115 Cey. J. Sci. (Bio. Sci.) 35 (2):115-136, 2006 A COMPARATIVE WOOD ANATOMICAL STUDY OF THE GENUS DIOSPYROS L. (EBENACEAE) IN SRI LANKA B. K. L. Wickremasinghe1* and Tissa R. Herat2 Department of Botany, The Open University of Sri Lanka, Nawala, Nugegoda, Sri Lanka. 2 Faculty of Applied Sciences, The Rajarata University of Sri Lanka, Mihintale, Sri Lanka. Accepted: 03 August 2006 1 ABSTRACT A comparative wood anatomical study of 28 Diospyros L. species described by earlier workers to be present in Sri Lanka was carried out. The diagnostic features of wood anatomical characters of the taxa concerned were evaluated by employing standard statistical methods such as Principal Component Analysis, Cluster Analysis and Discriminant Analysis. The Sri Lankan fraction of the genus revealed homogeneity with respect to their wood anatomical characters with a dependency range on environmental factors such as water availability. Further based on the wood anatomical characters an attempt was made to establish levels of specialization in order to understand the evolutionary relationships that exist within the species concerned. Even though the evolutionary relationships within the genus are vague, D. malabarica and D. atrata could be considered as possible ancestral forms from which the Sri Lankan endemic species of the taxon evolved. Key words: Diospyros, comparative wood anatomy. INTRODUCTION The history of the “ebony” woods goes back to the Tutankhamun’s period, the most famos Egyptian pharaoh ca. 3300 year before (Hora, 1981). Further, it is reported that, Sri Lanka (then Ceylon) was reputed for the best quality “ebony” woods throughout the world from very ancient times (Alston, 1931; Kostermans, 1977). However, proper scientific names of such taxa have been dated from the period of Dutch occupation in parts of Sri Lanka. For instance, Hermann (1717) reported a species, Highulhaenda [now known as D. ferrea (Willd.) Bakh.à which Burmann (1737) recognized as Highulhaenda folio myrti. Further, Linnaeus (1747) cited Higulhaenda folio myrti of Burmann in his Flora Zeylanica which according to Kostermans (1981) in the present concept is D. ferrea (Willd.) Bakh. Thwaites (1864) described 11 new species of Diospyros L. from Sri Lanka namely, D. acuminata Thw., D. acuta Thw. , D. affinis Thw., D. attenuata Thw., D. insignis Thw., D. moonii Thw., D. oblongifolia Thw. , D. oocarpa Thw. , D. oppositifolia Thw., D. quaesita Thw. He recognized two varieties within D. embryopteris [now known as D. malabarica (Desr.) Kostel] var nervosa and atrata which were considered endemic to Sri Lanka and also recognized and described a new species, Macreightia oblongifolia Thw. of the genus Macreightia of De Candolle (1844), which Kostermans (1981) later named as Diospyros oblongifolia (Thw.) Kosterm. Further, Thwaites (1864) recognized three varieties of Maba buxifolia var. ebenus Thw., var. microphylla Thw., and var. *Corresponding author’s e mail: bkwic@ou.ac.lk angustifolia Thw. and he was of the opinion that these three varieties were closely related or connected together by intermediate forms thereby representing a variable species with broad limits. Kostermans (1981) named these varieties as Diospyros ebenoides Kosterm., Diospyros nummulariifolia Kosterm. and Diospyros rheophytica Kosterm. respectively. Trimen (1893) recognized two genera, Maba Forst. and Diospyros L., and incorporated the then known taxa of Macreightia DC. within the genus Maba Forst. Wright (1904) published a comprehensive monograph of the Sri Lankan Diospyros L. and his work became a land mark study for the genus in Sri Lanka. He described morphological, anatomical and other taxonomical characters of 20 representative species of Diospyros L. in Sri Lanka. However, the genus Maba Forst. was excluded from the Ebenaceae. Worthington (1959) recognized 13 species of Sri Lankan Diospyros L., based on their morphological characters, of which three species were considered endemic to Sri Lanka, viz., D. crumenata Thw., D. quaesita Thw. and D. ovalifolia Wight. However, D. ovalifolia Wight is also reported from South India. (Kostermans, 1981; 1977) described two new endemic species, D. koenigii Kosterm, a species found only in Gannoruwa forest in Kandy District, and D. zeylanica Kosterm from Kanneliya forest near Hiniduma, Galle District. However, Kostermans (1981) in the Revision of the Ceylon Flora excluded this latter taxon stating that this B. K. L. Wickremasinghe and Tissa R. Herat plant represents probably a Stemonoporus (Dipterocarpaceae). Kostermans (1977) incorpora- ted other taxa recognized by Thwaites as Maba and Macreightia within Diospyros as described in first paragraph. In addition Kostermans (1981), described yet another variety of D. insignis, D. insignis var. parvifolia. Further he included Diospyros discolor Willd. (D. blancoi A.DC) [Velvet apple (E)], a species from the Philippines, now occasionally cultivated and naturalized in parts of the island, within the Flora of Sri Lanka totaling the number of species to thirty two. Jayasuriya (1998) added a further endemic species, D. pemadasai. All the revisions of the genus mentioned above have been based mainly on field observations and on morphological characters. Apart from the anatomical observations made by Wright in 1904 with then available facilities, no major anatomical study has been undertaken to determine the species limits to gain a better understanding of the Sri Lankan fraction of the genus. The present study is an attempt to investigate the wood anatomical characters and ecological parameters, with the idea of understanding the species limits of the Sri Lankan taxa of the genus Diospyros L. and comparing such findings with finding of other workers. MATERIALS AND METHODS Fresh wood samples of 28 species (Table 1), out of the 32 reported by Kostermans (1981) were collected from the field along with voucher herbarium specimens in triplicate. Collection of specimens was made to include wet, dry and intermediate zones. Further, species which were reported to be strictly confined to certain places were collected (Fig. 1). Four species were excluded since they were reported to be very rare or extinct (Kostermans, 1981). Collected specimens were identified by consulting Trimen (1893), Kostermans (1981) and comparing with the specimens deposited at the National Herbarium, Royal Botanic Gardens, Peradeniya (PDA). Wood samples were taken from straight branches as far as possible, to avoid inclusion of tension or pressure wood. An attempt was made to collect mature wood with maximum diameter whenever possible. Collected wood samples were cut into pieces of ca. 1 x 1 x 1 cm cubes and preserved in 70% alcohol. Radial, tangential and transverse sections were taken using a Sliding Microtome (E. Leitz Wetzla, Germany) at thickness ranging from 10 - 15 µm., stained in Safranin and Fast Green FCF series 116 following Sass (1958) and mounted in Canada balsam. Wood pieces of the size of match sticks were macerated using a solution of 30% hydrogen peroxide, distilled water and glacial acetic acid(1:4:5), then washed in running water for at least 24 hours, dehydrated, stained and mounted in Canada balsam following Johansen’s method (1940). Characters of wood for comparative anatomy were selected according to Tippo (1964) and Metcalf and Chalk (1989). Length of the vessel elements was measured including the tails, based on Chalk and Chattaway (1934). Tangential vessel diameters, vessel wall thicknesses and vessel frequencies were measured using cross sections. Individual pores were counted following Wheeler (1986). Measurements of vessel frequencies were based on 25 counts in an area of 2.8 mm2 field of view. Ray types were described according to Kribs (1935). The terminology used for microscopic features was based on the International Association of Wood Anatomists’ list of microscopic features for hardwood identification (IAWA Committee, 1989). Tangential wood sections were used to measure ray heights and ray frequencies while the lengths of fibers were measured using macerated tissues. Relative cell sizes and vessel frequency are given in accordance with Chattaway (1932). For quantitative data macerated wood, cross sections, radial sections and tangential sections were measured under the low power (x 10) and high power (x 40) of an Olympus PM – 10 DAD light microscope. Annual rainfall and mean annual temperatures were obtained from the Meteorological Department, Colombo. Altitudes of the areas where specimens were collected were determined using the altimeter supplemented with topographic sheets of the Survey Department (1992). Agro-ecological zones (Fig. 1) identified were based on the soil map of Sri Lanka (Irrigation Department, 1988). Statistical Analysis The basic statistical analysis was done according to procedures outlined by Snedecor and Cochran (1967) and Sokal and Rohlf (1981). For each sample from each species a mean value with 95% confidence interval of wood anatomical features was calculated. These values were tabulated with the species description citing mean ranges of Wood anatomical study of Diospyros 117 Figure 1. Distribution of Diospyros in Sri Lanka [modified after Kostermans ( 1981) ]. D. acuminata (2) D. acuta (3) D. affinis (4) D. atrata (5) D. attenuata (6) D. chaetocarpa (7) D. discolor (8) D. ebenoides (9) D. ebenum (10) D. ferrea (11) D. hirsuta (12) D. insignis (13) D. insignis var. parvifolia (14) D. koenigii (15) D. malabarica (16) D. melanoxylon (17) D. montana (18) D. moonii (19) D. nummulariifolia (20) D. oblongifolia (21) D. oocarpa (22) D.oppositifolia (23) D.ovalifolia (24) D. quaesita (25) D. racemosa (26) D. rheophytica (27) D. thwaitesii (28) D. walkeri. B. K. L. Wickremasinghe and Tissa R. Herat species and maximum and minimum intervals within parenthesis. For statistical analysis, mean values of each species were tabulated. Mean values derived were then used for fundamental statistical analysis such as mean comparison and analysis of variance. Further, Pearson product movement correlation was sought for the quantitative data to find out any interdependencies of anatomical characters. The same statistical techniques were employed to trace the relationships between ecological parameters such as altitude, mean annual rainfall and mean annual temperature. The correlation obtained was used to calculate r2 (coefficient of determination) to find dependency of the anatomical features over ecological parameters. The extent of influence of ecological parameters on wood anatomical features was sought in order to ascertain the degree of dependency between the two. The collected specimens were assigned to the climatic zones following Mueller-Dombois (1968) and further assigned to agro-ecological groups based on the soil map of Sri Lanka (Irrigation Department, 1988). Subsequently, mean vessel element lengths and mean fiber lengths were plotted separately with agro-ecological zones to detect whether there are any ecological patterns reflected. The codes used in investigating quantitative and qualitative characters and numerical codes used in qualitative analysis are given in Tables 2.A and 2.B. Multivariate analytical techniques, adopted by Sneath and Sokal (1973), Gauch (1986), Burley and Miller (1989) were attempted for further analysis. Principal Component Analysis (PCA) and Cluster Analysis (CA) were attempted in order to examine the taxonomic patterns of the data and to generate a classification system. In the course of statistical analysis, the species were coded with a number and an acronym as shown in Table 1. The original data were converted into standardized scores i.e. Z-scores (with 0 mean and standard deviation of 1) in order to give an equal weight to all the measurements. Phenetic similarity or dissimilarity was also calculated. The mesomorphy index (Carlquist and De Buhr, 1977) was calculated for the specimens since this index is reported to be related to the habitat preference. Another index, conductivity suggested by Carlquist (1984), was also calculated for each specimen because this index is believed to be a measure of vulnerability of vessels to extreme tensile forces created by extreme drought and cold conditions. Mesomorphy index and conductivity values were calculated by using the following 118 equations: Mesomorphy = Conductance = Vessel length x Vessel diameter ────────────── Vessels per mm2 (Vessel diameter)4 x 0.0001 ────────────── Vessels per mm2 All the computations and statistical analyses were done using the statistical program package SPSS/PC+ Statistics TM 4.0 and Windows version 6, (SPSS Inc., 1993). RESULTS AND DISCUSSION The occurrence of growth rings was not consistent among the species of Sri Lankan fraction of Diospyros L. Certain species showed very faint and discontinuous growth rings marked only by rather thick-walled fibers (Table 3). However, a number of non-endemic dry zone, intermediate and low country wet zone species lack in growth rings meanwhile some non-endemic dry, intermediate species and endemics of low country wet zone species showed discontinuous growth rings indicating that there was no considerable relationship between endemicity or with the ecological origin of the species with growth ring formation. The occurrence of growth rings in tropical woods is, generally, believed to depend on the environmental factors such as photoperiod, temperature and availability of water (Vliet, 1979; Kramer, 1964; Fahn et al., 1985). Therefore, it is evident that fluctuations in temperature, rainfall and altitudinal changes seem to be of little influence on the presence or absence of growth rings within the Sri Lankan species of the genus Diospyros L. Sri Lankan climate lacks the seasonality. However, wet and dry spells are well-recognized all over the island which is directly related to the mean annual rainfall. Therefore, occurrence of growth rings may be attributed to the soil moisture availability which in turn influences the variation in the plant growth.Further, whether the formation of growth rings is associated with the photoperiodism or is influenced by auxins as pointed out by Kramer (1964) is yet to be ascertained and require in-depth investigations. All the wood examined was non-storied. Individual vessel pores were oval to circular in cross sectional view. The pores were solitary or distributed in radially arranged multiples of varying number of pores (Fig. 2a, d, g). Wood anatomical study of Diospyros 119 Table 1. Species of Diospyros L. collected for the study and identified according to Kostermans (1981). endemicity, zonal distribution and acronyms used in the study. Species D. acuminata (Thw.) Kosterm. D. acuta Thw. D. affinis Thw. D. atrata Alston. D. attenuata Thw. D. chaetocarpa Kosterm. D. discolor Willd. D. ebenoides Kosterm. D. ebenum Koenig. D. ferrea (Willd.) Bakh. D. hirsuta L. D. insignis Thw. D. insignis var. parvifolia Kosterm. D. koenigii Kosterm. D. malabarica (Desr.) Kostel D. melanoxylon Roxb. D. montana Roxb. D. moonii Thw. D. nummulariifolia Kosterm. D. oblongifolia (Thw.) Kosterm. D. oocarpa Thw. D. oppositifolia Thw. D. ovalifolia Wight D. quaesita Thw. D. racemosa Roxb. D. rheophytica Kosterm. D. thwaitesii Beddome D. walkeri Wight Endemicity E E NE E E E NE E NE NE E NE E E NE NE NE E E E NE E NE E NE E E E Zonal Distribution W W D,I W W W W,I I D,I D,I W W W W D,I D D W D W D,I W D,I W W W W W Acronym Code DACUM DACUT DAFFI DATTR DATTN DCHAE DDISE DEBNO DEBNU DFERR DHIRS DINSI DINVP DKOEN DMALA DMELA DMONT DMOON DNUMM DOBLO DOOCA DOPPO DOVAL DQUES DRACE DRHEO DTHWA DWALK 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 E =Endemic, NE = Non-endemic, D = Dry zone, W = Wet zone, I = Intermediate zone Within the dry zone species such as D. montana Roxb., D. oocarpa Thw., D. affinis Thw. and D. ovalifolia Wight, the pores were arranged in multiples of 3-11 in a single group (Fig. 2g). In general, the dry zone species possess higher number of vessel groups per unit area than the intermediate and the wet zone species. Such observations agree with Carlquist (1984) was states that the increased number of groups of vessels indicates a xeric tendency of the habitat of a species. Metcalf and Chalk (1989) stated that, the number of vessels per square millimeter could be very significant as a taxonomic character only if very low or very high numbers of vessels are encountered. The present study reveals that the non-endemic dry zone species posses higher number of vessels per square millimeter than the endemic wet zone species. B. K. L. Wickremasinghe and Tissa R. Herat 120 Table 2. A. Codes used in the study for wood anatomical characters. Character Code Description Quantitative characters TVEL VDIA VWTH DFPO NGPA NSPA GSRT FLEN FDIA RPMM RHMM RHCE RWMM RWCE Vessel element length in µm Vessel Diameter in µm Vessel Wall Thickness in µm Distribution frequency of pores/mm2 Number of Groups of Pores per area/mm2 Number of Solitary Pores per area/mm2 Groups to Solitary Pores ratio Fiber length in µm Fiber Diameter in µm Rays per millimeter Ray Height in millimeter Ray Height in number of cells Ray Width in millimeter Ray Width in number of cells Qualitative characters VEDI PERP PITS TYLO GRRI FIBR RAUB TVRP PADS CRYS Vessel Distribution Perforation plates of vessel elements Pits whether minute or vestured Tyloses present or absent Absence or presence of Growth Rings Fibers thick or thin Rays, uniseriate or biseriate Type of vessel-ray pits Axial parenchyma distribution Crystals, Present or Absent However, such a relationship cannot be generalized for all the dry and wet zone species examined since some of the species, for instance D. rheophytica, a species which grows along the banks of rivers and rivulets in the low country wet zone, showed a higher value that does not fit into place along with the idea expressed above. Considering the percentage values of vessel elements per area, Wright (1904) attempted to correlate the higher percentages of vessel elements with thin-leaved trees and lower percentages with thick-leaved Sri Lankan species of Diospyros. From an ecological point of view, Wright (1904) was of the opinion that vessel distribution frequencies of the dry zone and the wet zone species could be related to the habitat. The results of the present study agree with Wright (1904) on the basis that the dry zone species show high frequencies of vessel distribution per area, compared with the wet zone. Working with the Sri Lankan Diospyros, Wright (1904) stated that there was no difference in the vessel dimensions between species growing in Mannar where more arid conditions occur, compared with those growing in Adams Peak Wilderness where the climatic conditions are quite different, especially in relation to the moisture availability. The tangential vessel diameters of the wood samples studied ranged from 48.4 to 107.2 µm. On this basis, the observations of the present study agree with Wright (1904). However, Metcalf and Chalk (1989) stated that vessel diameter can vary with samples which are taken from different places of the same tree. Further, Aloni and Zimmermann (1983) have pointed out that the size and frequency of vessels vary along the plant axis in response to concentration gradient of auxins. Kramer (1964) showed that wood formation depends to a great extent on the water availability of the habitat. As such vessel element diameters are of limited value in the delimitation of the taxa under consideration. However, these could be used to draw the generic limits of the Sri Lankan Diospyros. Wood anatomical study of Diospyros Bailey (1957) and more recently Baas (1976; 1982) believed that the length of vessel elements and other morphological characters such as perforation plates and types of pits reflect the level of specialization of a taxon and further recognized the evolutionary trends of vessel element lengths within angiosperm taxa which came to be known as Bailian trends. Metcalf and Chalk (1989) also stated that the vessel element length is more significant as a measure of phylogenetic specialization rather than as a diagnostic character for a taxon. It is the general opinion of Bailey (1957), Baas (1976; 1982) and Metcalf and Chalk (1989) that the less specialized plant taxa have 121 longer vessel elements than the specialized forms. On this basis, it is interesting to note that the nonendemic species of Sri Lanka have a shorter vessel element length than the endemic species found at wet and higher altitudes, while the taxa of the intermediate zone exhibit vessel element lengths of medium range. On this basis, the endemic Diospyros species of Sri Lanka are less specialized than the non-endemic taxa found in the dry zone. Whether these observations indicate a retrogressive evolutionary trend with respect to the Sri Lankan endemic taxa is interesting, however, and needs further investigation before conclusive decisions are made. Table 2. B. Code words and numerical codes used in qualitative analysis. Character (Abbreviations in parentheses) Character State Code Vessel Distribution (VEDI) Exclusively solitary Radial groups of 3-4 Radial or oblique Tangential arrangement Pore clusters 11 12 13 14 15 Perforation plate (PERP) Oblique Transverse Minute Vestured Present Absent Present Absent Thick-walled Thin-walled Present Absent Homogenous Heterogenous Vertically oriented Horizontally oriented Similar to Intervessel pits 11 12 11 12 11 12 11 12 11 12 11 12 11 12 11 12 Diffuse Paratracheal Apotracheal Vertical Present Absent or very rare 11 13 12 14 11 12 Pits (PITS) Tyloses (TYLO) Growth Rings (GRRI) Fibers (FIBR) Trachieds (TRAC) Rays (RAUB) Type of Vessel-Ray Pits(TVRP) Parenchyma distribution (PADS) Crystals (CRYS) 13 B. K. L. Wickremasinghe and Tissa R. Herat 122 Figure 2. Wood anatomical features of the genus Diospyros (a) Cross sectional view of D. ebenum x 105. (b) Radial longitudinal section of D. ebenum x 660, (c) Tangential longitudinal section of D. ebenum x 660, (d) Cross sectional view of D. moonii x 105, (e) Radial longitudinal section of D. moonii x 660, (f) Tangential longitudinal section of D. moonii x 660, (g) Cross sectional view of D. montana x 105, (h) Radial longitudinal section of D. montana x 660 and (i) Tangential longitudinal section of D. montana x 660. The thickness of vessel element wall ranged from 4.32 - 8.65 µm. The vessel wall thickness of the dry zone species is greater than that of Wet zone areas. Whether such characters observed in the dry zone taxa in relation to the water availability is an adaptive measure of conservation of water resources, needs further studies. Further, it is interesting to note that D. rheophytica Kosterm, a species restricted to the wet zone river banks, has the highest vessel wall thickness among the Sri Lankan species. Thus it is possible that further investigations on these lines may show a relationship between the vessel wall thickness and the availability of water. The terminology of wood fibers is controversial and yet to be resolved (Carlquist, 1986a; 1986b; Baas 1986). However, Metcalf and Wood anatomical study of Diospyros Chalk (1989) stated that non-septate fibers do not possess many features of taxonomic interest while septate fibers are often characteristic of large families and are useful as a taxonomic character. The wood fibers observed in the study are nonseptate and ranged from 376.00 (302.64 - 444.36) to 2055.00 (1564.66 - 2545.39) µm in length. The distribution of fiber length shows that the fiber length alone as a taxonomic character cannot be related to the level of specialization within the 123 genus studied. An index derived from the ratio between fiber length and vessel element length approaches to one in less specialized woods while higher values indicate higher levels of specialization in secondary xylem in angiosperms (Baretta-Kuipers, 1976). Results indicated a general tendency in which the non-endemic dry zone species of the genus are more specialized than thoseof endemic wet zone species (Table 3). Table 3. Occurrence of growth rings and the ray histology and ratio between fiber length and vessel element length in Sri Lankan species of Diospyros L. Code 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Species Diospyros acuminata (Thw.) Kosterm. Diospyros acuta Thw. Diospyros affinis Thw. Diospyros atrata Alston Diospyros attenuata Thw. Diospyros chaetocarpa Kosterm. Diospyros discolor Willd. Diospyros ebenoides Kosterm. Diospyros ebenum Koenig. Diospyros ferrea (Willd.) Bakh. Diospyros hirsuta L. Diospyros insignis Thw. Diospyros insignis var.parvifolia Kosterm. Diospyros koenigii Kosterm. Diospyros malabarica (Desr.) Kostel. Diospyros melanoxylon Roxb. Diospyros montana Roxb. Diospyros moonii Thw. Diospyros nummularifolia Kosterm. Diospyros oblongifolia (Thw.) Kosterm. Diospyros oocarpa Thw. Diospyros oppositifolia Thw. Diospyros ovalifolia Wight Diospyros quaesita Thw. Diospyros racemosa Roxb. Diospyros rheophytica Kosterm. Diospyros thwaitesii Beddome. Diospyros walkeri (Wight) Guerke Character Growth Rings Ray Histology FVRAT absent absent present absent present absent absent present present present absent absent uniseriate uniseriate uniseriate uniseriate uniseriate uniseriate uniseriate uniseriate uniseriate multiseriate uniseriate multiser iate 1.6 2.1 2.3 2.5 2.0 2.0 1.9 2.6 2.4 2.5 1.9 2.1 present present multiser iate uniseriate 2.0 2.2 absent present present absent uniseriate uniseriate uniseriate multiseriate 2.8 2.3 3.2 1.8 present uniseriate 2.1 present present present present present absent absent present absent uniseriate uniseriate uniseriate uniseriate uniseriate uniseriate uniseriate uniseriate uniseriate 2.2 2.6 1.8 2.1 2.1 2.4 1.7 1.4 1.7 B. K. L. Wickremasinghe and Tissa R. Herat The characteristic features of wood ray tissues are of importance in depicting an evolutionary sequence of specialization within angiosperms (Kribs, 1935). Barghoorn (1940) investigated the ontogenetic development and phylogenetic specialization of rays within the angiosperms and agreed with Kribs (1935) and stated that, heterogeneous Type I of Kribs’ which consists of a number of layers of procumbent cells with four upright or square marginal cells occurred in the extinct taxa of Benetitales, Pteridospermae, Cycadales and the primitive dicotyledonous species. On the other hand Kribs’ Heterogeneous Type III which exhibits procumbent cells with a single row of upright or square marginal cells is considered as a specialized form of rays. However, Barghoorn (1941) cautioned that when considering variations of ray structure for taxonomic purposes, ontogenic variations which might occur during development stages, should also be taken into account. Summarizing all the facts already mentioned on rays, Metcalf and Chalk (1989) further added that uniseriate rays are diagnostically important because of their restricted occurrence and that this feature is particularly characteristic of certain taxa that are phylogenetically advanced but it tends to be typical of individual genera rather than particular species. The present study reveals that, almost all the Sri Lankan Diospyros species possess the exclusively uniseriate rays (Table 3) and a few of them possess rays with multiseriate portions as wide as the uniseriate portions (Fig. 2 c, f) and this may be used to split the genus in to two broad groups. However, Wright (1904) recognized three groups on the same basis, (a). medullary rays are one cell in tangential breadth, (b) medullary rays are more than one cell in tangential breadth, and (c) presence of special radial groups of parenchyma. The present study partly agrees with that of Wright (1904) on the basis that the Sri Lankan taxa belong to exhibit the first two categories. However, the third category with special radial groups sensu Wright (1904) was not clearly distinguishable from the findings of the present study. Based on these lines, it is observed that the Sri Lankan species of Diospyros under consideration have xylem rays of the Kribs’ Heterogeneous Type I and or Heterogeneous Type II. Kribs Heterogeneous Type III which is considered to be the most advanced type (Barghoorn 1940; 1941a), was not observed in any Sri Lankan species studied. Further, Kribs’ Homogenous types of rays which are considered more evolutionarily advanced than the Heterogeneous types were also not observed. On this basis, the Sri Lankan taxa cannot be considered as a highly advanced group. With respect to the Sri 124 Lankan species of Diospyros the taxa posses Kribs’ Heterogeneous Type I rays as observed in the two endemic species D. atrata Alston and D. quaesita Thw., are less specialized than the taxa that posses Kribs’ Heterogeneous Type II rays observed in the non-endemic species D. affinis Thw., D. ebenum Koenig., D. oocarpa Thw., and the endemics D. attenuata Thw., D. ebenoides Kosterm., D. hirsuta L., D. koenigii Kosterm., D. moonii Thw., D. oblongifolia (Thw.) Kosterm., D. thwaitesii Beddome. However, six non-endemic taxa namely, D. discolor Willd., D. malabarica (Desr.) Kostel., D. montana Roxb., D. ovalifolia Wight, D. racemosa Roxb., D. melanoxylon Roxb. and the six endemic taxa namely, D. acuminata (Thw.) Kosterm., D. acuta Thw., D. insignis var. parvifolia Kosterm., D. oppositifolia Thw., D. rheophytica Kosterm. and D. walkeri (Wight) Guerke possess both Kribs' ray Types I and II. With respect to evolutionary advancement, the above mentioned taxa could be considered as intermediates. There seem to be no significant evolutionary trends within the Sri Lankan taxa of the genus with respect to Heterogeneous ray Types of Kribs. The nonendemic taxa with Kribs’ ray Type II are restricted to the dry zone while the endemics of the same types are found in the wet zone. Further, the endemic taxa with a mixture of Heterogeneous Type of Kribs I and II are distributed in the wet zone while the non-endemics mainly occur in the dry zone. On this basis, the Heterogeneous ray Types seem to have some relationship with environmental factors which needs further investigation. Kribs in 1937 regarded the absence of axial parenchyma as a primitive character (Metcalf and Chalk, 1989). Metcalf and Chalk (1989), summarizing the work on axial parenchyma distribution by Kribs further stated that the number of cells in a single strand of parenchyma could be considered as an index of level of specialization. Based on the classification of axial parenchyma by IAWA (1989), Sri Lankan Diospyros species showed banded parenchyma of reticulate, scalariform or marginal types only. Carlquist (1975) considered the presence or absence of axial parenchyma as an ecological variation rather than any evolutionarily significant feature. On this basis the presence of axial parenchyma within the Sri Lankan taxa is of limited value in the delimitation of the taxa concerned. The vessel elements of all the species studied possessed simple perforation plates, end-walls transverse or slightly oblique, with short tails. Wood anatomical study of Diospyros Hence, it is clear that there is no value of such characters in delimitation of the species. Intervessel pits of all the species observed were minute, alternate, vestured, circular or oval in shape (Fig. 3 a). Vessel-ray pits were also similar to intervessel pits in size and shape throughout the ray cells (Fig. 3 c), in all the specimens examined except D. ferrea (Willd.) Bakh., which possess large vesselray pits with much reduced borders and horizontally or vertically oriented as well as with normal pits (Fig. 3 d). On this basis, D. ferrea (Willd.) Bakh. stands out from the other species of Diospyros. Based on morphological characters Kostermans (1981) maintained D. ferrea (Willd.) Bakh. as a distinct species where as White (1993) believed that, based on wood anatomical characters 125 this taxon should be placed under a separate section of the genus Diospyros. According to the differences found with respect to vessel-ray pits, it could be suggested that D. ferrea (Willd.) Bakh. be separated into a Section of the genus as suggested by White (1993). The formation of tyloses has been considered as an indication of evolutionary primitiveness of angiosperms (Bonsen and Kucera, 1990). Tyloses were observed most of the species studied except D. chaetocarpa, D. acuta, D. affinis, D. attenuata, D. discolor, D. hirsuta, D. insignis, D. moonii and D. oppositifolia. In general, the dry zone species of Diospyros possess more tyloses than the wet zone species. Figure 3. Wood anatomical characteristics of the genus Diospyros, (a) Intervessel pits of D. insignis x 1320, (b) Crystals in rays of D. ebenum. x 1320, (c) Vessel -ray pits of D. ebenum x 1320 and (d) Vessel- ray pits of D. ferrea x 1320. B. K. L. Wickremasinghe and Tissa R. Herat Chattaway (1955; 1956) and IAWA (1989), have pointed out that, the absence or presence and if present, the type and location of crystals are of taxonomic importance. Prismatic or rhombohedral crystals were observed in ray cells (Fig. 3 b) in all the species of Sri Lankan Diospyros taxa studied. The abundance of crystals varied within the species as well as within specimens from the same species. However, in D. attenuata Thw., D. insignis var. parvifolia Kosterm., D. malabarica (Desr.) Kostel., D. rheophytica Kosterm. and D. walkeri (Wight) Guerke crystals were found very rarely. Crystals were mostly chambered. The occurrence of tyloses and crystals is of limited value in delimitation of the Sri Lankan species of the genus concerned. Multivariate statistical analyses were applied in the present study to trace the possible relationships between anatomical and ecological features. Sneath and Sokal (1973) and Sokal and Rohlf (1981) have shown the importance of multivariate statistical techniques in numerical Taxonomy. Using multivariate techniques, Jacobseni (1979), Robertse et al. (1980) Khidir and Wright (1982) and Hedren (1990) successfully solved the problems in variation between Allium cernuum and Allium stellatum, wood anatomy of South African Acacia, systematics of Graminae and the African complex of Justicia striata respectively. The same approach has been made by Somaratne and Heart (2001) and Pathirana and Heart (2004) in establishing the relationships between species of the genus Calophyllum and the genus Garcinia and to elucidate the possible relationships among the species of Diospyros in the present study. The dendrogram derived from Cluster Analysis (CA) based on quantitative wood anatomical features of the Diospyros L. species showed no considerable grouping tendency within the genus (Fig. 4). This finding indicates that taking wood anatomical features one at a time and taking as whole characters simultaneously led to different conclusions. However, based on the results of cluster analysis, it could be stated that wood anatomical features seem to be of limited value in delimitation of taxa within the genus Diospyros L. if they are treated individually. Further analysis with other vegetative anatomical data from leaf, young stem and node might be of importance. In the Principle Component Analysis (PCA), the characters such as pore distribution frequency 126 (DFPO), number of vessel groups per area (NGPA), vessel element length (TVEL) and fiber diameter (FDIA) are heavily loaded along the PCA Axis one and ray height in number of cells (RHCE) and in millimeters (RHMM), and ratio of vessel group to solitary (GSRT), highly influenced the axis two (Table 4). The total percentage variation obtained by the analysis is 71.4 % for the data recorded. The percentage variability explained by axis one and two is 42.3 %. The rest of the components are less significant and could be neglected. Further Table 4 indicates that these wood anatomical characters which are heavily loaded along the components could be considered as important features in separation of species within the genus. The scatter plots derived from the PCA with the idea of tracing the grouping tendency within the genus are shown in the Fig. 5 and 6. The scatter plots PC 1 Vs PC 2 (Fig. 5) indicate a weak tendency of grouping of specimens. Species such as D. insignis and D. walkeri were separated as two species. The analysis showed that the above mentioned species are to a certain extent, related wood anatomically and further, closely related to D. quaesita Thw. This agrees with Kostermans’ (1981) who observed similarity between these two species based on the colour of the wood and Trimen’s (1893) observation based on fruit characteristics. The occurrence of plumbagin, Scopaletin and a tri-oxygenated Coumarin in these two taxa, further confirmed their close relationship (Jayasinghe P., 1992, Unpublished data). The specimens of D. oppositifolia fall into an isolated position in scatter plots (No. 22 in Fig. 5 Fig. 6). This is the only species which possesses opposite leaves within the genus except D. melanoxylon which possesses sub-opposite leaves. In geographical distribution, it is highly restricted only to the top of southwest face of Haycock mountain in Galle district, at the altitude of 700 m. Wood anatomically, it is difficult to establish any living relationships between D. oppositifolia and the rest of the species within the genus. A group of species which consists of D. oocarpa, D. montana and D. ebenum (21, 17, 9 in Fig. 6) could be considered as closely related species, share the same habitats and are of arborescent habit. This could be attributed to the adaptability of the wood anatomical features to a particular habitat. Wood anatomical study of Diospyros 127 Rescaled Distance Cluster Combine 0 5 10 15 20 25 ┼─────────┼─────────┼─────────┼─────────┼─────────┼ DATTR DEBNU ATTNU DDISC DHIRS DACUT DEBNU DCHAE DEBNU DOVAL DATTR DOPPO DINSI DFERR DFERR DACUM DINVP DMOON DOOCA DOPPO DMONT DRACE DINSI DKOEN DMALA DMALA DOBLO DWALK DWALK DOBLO DOOCA DNUMM DACUM DOBLO DDISC DOOCA DQUES DINSI DEBNU DWALK DEBNO DHIRS DMONT DMALA DMELA DOVAL DINVP DRACE DRACE DEBNO DMALA DRHEO DTHWA DAFFI DATTR DAFFI DFERR ─┐ ─┤ ─┼─┐ ─┤ ├─┐ ─┘ │ │ ─┬─┘ │ ─┘ ├─────┐ ─────┤ │ ───┬─┘ │ ───┘ │ ───┐ │ ───┼─────┐ │ ───┘ ├─┤ ───┬───┐ │ │ ───┘ ├─┘ │ ───────┘ ├───┐ ─────────┐ │ │ ─────────┤ │ │ ─┬─┐ │ │ │ ─┘ │ │ │ │ ───┤ │ │ │ ───┼───┐ │ │ │ ───┘ ├─┼─┘ │ ───┬─┐ │ │ │ ───┘ ├─┘ │ │ ─────┘ │ │ ─┬───┐ │ │ ─┘ ├─┐ │ │ ─────┘ │ │ │ ───┐ │ │ │ ───┼─┐ │ │ ├─┐ ───┘ │ │ │ │ │ ─┬─┐ ├─┤ │ │ │ ─┘ │ │ │ │ │ │ ───┤ │ │ │ │ │ ─┐ ├─┘ │ │ │ │ ─┼─┤ ├─┘ │ │ ─┘ │ │ │ │ ───┤ │ │ │ ───┘ │ │ ├───┐ ─────┬─┤ │ │ │ ─────┘ │ │ │ │ ───────┘ │ │ │ ─┬─┐ │ │ │ ─┘ ├─────┐ │ │ ├───────────┐ ───┘ ├─────┘ │ │ │ ───┬─────┘ │ │ │ ───┘ │ │ ├───────┐ ─────────────────┘ │ │ │ ───────────┬─────────┘ │ │ ───────────┘ │ ├───┐ ─────┬───────────────────────────┘ │ │ ─────┘ │ ├───┐ ─────────────────────────┬───────────────┘ │ │ ─────────────────────────┘ │ │ ─────────────────────────────────────────────┘ │ ─────────────────────────────────────────────────┘ Figure 4. Dendrogram based on 13 wood anatomical characters of 58 Diospyros L. specimens. B. K. L. Wickremasinghe and Tissa R. Herat 128 Table 4. Principal Component Analysis of 13 quantitative wood anatomical characters of the genus Diospyros L. ─────────────────────────────────────── Component PC1 PC2 PC3 PC4 PC5 ─────────────────────────────────────── % variability explained 25.8 16.5 11.3 10.0 7.8 ────────────────────────────────────── Cumulative% 25.8 42.3 53.6 63.6 71.4 ─────────────────────────────────────── DFPO .85444 .01995 -.27438 .26724 -.01394 NGPA .82585 -.07640 -.22702 .26563 .03874 TVEL -.80365 -.10458 -.03101 .18217 .13401 FDIA -.68512 .22392 -.08122 .21299 -.15597 RHCE -.08323 .88622 .08822 -.06011 .07625 RHMM -.14091 .83826 .21480 .10138 .14959 GSRT -.36092 -.50267 .35952 .11301 .15506 RPMM .09078 .13689 .78069 .12307 -.27920 FLEN -.16790 .14691 .69696 .06438 .14092 NSPA .30555 .35471 -.54965 .28028 -.33151 VDIA -.05679 .05974 -.10736 -.84012 -.04924 VWTH .41839 -.32269 -.02985 -.43411 .33830 RWMM .03499 .17232 -.00043 .03343 .88867 ─────────────────────────────────────── Figure 5. PCA Scatter Diagram produced by plotting the first PC against the second PC for wood anatomical characters. Wood anatomical study of Diospyros 129 Figure 6. PCA Scatter diagram produced by plotting the first PC against the third PC for wood anatomical characters. D. ovalifolia which is widely distributed in the dry zone falls into a separate cluster (23 in Fig. 5 and Fig. 6) which is isolated from the rest of species. PCA of wood anatomical characters clearly indicates that D. ovalifolia is easily distinguished from the rest of the species of the genus. Further analysis shows that the rest of the species form a large cluster and reflects their wood anatomical similarities. The results of the Discriminant Analysis (DA) of quantitative wood anatomical data (Table 5) showed that, 8 out of 13 characters are significant at P < 0.05, indicating that these characters are more important in separating the species. Based on the outcome of the Discriminant Analysis, three discriminant scores were calculated for each specimen. Using the discriminant scores derived, two scatter plots were produced viz. discriminant function one with discriminant function two and discriminant function one with discriminant function three respectively (Fig. 7 and Fig. 8). The specimens of D. atrata and D. malabarica are grouped together showing their wood anatomical similarity (4 and15 in Fig. 5). D. atrata is an endemic and restricted to the intermediate and the wet zone of the island. According to Kostermans (1981), D. malabarica is widely distributed in the dry and the intermediate parts of Sri Lanka and also in India, Thailand, Indo-China and Western Malaysia. Based on the above facts, it could be tentatively concluded that D. malabarica would be the basic stock from which the other Diospyros species arose through the endemic D. atrata. Separate grouping pattern of D. oocarpa and D. montana, isolated positioning of D. ovalifolia and the same grouping pattern of D. walkeri were also reflected in both PCA and DA. The rest of the species falls into a single large group (Fig. 7 and Fig. 8) showing the ambiguity of their interrelationships. Results of the Discriminant Analysis of Qualitative data showed that out of 13 characters 7 were significant at P < 0.05 in the separation of the species within the genus (Table 6). The results indicate that intervascular pitting does not significantly vary within the genus and could be neglected in the analysis. Scatter plots derived from the Discriminant Analysis clearly showed that there is no particular grouping pattern in the data concerned. The results of the analysis with wood anatomical data and ecological origin assigned to the specimens showed that pore distribution frequency, fiber to vessel length ratio, number of groups per area, vessel element length and vessel wall thickness significantly (P < 0.05) vary with the ecological condition of the habitat (Table 7). The scatter plots derived from this analysis reflected B. K. L. Wickremasinghe and Tissa R. Herat only the dependence of wood anatomical characters on ecological conditions of the habitat. The intermediate zone specimens showed a tendency of merging with the dry zone specimens rather than the wet zone specimens. Based on the above findings it could be concluded that the wood 130 anatomical variation within the genus is of less importance in delimitation of the taxa. Further, these results are in compliance with findings which have already been discussed under individual characters. Figure 7. Scatter diagram produced by plotting Discriminant Function one against Discriminant Function 2 for quantitative wood anatomical characters Figure 8. Scatter diagram produced by plotting Discriminant Function one against Discriminant Function three or quantitative wood anatomical characters. Wood anatomical study of Diospyros 131 Table 5. Discriminant Analysis of 13 quantitative wood anatomical characters of the genus Diospyros L. Shows significance of quantitative wood anatomical characters used in Discriminant analysis (Wilks' Lambda (U-statistic) and univariate F-ratio with 27 and 29 degrees of freedom). ─────────────────────────────────────── Variable Wilks' Lambda F Significance ─────────────────────────────────────── DFPO .09906 9.7682 .0000 FDIA .21212 3.9894 .0002 FLEN .49479 1.0967 .4026 NGPA .16245 5.5377 .0000 NSPA .35452 1.9556 .0398 RHCE .19779 4.3563 .0001 RHMM .31332 2.3540 .0130 RPMM .73652 0.3842 .9927 RWMM .67344 0.5208 .9539 TVEL .20776 4.0958 .0002 VDIA .40418 1.5834 .1137 VWTH .31791 2.3045 .0149 GSRT .38641 1.7056 .0808 ─────────────────────────────────────── Table 6. Discriminant Analysis of 13 qualitative wood anatomical characters of the genus Diospyros L.. Shows significance of qualitative wood anatomical characters used in Discriminant analysis (Wilks' Lambda (U-statistic) and univariate F-ratio with 27 and 29 degrees of freedom). Variable Wilks' Lambda F Significance CRYS .29530 2.5631 .0073 FIBT .41912 1.4886 .1477 GRRI .25743 3.0982 .0018 PABA .00333 321.0351 .0000 PADI .07250 13.7418 .0000 PEPL .36352 1.8806 .0492 PITS is a constant. PIVE .67857 0.5088 .9593 RAGE .50893 1.0364 .4609 RASE .60455 0.7026 .8202 TRAC .22814 3.6339 .0005 TYLO .24449 3.3191 .0010 VEDI .24359 3.3353 .0010 ───────────────────────────── Oever et al., (1981) and Carlquist and Hoekman (1985) working with Symplocaceae and Staphylaceae respectively showed significant correlations among wood anatomical characters, such as vessel wall thickness to vessel diameter, vessel diameter to distribution frequency, vessel element length to fiber length and tracheary element length and vessel element length to ray height. B. K. L. Wickremasinghe and Tissa R. Herat 132 Table 7. Discriminant Analysis of wood anatomical characters against ecological zones of the genus Diospyros L. Shows significance quantitative wood anatomical data in relation to ecological origin of specimens ( Wilks' Lambda (U-statistic) and univariate F-ratio with 2 and 54 degrees of freedom). Variable Wilks' Lambda F Significance DFPO .53284 23.6723 .0000 FDIA .86836 4.0931 .0221 FLEN .96153 1.0802 .3468 FVRAT .77686 7.7554 .0011 NGPA .58267 19.3388 .0000 NSPA .91300 2.5727 .0857 RHCE .96929 .8555 .4307 RHMM .92733 2.1159 .1304 RPMM .90015 2.9949 .0584 RWCE .90975 2.6786 .0778 RWMM .97369 .7296 .4868 TVEL .67061 13.2620 .0000 VDIA .93179 1.9764 .1485 VWTH .76149 8.4569 .0006 Somaratne and Heart (2001) showed a higher significant correlation at P < 0.001, among vessel element length, fiber length and tracheary element lengths and significant correlation at P < 0.01 between height of ray and vessel element lengths; vessel diameter and fiber diameter in Calophyllum L. species in Sri Lanka. The correlation coefficients between fiber diameter and vessel element length and fiber diameter and number of vessel groups per area were significantly correlated at P < 0.05 in the present study. A correlation coefficient at P < 0.05 level was observed for fiber diameter and vessel wall thickness. These findings do not support the findings of the above mentioned work for the genus and further studies are needed to make conclusions. Evidence is available to a certain extent that there is a relationship between wood anatomical variation and latitudes and latitudes (Metcalf and Chalk 1989; Graff and Baas 1974; Oever et al., 1981). genus Diospyros L. in relation to altitude isnegligible. This may be due to the narrow altitudinal distribution of the species studied as all the specimens collected occurred in the area below 700 m in elevation. Graff and Baas (1974) pointed out that the elevation differences of the specimens should be at least more than 2000 m to obtain a significant dependency of wood anatomical characters on altitude. Significant correlation coefficients were obtained for mean annual rainfall and vessel element length (P < 0.05) and mean annual rainfall and vessel distribution frequency (P < 0.05). A strong correlation coefficient has also been observed among wood anatomical characters with mesomorphy and conductance. The results of the study showed a decrease in vessel element length and vessel diameter with decrease of water availability (Fig. 9 ). These are in agreement with findings of Carlquist and De Buhr (1977). Figure 10 shows the relationships among mesomorphy indices, conductance and mean annual rainfall. Wood anatomical study of Diospyros Figure 9. The relationship among rainfall, vessel element length and vessel diameter. Figure 10. The relationship between rainfall, mesomorphy and conductivity. 133 B. K. L. Wickremasinghe and Tissa R. Herat It is evident that there is a clear relationship between mesomorphy index and also with conductance i.e. an increase in mean annual rainfall causes an increased mesomorphy index and conductance. As these findings agree with the idea of Carlquist and De Buhr (1977), both mesomorphy indices and conductance values could be used to ascertain the ecological origin of a species of the genus under consideration. The relationships among agro-ecological zones versus mesomorphy and conductance values also show a linear pattern. Carlquist (1984) pointed out that vessel grouping and presence of vasicentric tracheids play an important role in water conductive efficiency and believed that in case of air embolism of vessel elements, the neighboring vasicentric tracheids which will resist the spreading of air bubbles along the lumen could be of importance in maintaining the function of the disabled vessel element. Present study reveals that, vessel grouping is greater in dry zone specimens rather than wet zone species. Further, appearance of tracheids is sporadic and it could be believed that these structures are of certain value in compensating for the disabled vessel elements. Lack of tracheids may be considered as a contributive factor which increases vessel grouping of the dry zone species of the genus. Further, it is observed that in the dry zone species, vessel groups consist of vessels with different diameters. Functionally this arrangement of vessels will be a mechanism to maintain the continuous water flow in case of air embolism of larger vessels which are susceptible to such conditions. This structural organization of the vessel arrangement could be correlated with the drought resistance ability of some species of the Diospyros taxa of Sri Lanka. Further studies on these lines are required for better understanding of the ecological adaptive ability of wood anatomical characters. As a whole, it should be pointed out that the wood anatomical characters of the taxa of the genus Diospyros are highly correlated with the environmental conditions of the habitat, especially with water availability. Characteristic features of vessel elements and fiber lengths cannot be used as an indication of evolutionary advancement of the Sri Lankan endemics of the genus. Such evolutionary trends cannot be traced among the endemic taxa in order to establish evolutionary relationships between the endemic and non-endemic Sri Lankan Diospyros species. Further studies on comparative anatomical studies of other vegetative and reproductive parts may be of importance in delimitation of Diospyros in Sri Lanka. ACKNOWLEDGEMENTS 134 Authors thank the colleagues who helped during field collections and in preparation of the manuscript, Mr. S. Somaratne for statistical anmalysis and the laboratory staff of the Department of Botany for laboratory work. REFERENCES Alston, A.G.H. (1931). Supplement to the Trimen’s Handbook to the Flora of Ceylon, Part VI. Dalau & Co. Ltd. London. Aloni, R., and Zimmermann, M.H. (1983). The Control of vessel size and density along the Plant Axis. Differentiation 24: 203-208. Baas, P. (1976). Some functional and adaptive aspects of vessel member morphology. In: P. Baas, A.J. Bolton & D.M. Catling (Eds.), Wood Structure in Biological and Technological Research, Leiden University, Leiden Botanical Series No. 3: 157181. Baas, P. (1982). Systematic, Phylogenetic, and Ecological Wood anatomy. In History and new Perspectives in Wood Anatomy. P. Bass, (Ed.), Martinus Nijhoff / Dr. W. Junk Publisher, The Hague, Pp.23-58. Baily, I. W. (1957). The potentialities and limitations of wood anatomy in the study of the phylogeny and classification of angiosperms. Journal of Arnold Arboratum 38: 243-251. Baretta-Kuipers, T. (1976). Comparative wood anatomy of bonnetiaceae, Theaceae and Guttiferae. In Wood Structure in Biological and Technological Research, P. Bass, A.J. Bolton and D. M. Catling (Eds.). Leiden University, Leiden Botanical Series No. 3: 76-101. Barghoorn, E. S. Jr. (1940). The Ontogenic development and phylogenetic specialization of rays in the xylem of Dicotyledons I, The primitive ray structure. American Journal of Botany 27(10): 918-928. Barghoorn, E. S. Jr. (1941a). The Ontogenic development and phylogenetic specialization of rays in the xylem of dicotyledons II, Modification of the multiseriate and uniseriate rays. American Journal of Botany 28 (4): 273-282. Wood anatomical study of Diospyros Bhargavan, S. and Ramachandran, M. (1984). Diospyros trichophylla Alston. (Ebenaceae) - A new record for India. Indian Journal of Forestry. Botanical. Survey of India 7 (3): 247-248. Bonsen, K.J.M. and Kucera, L. J. (1990). Vessel occlusion in plants: Morphological and evolutionary Aspects. International Association of Wood Anatomists Bulletin II (4): 393-399. Burley, J. and Miller, R. B. (1989). The application of statistics and computing in wood anatomy. In Anatomy of Dicotyledons, Vol.II, 2nd ed., C.R. Metcalf & L. Chalk, (Eds), Clarendon Press, Oxford, Pp.74-179. Burmann, J. (1737). Thesaurus Zelanicus exhibiens Plants in insule Zeylana Nascentes. Janssoniowassbergions & Salonem Schunten. Amsterdam, Pp.121. Carlquist, S. (1975). Ecological strategies of xylem evolution. University of California Press. Burkley, California. Carlquist, S. and De Buhr, L. (1977). Wood anatomy of Penaceae (Myrtales) : Comparative, phylogenetic, and ecological implications. Botanical Journal of Linnaean Society 75: 211277. Carlquist, S. (1984). Vessel grouping in Dicotyledon wood : Significance and relationship to imperforate tracheary elements. Aliso 10 (4): 505-525. Carlquist, S. (1986a). Terminology of imperforate tracheary element. International Association of Wood Anatomists Bulletin. n.s. 7(1) : 78-81. Carlquist, S. (1986b). Terminology of imperforate tracheary element. International Association of Wood anatomists Bulletin n.s. 7(2): 168-170. Carlquist, S. and Hoekman, D. A. (1985). Ecological wood anatomy of the woody Southern Californian Flora. International Association of Wood Anatomists Bulletin. n.s., 6 (4): 319-347. Chalk, L. and Chattaway, M. M. (1934). Measuring the length of vessel members. Tropical Woods 40: 19-26. Chattaway, M. M. (1932). Proposed standards for numerical values used in describing woods. Tropical Woods 29 : 20-28. 135 Chattaway, M. M. (1955). Crystals in woody tissue. Part I. Tropical Woods. 102 : 55-74. Chattaway, M. M. (1956). Crystals in woody tissue. Part II. Tropical Woods 104 : 100 -124. De Candolle, A. (1844). Ebenaceae, Prodromus Systematis Naturalis. Regni vegetablis. Sumptibus Fortin, Masson et Sociorum, Parisiis; 8: 209-243. Fahn, A., Burley, J., Longman, K.A., Mariaux, A., and Tomlinson P.B. (1985). Possible contribution of the wood wnatomy to the determination of the age of tropical trees, In ; A. Fahn, Werker, E., and P. Baas (eds.), Wood Anatomy and Identification of Trees and Shrubs from Israel and Adjacent Regions, The Israel Academy of Science and Humanities, Pp.31-54. Gauch, H.G. Jr. (1986). Multivariate Analysis in Community Ecology (Cambridge Study series in Ecology). Cambridge University Press, Cambridge, Pp. 298. Graff, N. A. van der and Bass, P. (1974). Wood anatomical variation in relation to latitude and altitude. Blumea 22: 101-121. Hedren, M. (1990) The Justicia eleganta complex in tropical Africa (Justicia. Sect. Hannieria, Acantheceae). Nordic Journal of Botany (Section of Tropical Botany) 10: 357-398. Hermann, P. (1717). Musaeum Zelanicum, sive catalogus plantarum. Lugduni Batavorum, Isaac Severinus, Pp.21 Hora, B. (1981). The Oxford Encyclopaedia of trees of the world. Oxford University Press. IAWA Committee (1989). IAWA list of microscopic features for hard wood Identification In; E.A. Wheeler, P. Baas and P.E. Gasson (Eds.), IAWA Bull. n.s. 10 (3): 219-332. Irrigation Department (Land Use Division). (1988). Soil Map of Sri Lanka (Agro-Ecological Regions), Department of Cartography ITC, Irrigation Department, Colombo. Jacobseni, T. D. (1979). Numerical analysis of variation between Allium cernum and Allium stellatum (Liliaceae). Taxon 28 (5/6): 517-523. Jayasuriya, A. H. M. (1998). A new species of Diospyros (Ebanaceae). Journal of South Asian Natural History 3 (1): 55-58. B. K. L. Wickremasinghe and Tissa R. Herat Johansen, D. A. (1940). Plant microtechnique. McGraw-Hill, USA. Kdhir, W. H. and Wright, K. (1982). Systematics of Graminae: A cluster analysis study. Taxon 31 (1): 9-36. Kostermans, A. J. G. H. (1977). Notes on Ceylonese Ebony Trees (Ebenaceae). Ceylon Journal of Science (Bio. Sci.) 12 (2): 89-109. Kostermans, A. J. G. H. (1981). Ebenaceae. In: M.D. Dassanayake and F.R. Fosberg, (Eds.), A Revised HandBook to the Flora of Ceylon, Vol. III: 1-54, Amerind, New Delhi. Kramer, P. J. (1964). The Role of Water in Wood Formation. In M. H. Zimmermann, (ed.), The Formation of Wood in Forest Trees Academic Press, New York, Pp.519-532. Kribs, D. A. (1935). Salient lines of structural ever, L. van der, Bass, P. and Zandee, M. (1981). Comparative wood anatomy of Symplocus and latitude and altitude of provenance. International Wood Anatomists Bulletin n.s. 2 (1): 2-24. Pathirana, P.S.K. and Heart, T.R. (2004). Comparative Vegetative Anatomical Study of the Genus Garcinia L. (Clusiaceae / Guttiferae) in Sri Lanka. Ceylon Journal of Science (Bio. Sci.) 32: 39-66. 136 Sokal, R. R. and Rohlf, F.R. (1981). Biometry. The Principles and Practice of Statistics in Biological Research, 2nd ed. W. H. Freeman & Co., New York, Pp.880. Somaratne, S. and Heart, T.R. (2001). Comparative vegetative Anatomical Study of the Genus Calophyllum L. (Clusiaceae) in Sri Lanka. Ceylon Journal of Sciences (Bio. Sci.) 28: 51-80. SPSS Inc. ( 1990). SPSS Statistical Data Analysis (SPSS/PC+ Statistics(TM) 4.0 for the IBM PC/XT/AT and PS/2). SPSS Inc., Michigan Avenue, Chicago, Illinois, USA, Pp.B1-E6. Survey Department ( 1992). Topographic Sheets (one inch) of Sri Lanka. Revised in 1973, Reprint 1992. Survey Department, Colombo. Thwaites, G.H.K. (1864). Enumeratio Plantarum Zeylaniae, Dalau & Co. Ltd. London, Pp.178-183. Tippo, O. (1964). The role of wood anatomy in phylogeny. American Midland Naturalist 36: 362372. Trimen. H. (1893). A Hand-Book to the Flora of Ceylon, Part III (Indian reprint, 1984). Bishen Singh Mahendra Pal Sing, Dehradun. 87-102. Vliet, G. J. C. M. van (1979). Wood anatomy of Combretaceae. Blumea 25: 141-223. Robertse, P. J., G. Venter and Janse van Rensberg (1980). The wood anatomy of the South African Acacia. International Wood Anatomists Bulletin n.s. 1 (3): 93-103. Wheeler, E.A. (1986). Vessels per Square Millimeter or Vessel Group per Square Millimeter? International Association of Wood Anatomists Bulletin. n.s. 7 (1) : 73-74. Sass, E. J. (1958). Plant Microtechnique, 3rd ed., Iowa State University Press, Ames, Iowa, U.S.A. White, F. (1993). Twenty two new and little-known Species of Diospyros (Ebenaceae) from New Caledonia with comments on section Maba. Bulletin de Museum Naturel Histoire, Paris, 4e Ser., 14, Section B, Adansonia, no 2, 179-222. Sneath, P. H. A. and Sokal, R.R. (1973). Numerical Taxonomy. W. H. Freeman & Co., San Francisco, Pp.573. Snedecor, G. W. and Cochran, W.G. (1967) Statistical Methods, 6th ed. Iowa State University Press, Ames, Iowa, U.S.A, Pp.593. Worthington, T.B. (1959). Ceylon Trees. The Colombo Apothecaries Co. Ltd., Colombo. Wright, H. (1904). The genus Diospyros in Ceylon; Its Morphology, Anatomy and Taxonomy. Annals of Royal Botanic Gardens, Peradeniya. II: Part I. 1-106, Part II. 132-210.