FOLIAR EPIDERMAL , STEM AND PETIOLE ANATOMY OF MEGHALAYAN DIOSCOREA L . ( DIOSCOREACEAE ) AND ITS SYSTEMATIC IMPLICATION

A comparative anatomical study of eight species of Dioscorea L. from Meghalaya, North East India was carried out in order to evaluate the taxonomic significance of anatomical characters to differentiate the species. Characters were coded and analyzed by PCA and cluster analysis. The combination of selected qualitative and quantitative anatomical characters of foliar epidermis, stem and petiole were significant for identification of species. The characters that contributed most to the separation of the species were type of stomata, length of stomata, stomatal index, leaf epidermal hairs, stem epidermal hairs, layer of stem sclerenchyma, number of vascular bundle in outer ring and inner ring of stem, paired or unpaired metaxylem, presence of phloem at both ends or at one end, presence or absence of starch grain in stem, petiole epidermal hair and presence or absence starch grain and crystal in petiole. An indented dichotomous key based on anatomical characters was constructed to distinguish and identify the species. Introduction Dioscorea L., with about 602 species (Coursey, 1967) in the tropical and subtropical regions, regarded as the core genus of the family Dioscoreaceae is a pivotal taxon in the evolution of Liliopsida and occupies a basal position among all extant monocotyledonous plants (Chase et al., 2006; Dahlgren, 1989). About 50 species of Dioscorea in India (Anonymous, 1952) and approximately 28 species are distributed in North East India (Sharma and Hore, 1995). Meghalaya, one of the eight hot spots of North East India has wide range of distribution of Dioscorea species in wild habitat. Only a few species of Dioscorea are consumed as food and also used as medicines by the indigenous people of this area in spite of its wide range of distribution. Identification of Dioscorea species has always presented a challenge to taxonomists due to its morphological diversity, dioecy and small flowers. Kunth (1924) divided Dioscorea into four subgenera based on seed morphology. Burkill (1960) divided the genus into 23 sections based on seed characters, tuber characters and male inflorescence morphology. Coursey (1967) divided the genus into 70 sections based on underground tuber. Several authors studied the floristics of the genus including Miege (1968), Milne-Redhead (1975), N’Kounkou (1993), Tellez and Schubert (1994), Miege and Sebsebe (1998) and Ding and Gilbert (2000). Morphological approaches provide a workable system of taxonomy, yet it cannot be denied that data from other field like anatomy, cytology, palynology, molecular studies etc., provide evidences for the accurate delimitation of the species. Anatomical parameters play an important role in plant taxonomy (Metcalfe and Chalk, 1957). Anatomical characters have proved to be more useful for delimitation of higher taxonomic ranks, such as genera and families. Few researchers studied foliar epidermal anatomy of different species of Dioscorea (Abdulrahman et al., 2009; Corresponding author. Email: nilofersheikh83@gmail.com DOI: http://dx.doi.org/10.3329/bjpt.v24i1.33033


Introduction
Dioscorea L., with about 602 species (Coursey, 1967) in the tropical and subtropical regions, regarded as the core genus of the family Dioscoreaceae is a pivotal taxon in the evolution of Liliopsida and occupies a basal position among all extant monocotyledonous plants (Chase et al., 2006;Dahlgren, 1989).About 50 species of Dioscorea in India (Anonymous, 1952) and approximately 28 species are distributed in North East India (Sharma and Hore, 1995).Meghalaya, one of the eight hot spots of North East India has wide range of distribution of Dioscorea species in wild habitat.Only a few species of Dioscorea are consumed as food and also used as medicines by the indigenous people of this area in spite of its wide range of distribution.Identification of Dioscorea species has always presented a challenge to taxonomists due to its morphological diversity, dioecy and small flowers.Kunth (1924) divided Dioscorea into four subgenera based on seed morphology.Burkill (1960) divided the genus into 23 sections based on seed characters, tuber characters and male inflorescence morphology.Coursey (1967) divided the genus into 70 sections based on underground tuber.Several authors studied the floristics of the genus including Miege (1968), Milne-Redhead (1975), N'Kounkou (1993), Tellez and Schubert (1994), Miege and Sebsebe (1998) and Ding and Gilbert (2000).
Morphological approaches provide a workable system of taxonomy, yet it cannot be denied that data from other field like anatomy, cytology, palynology, molecular studies etc., provide evidences for the accurate delimitation of the species.Anatomical parameters play an important role in plant taxonomy (Metcalfe and Chalk, 1957).Anatomical characters have proved to be more useful for delimitation of higher taxonomic ranks, such as genera and families.Few researchers studied foliar epidermal anatomy of different species of Dioscorea (Abdulrahman et al., 2009;Aina and Atumeyi, 2011;Shah and Gopal, 1972), however, no concrete data of anatomical analysis used in taxonomic delimitation of this genus have been conducted earlier.Therefore, in the present study an attempt has been made to investigate the anatomical features of foliar epidermis, stem and petiole of different species of Meghalayan Dioscorea and to single out distinctive anatomical characters potentially useful for infrageneric classification.

Plant material
Fresh

Foliar epidermal anatomy
For foliar epidermal study, fresh leaf epidermal peelings from both the surfaces (adaxial and abaxial) and transverse section of the leaves for epidermal study were made by hand with the help of sharp razor and forceps.Ten individual per species were studied.The epidermal peelings and the transverse section of the leaves were cut into suitable size, taken on a clean slide, stained with 5% aqueous safranine, mounted in 50% glycerin and sealed the margins of cover slips with DPX.The prepared slides were observed under a light microscope at x40 magnification and photographs were taken.To calculate the stomatal index, following formula was used:

I= X 100
Where I = stomatal index, S = number of stomata per unit area, and E = number of epidermal cells per unit area.The terminology adopted by Metcalfe and Chalk (1950) and Metcalfe (1961) was followed to describe the stomatal types.

Stem and petiole anatomy
For the anatomical study of stem and petiole, transverse section of stem and petiole of 10 individual plants per species were made using sharp blade from the fresh material.Temporary slides were prepared following O'Brien et al. (1964) using Toluidine blue.The stained sections were observed under a light microscope at x10 magnification and Photographed.Presence and absence of starch grain in stem and petiole was separately analyzed by Lugol's solution (Jensen, 1962).The anatomical characters observed under microscopes were recorded.

Anatomical data recording and statistical analysis
A total of 27 characters including both qualitative and quantitative were observed (Table 1).The characters or traits that appeared in more than one state and differed among the examined species were coded and assessed through principal component analysis (PCA).Cluster analysis (CA) was also performed in order to group the species based on the similarity of anatomical characters or traits.Both PCA and CA were performed using XLSTAT ver.2015.4.01.22283 statistical software.

Foliar epidermal anatomy
The upper (adaxial) and lower (abaxial) epidermis has one layer of cells and each is covered by a cuticle.The outer epidermis is also characterized by the presence of hairs in some species.Stomata are restricted only to the lower surface (hypostomatic).Four types of stomata complex were identified namely anamocytic, tetracytic, paracytic and anisocytic (Fig. 1).D. alata possessed anisocytic, anamocytic and tetracytic stomata; D. pubera possessed paracytic and tetracytic stomata; D. belophylla, D. pentaphylla and D. bulbifera had paracytic, anisocytic and anamocytic types of stomata; D. glabra with anamocytic and tetracytic stomata; D. lepcharum and D. oppositifolia possessed all the four types of stomata.In the present analysis, the stomatal index ranges from 15.3 in D. alata to 24.6 in D. bulbifera.

Stem anatomy
Transverse sections of stems are generally circular with longitudinal ridges or wings (Fig. 2).The epidermis consists of thin-walled rectangular, cuboidal or rounded cells.The epidermal cells are 1-cell layer thick.The cuticle is generally thin.In D. pubera some epidermal cells contained numerous hairs.The cortex lying just beneath the epidermis is composed of 3-7 layers of cells.The cortical cells are of various size and shapes.The inner boundary of cortex is located at a zone of sclerenchyma with 2-many cells.The pith occupies the central position and composed of thinwall hexagonal parenchyma cells.The vascular bundles of the stem are arranged in two concentric circles.The vascular bundles of the outer circle are smaller than the inner, with 2 metaxylem vessels together with 1 phloem unit at the middle in the bundles.The bundles of inner circle have mostly paired metaxylem vessels with phloem unit present at both ends in all species except but in D. bulbifera, where the bundles of inner circle have unpaired metaxylem with phloem unit at one end.The number of vascular bundles in transverse section varies within the species.

Petioles anatomy
The cortex is composed of collenchymatous tissues.The vascular bundles are arranged in a ring and are basically collateral with the presences of 2-3 phloem units in each bundle.The number of vascular bundles in each petiole is also variable but constant in each species (Fig. 3).

Cluster analysis
The result of cluster analysis based on unweighted pair group average method (UPGMA) initially produced two main clusters, cluster A with a single species D. belophylla and the other cluster B at 0.507 level of similarity (Fig. 4).

Principal component analysis
The first seven principal components explained about 100% of the variation, while the first component (PC1) alone counted for 29.26% of the variability (Table 2).Characters such as tetracytic type of stomata, presence or absence of leaf epidermal hair, length of stomata, presence of stem epidermal hair, number of vascular bundle present in outer ring of stem, number of schlerenchyma of in stem, presence of epidermal hair in petioles has the highest loading on PC1.The second component (PC2) explained 20.84% of the total variation was highly correlated with paracytic type of stomata, number of stomata, layers of cortex in stem, presence of paired or unpaired metaxylem in stem, presence of phloem at one end or at both ends and number of vascular bundles in petioles.The remaining component explained less variability.To assess the scores of the individual species with its characters or traits, PC1 and PC2 were plotted (Fig. 5).The group of species on the left side of the biplot is separated from the species on the right side of biplot due to certain combination of anatomical characters, viz., anamocytic and paraytic types of stomata, presence of starch in petiole, number of vascular bundle in petiole, presence of metaxylem paired or unpaired in stem, presence of phloem at one end or at both ends in stem, whereas species on the right side of the plot is represented with characters such as presence of foliar epidermal, stem epidermal and petiole epidermal hairs, number of petiole schlerenchyma layers, number of stomata per unit area etc.The finding was consistent with the separation of species into four major groups by UPGMA clustering (Fig. 4).

Discussion
The use of anatomical characters or traits for taxonomic studies has proved useful for identification of fragmented plant and herbarium specimens (Metcalfe and Chalk, 1957).Anatomy can provide useful information for establishing interrelations between taxa at the species and supra species levels.Sometimes it can also help in individual identifications.The internal structure of leaf is more affected by environmental factors and thus is of little value for delimiting taxonomic groups.Other characters of leaf, such as the epidermis and stomata have proved to be much more reliable for taxonomic consideration in many genera (Uphof, 1962;Dickison, 2000;Yang and Lin, 2005;Strgulc-Krajsek et al., 2006).The petiole structure is of considerable taxonomic importance in many genera, since it is less affected by environmental changes (Metcalfe and Chalk, 1957).The result of the present study allows the selection of some diagnostic anatomical characters for the identification of Meghalayan Dioscorea species.Stebbin and Khush (1961) and Ayensu (1972) reported that all stomata in Dioscorea species were anomocytic which was found in consistent with Shah and Gopal (1972) and Abdulrahaman et al. (2009).Our results showing presence of different types of stomata including paracytic, tetracytic and anomocytic etc were found congruent with those of Shah and Gopal (1972) and Abdulrahaman et al. (2009).The present study reveals an anamocytic character trait contributes much towards variability.Presence of starch grain in stem and petiole is also an important character trait for differentiating among the species (Onwuene, 1978).From the PCA analysis, the character trait i.e. presence of starch grain in petiole was another important trait which shows high rate of variability.Hence the foregoing analysis and presented illustration clearly show the importance of anatomical data as an additional tool in the taxonomy of the genus and also contributes to the identification of species.The results provided some useful features for future phylogenetic and taxonomic studies.Based on these features, an artificial indented dichotomous key is presented hereunder to delimit the species.
Within the cluster B, two subcluster C and D are separated at similarity level of 0.54.Subcluster C with D. bulbifera is separated from D. pentaphylla and D. oppositifolia.D. pentaphylla and D. oppositifolia forms a clade at 0.74 level of similarity.Similarly, subcluster D with D. glabra and D. pubera are separated from D. alata and D. lepcharum which forms a clade at 0.72 level of similarity.

Fig. 5 .
Fig. 5.A biplot based on first and second principal components of anatomical characters for eight Dioscorea species.

Dioscorea based on anatomical characters.
Layer of schlerenchyma in stem is more than 5; absence of starch grain in stem; vascular bundle in outer ring in stem is more than 9 D. belophylla -Layer of schlerenchyma in stem is less than 5; presence of starch grain in stem; vascular bundle in outer ring in stem is less than 9 4 4. Inner ring of vascular bundle in stem is less than 6; metaxylem in stem is paired; phloem on both ends of vascular bundle D. pentaphylla -Inner ring of vascular bundle in stem is more than 6; metaxylem unpaired; phloem at one end of vascular bundle D. bulbifera 5. Presence of starch grain in stem; outer ring of vascular bundle more than 9 6 -Absence of starch grain in stem; outer ring of vascular bundle less than 9 7 6.Inner ring of vascular bundle more than 9; petiole outline mainly pentagonal; layer of cortex in petiole less than 4 D. alata -Inner ring of vascular bundle less than 9, petiole outline crescent shape; layer of cortex in petiole more than 4 D. lepcharum 7. Inner ring of vascular bundle in stem less than 8; presence of paracytic stomata; absence of petiole crystal, layer of cortex in petiole is more than 4 D. oppositifolia -Inner ring of vascular bundle more than 8; absence of paracytic stomata; presence of petiole crystal, layer of cortex in petiole is less than 4 D.glabra