (in millimetre) vern2_d2 Second measure of diameter at POM2 (in millimetre) vern3_d2 Third measure of diameter at POM2 (in millimetre) calc_g2 Girth at POM2 (in centimetre; calculated using the averaged value as diameter from the three measurements)Remarks Remarks and notes 05_Manamboli_vernier.csvThis file has the girth measurement of trees with lianas where digital vernier calipers were used to measure stem diameter since it was not possible to measure stem girth using measuring tape.site: Site name (Man = Manamboli)cno: Census Numberymd: Date in DD-MM-YYYY format (Day Month Year)gno: Grid Numbertno: Unique tag number for the plantpno: Pole Number (unique alphabetic code for each stem of multi-stemmed individuals)sps: Species name as codesvern1_d1 First measure of diameter at POM1 (in millimetre) vern2_d1 Second measure of diameter at POM1 (in millimetre) vern3_d1 Third measure of diameter at POM1 (in millimetre) calc_g1: Girth at POM1 (in centimetre; calculated using the averaged value as diameter from the three measurements)vern1_d2 First measure of diameter at POM2 (in millimetre) vern2_d2 Second measure of diameter at POM2 (in millimetre) vern3_d2 Third measure of diameter at POM2 (in millimetre) calc_g2 Girth at POM2 (in centimetre; calculated using the averaged value as diameter from the three measurements)Remarks Remarks and notes 06_Candura_Height_data.csvThis contains data on the heights of individual trees in plot as measured in 2018.site: Site name (Can = Candura)ymd: Date in YYYY/MM/DD format (Year Month Day)gno: Grid Number: tno: unique tag number: pno: Pole Number (unique alphabetic code for each stem of multi-stemmed individuals)sps: Species name as codes: lx: X coordinate of tree in the 10 m × 10 m subplot (in metres)ly: Y coordinate of tree in the 10 m × 10 m subplot (in metres)height: Height of tree in metresremarks: Remarks: and notes 07_Manamboli_Height_data.csvThis contains data on the heights of individual trees in plot as measured in 2018.site: Site name (Man = Manamboli)ymd: Date in YYYY-MM-DD format (Year Month Day)gno: Grid Number: tno: unique tag number: pno: Pole Number (unique alphabetic code for each stem of multi-stemmed individuals)sps: Species name as codes: lx: X coordinate of tree in the 10 m × 10 m subplot (in metres)ly: Y coordinate of tree in the 10 m × 10 m subplot (in metres)height: Height of tree in metresremarks: Remarks: and notes 08_Species_name_match.csvThis file provides the combined list of species codes updated taxonomy and successional guild. Scientific names were updated to current taxonomy using the species name matching tool of the Global Biodiversity Information Facility, GBIF (www.gbif.org).sps: Species codesquery : Scientific name of the plant at the time of data collection: scientificName: : with auther citation: key: GBIF keyrank: Taxonomic rank or level of identification (GENUS, SPECIES)kingdom: Taxonomic Kingdom (plants) provided by GBIF name matching tool: phylum: Taxonomic Phylum provided by GBIF name matching toolclass: Taxonomic Class provided by GBIF name matching toolorder: Taxonomic Order provided by GBIF name matching toolfamily: Taxonomic Family provided by GBIF name matching toolgenus: Taxonomic Genus provided by GBIF name matching toolbotanical_name: Updated scientific name of the species provided by GBIF name matching toolhabt_new: Successional guild of the species (Mature = mature forest species; Secondary = secondary successional species; Int - Introduced species) 09_R_scrpt_for_manuscript.RText file with code in the R statistical and programming environment (www.r-project.org). ReferenceMarthews TR, Riutta T, Oliveras Menor I, Urrutia R, Moore S, Metcalfe D, Malhi Y, Phillips O, Huaraca Huasco W, Ruiz Jaén M, Girardin C, Butt N, Cain R and colleagues from the RAINFOR and GEM networks (2014). Measuring Tropical Forest Carbon Allocation and Cycling: A RAINFOR-GEM Field Manual for Intensive Census Plots (v3.0). Manual, Global Ecosystems Monitoring network, http: //gem.tropicalforests.ox.ac.uk/.

Tropical rainforest remnants in human-modified landscapes exhibit varying levels of degradation, from highly degraded open-canopied and invasive plant-invaded forests to closed-canopy forests that appear structurally intact. The former are frequently identified as being in a state of arrested recovery, and targeted for restoration, but restoration needs and opportunities in the latter remain underexplored. Using tree and seedling data from 105 plots spanning a canopy cover gradient in rainforest fragments and 19 plots in relatively-intact "reference" rainforests in India's Western Ghats mountains, we show that the floristic composition, conservation significance, and carbon stocks of closed-canopy fragments (CC) more closely resemble open-canopy fragments (OC) than reference rainforests (RR). For example, densities of old-growth forest species, endemic/threatened species, and carbon stocks, increased from 15 %, 28 %, and 22 % of reference values in OC to 32 %, 46 %, and 35 % in CC, respectively, while tree community similarity to RR showed no increase from OC to CC (15 %). Seedlings mirrored this pattern, offering little indication of natural recovery in closed-canopy fragments. Further, we show using simulations that seedling enrichment in closed-canopy fragments can initiate varying levels of floristic and functional recovery towards reference targets. Our findings illustrate that seemingly structurally-intact tropical rainforest fragments can remain arrested in a floristically degraded condition, and represent worthwhile targets for ecological restoration. Such forests expand opportunities for restoring conservation-priority and high carbonstoring species using relatively low-cost methods (e.g., enrichment planting), which can complement intensive restoration of highly degraded forests and minimally-assisted natural recovery of less-fragmented forests.

Evidence for climate-change related alteration in distributions and ranges of forest trees is accumulating, but information from Asian tropical forests, particularly for threatened and endemic species, remains limited. Here, we examine landscape-level distribution-abundance patterns of 11 endemic and threatened tropical rainforest tree species in the Anamalai Hills and model their distribution and responses to future climate scenarios in the Southern Western Ghats (SWG), India. Six of the primarily low- and mid-elevation species were more abundant in protected reserves than forest fragments. Occurrence data from the Anamalais and SWG (N = 3004, range: 41-706 per species) were used to model distributions using maximum entropy (maxent) species distribution modelling in R. Maxent model performances indicated excellent fits for 9 of 11 species (AUC>0.90) with precipitation and temperature variables showing higher permutation importance. There was high interspecies variability in range size (197-12,221 km2) and niche width (0.04-0.50). Models of distribution under future climate in 2061-2080 predict range reductions in six species (including near-extinction for two species), increases for three species, and no substantial change for two species. Predicted southward and westward shifts in ranges and persistence in parts of the SWG indicate the importance of identifying and conserving micro-climatic zones or refugia to ensure the persistence of tree species under anticipated climate change.

This dataset contains compiled Fruit, seed dispersal, and life history traits of tropical rainforest trees of the Anamalai Hills, Western Ghats, India. The list of species included are from the following two related publications: - Muthuramkumar, S., Ayyappan, N., Parthasarathy, N., Mudappa, D., Raman, T.R.S., Selwyn, M.A. and Pragasan, L.A. (2006), Plant Community Structure in Tropical Rain Forest Fragments of the Western Ghats, India. Biotropica, 38: 143-160. https://doi.org/10.1111/j.1744-7429.2006.00118.x - Osuri, A., Chakravarthy, D., Mudappa, D., Raman, T., Ayyappan, N., Muthuramkumar, S., & Parthasarathy, N. (2017). Successional status, seed dispersal mode and overstorey species influence tree regeneration in tropical rain-forest fragments in Western Ghats, India. Journal of Tropical Ecology, 33(4), 270-284. doi:10.1017/S0266467417000219 The present dataset is an expanded and updated version of the related dataset available at https://doi.org/10.5061/dryad.vd0nn Species traits information was collated from BIOTIK (http://www.biotik.org/), Flowers of India (http://www.flowersofindia.net/), India Biodiversity Portal (http://indiabiodiversity.org/), Global wood density database (https://doi.org/10.5061/dryad.234/1) and Osuri et al. (2014): https://doi.org/10.1017/S0266467417000219. We also referred to the following previous studies that provided information on the successional status of rain-forest species in the Western Ghats (Chetana 2013, Pascal 1988, Raman et al. 2009, Sreejith 2005). References: CHETANA, H. C. 2013. Assessing the ecological processes in abandoned tea plantations and its implication for ecological restoration in the Western Ghats, India. PhD thesis, Manipal University. OSURI, A. M., KUMAR, V. S. & SANKARAN, M. 2014. Altered stand structure and tree allometry reduce carbon storage in evergreen forest fragments in India’s Western Ghats. Forest Ecology and Management 329: 375–383. PASCAL, J. P. 1988. Wet evergreen forests of the Western Ghats of India: Ecology, structure, floristic composition and succession. Institut Français de Pondichéry, Pondicherry. RAMAN, T. R. S., MUDAPPA, D. & KAPOOR, V. 2009. Restoring rainforest fragments: survival of mixed-native species seedlings under contrasting site conditions in the Western Ghats, India. Restoration Ecology 17:137–147. SREEJITH, K. A. 2005. Ecological and ecophysiological studies on the successional status of tree seedlings in tropical wet evergreen and semi-evergreen forests of Kerala. PhD thesis, Forest Research Institute, Dehradun. Geographic Coverage: 1. Location/Study Area: Valparai Plateau, Tamil Nadu, India; Anamalai Tiger Reserve, Tamil Nadu, India 2. GPS coordinates: Valparai Plateau (10°15'- 10°22'N, 76°52' - 76°59'E); Anamalai Tiger Reserve (10°12' - 10°35'N, 76°49' - 77°24'E) Temporal Coverage: 1. Begins: 2003-03-01 (Year, Month, Day) 2. Ends: 2022-08-16 (Year, Month, Day) Besides the README.txt file, the dataset includes two comma-delimited text (csv) files with the data in columns as explained below: Anamalai_tree_traits.csv gbif_namematch.csv Anamalai_tree_traits.csv old_code: Species codes used at the time of data collection (manuscript Muthuramkumar et al. 2006: https://doi.org/10.1111/j.1744-7429.2006.00118.x) osuri_code: Revised species codes if used in the Osuri et al. 2017 paper in Journal of Tropical Ecology 33:270-284 (https://doi.org/10.1017/S0266467417000219) and related dataset (https://doi.org/10.5061/dryad.vd0nn) or else indicated as NA current_code: Species codes used at present in relation to current taxonomy and name as in ‘species’ column old_speciesName: Species names used at the time of data collection (Muthuramkumar et al. 2006: https://doi.org/10.1111/j.1744-7429.2006.00118.x) verbatimScientificName: Revised species names used to match names to current taxonomy (next column) species: Accepted species scientific name at present as determined by running the name in the previous column through the Global Biodiversity Information Facility lookup tool (GBIF: https://www.gbif.org/tools/species-lookup) family: Plant Family that the taxon belongs to genus: Genus the taxon belongs to fruit_type: Type of fruit fleshy_dry: Whether fruit is a dry fruit or fleshy, with aril or other parts seed_size: Species seed size: L = Large (>3 cm); M = Medium (1-3 cm); S = Small (<1 cm) disperser: Categories indicating seed dispersal mode: Bird, mammal, bird and mammal (Mammal_bird), gravity, wind, or unknown new_disp: Composite category based on disperser and seed size (taking seed size as S and L) habitat: Habitat affinity category: EG_edg - evergreen forest edge; EG_for - evergreen forest; Dec_for - deciduous forest; Int – Introduced species; Unknown – Unknown habt_new: Habitat affinity new category: Mature – mature forest; Secondary – secondary forest, Int - Introduced species; Unknown – unknown ad_ht: Species maximum adult height (m) wden: Species wood density (g cm^-3); NA - not available; sourced from Global wood density database (https://doi.org/10.5061/dryad.234/1) remarks: Notes or remarks gbif_namematch.csv verbatimScientificName: Scientific name as used for the GBIF lookup scientificName: Name returned by GBIF species lookup tool (https://www.gbif.org/tools/species-lookup) key: GBIF key returned by the species lookup tool matchType: Match type returned by the species lookup tool confidence: Confidence returned by the species lookup tool status: Indicates whether the verbatimScientificName is an Accepted name or a Synonym rank: Indicates whether the taxon was identified to Family, Genus or species kingdom: Kingdom of taxon phylum: Phylum of taxon class: Class of taxon order: Order of taxon family: Family of taxon genus: Genus of taxon species: Species name (accepted and current species name) canonicalName: Canonical form of species name returned by species lookup tool authorship: Author of taxon

Dipterocarpus bourdillonii, a Critically Endangered tree species endemic to the Western Ghats, India, has hitherto been reported mainly from the states of Kerala and Karnataka on the western slopes of the mountain range. In Tamil Nadu, this species has been reported to occur in two locations, but no population details have been documented and the species has neither been listed in state floras nor in a recent compendium of plant species. The present study documents the occurrence of a population of the species, with at least 40 individuals, in the Anamalai Tiger Reserve, Tamil Nadu, extends the known upper limit of its altitudinal range to 733 m, and suggests further surveys and in situ conservation efforts.

DESCRIPTION This dataset includes vegetation plot data on trees, lianas, understorey plants, and regeneration, and related data and species name matching files in five rainforest sites collected in 2003 as part of the following study: MUTHURAMKUMAR, S., AYYAPPAN, N., PARTHASARATHY, N., MUDAPPA, D., RAMAN, T. R. S., SELWYN, M. A. & PRAGASAN, L. A. 2006. Plant community structure in tropical rain forest fragments of the Western Ghats, India. Biotropica 38: 143–160. DOI: 10.1111/j.1744-7429.2006.00118.x The regeneration data were analysed and presented in the following publication and related dataset: OSURI, A. M., CHAKRAVARTHY, D., MUDAPPA, D., RAMAN, T. R. S., AYYAPPAN, N., MUTHURAMKUMAR, S. & PARTHASARATHY, N. 2017. Successional status, seed dispersal mode and overstorey species influence tree regeneration in tropical rain-forest fragments in Western Ghats, India. Journal of Tropical Ecology 33(4): 270-284. DOI: 10.1017/S0266467417000219 OSURI, A. M., CHAKRAVARTHY, D., MUDAPPA, D., RAMAN, T. R. S., AYYAPPAN, N., MUTHURAMKUMAR, S. & PARTHASARATHY, N. 2017. Data from: Successional status, seed dispersal mode and overstorey species influence tree regeneration in tropical rain-forest fragments in Western Ghats, India, Dryad, Dataset, https://doi.org/10.5061/dryad.vd0nn CONTACTS CONTACT #1 1. Name: T. R. Shankar Raman 2. Work Address: Nature Conservation Foundation, 1311, 12th A Main, Vijayanagar 1st Stage, Mysuru 570017, Karnataka, India 3. Work Phone: +91 821 2515601 4. Email address: trsr@ncf-india.org 5. ORCID: https://orcid.org/0000-0002-1347-3953 CONTACT #2 1. Name: Divya Mudappa 2. Work Address: Nature Conservation Foundation, 1311, 12th A Main, Vijayanagar 1st Stage, Mysuru 570017, Karnataka, India 3. Work Phone: +91 821 2515601 4. Email address: divya@ncf-india.org 5. ORCID: https://orcid.org/0000-0001-9708-4826 CONTACT #3 1. Name: Anand M. Osuri 2. Work Address: Nature Conservation Foundation, 1311, 12th A Main, Vijayanagar 1st Stage, Mysuru 570017, Karnataka, India 3. Work Phone: +91 821 2515601 4. Email address: aosuri@ncf-india.org 5. ORCID: https://orcid.org/0000-0001-9909-5633 CONTACT #4 1. Name: N. Ayyappan 2. Work Address: French Institute of Pondicherry, No. 11, Post Box No. 33, Saint Louis Street, Pondicherry – 605 001, India. 3. Work Phone: + 91- 413-2231616 4. Email address: ayyappan.n@ifpindia.org 5. ORCID: https://orcid.org/0000-0003-4383-557X CONTACT #5 1. Name: S. Muthuramkumar 2. Work Address: V.H.N.S.N. College, 3/151-1, College Road, Virudhunagar - 626001, Tamil Nadu, India. 3. Work Phone: + 91-4562-280154 4. Email address: muthuramkumar@vhnsnc.edu.in 5. ORCID: https://orcid.org/0000-0002-7791-8499 CONTACT #6 1. Name: N. Parthasarathy 2. Work Address: Department of Ecology and Environmental Sciences, Pondicherry University, R Venkat Raman Nagar, Kalapet, Pondicherry 605014, India 3. Work Phone: + 91-413-2654326 4. Email address: parthapu@yahoo.com 5. ORCID: https://orcid.org/0000-0002-4172-5441 KEYWORDS Anamalai hills; biodiversity hotspot; disturbance; endemics; fragmentation; lianas; plant conservation; tree diversity; tropical rain forest; understory plants. GEOGRAPHIC COVERAGE 1. Location/Study Area: Valparai Plateau, Tamil Nadu, India; Anamalai Tiger Reserve, Tamil Nadu, India 2. GPS coordinates: Valparai Plateau (10°15'- 10°22'N, 76°52' - 76°59'E); Anamalai Tiger Reserve (10°12' - 10°35'N, 76°49' - 77°24'E) TEMPORAL COVERAGE 1. Begins: 2003-03-01 (Year, Month, Day) 2. Ends: 2003-04-30 (Year, Month, Day) METHODS Methods involved systematic vegetation plots for trees, lianans and understorey plants as described in Muthuramkumar et al. 2006 (Biotropica 38: 143–160. DOI: 10.1111/j.1744-7429.2006.00118.x) and for tree and woody regeneration as described in Osuri et al. 2017 (Journal of Tropical Ecology 33(4): 270-284. DOI: 10.1017/S0266467417000219). The vegetation sampling methods are briefly described below. The present study was conducted in five tropical wet evergreen forest fragments located on the Valparai plateau (Fig. 1): Akkamalai (AK, 2600 ha), Upper Manamboli (UM, 100 ha), Lower Manamboli (LM, 100 ha), Tata Finlay (TF, 32 ha), and Injipara (IP, 18 ha). In each site, vegetation was sampled in randomly placed noncontiguous plots of 20 × 20 m located at least 50 m apart and at least 20 m into the fragment interior from the edges, major trails, or roads. We sampled 20 plots each in IP, TF, and LM, and 25 plots each in UM and AK. Within each plot, all trees ≥30cm girth at breast height (gbh, at 1.3 m; corresponding to DBH of 9.55 cm) and lianas ≥1 cm diameter at breast height (DBH) were identified to species, counted, and their girth/diameter measured. For multi-stemmed trees bole girths were measured separately, basal area calculated and summed. Each 20 x 20 m plot was divided into four 10 × 10 m quarters. For understory plants, 2 × 2 m quadrats were laid at the four corners of the 20 × 20 m plot (one in each of the corresponding four quarters) and all shrubs, undershrubs, herbs, ferns, and small twiners found within the quadrats were enumerated and identified. The regeneration sampling was done in a 5 × 5-m plot (0.0025 ha) placed at the outer corner of the first (south-west) quarter of the 20 x 20 m plot. Within each regeneration plot, we identified, counted and measured all tree saplings >1 cm diameter at breast height (dbh, at 1.3 m) and <9.55 cm dbh (equivalent to <30 cm girth at breast height, gbh). Woody shrubs of 1–9.55 cm dbh were alsorecorded in the regeneration plots (but these were excluded in the Osuri et al. 2017 analysis). For vegetatively propagating plants a clump of stems that is basally connected was considered as one individual. Canopy height was measured with a range finder and canopy closure was measured using a spherical densiometer. Vouchers were identified with regional flora and confirmed with the Western Ghats collections available in the herbarium of Salim Ali School of Ecology, Pondicherry University, from our previous works in the region. FILES INCLUDED Besides the 00_README.txt file that contains this metadata, the dataset includes the following 11 files, whose details and contents are explained below. 01_all_sites.csv Description: The file contains details of the five study sites (three continuous forest and two forest fragment sites). Note: Current Name of TF (Tata Finlay) site is Old Valparai, current name of Akkamalai (AK) is Iyerpadi-Akkamalai complex. Sites and codes correspond to the Muthuramkumar et al. 2006 paper (https://doi.org/10.1111/j.1744-7429.2006.00118.x). Column names and descriptions: eventDate: Date range when sampling was carried out in the sites old_sitename: Name of the site as used in the Muthuramkumar et al. (2006) paper (https://doi.org/10.1111/j.1744-7429.2006.00118.x) sitecode: Site code as used in the Muthuramkumar et al. (2006) paper (https://doi.org/10.1111/j.1744-7429.2006.00118.x) site: Site name as at present and used in this dataset decimalLatitude: latitude in decimal degrees North decimalLongitude: longitude in decimal degrees East geodeticDatum: Geodetic Datum WGS 84 coordinateUncertaintyInMeters: Uncertainty in metres of the GPS location (as only one location available for entire site where points were distributed) type: Indicates whether site was continuous rainforest or rainforest fragment Area_ha: Area in hectares Altitude_min_m: Minimum altitude in metres of sampled plots Altitude_max_m: Maximum altitude in metres of sampled plots Ownership: Whether site is in privately owned land or within state-protected reserve Average_canopy_height_m: average canopy height in metres Canopy_closure_%: estimated canopy closure in percentage Nearby_plantations: Adjoining plantations 02_all_trees_adult_data.csv Description: The file contains records of all adult trees >= 30 cm girth at breast height of 1.3 m (gbh) recorded within 20 m x 20 m plots across three continuous forests and two forest fragments. Note: Same as in the Osuri et al. (2017) dataset (https://doi.org/10.5061/dryad.vd0nn), with Tithonia diversifolia added back in Injipara and data from one additional site (Manamboli Lower) added back from the original dataset corresponding to the Muthuramkumar et al. 2006 paper (https://doi.org/10.1111/j.1744-7429.2006.00118.x). Column names and descriptions: x: Row index site: Name of forest site plot_no: An unique plot number assigned to each 20m x 20m adult tree plot within each site q_no: An unique number assigned to each of four 10m x 10m quarters within each adult plot t_no: An unique number assigned to each individual tree within each site. old_code: Species codes used at the time of data collection (refer to Appendix A of the main paper for full species names, and the 06_all_species_names.csv file with this dataset) osuri_code: Revised species codes used in the Osuri et al. 2017 paper in Journal of Tropical Ecology 33: 270-284 (https://doi.org/10.1017/S0266467417000219) and related dataset (https://doi.org/10.5061/dryad.vd0nn) current_code: Species codes used at present gbh_1 to gbh_16: Girth at breast height of single- (gbh_1) and multi-stemmed (gbh_2 – gbh_16) individuals, measured in centimetres (cm) P_ID: Unique plot ID created by combining columns site and plot_no 03_all_liana_data.csv Description: The file contains records of all lianas >= 1 cm diameter at breast height of 1.3 m (dbh) recorded within 20 m x 20 m plots across three continuous forests and two forest fragments. Note: Lianas were not included in the Osuri et al. (2017) dataset (https://doi.org/10.5061/dryad.vd0nn). Column names and descriptions: x: Row index site: Name of forest site plot_no: An unique plot number assigned to each 20m x 20m adult tree plot within each site q_no: An unique number assigned to each of four 10m x 10m quarters within each adult plot t_no: An unique number assigned to each individual tree within each site. old_code: Species codes used at the time of data collection (refer to Appendix A of the main paper for full species names, and the 06_all_species_names.csv file with this dataset) osuri_code: Indicated as NA since these data were not used in the Osuri et al. 2017 paper in Journal of Tropical Ecology 33: 270-284 (https://doi.org/10.1017/S0266467417000219) and related dataset (https://doi.org/10.5061/dryad.vd0nn) current_code: Species codes used at present dbh_1 to dbh_11: Diameter at breast height of single- (dbh_1) and multi-stemmed (dbh_2 – dbh_11) individuals, measured in centimetres (cm) P_ID: Unique plot ID created by combining columns site and plot_no 04_all_herbs_data.csv Description: The file contains records of all understorey plants (shrubs, undershrubs, herbs, ferns, and small twiners) recorded in 2 m × 2 m quadrats laid at the four corners of each 20 m × 20 m plot in three continuous forests and two forest fragments. Note: Understorey plants were not included in the Osuri et al. (2017) dataset (https: //doi.org/10.5061/dryad.vd0nn). Column names and descriptions: x: Row index site: Name of forest site plot_no: An unique plot number assigned to each 20m x 20m plot within each site corner_no: An unique number assigned to each of four 2 m x 2 m quadrat laid at the four corners of the 20 m x 20 m plot t_no: A number assigned to each individual species recorded within the corner plot. old_code: Species codes used at the time of data collection (refer to Appendix A of the main paper for full species names, and the 06_all_species_names.csv file with this dataset) osuri_code: Indicated as NA since these data were not used in the Osuri et al. 2017 paper in Journal of Tropical Ecology 33: 270-284 (https://doi.org/10.1017/S0266467417000219) and related dataset (https://doi.org/10.5061/dryad.vd0nn) current_code: Species codes used at present count: Number of individuals counted (for vegetatively propagating plants a clump of stems that was basally connected was considered as one individual) P_ID: Unique plot ID created by combining columns site and plot_no 05_all_regeneration_data.csv Description: The file contains records of woody seedlings and saplings (1-5 cm diameter at breast height at 1.3 m, dbh) and larger-stemmed trees (>5 cm dbh) recorded within single 5 m x 5 m regeneration plots nested within 20 m x 20 m plots. Plots were located in three continuous forests and two forest fragments. Data were filtered during analysis in Osuri et al. (2017, Journal of Tropical Ecology) to retain only seedling and saplings, defined as individuals with effective diameter <=5 cm. Note: Same as in the Osuri et al. (2017) dataset, with Tithonia diversifolia added back in Injipara from original dataset; and data from one additional site (Manamboli Lower) added back from the Muthuramkumar et al. 2006 dataset. Column names and descriptions: x: Row index site: Name of forest site plot_no: An unique plot number assigned to each 20m x 20m adult tree plot within each site q_no: The 5 m x 5 m plot was placed in the SW corner of the 20 m x 20 m plot in this q_no which indicates one of the four 10 m x 10 m quarters of the 20 m x 20 m plot, where each quarter was given a unique number in each site t_no: An unique number assigned to each individual seedling, sapling or tree within each site. old_code: Species codes used at the time of data collection (for full species names refer to 06_all_species_names.csv file with this dataset) osuri_code: Revised species codes used in the Osuri et al. 2017 paper in Journal of Tropical Ecology 33: 270-284 (https://doi.org/10.1017/S0266467417000219) and related dataset (https://doi.org/10.5061/dryad.vd0nn) current_code: Species codes used at present dbh_1 to dbh_12: Diameter at breast height of single- (dbh_1) and multi-stemmed (dbh_2 – dbh_12) individuals, measured in centimetres (cm) eff_dbh: Effective diameter at breast height (cm)- calculated as ((dbh)^2 +(dbh_1)^2 +...+(dbh_12)^2)^(1/2), P_ID: Unique plot ID created by combining columns site and plot_no 06_all_canopy_readings.csv Description: The file contains canopy-related measurements taken in each 20 m × 20 m plot in three continuous forests and two forest fragments. Note: Units of light meter reading were not recorded Column names and descriptions: site: Name of forest site plot_no: An unique plot number assigned to each 20m x 20m adult tree plot within each site reading: A number assigned to the 1 to 4 readings taken in each plot light: Light measurement taken with a light meter in the plot canopy_openness: Canopy openness (scored from 0-100%) using a spherical densiometer (Canopy cover = 100 - canopy openness) P_ID: Unique plot ID created by combining columns site and plot_no 07_all_extracanopy_trees_data.csv Description: The file contains records of additional trees outside the 5 x 5 m plot whose canopy was overhead of the plot. Note: Species codes are used to denote presence (not count of stems) of that species in the overhead canopy. Column names and descriptions: site: Name of forest site plot_no: An unique plot number assigned to each 20m x 20m adult tree plot within each site q_no: The 5 m x 5 m plot was placed in the SW corner of the 20 m x 20 m plot in this q_no which indicates one of the four 10 m x 10 m quarters of the 20 m x 20 m plot, where each quarter was given a unique number in each site old_code: Species codes used at the time of data collection (for full species names refer to 06_all_species_names.csv file with this dataset) osuri_code: Revised species codes used in the Osuri et al. 2017 paper in Journal of Tropical Ecology 33: 270-284 (https://doi.org/10.1017/S0266467417000219) and related dataset (https://doi.org/10.5061/dryad.vd0nn) current_code: Species codes used at present P_ID: Unique plot ID created by combining columns site and plot_no 08_all_species_names.csv Description: This file provides species codes and species scientific names as originally used in the Muthuramkumar et al. 2006 paper (https://doi.org/10.1111/j.1744-7429.2006.00118.x), and as matched with the Global Biodiversity Information Facility (GBIF) species name matching tool Note: For plots that had no species occurrences (old_code = No herbs, Noliana), NA has been used for other columns Column names and descriptions: group: Code indicating main dataset group where species occurs (tree and regeneration data, liana data, understorey plants data) old_code: Species codes used at the time of data collection (refer to traits data file for full species names) osuri_code: Revised species codes if used in the Osuri et al. 2017 paper in Journal of Tropical Ecology 33: 270-284 (https://doi.org/10.1017/S0266467417000219) and related dataset (https://doi.org/10.5061/dryad.vd0nn) or else indicated as NA current_code: Species codes used at present original_name: Scientific name of plant species as used at the time of the original publication (Muthuramkumar et al. 2006) original_fullname: Scientific name and authorship of plant species as used at the time of the original publication (Muthuramkumar et al. 2006) original_family: Family of the plant species as used at the time of original publication GBIFname: Scientific name as matched by GBIF species name matching tool key: GBIF name matching tool key number matchType: Type of match confidence: Confidence returned by name matching tool status: Whether accepted name or synonym rank: Taxanomic rank (level) to which identified kingdom: Taxonomic Kingdom phylum: Taxonomic Phylum class: Taxonomic Class order: Taxonomic Order family: Taxonomic Family genus: Taxonomic Genus species: Taxonomic Species canonicalName: Canonical part of scientific name matched by GBIF authorship: Authorship of scientific na

Restoration of canopy cover through tree planting can assist in overcoming barriers to natural regeneration and catalyze recovery of degraded tropical forests. India has made international pledges to restore millions of hectares of degraded forests by 2030, but lacks empirical research on regeneration under different types of planted and natural overstories to guide this mission. We conducted a field study (65 plots of 25 m2) to examine the influence of overstory type and canopy cover on naturally regenerating tree seedlings across degraded rainforests (DRs), mixed‐native species ecological restoration (ER) sites, monoculture eucalypt plantations (MP), and mature “benchmark” rainforests (BR) in the Western Ghats mountains of peninsular India. ER had higher native tree seedling densities and recovered community composition toward BR levels compared to DR, while communities in MP shifted in the opposite direction. Densities of native late‐successional species increased with canopy cover (particularly in ER), but greater canopy cover was also associated with increases in alien species, a few of which are shade‐tolerant. Further, in a nursery experiment comprising four rainforest species, seed germination and early survival increased with shade, but did not vary across soils originating from DR, ER, and MP. Our findings show that while improving canopy cover is important, doing so by planting diverse native species, and controlling invasive alien species, can benefit rainforest recovery in DR fragments. Conversely, planting non‐native monocultures in degraded forests, which is a prevalent practice in India, could prove counterproductive for forest recovery in the long term.

In the Western Ghats, India, we study how different intensities of tea cultivation influence birds. We compared bird communities in conventional monoculture tea and mixed- shade tea plantations, both of which use agrochemicals, with organic tea plantations, a rainforest fragment, and continuous rainforest within the Anamalai Tiger Reserve. In 225 point count surveys, overall bird species richness and abundance were lowest in conventional tea and up to 33% higher in organic tea. Mixed-shade tea had 40% higher species richness (including 15 canopy and 4 shrub and mid-storey species - primarily frugivores, nectarivores and insectivores), and 83% higher bird abundance than conventional tea, with a greater proportion of forest-affiliated birds and similarity in species composition with forest sites. The rainforest fragment and continuous rainforest had a higher proportion, richness and abundance of forest-affiliated birds and fewer open-country birds, unlike tea plantations where the pattern was reversed. Habitat associations of 62 bird species in indicator species analysis revealed similar patterns. Thus organic tea is better than conventional tea for birds, but mixed-shade tea is even better, although still poorer than forests. Retaining or promoting native shade trees in tea plantations will increase bird diversity and abundance, including of forest-affiliated species and support landscape-level bird conservation.

Abstract The Nilgiri tahr Nilgiritragus hylocrius is an Endangered species of mountain ungulate endemic to the Western Ghats of India, a biodiversity hotspot. Habitat fragmentation, hunting and a restricted range are the major threats to this species. Although several surveys have assessed the species’ status, a population estimate based on a scientifically robust method is needed. We used the double-observer method to estimate the population of the Nilgiri tahr in the Anamalai Tiger Reserve, a protected area in the Western Ghats. We walked 257 km of transects across the Reserve, covering 36 grassland blocks (i.e. clusters of montane grasslands that were relatively separate from each other). We counted a minimum of 422 individuals in 28 groups, and estimated the tahr population in the study area to be 510 individuals (95% CI 300–858) in 35 groups. The male:female ratio was 0.71 and the young:female ratio was 0.56. Comparing our estimate with previous surveys suggests that the Nilgiri tahr population in Anamalai Tiger Reserve is stable. We found the double-observer survey method to be appropriate for population estimation and long-term monitoring of this species, and make recommendations for improved field protocols to facilitate the implementation of the method in the tropical mountains of the Western Ghats. Our findings suggest that the Reserve harbours 20–25% of the global population of the Nilgiri tahr, highlighting the area's importance for the conservation of this species.

Asian hornbills are known to forage and breed in fragmented rainforests and agroforestry plantations in human‐modified landscapes adjoining contiguous protected forests. However, the factors influencing year‐round hornbill abundance, demography and tracking of key food resources such as wild fig Ficus fruits in modified habitats and protected forests remain poorly understood. We carried out monthly surveys of two species of high conservation concern, the Vulnerable Great Hornbill (GH, Buceros bicornis ) and the endemic Malabar Grey Hornbill (MGH, Ocyceros griseus ) for 15 months and monitored ripe fig fruit availability for 12 months along 11 line transects (total length 24 km) in shade‐coffee plantations and adjoining continuous rainforests in a protected area (PA) in the Anamalai Hills, Western Ghats, India. Both hornbill species used plantations and the PA year‐round but distance sampling density estimates were higher in the PA in both nesting (GH by 57%; MGH by 50%) and non‐nesting (GH by 53%; MGH by 144%) seasons. Relative to estimates from 2004 to 2005, mean GH density appeared stable or increasing, whereas MGH had declined by 39% in the PA and by 56% in plantations. Monthly encounter rate of both hornbills tended to be higher in the PA and that of MGH was also positively related to the density of fig trees with ripe fruit. Sex ratios of observed adult birds in the non‐nesting season were relatively even (GH) or slightly female‐biased (MGH), but became male‐biased in both species during the nesting season when females were confined in tree‐cavity nests. We used change in the adult sex ratio of observed birds from the non‐nesting to nesting season to estimate an index of the proportion of adult pairs breeding at any point within the season, providing the first such estimates for any hornbill species. The proportion of breeding pairs was higher in the PA (GH – 56%, MGH – 64%) than in the plantations (GH – 33%, MGH – 30%). Although hornbills use shade‐coffee plantations year‐round, partly due to fig fruit availability, differences in hornbill density and breeding incidence, as assessed from the sex ratios of observed adult birds, indicate that plantations are a sub‐optimal habitat for both species.

Seed predation is one of the key ecosystem processes governing the plant population and community structure in forests. Forest fragmentation and habitat loss have been shown to affect seed predation, leading to altered tree recruitment. However, the effects of fragmentation and habitat loss on seed predation are highly variable and context-specific, with limited information from South Asia. For four rainforest tree species in a production landscape of tea and coffee, we examined the influence of forest type (benchmark, N = 3, and fragmented, N = 5, forests) and seed size on the proportion of seeds, 1) predated by vertebrate seed predators, 2) predated by invertebrate seed predators, and 3) removed by vertebrate seed predators. The seed fates were tracked by placing the seeds in four 1 x 1 m plots under the canopy of the tree (N >= 11 individuals per tree species and 48 trees overall), and camera traps were placed under a subset of trees (N = 29, trap-nights = 712). Overall, we recorded nine species of seed predators, and that the seed predator species composition differed among the four tree species. However, there was no significant difference in either seed predator species composition or their visitations across the forest types. The seeds were more likely to be predated (on average 6% higher predation) by vertebrate seed predators in the fragments than in the benchmark forests. Medium-seeded species were more likely to be predated by vertebrate seed predators than large-seeded species (on average 41% higher predation). Invertebrate seed predation and seed removal by vertebrate seed predators had a weak relationship with forest type, with higher invertebrate seed predation in the fragments and higher seed removal in the benchmark forests. With altered seed dispersal patterns resulting from habitat fragmentation and habitat loss, differential seed predation can alter regeneration patterns, thereby influencing adult tree communities in fragmented forests.

Forest fragmentation and habitat loss are major disruptors of plant–frugivore interactions, affecting seed dispersal and altering recruitment patterns of the dependent tree species. In a heterogeneous production landscape (primarily tea and coffee plantations) in the southern Western Ghats, India, we examined effects of surrounding forest cover and fruit crop size on frugivory of four rainforest bird‐dispersed tree species ( N = 131 trees, ≥30 trees per species, observed for 623 hr). Frugivore composition differed among the four tree species with the large‐seeded Canarium strictum and Myristica dactyloides being exclusively dependent on large‐bodied avian frugivores, whereas medium‐seeded Persea macrantha and Heynea trijuga were predominantly visited by small‐bodied and large‐bodied avian frugivores, respectively. Using the seed‐dispersal‐effectiveness framework, we identified effective frugivores and examined their responses to forest cover and fruit crop size. Results were idiosyncratic and were governed by plant and frugivore traits. Visitations to medium‐seeded Persea had a positive relationship with forest cover but the relationship was negative for the large‐seeded Myristica . In addition, two of the three effective frugivores for Persea responded to the interactive effect of forest cover and fruit crop size . Frugivore visitations to Heynea were not related to forest cover or fruit crop, and there were too few visitations to Canarium to discern any trends . These results highlight the context‐specific responses of plant–frugivore interactions to forest cover and fruit crop size influenced by plant and frugivore traits.

Abstract Ecological restoration is a leading strategy for reversing biodiversity losses and enhancing terrestrial carbon sequestration in degraded tropical forests. There have been few comprehensive assessments of recovery following restoration in fragmented forest landscapes, and the efficacy of active versus passive (i.e., natural regeneration) restoration remains unclear. We examined 11 indicators of forest structure, tree diversity and composition (adult and sapling), and aboveground carbon storage in 25 pairs of actively restored (AR; 7–15 yr after weed removal and mixed‐native tree species planting) and naturally regenerating (NR) plots within degraded rainforest fragments, and in 17 less‐disturbed benchmark (BM) rainforest plots in the Western Ghats, India. We assessed the effects of active restoration on the 11 indicators and tested the hypothesis that the effects of active restoration increase with isolation from contiguous and relatively intact rainforests. Active restoration significantly increased canopy cover, adult tree and sapling density, adult and sapling species density (overall and late‐successional), compositional similarity to benchmarks, and aboveground carbon storage, which recovered 14–82% toward BM targets relative to NR baselines. By contrast, tree height–diameter ratios and the proportion of native saplings did not recover consistently in actively restored forests. The effects of active restoration on canopy cover, species density (adult), late‐successional species density (adult and sapling), and species composition, but not carbon storage, increased with isolation across the fragmented landscape. Our findings show that active restoration can promote recovery of forest structure, composition, and carbon storage within 7–15 yr of restoration in degraded tropical rainforest fragments, although the benefits of active over passive restoration across fragmented landscapes would depend on indicator type and may increase with site isolation. These findings on early stages of recovery suggest that active restoration in ubiquitous fragmented landscapes of the tropics could complement passive restoration of degraded forests in less fragmented landscapes, and protection of intact forests, as a key strategy for conserving biodiversity and mitigating climate change.

Loss of mature tropical forests to agricultural expansion often creates landscapes with forest fragments embedded within a matrix of human-modified habitats and land uses. Such habitat fragmentation may be detrimental to species with specialized habitat and foraging requirements and their ability to persist in such landscapes may depend on their adaptability to habitat modification. Great Hombills Buceros bicornis, among the largest birds in Asian tropical rainforests, depend on large trees for nesting and a diverse array of patchily distributed fruiting trees. In the human-modified landscape of the Anamalai Hills, India, we compared the breeding biology and nesting behaviour of Great Hombills in contiguous rainforest (N=3 nests) and in modified habitat consisting of coffee plantations and rainforest fragments (N=5 nests). The nesting cycle of seven of the eight nests monitored varied between 114 and 130 days. Nest provisioning behaviour was similar in contiguous forest and modified habitat in terms of visitation and food delivery rates, but visitation tended to be higher and food delivery rate lower during the nestling phase than during incubation. As expected, tree density and native food plant diversity were lower in modified habitat than in continuous forest. The diversity of food provisioned was lower in modified habitat with a 57.5% dietary overlap with contiguous forest. Hombills in the modified habitat of coffee plantations used non-native tree species for nesting and foraging, indicating their adaptability to modified landscapes.