Abstract: Plant species composition in Ba Be National Park, Vietnam identified five plants that were rare and endangered and at serious risk of disappearing from the park. However, only Sinocalamus mucclure (string bamboo) and Markhamia stipulata had sufficient numbers to warrant molecular analysis. ISSR (Inter Simple Sequence Repeats) and RAPD-PCR (Random Amplified Polymorphic DNA-Polymerase Chain Reaction) were used to determine genetic differences between populations of Sinocalamus mucclure and Markhamia stipulata and whether genetic diversity was correlated to geographic distance. Disturbance in the area was evident, however, biogeography features usually associated with human disturbance was a minor contribution to species fragmentation. Analysis of similarity in biogeography and vegetation parameters demonstrated that similarity between the sites was between 65 and 85% and introduced plant species to the park was low. Trees dominated the canopy and species richness varied between the sites. Multi-Dimensional Scaling (MDS) of molecular data showed significant differences in populations of both plants and that substantial genetic variation between individuals of both species were present. ISSR revealed slightly less genetic diversity in both species (70-71% polymorphism) than RAPD-PCR (65-71% polymorphism). In conclusion, ISSR appeared to be more discriminatory than RAPD-PCR but both were versatile, sensitive and cost effective methods for use in genetic diversity and conservation genetics.
INTRODUCTION
Tropical and sub-tropical forests cover only about 7% of land on earth yet contain up to 50% of all plant species. These regions are important areas of biodiversity, containing many endemic vascular plants, yet the nature and integrity of these important ecological zones are being impacted on at a greater rate than ever (Breckle, 2002). This high diversity is in part due to steep ecological gradients, including microclimatic conditions, sharply defined ecotones and a lower amount of anthropogenic disturbance compared to temperate and dry forests (Collins et al., 1991; Bruner et al., 2011). The distribution of plant species within tropical and sub-tropical National Parks have been subjected to less human impact and are less disturbed and fragmented (Margules and Pressey, 2000) but despite this few studies related to fragmentation in these forests have been done.
In Vietnam, total forest cover has declined steadily throughout the 20th century this decline has accelerated in recent decades. For example, out of the 19 million hectares that could be classified as forests, only 3 million hectares are considered to possess ‘well-stacked’ healthy forests and old growth forests are estimated to be only 2 million hectares. In the north and northwest regions of Vietnam, forest cover has been reduced from 95% in 1943 to between 14 and 24% in 1995 (Quy, 2007). Conservation of these remaining forests is essential and that rare and endangered plants must form a solid basis for conservation strategies.
The two rare and endangered species under detail investigation in this study have their highest frequency in Ba Be National Park and represent small fragmented populations with a high risk of local extinction. String bamboo Sinocalamus mucclure and Markhamia stipulata are currently restricted to the limestone ridges of the National Park and only string bamboo has apparently been recorded elsewhere; in a localized region of southern China. Unfortunately, S. mucclure and M. stipulata are now primarily distributed where it is convenient for local tribal people to collect them for construction and medicine (Tran et al., 2001). Land clearance has also led directly to the endangerment of both species, through the division of previously continuous populations into smaller fragmented areas separated by inhospitable terrain.
The risk of extinction now faced by the remaining populations are compounded by the species own biology, i.e., they require a specialized habitat, they have poor seed dispersal mechanisms and are slow growing after harvest (Ministry of Agriculture and Rural Development Vietnam, 2003). These factors make the species particularly susceptible to inbreeding, with subsequent potential loss of genetic variation, accumulation of deleterious alleles and inbreeding depression (New, 2000). Previous studies have found that a loss of genetic diversity decreases the ability of wild populations to survive climatic extremes, pollutants, pests, diseases and exploitation (Frankham, 1995; Lynch et al., 1995; Frankham et al., 2000). According to Hopper and Coates (1990), the maintenance of genetic diversity and heterozygosity in natural populations may provide the best general strategy for ensuring the survival of most organisms. A number of authors (Dudash, 1990; Agren and Schemske, 1993; Caro and Laurenson, 1994; Hauser and Loeschcke, 1994) have claimed that there is no evidence that populations in the wild suffer from inbreeding depression and that catastrophes, over-exploitation and demographic or environmental factors are more important causes of extinction than inbreeding. Frankham et al. (2000), however makes a valid point, in that extinctions may be incorrectly attributed to non-genetic factors alone, when it is the interaction between genetic and non-genetic factors that is important.
A direct way of measuring genetic diversity is by the use of molecular analyses. Such techniques have the potential to reveal an immense number of characters, though they may also vary in the way they resolve genetic differences, in the type of data they generate and in the taxonomic levels at which they are best applied (Karp et al., 1996). One method, RAPD-PCR amplifies random genomic DNA sequences using single, short arbitrary primers and these can be used as genetic markers. The RAPD technique targets numerous loci in the genome which makes this method particularly attractive for analysis of genetic distance and similarity between closely related species (Persson and Gustavsson, 2001; Crockett et al., 2002). A second method to compensate for some weakness in RAPD-PCR, such as inconsistency between samples and dominancy of the loci targeted has been developed using ISSR primers (Raina et al., 2001; Martins et al., 2003). In the few reports comparing the two methods, RAPD bands appeared to be more reliable, or just as good a descriptor of relatedness than ISSR (Arif et al., 2009; Tripathi and Goswami, 2011). However, in other reports ISSR was considered more reliable than RAPD-PCR (Goulao and Oliveira, 2001; Pu et al., 2009; Goswami and Tripathi, 2010).
The objectives of this study were to identify plant species at risk in a fragmented landscape of Ba Be National Park and if possible to identify rare and endangered species which require more detail investigation at the molecular level. Other objectives were to determine which of the two multilocus DNA fingerprinting methods revealed more genetic information in the rare and endangered S. mucclure and M. stipulata. To determine how species abundance and species richness changed with biogeography elements at sites where both these endangered species were present. Finally, does the fragmentation of both species lead to genetic differences between populations and do the present populations of either species represent a significant storage of genetic variation.
MATERIALS AND METHODS
Study site and vegetation sampling: Ba Be National Park, Vietnam is in the Bac Kan province. It lies at an altitude of between 250 and 460 m above sea level, with variable and mostly limestone derived top soils, 80% average humidity, 20-23°C average temperatures and 1300-1800 mm of precipitation per year (Tran et al., 2001). Sites for this study were selected from areas both in the buffer zone and the outside zone of the National Park (Fig. 1). Complete species composition was recorded in the six vegetation sites using 10x100 m2 plots in a 100 m linear site profile during 2005. In addition, each of the 100 m2 plots were divided into 25 cells and all species present in each cell were recorded. In other sites, identified with enough individuals of either S. mucclure or M. stipulata for molecular analysis, only these plants were collected (Fig. 1). Taxonomic classification of all species was based on Pham (1999) for families and genera.
Biogeography data on soil, disturbance, vegetation and physical data was collected at all sites according to the list given in Table 1 (Le Brocque and Buckney, 2003).
Plant material collected: String bamboo, Sinocalamus mucclure Hay (sometimes classified as Ampelocalamus sp.) is a monocotyledon plant with a trailing root system, copious branching and rhizocorms. Typically the stem is 10-12 m long (some are up to 18 m) and when the stem is young it is pink to green in colour but when old it is yellow. Leaves are unifoliate, small, serrated at the margins with parallel-veins; the leaf/peduncle/rachis develops into a small young stem. The flower is small, pink, six stamens, pistillidium and flowers in November or December (Institute of Forest Resources and Environment, 2004). String bamboo regenerates slowly after flowering and the plant is used for making rope, baskets, musical string, bonsai and bamboo shoots.
| Fig. 1: | (a) Location of Ba Be National Park, Vietnam. (b) Study site of Sinocalamus mucclure and Markhamia stipulata, location and number of populations sampled and location and number of species composition (vegetation) sites |
| Table 1: | Biogeography data on soil, disturbance, vegetation and physical data of all the sites |
As well, it is used in medicine for treatment of extreme fever and heat dehydration (Pham, 1999).
Markhamia stipulata (RoxB.) Steen is a medium size tree approximately 25-30 m high, with a trunk diameter that may be up to 50 cm. The bark is grey, with many thin inner layers that are yellow. It has low separate branches, containing bifoliate pinnate leaves, opposite and 40-45 cm long. The flower is bisexual/hermaphrodite, large and irregular. The calyx is bell-shaped, the corolla is gamopetalous, brown-yellow in colour, bilabiate and stamens have two calycine-stamens in a compound ovarium (Institute of Forest Resources and Environment, 2004). M. stipulata is often used as an attractive garden specimen. The yellow wood has a pungent smell, the wood is used for making furniture and is also used for house construction (Pham, 1999).
Individual S. mucclure and M. stipulata plants within populations were selected at random, provided they satisfied the criteria of healthy growth. A representative sample for genetic variation in each area or site needed to contain at least 18 healthy individuals per population for reliable statistical analysis. Sampling areas for S. mucclure and M. stipulata were small in size but distances between populations were large (Fig. 1). Each leaf or shoot bud sample collected was wrapped in aluminium foil and placed in a container of liquid nitrogen to be stored at -80°C.
Extraction of DNA and PCR: Leaf slices (or shoot buds) still frozen in liquid nitrogen were extracted with CTAB isolation buffer and incubated at 60°C for 30 min (Magel et al., 2002). Proteinase K (10 mg mL-1)(Sigma) was added and incubated for 90 min at 37°C. The mixture was extracted with equal volumes of chloroform: isoamylalcohol (24:1) and the nucleic acids precipitated with the addition of 2/3 the volume of isopropanol. The pellet was washed with chilled 70% ethanol, the DNA dissolve overnight in TE buffer and treated with RNase A (10 mg mL-1) (Sigma) for 60 min at 37°C (De Filippis and Magel, 1998). Concentrations of DNA were estimated using a Pharmacia GeneQuant DNA/RNA calculator and confirmed by comparison with a known standard 100 bp DNA molecular weight ladder (Promega). The purity of the DNA was estimated by the ratio of absorbance readings at 260:280 nm.
| Table 2: | Summary of the primers used in ISSR and RAPD PCR reactions for comparative studies of Sinocalamus mucclure and Markhamia stipulata in Ba Be National Park, Vietnam |
| The anchored bases are underlined in the ISSR primers and primer sequence is from the 5’ to 3’ direction | |
PCR was carried out on DNA samples using 20 RAPD primers and 20 anchored ISSR primers with a G+C content of at least 50%. A list of the primers used, molecular weight range amplified and total number of bands assessed is given in Table 2. Each PCR reaction was performed in a final volume of 25 μL containing 2.5 μL 10x Taq Polymerase Reaction Buffer (Biotech), 0.6 M betaine, 2 mM each of dATP, dTTP, dCTP, dGTP (Biotech), 20 ng genomic DNA, 2.5 mM MgCl2 and 1 unit Taq DNA polymerase (Biotech) per reaction; plus 0.5 mM RAPD primer or 1.0 mM ISSR primer. Amplification reactions were carried out in a Biometra Personal Cycler programmed for: 1 cycle at 94°C for 5 min, then 40°C for 1 min; followed by 40 cycles each consisting of 1 min at 94°C, 1 min at 40°C and 2 min at 72°C and a final extension at 72°C for 5 min. Amplified DNA fragments were separated in 2% agarose gels with TBE buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA, pH 8.0) at 100 volts and stained for DNA with ethidium bromide (10 mg mL-1). The gels were then viewed on an Uvitec Transilluminator (M020616, Integrated Sciences) and photographed using a Kodak camera (EDAS 290).
Data analysis: PRIMER Version 5.2.7 (Clark and Gorley, 2001) was used to understand patterns of variation within and amongst populations and individuals by converting a set of variables into a few dimensions so that individual variations are condensed into a set of two axes (i.e., MDS). Such graphical analysis helped to identify the individuals and species which tended to cluster together. Cluster procedure was an average linking one and all similarities used were Bray-Curtis. The ANOSIM and SIMPER facilities in PRIMER were used to determine species contribution and percent similarity at each site using vegetation and biogeography data. Species Richness was determined by the total number of species per site irrespective of frequency but this may be dependent on sample size. As a result, Margalef index for species richness, Shannons diversity index and Pielou evenness index were also determined which are less dependent on the total number of individuals (Clark and Gorley, 2001).
Amplified DNA bands in RAPD or ISSR reactions were visually scored ‘1’ for the presence and ‘0’ for the absence of bands and the binary data used for statistical analysis. Faint bands and bands of doubtful reproducibility were ignored and duplicate PCR reactions were performed. The absence or presence of a band had to be consistent in both samples. To further reduce the possibility of comparing non-homologous bands, a known sample with the most number of amplified bands was included in all gels. Data was analyzed with PopGene Version 1.31 (Yeh et al., 1999) and G-statistic (Gst), MDS and dendrograms were used to estimate genetic differences between individuals and populations. The G-statistic estimates genetic differentiation and reduction in the number of heterozygote loci based on Nei’s regular and unbiased genetic measures (Yeh et al., 1999). Mantel test statistic (r) (Mantel, 1967) was used to determine correlation between geographic and genetic distances using the program IBD Isolation by Distance Version 1.5.2.
RESULTS
Vegetation analysis: MDS analysis of all data collected from species composition and abundance (Fig. 2) identified 13 plants which were distinctly different (i.e., separated) from the rest of the cluster of plant species and this enabled us to concentrate our molecular analysis on two endangered species from this differentiated group. Amongst the plants separated by MDS were important wood species, where the wood was of high value for construction and carving (Chukrasia tabularis, Markhamia stipulata, Toona sureni, Dimocarpus fumatus). C. tabularis and M. stipulata were considered rare and endangered, however, C. tabularis was present in insufficient numbers for molecular analysis. A second group of plants consisted of species with poor quality wood which were often collected for firewood (Aprosa microcalyx, Dracontomelon duperreanum, Polyalthia petelotti, Polyalthia sp.) and none were rare. A third group of species (Crassocephalum crepidioides, Burretiodentrom tonkinensis, Garcinia fragarioides, Sinocalamus mucclure) were mainly collected for use as a vegetable or for medicine. G. fragarioides, B. tonkinensis and S. mucclure were considered rare and endangered but only S. mucclure was present in sufficient numbers for molecular analysis. The final species identified in MDS was Streblus ilicifolius and it was separated mainly because of its frequency and distribution over all six sites; contributing on average 20% to total species present (Fig. 2).
Comparison of all sites based on biogeography and vegetation criteria showed that the sites had a narrow range of similarity between 65 and 85%.
| Fig. 2: | Multi dimensional scaling (MDS) illustrating species forming vegetation clusters, and which species can be considered isolated from the clusters, using all abundance and biogeographic data obtained from all the sites sampled |
| Table 3: | Summary data of life-form and species richness measurement for the six species composition (vegetation) sites analysed in Ba Be National Park, Vietnam. |
| Fig. 3(a-d) : | Multi dimensional scaling (MDS) illustrating genetic information provided by (a), (c) RAPD-PCR and (b), (d) ISSR for populations of Sinocalamus mucclure (a and b) and Markhamia stipulata (c and d). MDS was non-metric and between 104 and 144 bands were scored |
Contribution by criteria normally associated with human disturbance, such as amount of deforestation, close to tracks, villages nearby and weeds/grasses contributed less than 10% to total dissimilarity. The S. mucclure sampled sites were mainly characterized by the amount of large stone outcrops, shallow soils, slope and water availability. On the other hand, liana cover, altitude, humus cover and alkaline pH mainly characterized the M. stipulata sampled sites. Therefore, influence by human disturbance as measured by our criteria appeared not be a high contributing factor in the fragmentation of these two species.
In the total of over 2000 plants identified at the vegetation sites, only 6 plant species had been introduced to Vietnam and a further 13 species were considered endemic to Vietnam. Species richness identified site 1 as the most diverse both in species frequency and richness, while site 6 was the least diverse in both categories (Table 3). Site 2 was nearly as low in species richness as site 6 and this was independent on which index of diversity was used, while sites 3, 4 and 5 were similar to each other in species richness and diversity (Table 3). Trees were the dominant vegetation type on all sites, however none of the life-form categories were good indicators of species richness and Pielou’s evenness index was in all sites fairly constant at over 0.9.
| Table 4: | Summary of DNA polymorphism, population statistics and Mantel test for Sinocalamus mucclure and Markhamia stipulata in Ba Be National Park, Vietnam |
| G-statistic is identical to the F-statistic when only two possible alleles are scored at any locus, or where analysis of presence or absence of bands is conducted. G-statistic however does not require knowledge of genotype frequencies and is a measure of the reduction in the numbers of heterozygote loci and genetic differentiation. Mantel statistic (r) was used to determine correlation between geographic and genetic distances | |
| Fig. 4(a-d) : | Average linkage dendrogram illustrating genetic relationship for Sinocalamus mucclure in analysis by (a) RAPD-PCR and (b) ISSR and Markhamia stipulata in analysis by (c) RAPD-PCR and (d) ISSR. Cluster analysis of all individuals was used and between 104 and 144 bands were scored |
The habitat sites closest to M. stipulata contained higher species richness than the sites closest to S. mucclure.
RAPD-PCR: The number of polymorphic bands generated by the 20 RAPD primers between populations is given in Table 4 and there were on average between 139 and 144 band difference detected in comparison of populations by RAPD-PCR. In other words, between 65 and 71% RAPD polymorphisms were present between populations (Table 4). MDS showed more clearly differences between individuals and grouping of site populations from the RAPD polymorphic data (Fig. 3). In S. mucclure populations 7 to 10 were grouped together and another grouping consisted of populations 1 to 6. It is worth noting that population 9 appeared only partly grouped with populations 7, 8 and 10 (Fig. 3a). This group division was confirmed when the dendrogram of the data was constructed, but population 9 was still not clearly separated (Fig. 4a). In M. stipulata, only populations 2 and 3 were grouped together. Two other groupings consisted of populations 1 and 4. It is also worth noting that population 4 appeared to be grouped well apart from the rest (Fig. 3c). This group division was also apparent when the dendrogram of the data was constructed (Fig. 4c).
ISSR: Amplification by ISSR primers required a concentration of 3.0 mM MgCl2 for consistent band patterns and 1.5 units of Taq polymerase were found optimum for the ISSR primers. The number of polymorphic bands generated by the 15 ISSR primers is shown in Table 4. There were on average between 104 and 114 band differences detected in comparison of populations by ISSR; in other words 70-71% polymorphism were present in the populations (Table 4). ISSR primers MS1, MS2 and MS4 produced more bands per fingerprint than some RAPD primers alone but ISSR primers MS9 and MS10 produced fewer bands. However, primers MS7 and MS9 appeared to produce not as many low molecular weight bands as did other ISSR primers, or some RAPD primers. MDS showed clear differences between site populations from the ISSR polymorphic band data (Fig. 3). In S. mucclure populations 7, 8 and 10 were grouped together. Another grouping consisted of populations 1 to 6 and using ISSR data, population 9 appeared not to be grouped with any other population (Fig. 3b). This group division was confirmed when a dendrogram of the data was constructed (Fig. 4b). In M. stipulata only populations 2 and 3 were grouped together. Two other separate groupings consisted of populations 1 and 4 (Fig. 3d). This group division was confirmed when a dendrogram of the data was constructed (Fig. 4d).
DISCUSSION
Floristic structure: Analysis of similarity showed that differences in floristic structure between the six species composition sites could be accounted for by variation in the frequency of many species, rather than the presence of a limited number of dominant species, or unique species. Streblus ilicifolius was contributing on average about 20% to total species present but it was distributed over all six sites. The two endangered species studied were present as fragmented populations with low number of individuals at many sites and both appeared to be associated with distinct groupings of other plants. This suggests that conservation and re-planting of these species in forests might need to consider the species already present in the planting area. Nevertheless, the claim that rare and endangered species have a restricted distribution in the wild, as found in the study of Vaxevanidou et al. (2006) was not fully supported by this study. The results may indicate that there is no distinct identifiable gradient of increasing disturbance leading to species decline across the sites examined. Each site is experiencing disturbance to some degree but in the forms of various levels of moisture, forest exploitation, different physical conditions and soil characteristics.
Secondary, forest sites examined as part of a larger study were still showing their history of shifting cultivation, of slash and burn for expansion of agriculture or for collection of forest products. This form of forest disturbance can be complex and may contribute in subtle ways to variation in vegetation and environmental factors (Le Brocque and Buckney, 2003). In the sites examined the major biogeography criteria for differentiation of forest types was shown to be soil pH and structure and water availability; as was also reported by King and Buckney (2001). Clearly, no single environmental variable adequately explains the major vegetation patterns observed and the distribution of endangered species, but a most important factor appeared to be soil characteristic.
Molecular screening: RAPD-PCR is a simple technique that can be applied where no knowledge of the target DNA is available and no information about the region amplified is possible, but it is likely that middle or highly repetitive DNA is primarily targeted (Persson and Gustavsson, 2001). The results from this study suggest that RAPD-PCR was able to identify genetic differences between individuals and populations which have been isolated/disturbed over time and distance and this has also been reported for other plants (Cabrita et al., 2001; Heider et al., 2007; Tripathi and Goswami, 2011). These results must be considered and evaluated with some caution since the nature of the RAPD-PCR method means that normally it over estimates genetic differences. This over estimation appears consistent with this study where RAPD-PCR was able to detect large numbers of polymorphic band differences between populations than ISSR (Prathepha, 2000; Paul et al., 2010).
ISSR analysis also requires no prior knowledge of the target DNA, nor are specific ISSR primers needed for each species or population tested. ISSR makers may be co-dominant, however, it is still uncertain if the ISSR loci targeted in our study were co-dominant (Arif et al., 2009). Comparison of primers targeting six and eight two base repeats, respectively, demonstrated that each had different abilities to target sites on DNA. It is likely that the ability of an ISSR primer to amplify a sequence in the DNA depends on how abundant the sequence repeat is. Differences in the frequency of repetition of dinucleotide microsatellites have been reported in other species of plants (Whelan et al., 2000; James and McDougall, 2007). In this study, with a number of different anchored ISSR primers, the bases in the repeat sequence, the number of repeats and G+C content were important; especially the GC content which is best kept at values above 50%. In general, dinucleotide repeats amplified slightly more DNA bands than trinucleotide repeats, so the longer the repeat sequence is, the fewer the bands amplified. There is evidence in the literature that ISSR and other multi-locus microsatellite-fingerprinting methods, capture only some of the polymorphisms associated with microsatellites and ISSR repeats (Hayden and Sharp, 2001) and this must be considered in interpretation of data (Goswami and Tripathi, 2010).
Genetic variation and fragmentation: The G-statistic demonstrated that both S. mucclure and M. stipulata contained moderate to large genetic differentiation but comparisons between the two species showed consistently lower G-values in M. stipulata (Table 4). The G-statistic values for ISSR analysis (0.16-0.18) showed slightly higher values than with RAPD-PCR data (0.14-0.15). These values suggest similarities detected between individuals at any one site in this study were moderate (but not low) and indicated that significant genetic differences were present in individuals. Similarity between what might have been considered relatively unrelated and distant sites were still only moderate (De Filippis et al., 1996; Cumagun et al., 2007). In summary, results from the molecular data showed that a considerable amount of genetic variation within and between populations was present at the fragmented sites sampled. Similar, results were found in a study of the legume Flemingia macrophylla in North Vietnam, in which the authors discovered that the genetic composition of lowland accessions differed significantly from those in upland regions (Heider et al., 2007).
It is often the case that variability in smaller populations of plants can appear to contain nearly the same amount of variation than is contained in larger populations (Young and Clarke, 2000). Therefore our data was in agreement with the findings of Kelly et al. (2000) who discovered that small populations of the New Zealand mistletoe (with less than 10 individuals) did not show significantly reduced diversity compared with larger populations (up to 200 individuals sampled). Similarly, Fleishman et al. (2001) noted that small populations of Cordylanthus palmatus were just as genetically diverse as larger populations. Care should be taken in drawing conclusions from the latter study, however, since it is not known how many of the seedlings sampled would have reached maturity (Lesica and Allendorf, 1991).
Genetic diversity and distance: A significant correlation was observed between populations and geographic distance as determined by the Mantel test and distance appeared to be a major contributing factor to changes in genetic diversity. This relationship was apparent for both species, although in using RAPD-PCR data for S. mucclure this relationship was not significant (Table 4). Plants from the Botanic Garden site for M. stipulata (population 1) had high genetic variation, but were also correlated with distance to other populations and may represent a method of temporal genetic transfer amongst closely related species (Falconer and Mackay, 1996). Similar results were found in a study of the annual Lesquerella fendleri in which the author discovered that the genetic composition of plants differed according to geographic distance and how far these populations had been transported away from their point of origin (Schneller, 1998). This temporal genetic transfer is entirely possible in view of the biology and past collection and movement of M. stipulata.
CONCLUSION
The results from this study suggest that small population fragments of S. mucclure and M. stipulata in Ba Be National Park are well worth conserving as part of an overall strategy to maintain genetic diversity in both species. Management to ensure better conservation of both species however must be more pro-active and the information must be presented to management of the National Park so that they understand the significance of the findings. Promote ex situ conservation in the buffer zone and in situ monitoring of the remaining fragmented populations. Preserve as much as possible the shallow soils of the park. Educate the indigenous population not to collect material from the park and develop a plan to invest in methods of propagation that will allow growing of valuable species in domestic situations. Domestic plants could be harvested with ease and without the need for wild collections, eventually providing a wider genetic basis for the plants. Finally, the conservation of both these endangered plants would benefit considerably by interconnecting the present fragmented populations via corridors to help guarantee continuing genetic exchange.