The study in ecology of the indochinese silvered langur (trachypithecus germaini milne - Edwards, 1876) at the chua Hang karst mountain, in Kien Luong district, Kien Giang province

ng. Feeding ecology studies have not been conducted in the wild and quantitative studies of locomotion have been done only in captivity. Besides the present study, there has been only one other project in Vietnam which has gone beyond survey and census reporting of wild the silvered langurs. 1.3. Feeding ecology of colobinae and Trachypithecus Colobines are folivorous, though their diet may be supplemented with flowers, fruits, seeds, and the occasional insect. Unlike the other subfamily of Old World monkeys, the Cercopithecinae, colobinae possess no cheek pouches. To aid in digestion, particularly of hard-to- 5 digest leaves, they have multichambered, complex stomachs, making them the only ruminant primates. The microorganisms in the forestomach, probably digest most of these compounds. The food selection strategies of colobinae correlated to the five models: (1) maximum of energy-rich food uptake; (2) maximum of protein-rich food uptake; (3) minimum of plant secondary compounds uptake; (4) limit of fiber content uptake; (5) nutrition balance. Furthermore, the food selection of langur species was affected by presence and abundance of favorite foods and other foods, property of food sources at habitats. The previous studies on the langur species, belonging Trachypithecus, showed leaves account over 50% of the diet. The plant species of food component was diversity, include trees, shrubs, vines, and epiphytes. In addition, food selective behavior of the langurs varied follow the season and food abundance in the habitats. Therefore, study on feeding ecology play an important role for indicating the suitable food selection model for the Indochinese silvered langur, which will be used for the silvered langur conservation. CHAPTER 2. STUDY SITE, STUDY PERIOD AND METHODS 2.1. Study site Research was conducted on T. germaini and vegetation at Chua Hang Karst Mountain of the Hon Chong Karst area in Kien Luong District, Kien Giang Province (10o08’11” N and 104o38’21”). The primary study site is a 56.5 ha karst hill, Chua Hang, which rises from one to 180.7 m above sea level. From May 2007 to May 2016, mean minimum and maximum temperatures were approximately 21.5oC and 31.5oC, respectively. Total annual precipitation at the site during the 6 study period was 2,156.62 mm, with 90% of rain falling between May and October (wet season). 2.2. Study period The study was performed from September 2013 to February 2017. The silvered langur behavior, feeding behavior observation and activity budget record was performed from September 2013 to August 2015. The vegetation characterisations and phenology were surveyed from March 2015 to February 2016. The plant food samples for chemical composition analyze, home range and population size data were collected during the study period. The study data were analyzed from September 2013 to February 2019. 2.3. Study methods 2.3.1. Vegetation characterisations These habitats were analyzed using quadrat and line transect sampling methods. Quadrat sampling used 1 m x 1 m plots (170 plots) for documenting shrub, vine, epiphyte, and parasitic plant species in the cliff, slope, and peak habitats and 5 m x 5 m plots (3 plots) to document plant species in the adjacent mangroves. In addition, we collected data for trees with diameter at breast height over 5 cm in four line transects (70-100 m x 2 m) in the slope habitat. These data were used to provide phenology and vegetation characterisations, including percentage of canopy cover, density, Importance Value Index (IVI), and Simpron and Margalef Index. Plant samples were catergorized using comparison method with plant sepciments at Southern Institute of Ecology and Pham Hoang Ho’s plant species lists (1999). 2.3.2. Phenology The percentage of young leaves, mature leaves, buds, flowers, and fruits on food plant at the habitats were estimated as decribed by 7 Chapman et al. (1992). The percentage ratios were devided into five level: 0= Not present on plant; 1= 0-25%; 2= 26-50%; 3 = 51%-75%; 4= 76-100% present on all branchs of plant. Study periods for phenology were about 1-2 days each month throughout the study year. 2.3.3. Behavioral observation and recorded data The focal-animal sampling method (Altmann, 1974) was used to observe and record all occurrences of behavior in the specified categories of resting, looking around, socializing, travelling, and feeding on each observation day from 6:00 am until 6:00 pm. Focal animals were chosen randomly each day, based on which langur group was encountered. Data were collected by using focal animal sampling for a ten-minute interval with 30 second instantaneous recording. During each 30 second interval we recorded activity, when the langur was feeding, and the species and plant part ingested (fruits, seeds, flowers, young leaves, mature leaves, buds, petioles, and others). These data were used to calculate the proportion of the day spent feeding and time spent feeding on different foods and plant species. The observation was performed for three consecutive days of each month from May 2014 to May 2016. 2.3.4. Home range and population size Using GPS and compass to mark the habitat of the silvered langurs. Determine the center position of the the langur groups every 15 minutes or when the langur groups move out a distance over 50 m. The data was processed by Mapinfo 9.5 software. The home range of the silvered langurs is calculated using the minimal convex polygons method. The core zones was identified to account for 75% of the langur-recorded occurrence and the adjacent region accounted for 25% of the langur-recorded occurrence. 8 2.3.4. Nutritional analyses The plant species consumed were tested for chemical composition and nutritional quality. The plant species recorded in habitats were devided into two categories, eaten and uneaten species, then subdivided eaten plants into two groups 1) frequently consumed species, and 2) less frequently consumed species. Samples were collected and analyzed for crude protein, neutral detergent fiber (NDF) with residual ash, acid detergent fiber (ADF), lignin, lipid, condensed tannin, total sugar, calcium, crude ash, and water content. Soil samples in the habitats were collected at the same location with the plant samples. Soil samples were collected at 30 cm and 60 cm deep for analyze pHwater, pHKCl, crude ash content, total nitrogen and organic carbon. 2.3.5. Statistical analyses The GPS spots of line transect, the langur groups location was analyzed by using Mapinfo 10.05 and ArcGIS 10.3 softwares. The feeding behavior data were managed in Microsoft Excel 2013 and statistically analyzed using SPSS 20 for Windows, with significance level set to 0.05. Chi Square test was used to test for differences in feeding behaviour and annual dietary composition for the Indochinese silvered langur. Mann-Whitney U-tests was used to test for significance in dietary percentages across seasons. Kruskal-Wallis test was used to compare the monthly dietary percentages. Nutritional components in leaf samples were analyzed, including crude protein, NDF, ADF, lignin, condensed tannin, Ca, lipid, total sugar, crude ash, and water content. The Shapiro-Wilk test was used to check whether the leaf chemistry data are normally distributed. The 9 results indicate water content, NDF, ADF, protein, total sugar, and NFC are normally distributed. Therefore, parametric tests were used to analyze these chemical categories. Lignin, tannin, lipid, Ca, crude ash, hemicellulose, protein/NDF and protein/ADF values were not normally distributed, so non-parametric tests were used to analyze these categories. Differences in leaves chemistry between two groups (eaten leaves vs. uneaten leaves, most frequently consumed leaves vs. less frequently consumed leaves, and less frequently consumed leaves vs. uneaten leaf samples) were analyzed by the Mann Whitney U-test or Welch’s t-test. Differences in leaf chemistry among the three categories (most frequently consumed leaves, less frequently consumed leaves, and uneaten leaf samples) were analyzed by Kruskal-Wallis test or ANOVA test. P-values reported by these tests were corrected using false-discovery rate (FDR) (Benjamini & Hochberg, 1995). In addition, generalized linear model (GLM) was used to explore the feeding behavior of the Indochinese silvered langur in the Chua Hang karst area. All model fittings were done in R Studio version 3.5 (R Development Core Team, 2018). CHAPTER 3. RESULTS AND DISCUSSION 3.1. The Indochinese silvered langur population size The Kien Luong Karst Area consists of 21 small karst hills which are isolated from one another by cultivated land and human settlements. Among these karst hills, the silvered langurs were found only on four, with a total of 286 individuals, with the largest population on Chua Hang hill. The population on Chua Hang Karst Mountain have 174 individuals, include 74 adults, 50 juveniles, and 10 infants. Comparion with the previous surveys, the number of silvered langurs at Chua 10 Hang Karst Mountain increased over time (Fig. 3.1). Among 134 the silvered langur individuals, there are 17 adult males, 23 adult females, 8 juvenile males, 12 juvenile females, 10 infants, and 34 adult individuals and 29 juvenile which did not indicated gender. So, the male:female ratio is 1:1.3; adult:juvenile ratio is 1:0.7; adult:juvenile:infant ratio is 7.4:5:1. The silvered langur population on Chua Hang Karst Moutain divided into 6 groups with highest number of individual/group of 42 individuals and lowest number of individual/group of 15 individuals. The ratios of gender and old of the groups are diferent (Table 3.1). Figure 3.1. The number of the silvered langur individuals at Chua Hang Karst Mountain Among the 6 langur groups, each group have one or more organizing forms of subgroups. In group 1, the individuals were organized into big subgroup with over 22 individuals in the wet season to eat leaves, fruits and young buds together. However, in the dry season, they were reorganized into the second subgroup form with about 8 to 15 individuals; the third subgroup form with 3-4 members of one family; the fourth subgroup form with 3-4 adult langurs and many infants; and 0 20 40 60 80 100 120 140 160 2000 2010 2016 Present study N u m b er o f th e la n g u r in d iv id u as 11 the fifth subgroup form with 2-5 adult langurs. The group 5 was also organized similar the group 1, except the fourth subgroup form. The group 2, 3, 4 and 6 were follow the second subgroub form and the fifth subgroup form and did not diferent between dry and wet season. Table 3.1. The number of individuals of the silvered langur at Chua Hang Karst Moutain Group Group size Individuals AM AF JM JF IF AU JU 1 42 7 9 5 6 4 6 5 2 16 2 4 0 1 1 3 5 3 17 2 3 1 1 1 4 5 4 15 3 2 0 1 1 3 5 5 28 2 4 1 2 1 11 7 6 16 1 1 2 1 2 7 2 Total 134 17 23 9 12 10 34 29 AM- adult male, AF- adult female, AU- adult not indicate gender, JM- juvenile male, JF- juvenile female, IF- infant. 3.2. Home range and habitat use The Indochinese silvered langurs distributed on the area 47.4 ha of Chua Hang Karst Mountain, 1.95 ha of other adjacent area and 0.79 ha of adjacent mangroves. Total of distributed area of the silvered langur is about 50 ha. The home range of the silvered langur is about 36.8 ha, accouting for 74% of the distributed area. The 13.2 ha of remaining area, almost at the peak with above 110m altitude, dit not recorded the present of the silvered langurs. In the home range area (36.8 ha), the core area (accounting for 75% of the sites which were recorded the present of the langurs) is 55 ha, the edge area (accounting for 25%) is 31.3 ha (Fig. 3.10). 12 Figure 3.10. Location and distributed area of the silvered langurs at Chua Hang Karst Mountain Table 3.2. Density (individual/ha) of the silvered langur groups at Chua Hang Karst Mountain Group Individuals Home range (ha) Density (individuals/ha) 1 and 6 59 5.11 0.09 2 16 3.68 0.23 3 17 1.35 0.08 4 15 3.94 0.26 5 28 4.34 0.15 The silvered langur groups live the same karst habitat, however distributed density is different between the groups (Table 3.2). In addition, each groups had the distinct home range and habitat use, some case the home ranges were overlapped between the groups which 13 closed together. The home range of group 1 and 6 was overlapped 60%; other groups were not significant. The home range area of the silvered langur groups was different between seasons. 3.3. Vegetation structure The vegetation on Kien Luong Karst area devided into four habitats: the cliff habitat, the slope habitat, the peak habitat, and the adjacent mangroves. The total number of plant species recorded in the Chua Hang Karst Mountain area was 185 belonging to 61 families, including 22% trees, 16% small trees, 20% shrubs, 24% vines, and 7% epiphytes. The plant species component was variety between the habitats. 3.4. Phenology During the study period, young and mature leaves were the most abundant parts available throughout the seasons. The availability of which fluctuated between 23.8% and 57.7% and 27.3% and 58.1% for young leaves and mature leaves, respectively. The availability of young leaves peaked in March, but by September most of them were mature. Buds were available at low levels throughout the year, fluctuating between 0.4% and 7.4%. Reproductive plant parts (flowers and fruits), with abundance consistently lower than vegetative parts, also fluctuated monthly in abundance. Flowers and fruits occurred at low-levels year round, between 0.0% and 8.8% and 5.0% and 14.7% for flowers and fruits, respectively. 3.5. Feeding behaviour of the Indochinese silvered langur The total recorded time in this two-year study was 320.44 hours, with 17,040 feeding bouts recorded (142 hour) for the Indochinese silvered langur on Chua Hang Karst Mountain. Feeding dominated the activity budget, accounting for about 45.0% of the total recorded time, 14 followed by resting at 25.0%, looking around at 14.3%, travelling at 8.7% and socializing at 5.7% (Fig. 3.20). Figure 3.20. Activity budget of the Indochine silvered lagur While the activity budget data of Indochinese silvered langurs differed significantly, P<0.05 and P<0.001, respectively, between the age-sex classes, the dietary data did not differ across age-sex classes (χ2= 1.5283, df = 6, p-value = 0.9576). Thus, the data were pooled and analyzed together to assess food selection of the species. Moreover, the plant sampling for chemical analyses was independent of the age- sex class of the animals consuming the plant items, so pooling dietary data is reasonable. 3.6. The annual dietary composition of the Indochinese silvered langur The diet of Indochinese silvered langurs was principally composed of young leaves (58.0%), followed by fruits (22.7%), mature leaves (9.5%), flowers (4.7%), buds (3.3%), petioles (1.2%) and others (0.5%) of 62 plant species, belonging 37 families (Fig. 3.28). Travelling 8.7% Feeding 45.0% Resting 25.0% Socializing 5.7% Observing 14.3% Other 1.3% 15 Figure 3.28. Percentage of plant parts in the langur’s food composition The plant species composition on the silvered langur diet was different between the wet season (27 species) and dry season (23 species) (χ2=364.1; df=7; p<0.01). The Indochinese silvered langur spent 72% of their annual feeding time eating leaves, including young leaves (78%, SD=7.4%), mature leaves (16%, SD=4.8%), buds (4%, SD=7.6%) and petioles (2%, SD=1.8%). Despite little variation between seasons (χ2=351,2; df=5; p<0,05), monthly variation in feeding records of leaves is evident (χ2=3177,4; df =55; p<0.05). Young leaves are the most important food and dominated throughout the year. Feeding on young leaves peaked at 85.1% in July, and was at its lowest at 58.7% in November. Mature leaves and young leaves belonging to the same plant species were eaten together during the months when there were insufficient amounts of young leaves available. While mature leaves were eaten in every month, the Indochinese silvered langur’s consumption of mature leaves varied quite extensively, peaking at 27.8% in August, and descending to a low of 11.9% in July, when young leaf consumption was at its highest. The highest level of petiole consumption occurred in May at 6.4%. In 16 all other months they accounted for less than 3% and they were not selected at all in some months (Fig. 3.29, and Fig. 3.31). Figure 3.29. Percentage of plant parts in the silvered langur’s food composition between months Figure 3.31. Seasonal feeding of plant parts 3.7. Plant chemistry and plant food selection 3.7.1. Plant chemistry The chemical component data of 28 plant food samples showed in Table 3.16 indicated that protein content of the silvered lagurs diet was lower than the level requirement which was suggested by National Research Council (2003) (12-22%). The fiber content (NDF and ADF) 3.2 59.5 9.0 24.5 3.3 0.5 6.4 57.0 10.1 20.9 2.7 2.9 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 Flowers Young leaves Mature leaves Fruits Buds others % Dry season Wet season 17 was higher than the requirement levels (5-10% and 10-30%, respectively). The Ca and lipid content were suitable (0.8% and 0.5- 2%, respectively). Table 3.16. Chemical component of the plant food samples (leaves, fruits, and flowers, n=28) Component (%) Mean (N=28) SD Min Max Water 73.68 8.63 57.77 89.80 Protein 5.58 3.98 0.88 15.9 ADF 37.74 9.58 18.8 56.4 NDF 45.76 11.68 22.60 76.30 Lignin 33.89 19.74 7.29 61.50 Tannin 3.44 3.98 0.42 17.6 Sugar 6.80 3.91 2.00 14.87 Lipid 1.24 1.24 0.07 4.39 Ca 0.97 0.069 0.20 3.71 Ash 5.43 3.59 2.10 16.33 The leaves of the 58 plant species eaten by the Indochinese silvered langurs, were divided into two groups, most frequently consumed leaf group including 13 plant species (plant species which were fed upon throughout the year and feeding records were always over 2%) and less frequently consumed leaf group including 45 plant species (<2% of feeding record). For the chemical component analysis, we selected twenty young leaf samples at random from the two leaf categories in the habitats, including ten most frequently consumed leaf samples and six less frequently consumed leaf samples. In addition, four uneaten leaf samples, those have high canopy cover and/or IVI in habitats, were analyzed. The data in Table 3.17 showed the percentage of 18 chemical components of the eaten leaves (most frequently consumed leaves and less frequently consumed leaves) and uneaten leaves. Table 3.17. Nutrient composition and defensive compound content in eaten and uneaten leaf samples from Chua Hang Karst Mountain Parameter (%) Eaten leaves N=16 Uneaten leaves N=4 Mean Sd Mean Sd Water 74.3 8.5 71.3 12.4 NDF 44.1 13.2 44.2 13.7 ADF 36.1 9.6 38.1 13.9 Crude Protein 5.6 4.4 13.1 5.2 Lignin 24.8 18.0 51.5 11.3 Condensed tannin 2.6 3.1 8.7 3.0 Total sugar 7.8 3.6 4.2 0.8 Lipid 1.1 1.3 3.1 1.4 Ca 1.0 0.9 1.2 0.48 Ash 5.3 3.4 11.7 3.79 CP/ADF 0.14 0.1 0.35 0.29 The mean nutrient content of leaves selected by the Indochinese silvered langurs in Chua Hang Karst Mountain was: water 74.3%, NDF 44.1%, crude protein 5.6%, lipids 1.1%, ash 5.3%, total sugar 7.8%, Ca 1.0% and with a protein: fiber ratio of 0.14. Leaves eaten by the Indochinese silvered langur (N=16) were lower in crude protein, lipid, and ash, but higher in total sugar and water content than leaves not selected (N=4). The condensed tannin of eaten leaves was lower than uneaten leaves (P=0.014). In addition, within the fiber composition (including NDF, ADF and lignin), NDF and ADF content 19 did not differ between eaten leaves and uneaten leaves, while lignin content was lower in eaten leaves than that in uneaten leaves (P=0.007). The flower of 7 plant species which were eaten by the silvered langur, including Sterculia stigmarota, Amphineurion marginatum, Bauhinia bracteata, Cayratia trifolia, Ampelocissus martini, Phyllathus reticulatus, content 73.19% water, 5.73% protein, 1.36% lipid, 4.38% ash, 7.45% sugar, 0.63% Ca and ratio CP/ADF 0.14. In addition, the silvered langur ate 23 fruit species. The chemical component of eaten fruits include 73.19% water, 5.73% protein, 1.36% lipid, 4.38% ash, 7.45% sugar, 0.63% Ca, and ratio CP/ADF 0.14. 3.7.2. The relation of plant chemistry and the silvered langur’s food selection In order to indicate whether there is a relationship between leaf chemistry and food selection, we compared the chemical composition of eaten leaves (N=16), including 10 most frequently consumed leaves and 6 less frequently consumed leaves, to uneaten leaves (N=4). Differences between two food categories (eaten leaves vs. uneaten leaves) were found to be significant for lipid (U=9; P=0.029), lignin (U=5; P=0.007), condensed tannin (U=5.5; P=0.014), ashes (U=7; P=0.021), crude protein (t=-2.66; df=4.1712; P= 0.050), and total sugar (t=3.6381; df=17.993; P=0.001) using the Mann-Whitney U test or Welch’s t-test (Table 3.18). In addition, the most frequently consumed leaves were only significant different in lignin content from the less frequently eaten consumed leaves (U=10.0; P= 0.03). The crude protein content differed significantly across the three food categories (F-value=4.314; df=2; P=0.03). 20 Table 3.18. Comparison of nutrient and defensive compound content in eaten leaf samples and uneaten leaf samples from Chua Hang Karst Mountain Leaf samples Parametric test Non-parametric test Most frequently consumed leaves, less eaten and uneaten leaf samples Anova test: protein (F-value=4.314; df=2; P=0.03) Eaten leaf samples vs. uneaten leaf samples Welch’s t-test: crude protein (t= -2.66, df=4.1712, P=0.05); total sugar (t=3.63, df=17.993, P=0.001) Mann-Whitey U test: lignin (U= 5.0, P=0.007, FDR corrected); ash (U= 7.0, P=0.020); lipid (U= 9.0, P=0.020); condensed tannin (U= 5.5, P=0.013) Most frequently consumed leaves vs. less frequently consumed leaves Mann-Whitey U test: lignin (U=10.0, P= 0.03) P-values for multiple comparisons were adjusted by the false-discovery rate (FDR) correction method (Benjamini & Hochberg, 1995) In addition, generalised linear model (GLM) results showed that the model that best explained the relationship of leaf chemical components and feeding records include 5 factors: NDF, lignin, lipid, total sugar, and calcium. The top model showed significant correlation of feeding records for NDF (P<0.001), lignin (P<0.000) and Ca (P<0.05) but not for the other factors (total sugar: P=0.52; lipid: P=0.45). The positive correlation of the feeding records with NDF indicate that they spend more time eating leaves containing high NDF. However, GLM model for the relationship of leaf choice based on leaf chemical properties did not identify the most significant factors explaining the relationship of leaves choice and leaf chemical 21 components. The five factors (water content: P=0.70, ADF: P=0.998, condensed tannin: P=0.183, lipid: P=0.543, and lignin: P=0.523) did not significant correlate with leaves choice (Table 3.19). Table 3.19. Best-fit models for the effect of leaf chemical properties on feeding records using GLM with NDF, total sugar, lignin, lipid and calcium as explanatory variables. Type Estimate Standard error z- value Pr(>|z|) Significant AIC=-78.285 (Intercept) -5.18432 0.876649 -5.914 3.34E-09 *** NDF 0.056421 0.015199 3.712 0.000206 *** Sugar 0.005839 0.054975 0.106 0.915418 Lignin -0.06951 0.015834 -4.39 1.14E-05 *** Lipid 0.146817 0.194134 0.756 0.449491 Ca 0.529181 0.214671 2.465 0.013698 ** *: at level of 0.1, ** at level of 0.05; *** at level of 0.01 3.8. Soil chemistry The Chua Hang Karst Mountain have soil surface layer thin. From 3 to 30 cm deep, soil was clotted and mixed to small pebbles and plant roots. At the over 30 cm deep, soil contain many big pebbles. Soil chemistry and composition on the Chua Hang Karst Mountain differ slightly between habitats (Table 3.23). Soil classification follow sand:silt:clay ratio indicated