MICROHABITAT INFLUENCE ON GROWTH DISTRIBUTION PATTERN OF RAMIN ( ) IN SIAK, RIAU Gonystylus bancanus PROVINCE** DIDIT OKTA PRIBADI* and YAYAN WAHYU C. KUSUMA Center for Plant Conservation, Botanic Gardens, Indonesian Institute of Sciences, Bogor, Indonesia R 4eceived 12 April 2012/ Accepted 31 December 201 ABSTRACT Plant growth distribution patterns are influenced by habitat characteristics, ability of adaptation and association with other plant or animals. The influence of those factors, especially habitat characteristic, needs to be species defined to support plant conservation management. This study was aimed to: 1) measure plant growth dependence on their microhabitat; 2) define microhabitat variables that significantly influence the growth; and 3) develop suitable conservation measures at species level. Ramin ( ) is one of major timber species that has been facing Gonystylus bancanus high exploitation in Indonesia. This species is usually found on specific “peat swamps” ecosystem. Data were collected through primary surveys in Riau Province and analyzed by clustering the adult based on total height and basal area variables and describing the distribution pattern of cluster. Then, Discriminant Function Analysis the (DFA) was used to overlay the cluster with the distribution of microhabitat characteristic consist altitude, slope, ing soil humidity, soil pH, peat depth and canopy cover (measured in percentage). The results showed that distribution of microhabitat matched with 67.4% of height distribution and 78.3% of width distribution of tree basal area. Altitude and canopy cover percentage had significant correlation with total height distribution (α=0.05). Meanwhile, altitude, canopy cover and slope had significant correlation with basal area (α=0.1). However, peat depth variable showed an interesting pattern since shallower peat depth was followed by wider basal area. High correlation between plant growth and its microhabitat suggested that to conserve , conservation offered better strategy G. bancanus in-situ than conservationex-situ . Keywords: Discriminant Function Analysis, microhabitat, peat swamp forest, plant growth Gonystylus bancanus, distribution, ramin INTRODUCTION Ramin ( ) wood is one of the Gonystylus bancanus most valuable timbers in Indonesia. It has various local names such as gaharu buaya, medang keladi or pulai miang. It is known as an expensive wood which is similar to merbau ( spp., especially Intsia Intsia bijuga Intsia palembanica Shorea and ), meranti ( spp.), bangkirai ( spp. and spp.), Shorea Hopea eboni ( spp.) and cendana (Diospyros Santalum album L.). People usually use it as material for construction, making doors and windows and producing various kinds of furniture. Ramin wood is produced from genus , a Gonystylus membe r of T hymelaeac eae f amily and Gonystyloideae subfamily (Sidiyasa 2005). Approximately, 30 species of are Gonystylus distributed in Brunei Darussalam, Fiji, Indonesia (Kalimantan and Sumatera), Malaysia (Peninsular Malaysia, Sabah and Sarawak), Singapore, Nicobar Island, Solomon Island and the Philippines (Soerianegara . 1994). There are et al ten species which are found in Indonesia namely G. affinis G. bancanus G. Radlk., (Miq.) Kurz., brunnescens G. confusus G. Airy Shaw, Airy Shaw, forbesii G. keithii G. macrophyllus Gilg., Airy Shaw, (Miq.) Airy Shaw, Hook f., G. maingayi G. velutinus Airy Shaw, and Airy Shaw (Shaw G. xylocarpus 1954; Purba 2002).et al. BIOTROPIA Vol. 22 No. 1, 2015: 1 - 10 DOI: 10.11598/btb.2015.22.1.284 1 * Corresponding author : diditpribadi@yahoo.com **This paper was presented at Association for Tropical Biology and Conservation (ATBC) Conference 2010 on Tropical Biodiversity: Surviving the Food, Energy and Climate Crisis (19-23 July 2010). Sanur-Denpasar, Bali, Indonesia mailto:diditpribadi@yahoo.com BIOTROPIA Vol. 22 No. 1, 2015 2 Gonystylus bancanus is the most valuable ramin wood traded in international market. It can be found in Peninsular Malaysia, Sabah, Sarawak, Sumatera, Bangka, Kalimantan and Brunei Darussalam (Lim 2004; Sidiyasa 2005). It et al. grows in peat swamp forest having wet climate and waterlogged soil. It mostly becomes dominant tree at peat depth of 350-600 mm (Istomo 2006; Rostiwati . 2007).et al In the early 1980s, Indonesia had become the biggest exporter of ramin followed by Sarawak an d Pe n i ns ul ar Ma lay si a ( S oe r i an e g a r a et al. 1994). In the 1970s the annual harvested volume of ramin in Indonesia was about 1.5 million m , and it decreased drastically to 131,000 3 m in the year of 2000 (Lim . 2004) which 3 et al indicated that exploitation of ramin wood had exceeded their regeneration capacity (Bismark et al. 2005). Therefore, ramin protection and conservation efforts have been started by including spp., especially , in Gonystylus G. bancanus Appendix II CITES in 2004 (Samedi 2005). It means that the trade of spp., especially Gonystylus G. bancanus, has been restricted by certain quota every year. Trade restriction is among policies which are expected to reduce ramin wood's exploitation. However, since illegal logging activities and the loss of forestland are still continuing in Indonesia, it should be coupled with direct conservation measures in the field. Bogor Botanical Garden is one of the Indonesian conservation institutions having responsibility to conserve such kind of plants. Therefore, this research becomes important to analyze the growth pattern of G. bancanus in its habitat, thus conservation management could be defined properly. Plant growth distribution patterns are influenced by habitat characteristics, ability of adaptation and association with other plant species or animals. Recently, studies on the natural distribution of plants and animals species are growing not only to assess the impact of anthropogenic effect or climate change on their habitat, but also to define appropriate conservation management of the species (Guissan & Thuiller 2005; Sinclair . 2010). et al Ever y species has different strength of relationship with its habitat, thus some species have endemism distribution and the others have cosmopolitan distribution. The endemicity level of a species will determine the effectiveness of conservation management effort either in ex-situ or conservation.in-situ Based on the above reasons, the aims of this research were: 1) to measure growth dependence of on their microhabitat; 2) to define G. bancanus microhabitat variables that significantly influence the growth of ; and 3) to formulate G. bancanus suitable conservation measures for . G. bancanus Few parts of the research had been orally presented during ATBC (Association for Tropical Biology and Conservation) 2010 Conference in Bali. MATERIALS AND METHODS Site Description The research was conducted in Danau Pulau Besar/Danau Bawah Wildlife Sanctuary that encompasses an area of 25,000 ha. This area is located in Dayun Village, Siak District, Siak Municipality, Riau Province. Geographically, this area lies between 00 35'-00 45' N and 102 10'-0 0 0 102 19' E (IPB & UKSDA Riau 2000). It is 0 managed by Balai Besar Konservasi Sumber Daya Alam, Riau Province. Generally, this area consists of peat swamp area which has flat topography, low altitude from 2 to 6 m above sea level and slope ranging from 00 to 3 (IPB & UKSDA Riau 2000). It is situated in 0 an alluvium area which was produced by Siak River sedimentation, thus it contains high organic and inorganic matters. Based on climate type of Schmidt-Fergusson, this area is classified into type A which has rainfall from 2,200 to 2,600 mm per year. Temperature of the area is 26.2 C per year on 0 average, with maximum temperature of 32.4 C 0 and minimum temperature of 21.7 C. Air 0 humidity in this area can reach 84% on average with maximum humidity of 97% and minimum humidity of 60% (IPB & UKSDA Riau 2000). Research Time and Location The research was conducted from 5 to 17 December 2007. The location of the study site was Danau Pulau Besar/Danau Bawah Wildlife Sanctuary, Siak Municipality, Riau Province. Microhabitat influence on growth distribution pattern of ramin ( ) Didit Okta PribadiGonystylus bancanus – et al. Survey Methods Field survey was conducted to collect primary data which consisted of spatial distribution of G. bancanus, total height and diameter breast height o f a d u l t t r e e , i n c l u d i n g m i c r o h a b i t a t characteristics in each location of G. bancanus comprising altitude, slope, soil humidity, soil pH, peat depth and canopy cover (in percentage). To investigate the microhabitat characteristics, unbound plots of ca. 100 m were applied with at 2 least one individual of centered within G. bancanus the plot. The sample was taken based on purposive random sampling method because the research was only focused on as the G. bancanus target species. Data Analysis Data were analyzed by clustering the adult based on total height and basal area variables. The criteria used to define the cluster were as follows:  Cluster 1: more than average + standard deviation  Cluster 2: average ± standard deviation  Cluster 3: less than average – standard deviation Furthermore, Discriminant Function Analysis (DFA) was used to overlay the cluster with the distribution of microhabitat characteristics consisted of altitude, slope, soil humidity, soil pH, peat depth and canopy cover. DFA is one of multivariate analyses used to evaluate the accuracy of cluster formation based on several quantitative variables. In this study, cluster of growth formation based on height and width of basal area was evaluated by microhabitat characteristics. The correlation between cluster of growth (based on height and width of basal area) and microhabitat characteristics were used to estimate influence value of microhabitat characteristics on growth distribution of . In addition, DFA was G. bancanus also used to discover which microhabitat characteristics best explained differences on distribution of the species (McGarigal . 2000). et al When correlation between the plant and its habitat is relatively high, it means that the adaptation of the plant in different habitat is relatively low. RESULTS AND DISCUSSION Based on height variable, was G. bancanus divided into three clusters: 1) Cluster 1 consisted of 9 individuals with 32-35 m of total height; 2) Cluster 2 consisted of 29 individuals with 22-30 m of total height; 3) Cluster 3 consisted of 8 individuals with 15-20 m of total height. Based on basal area, was divided into three G. bancanus clusters: 1) Cluster 1 consisted of 7 individuals with 2,428-2,307 cm ; 2) Cluster 2 consisted of 31 2 individuals with 594-1,698 cm ; 3) Cluster 3 2 consisted of 8 individuals with 154-330 cm .2 Each cluster of height and basal area had microhabitat characteristics that were shown in Table 1 and 2, respectively. There were 6 measured microhabitat characteristics i.e. altitude, slope, soil humidity, soil pH, peat depth and canopy cover. Each cluster had respective mean and standard deviation for each microhabitat characteristic. Based on the above results, DFA was conducted to evaluate whether clustering of height matched with distribution of microhabitat characteristics. T he result showed that distribution of microhabitat characteristics only 67.4% matched with cluster of height. However, f ur the r DFA an aly sis on mic rohabit at characteristics expressed that the individuals of G. bancanus could be grouped only in 2 clusters with detailed explanation as follows (Table 3): a. Nine individuals member of Cluster 1 of height should be included in Cluster 2 of microhabitat characteristics; b. From 29 individuals member of Cluster 2 of height, 26 individuals should be included in Cluster 2 of microhabitat characteristics, while the other 3 individuals should be included in Cluster 3 of microhabitat characteristics; c. From 8 individuals member of Cluster 3 of height, 3 individuals should be included in Cluster 2 of microhabitat characteristics, while the other 5 individuals should be included in Cluster 3 of microhabitat characteristics. Since the correlation between distribution of height with microhabitat characteristics was only 3 BIOTROPIA Vol. 22 No. 1, 2015 4 Table 1. Microhabitat characteristics of each cluster for total height Group Statistics Total height Mean Standard deviation Valid N (L istwise) Unweighted Weighted 1 Alt 23.8889 1.90029 9 9.000 Slope 12.2222 6.96020 9 9.000 SRH 94.4444 11.30388 9 9.000 SPH 5.3556 .46934 9 9.000 PD 119.8889 26.47666 9 9.000 CC .0000 .00000 9 9.000 2 Alt 23.1724 2.81665 29 29.000 Slope 12.7241 5.94406 29 29.000 SRH 95.7931 10.23083 29 29.000 SPH 5.4069 .59036 29 29.000 PD 112.3966 29.87510 29 29.000 CC 18.2759 24.79472 29 29.000 3 Alt 18.7500 4.97853 8 8.000 Slope 9.6875 4.78791 8 8.000 SRH 94.3750 10.50085 8 8.000 SPH 5.1125 .41897 8 8.000 PD 121.0625 34.62290 8 8.000 CC 55.6250 29.20830 8 8.000 Total Alt 22.5435 3.55094 46 46.000 Slope 12.0978 5.95410 46 46.000 SRH 95.2826 10.26897 46 46.000 SPH 5.3457 .54353 46 46.000 PD 115.3696 29.69824 46 46.000 CC 21.1957 28.65862 46 46.000 Notes: Alt = altitude; Slope = slope; SRH = soil humidity; SPH = soil pH; PD = peat depth; CC = canopy cover 67.4%, it was predicted that there were other factors affecting plant total height distribution such as genetic factors, age differences, or even other physiological factors (Ryan . 2006). et al Although the influence of age differences was minimized by selecting only adult plants which had similar appearance, however, the age of the plants was not measured quantitatively (e.g. tree rings approach) study. in this Another result of DFA showed that there were only two variables having significant influence on the tree height distribution consisting altitude and canopy cover (Table 4). Based on information from Table 1, it was described that increasing altitude was followed by increasing tree total height, while decreasing canopy cover was followed by increasing total height. Figure 1 shows that Cluster 2 (higher tree height) was located in higher altitude and lower canopy cover, whereas Cluster 3 (lower tree height) was located in lower altitude and higher canopy cover. This explained light sufficiency is important to support the height growth of . Higher G. bancanus altitude and less canopy cover gave more open space to the plant for getting more light intensity. No, 5 Table 2. Microhabitat characteristics of each cluster for basal area Group Statistics Basal area Mean Standard deviation Valid N (L istwise) Unweighted Weighted 1 Alt 23.8571 2.85357 7 7.000 Slope 17.8571 4.94734 7 7.000 SRH 95.7143 11.33893 7 7.000 SPH 5.2000 .46547 7 7.000 PD 106.0571 38.28607 7 7.000 CC 4.2857 11.33893 7 7.000 2 Alt 23.1613 2.55730 31 31.000 Slope 11.2903 5.74424 31 31.000 SRH 96.0645 9.93960 31 31.000 SPH 5.4194 .58276 31 31.000 PD 113.8097 27.25907 31 31.000 CC 15.4839 24.60877 31 31.000 3 Alt 19.0000 5.31843 8 8.000 Slope 10.1875 5.02805 8 8.000 SRH 91.8750 11.31923 8 8.000 SPH 5.1875 .42908 8 8.000 PD 129.5625 30.02075 8 8.000 CC 58.1250 24.19231 8 8.000 Total Alt 22.5435 3.55094 46 46.000 Slope 12.0978 5.95410 46 46.000 SRH 95.2826 10.26897 46 46.000 SPH 5.3457 .54353 46 46.000 PD 115.3696 29.69824 46 46.000 CC 21.1957 28.65862 46 46.000 Note: Alt = altitude; Slope = slope; SRH = soil humidity; SPH = soil pH; PD = peat depth; CC = canopy cover Table 3. Evaluation result for cluster of total height membership based on microhabitat characteristics by using DFA Classification Results a Total height Predicted group membership 1 2 3 Total Original count 1 0 9 0 9 2 0 26 3 29 3 0 3 5 8 % 1 .0 100.0 .0 100.0 .0 89.7 10.3 100.0 3 .0 37.5 62.5 100.0 Notes: 1. = 67.4% of original grouped cases were correctly classifieda 2. cluster of height are represented by row, while cluster of microhabitat characteristics are represented by column No. Microhabitat influence on growth distribution pattern of ramin ( ) Didit Okta PribadiGonystylus bancanus – et al. 2 BIOTROPIA Vol. 22 No. 1, 2015 6 Table 4. Significant influences from microhabitat characteristics on tree total height distribution Tests of Equality of Group Means Wilks' Lambda F df1 df2 Significance level Alt .748 7.236 2 43 .002 Slope .964 .811 2 43 .451 SRH .996 .093 2 43 .911 SPH .959 .918 2 43 .407 PD .982 .386 2 43 .682 CC .627 12.772 2 43 .000 Figure 1. Plot of cluster of tree height based on altitude and canopy cover variables Simultaneously, DFA was also conducted to evaluate whether clustering of basal area matched with distribution of microhabitat characteristics. The result showed that distribution of microhabitat characteristics only matched 78.3% with the cluster of basal area. Detailed explanation of the result was as follows (Table 5): a. From 7 individuals member of Cluster 1 of basal area, 2 individuals should be included in Cluster 1 of microhabitat characteristics, while the other 5 individuals should be entered in cluster 2 of microhabitat characteristics; b. From 31 individuals member of Cluster 2 of basal area, 1 individual should be included in Cluster 1 of microhabitat characteristics, 28 individuals should be included in Cluster 2 of microhabitat characteristics, while the other 2 individuals should be included in Cluster 3 of microhabitat characteristics; c. From 8 individuals member of Cluster 3 of basal area, 2 individuals should be included in Cluster 2 of microhabitat characteristics, while the other 6 individuals should be included in cluster 3 of microhabitat characteristics. Note: Significance level of respective variable is shown by value in the sixth column which is less than 0.05 7 Distribution of microhabitat characteristics had higher correlation with variable of basal area than with variable of height. It described that 78.3% of basal area distribution were influenced by microhabitat characteristics and the rest was influenced by other factors such as genetic factors or age differences. As mentioned before, we did not measure the influence of age differences. Table 5. Evaluation result for cluster basal area width membership based on microhabitat characteristics by using DFA Classification Resultsa Basal area Predicted Group Membership 1 2 3 Total Original Count 1 2 5 0 7 2 1 28 2 31 3 0 2 6 8 % 1 28.6 71.4 .0 100.0 2 3.2 90.3 6.5 100.0 3 .0 25.0 75.0 100.0 Notes: 1. = 78.3% of original grouped cases were correctly classifieda 2. Cluster of basal area are represented by row, while cluster of microhabitat characteristics are represented by column Table 6. Significant influences of microhabitat characteristics on basal area distribution Tests of Equality of Group Means Wilks' Lambda F df1 df2 Significance level Alt .781 6.035 2 43 .005 Slope .823 4.609 2 43 .015 SRH .976 .525 2 43 .595 SPH .961 .870 2 43 .426 PD .942 1.319 2 43 .278 CC .623 12.995 2 43 .000 Note: Significance level of respective variable is shown by value in the sixth column which is less than 0.05 Figure 2. Plot of cluster of basal area based on altitude, slope and canopy cover Microhabitat influence on growth distribution pattern of ramin ( ) Didit Okta PribadiGonystylus bancanus – et al. BIOTROPIA Vol. 22 No. 1, 2015 8 Another result of DFA showed that there were three variables having significant influence on basal area distribution consisting of altitude, slope and canopy cover (Table 6). Based on information from Table 2, the increasing altitude, slope and decreasing canopy cover were followed by increasing basal area. This could also be seen from Figure 2 where Cluster 1 (the widest basal area) was located in higher altitude, higher slope and lower canopy cover, whereas Cluster 3 (the narrowest basal area) was located in lower altitude, lower slope and higher canopy cover. Peat depth variable showed an interesting pattern since shallower peat depth was followed by wider basal area, although not significant. The above result indicated that the distribution of microhabitat characteristics had 67.4% correlation, with distribution of G. bancanus vertical growth. Microhabitat characteristics which could support vertical growth of G. bancanus were higher altitude and decreasing canopy cover. This meant that needs G. bancanus sufficient light intensity to support its vertical growth. Therefore, sufficient light intensity should be provided if would be G. bancanus cultivated in conservation to produce ex-situ longer log. A research by Jans . (2012) showed et al that the seedling growth of was the G. bancanus highest in partial sunlight and reduced with decreasing light intensity. Afterwards, big-sized tree will flourish in full sunlight (Partomihardjo et al. 2008). The result also indicated that distribution of micr ohabitat characte ristics had higher correlation with distribution of basal area than with vertical growth. The correlation of microhabitat characteristics with basal area was 78.3%. Microhabitat characteristics supporting basal area growth of were higher G. bancanus altitude, increasing slope, decreasing canopy cover, while level and shallower peat depth could also support basal area growth. Peat swamp ecosystem is usually developed in a basin area. Therefore, the result actually indicated that would have maxim basal area if G. bancanus um they grew at the edge of the basin area which had higher altitude, higher slope, lower canopy cover and shallower peat depth. The result followed Istomo (1998) and Bismark . (2005) where et al deeper peat depth gave higher density, thus shallower peat depth gave lower density and higher basal area. It showed that the range of their habitat actually was relatively narrow. The distribution of adult at the edge of the G. bancanus basin area, recorded by GPS, in Danau Pulau Besar/Danau Bawah Wildlife Sanctuary could be seen in Figure 3. From the correlation between microhabitat characteristics and distribution of vertical growth and basal area, it could be estimated that planting of without any treatments in the G. bancanus outside area of their habitat would not be optimal. The survival rate would be low, and if they were successfully grown, the growth rate would be very slow s . Ismail in term of height and tree basal area et al. G. bancanus (2007) found that planting of in non-peat swamp area after 11 years resulted to low survival rate (52%), low average of total height (8.16 m), and low average diameter breast height (9.1 cm). Therefore, even though G. bancanus can be planted in non-peat swamp areas, limiting factor on utilization will be the relatively small diameter and log length. New or existing technology that can utilize smaller size logs may solve the problem (Ismail . 2007). Apparently, et al planting of is not easy and its growth G. bancanus is relatively very slow. According to the above discussion, although there is a possibility to conserve in an G. bancanus ex-situ in-situ conservation area, but the role of conservation becomes more important to conserve in an optimal condition. G. bancanus Therefore, massive exploitation and land use change in peat swamp ecosystem become two important challenges in conserving . G. bancanus However, a technolog breakthrough to y cultivate outside their habitat is G. bancanus still needed to strengthen the conservation of this species. As suggested by Krigas . (2010) et al this kind of study can provide valuable information to facilitate the transfer from wild habitats to man-made habitats like botanical garden. This transfer is important to produce material for reintroduction and reinforcement of this species when its existence in nature becomes increasingly rare (Guerrant 2004). et al. Bogor Botanical Garden has been planting seedlings of since year 2007, but G. bancanus their growth are still insignificant despite their survival. Figure 3. Distribution of within the study area in Riau Province in 2008 G. bancanus CONCLUSIONS High correlation between the distribution of tree growth and microhabitat characteristics suggested that conservation offers better in-situ strategy than conservation to conserve ex-situ G. bancanus , especially when an appropriate planting method is not yet developed. ACKNOWLEDGEMENTS We thanked our colleagues (Yupi Isnaini, Enda Suhenda and M. Madhari) for their assistance in various aspects of this study. We also thanked the people in DPB-DW (Sutrisno, Jumaat, Beni Ishak Silalahi) for their hospitality, and BKSDA Riau for the permission support. The study was funded by DIPA - Center for Plant Conservation Bogor Botanical Gardens-LIPI. REFERENCES Bismark M, Kalima T, Wibowo A, and Sawitri R. 2005. Growing Stock, Distribution and Conservation of Ramin in Indonesia. Technical Report No. 01. ITTO PRO.87/03 Rev. 2 (F). Identification of Gonystylus spp. (Ramin) Potency, Distribution, Conservation and Plantation Barrier. Ministry of Forestry, Center for Forest and Nature Conservation Research and Development in cooperation with International Tropical Timber Organization. Guerrant EO, Havens K, Maunder M. 2004. Ex-situ Plant Conservation: Supporting Species Survival in The Wild. Washington DC(US):Island Press. Guisan A, Thuiller W. 2005. Predicting species distribution: offering more than simple habitat model. Ecol Lett 8: 993-1009. Institut Pertanian Bogor, Unit Konservasi Sumberdaya Alam Riau. 2000. Laporan Akhir Rencana Pengelolaan Suaka Margasatwa Danau Pulau Besar/Danau Bawah Kabupaten Siak Propinsi Riau. Bogor (ID):Unit Konservasi Sumberdaya Alam Riau dan Fakultas Kehutanan Institut Pertanian Bogor. Ismail P, Shamsudin I, Abdul Rahman K, Hashim WS, Ismail H. 2007. Planting of in Gonystylus bancanus non peat swamp area. JTFS 19 (1): 50-6. Istomo. 1998. Penyebaran pertumbuhan pohon ramin ( (Miq.) Kurz.) di hutan rawa Gonystylus bancanus gambut: studi kasus di HPH PT. Inhutani III Kalimantan Tengah. Laboratorium Ekologi Hutan. Buletin Manajemen Hutan. p 33-9. Istomo. 2006. Kandungan fosfor dan kalsium pada tanah dan biomassa hutan rawa gambut (studi kasus di wilayah HPH PT. Diamond Raya Timber, Bagan Siapi-api, Provinsi Riau). JMHT 12 (3): 40-57. Jans WWP, Dibor L, Verwer C, Kruijt B, Tan S, Van Der Meer PJ. 2012. Effects of light and soil flooding on the growth and photosynthesis of ramin (Gonystylus bancanus) seedlings in Malaysia. JTFS 24(1):54-63. Krigas N, Mouflis G, Grigoriadou K, and Maloupa E. 2010. Conservation of important plants from the ionian islands at the Balkan Botanic Garden of Kroussia, N Greece: using GIS to link the collection data in-situ with plant propagation and cultivation. ex-situ Biodivers Conserv 19:3583-603. 9 Microhabitat influence on growth distribution pattern of ramin ( ) Didit Okta PribadiGonystylus bancanus – et al. Lim T W, Soehartono T, Keong CH. 2004. Framing the Picture: An Assessment of Ramin Trade in Indonesia, Malaysia and Singapore. Malaysia(MY): TRAFFIC Southeast Asia. McGarigal K, Cushman SA, Stafford S. 2000. Multivariate Statistics for Wildlife and Ecology Research. New York(US): Springer Verlag. Partomihardjo T, Prajadinata, S, Hidayat A. 2008. Current Status of Ramin Seeds Source in Sumatra. Technical Report ITTO PROJECT PO 426/06 Rev. 1 (F). Indonesia: Ministry of Forestry. Purba C, Suparno, Hariyono, Zulfahmi, Seventri AD. 2002. Laporan Studi Perdagangan Domestik dan Internasional Kayu Ramin Sebagai Bagian dari Pengumpulan Informasi Dasar dalam Kerangka Pelestarian Jenis Ramin di Indonesia. Bogor(ID): Forest Watch Indonesia/Telapak. Rostiwati T, Murniati, Hendromono. 2007. Growth of ramin ( (miq.) Kurz.) plantation Gonystylus bancanus on various peat swamp forests in Indonesia. IJFR 4 (2): 73-81. Ryan MG, Phillips N, Bond BJ. 2006. The hydraulic limitation hypothesis revisited. Plant Cell Environ (29): 367–81. Samedi. 2005. Kontrol perdagangan ramin ( spp.) Gonystylus internasional. Dalam: Semiloka Nasional Konservasi dan Pembangunan Hutan Ramin di Indonesia. Prosiding. 28 September 2005; Bogor (ID): Puslitbang Hutan dan Konservasi Alam-ITTO. pp 60-72. Shaw A. 1954. Gonystilaceae. In: van Steenis, CGGJ editor. Flora Malesiana I, Vol. (4). Jakarta(ID): Noordhoff- Kolff N. V. p 350-65. Sidiyasa K. 2005. Potensi botani, ekonomi dan ekologi ramin ( spp.). Dalam: Semiloka Nasional Gonystylus Konservasi dan Pembangunan Hutan di Ramin Indonesia Melalui Regulasi Perdagangan dan Pemacuan Alih Teknologi Konservasi, Penanaman dan Tehnik Silvikultur. Prosiding: 28 September 200 5; B og or (ID): P usat Peneliti an dan Pengembangan Hutan dan Konservasi Alam bekerja sama dengan ITTO dan PPD. p 9-34. Sinclair SJ, White MD, Newell GR. 2010. How seful re u a s d m mpecies istribution odels for anaging b f ciodiversity under uture limates? Ecol Soc 15(1): 8 [online] URL: http://www.ecologyandsociety. org/vol15/iss1/art8/ Soerianegara I, Sambas EN, Martawijaya, Sudo S, Groen LE. Gonystylus Teijsm. & Binnend. In: 1994. Prosea: Plant Resource of South-East Asia 5 (1) Timber Trees: Major Commercial Timber. Edited by Soerianegara I and RHMJ Lemmens. Wageningen (NL):PROSEA, Pudoc. p 221-30. 10 BIOTROPIA Vol. 22 No. 1, 2015 http://www.ecologyandsociety.