MATERIALS AND METHODS

Study area: Collection of plant materials from 5 provinces in Indonesia, namely East Java, Central Java, West Java, Lampung and West Nusa Tenggara was conducted in January-March, 2019. Preparation of microscopic slides and data collection were carried out at the Plant Structure and Development Laboratory, Universitas Gadjah Mada in April and May, 2019. Data analysis was carried out in June, 2019.

Material collection: Materials used in this study were leaves of A. malaccensis and G. versteegii collected from different localities in Indonesia (Table 1). Plant samples are obtained from various sources including Indonesian Agarwood Association, Indonesian Agarwood Nursery Association, Bogor Botanical Garden, Agarwood Nursery and private collection from communities.

Experimental design: Qualitative and quantitative morphological characters of leaves were observed on fresh leaves, while venation and micromorphological data were examined form microscopic slides of cleared leaves.

Table 1:	List of Aquilaria malaccensis and Gyrinops versteegii samples
Ama: A. malaccensis, Gve: G. versteegii

Table 2:	List of leaf architecture characters

The microscopic slides were prepared using leaf clearing method of Ruzin²¹ with modifications. Fresh leaves were cut into 3 sections representing the upper, middle and lower laminar parts and soaked in 10% NaOH solution for 24-48 h. After 3 times rinsing with distilled water, the leaf fragments were then bleached in 5.25% sodium hypochlorite until the leaves became colorless. The leaf fragments were washed twice in water for 15 min, followed by dehydration in 70% ethanol overnight. The next process was incubating the leaf fragments in 50% chloral hydrate solution at 70°C and maintaining the temperature by placing the container on the hot plate until the leaf fragments became transparent. The staining process was done by soaking leaves in 1% safranin for 30-60 min. The subsequent dehydration process was done in a series of 70% ethanol, absolute ethanol and ethanol-xylol solution (1:1) each for 15 min, followed by final incubation at ethanol-xylol solution (1:1) for 30 min. The fully cleared leaf fragments were then mounted on glass slides using Canada balsam.

Observation: Observation on leaf venation and micro-morphological characters was performed using BOECO BM-180 binocular microscope at magnification of 4×10. The microscope was equipped with OptiLab microscope camera for digital image capturing. Twenty seven characters were used in the characterization of leaf architecture (Table 2).

Measurement: The leaf architecture data were recorded based on manual of leaf architecture²² with modifications and several additional characters developed by the authors. Leaf micromorphological characters examined in this study were trichomes and calcium oxalate. General features on leaf architecture of A. malaccensis and G. vesteegii were described accordingly.

Analysis: Data on leaf architecture was subjected to cluster analysis and principal component analysis to reveal the grouping pattern and role of the characters in distinguishing species. These multivariate analysis were performed using²³ MVSP version 3.1.

RESULTS

Laminar shape of A. malaccensis was oblong or ovate, symmetrical, with obtuse base, acuminate apex, entire margin and glabrous surface. The leaf texture was papyraceous or coriaceous, with laminar size varied from 1.382-2.646 mm². The leaf venation was pinnate, reticulodromus, irregular secondary vein spacing, the 3rd degree veins category was random reticular and the 4th degree veins were dichotomizing (Fig. 1). The areolation was moderately developed, the type marginal ultimate venation was fimbrial vein, with free ending ultimate veins was unbranched or 1-branched (Fig. 2). There were variations on midrib width, marginal vein width and the outermost laminar margin width (Fig. 3). Trichomes in lamina, midrib and leaf margin were simple and non-glandular. Two forms of calcium oxalate crystal were found on lamina, sand-type and raphide-type (Fig. 4). While all samples had calcium oxalate crystals on their lamina, this was not the case in midrib and leaf margin. Only leaves sample from Lampung showed the occurrence of calcium oxalate crystals in lamina, midrib and leaf margin.

Leaves of G. versteegii were elliptic or lancet, symmetrical, with obtuse base, acuminate apex, entire margin, glabrous surface and papyraceous or coriaceous texture.

Fig. 1(A-B):	(A-B) Venation pattern of Aquilaria malaccensis, showing 4 vein categories
	a: 1st degree vein, b: 2nd degree vein, c: 3rd degree vein, d: 4th degree vein

Fig. 2(A-D):	(A-D) Areolation and marginal vein characters of Aquilaria malaccensis
	a: A moderately developed areolation, b: Unbranched free ending ultimate veins, c: 1-branched free ending ultimate veins, d: Fimbrial vein marginal ultimate venation

There were no variations on qualitative characters of leaf venation in all samples, whereas the quantitative characters showed moderate variations on midrib width, marginal vein width and the outermost laminar margin width. The venation of G. versteegii was pinnate, reticulodromus, irregular spacing on 2nd degree veins, random reticulate on 3rd degree veins and dichotomizing 4th degree veins (Fig. 5). The areolation was moderately developed, with 1-branched free ending ultimate veins and a fimbrial vein type on the marginal ultimate venation (Fig. 6). Variations on midrib width, marginal vein width and the outermost laminar margin width were shown in Fig. 7. There was a notable variation on the occurrence of trichomes, in which only the sample form Probolinggo had simple, non-glandular trichomes on lamina, midrib and leaf margin.

Fig. 3(A-D):	(A-D) Variations on midrib width, marginal vein width and the outermost laminar margin width of Aquilaria malaccensis leaves
	a.1: Narrow midrib, a.2: Wide midrib, b.1: Narrow marginal vein, b.2: Wide marginal vein, c.1: Narrow outermost laminar margin, c.2: Wide outermost laminar margin

Fig. 4:	Calcium oxalate crystals in the lamina of Aquilaria malaccensis

The other five samples had no trichomes. The calcium oxalate crystals were found in lamina of all samples. Among six samples of G. versteegii, only leaves sample from Nganjuk had no calcium oxalate crystals on leaf margin, whereas leaves sample from Banjarnegara had calcium oxalate crystals on their midrib. The types of calcium oxalate crystals were similar to those found in A. malaccensis, namely sand-type and raphide-type (Fig. 8).

Fig. 5(A-B):	(A-B) Venation pattern of Gyrinops versteegii, showing 4 vein category
	a: 1st degree vein, b: 2nd degree vein, c: 3rd degree vein, d: 4th degree vein

Fig. 6(A-C):	(A-C) Areolation and marginal vein characters of Gyrinops versteegii
	a: A moderately developed areolation, b: 1-branched free ending ultimate veins, c: Fimbrial vein marginal ultimate venation

Fig. 7(A-D):	(A-D) Variations on midrib width, marginal vein width and the outermost laminar margin width of Gyrinops versteegii leaves
	a.1: Narrow midrib, a.2: Wide midrib, b.1: Narrow marginal vein, b.2: Wide marginal vein, c.1: Narrow outermost laminar margin, c.2: Wide outermost laminar margin

Fig. 8:	Calcium oxalate crystals in the lamina of Gyrinops versteegii

Cluster analysis on 27 characters of leaf architecture on A. malaccensis and G. versteegii showed clear separation of these two species as shown in dendrogram (Fig. 9). Comparing the grouping patterns of the two species, it was obvious that A. malaccensis showed higher diversity than G. versteegii as indicated by the taxonomic distance between their respective samples. Results of principal component analysis revealed seven characters which had considerable role in the grouping of samples as indicated by the loading value of higher than 0.3 (Table 3). These seven leaf architecture characters were leaf shape, laminar size, midrib width, trichomes on lamina, trichomes on midrib, calcium oxalate crystals density on lamina and length of raphide-type calcium oxalate crystals.

Fig. 9:	Dendrogram of Aquilaria malaccensis and Gyrinops versteegii samples based on 27 leaf architectural characters used in this study
	Ama: A. malaccensis, Gve: G. versteegii

Table 3:	Loadings of 27 leaf architectural characters resulted from principal component analysis
PC1: First principal component, PC2: Second principal component generated from principal component analysis (PCA) method

The scatter plot generated from principal component analysis (Fig. 10) confirmed the separation of the two species based on these seven characters.

DISCUSSION

From a total 27 leaf architecture characters examined in A. malaccensis and G. versteegii, eight characters which represented morphology only showed slight variation. Variation in leaf shape within a species was a common phenomenon as reported by Zahidi et al.²⁴ on Argania spinosa and Iroka et al.²⁵ on Stachytarpheta jamaicensis and S. angustifolia. The role of leaf morphology for species identification and differentiation of species in a particular genus has been pointed out by Oliveira et al.²⁶ on Psidium, Baroga and Buot Jr.¹⁷ on Terminalia and Abozeid et al.²⁷ on Vicia. Within-species variation on leaf quantitative characters observed in this study such as laminar size and midrib width was quite common for samples collecting from different localities, considering the nature of quantitative characters which are commonly influenced by environmental factors compared to qualitative characters^28,29. Variation on laminar size was found in other species, such as in Ficus deltoidea³⁰ and Corchorus olitorius³¹.

Leaf venation characters in all samples of G. versteegii showed no variation, whereas in A. malaccensis only one out of eight characters varied, namely the free ending ultimate vein. Two types of free ending ultimate vein were found, unbranched or 1-branched. The consistency and uniformity of leaf venation characters supports their use as a taxonomic evidence for identification at the species level. The importance of leaf venation in species identification has been noted by Sharma et al.¹⁴ on Mangifera indica and for differentiating problematic species as noted by Baltazar and Buot Jr.³² on Coffea liberica. Variation on free ending ultimate vein in A. malaccensis observed in this study was found in other plants, namely in Spilanthes³³ and Physalis³⁴. Results of this study were in line with previous studies on leaf venation patterns on various plant species. Variation on midrib characters was mentioned by Salvana and Buot Jr.³⁵ on three species of Hoya. Moreover, the role of midrib outline as diagnostic character has been pointed out by Mantovani et al.³⁶ for the taxonomy of Anthurium.

Fig. 10:	PCA plot of A. malaccensis and G. versteegii samples based on 27 leaf architectural characters used in this study
	Ama: A. malaccensis, Gve: G. versteegii

Variations on marginal vein characters as showed in this study were in line with those reported in Mortoniodendron³⁷ and Psidium²⁶. Moreover, Mishra et al.³⁸ noted the functional aspect of marginal venation patterns in Coffea arabica for adaptation to the habitat.

A number of studies have proven that leaf architecture was useful for species identification as reported in Shorea³⁹, Bauhinia⁴⁰ and Psidium²⁶. In addition, taxonomic significance of leaf venation for classification has been highlighted by Sun et al.⁴¹ in Dioscorea. Characterization of leaf vein pattern is also important for medicinal plants in the recognition of botanical origin and quality control of plant materials⁴², since there are a lot of medicinal herbs are sold in the form of dried and fragmented leaves. Correct identification of such commercial product is important to avoid the problem of misidentification and adulteration.

The pattern of samples grouping on the dendrogram (Fig. 9) showed clear separation between A. malaccensis and G. versteegii. This result indicated that the 27 leaf architectural characters used in this study were useful as reliable taxonomic evidence for species recognition and identification. The taxonomic value of leaf architecture is particularly important for identification of sterile specimens when reproductive structures such as flowers, fruits and seeds are unavailable^16,17,39.

Based on the principal components analysis, three leaf architectural characters showed major role in differentiating the two species, namely leaf shape, laminar size and midrib width as indicated by seven highest character loading the first principal component (PC1). Accordingly, these three characters considered as distinguishing characters between A. malaccensis and G. versteegii as shown in the grouping of samples in PCA scatter plot (Fig. 10). The application of cluster analysis and principal component analysis in this study was proven to be useful in determining diagnostic characters for distinguishing species. The use of multivariate analysis methods as tools for determining useful characters for identification and delineating taxa has been reported in leaf architectural studies of Capparaceae⁴³, Hoya⁴⁴ and leaf morphology of Vicia²⁷.

The micromorphological features used in this study, the trichomes and calcium oxalate crystals characteristics, both showed considerable variation. The role of trichomes character in plant taxonomy, especially for identification and delineation of species has been proven in various taxa, such as in Abutilon⁴⁵. Previous studies on leaf micromorphology also indicated within-species variations on trichomes characteristics, such as on Acinos graveolens⁴⁶ and Salvia nemorosa⁴⁷.

Focusing on the micromorphology examined in this study, from eight micromorphological characters, three of them have a role in distinguishing species and even and between samples within the same species. These three characters were calcium oxalate density on lamina, length of raphide-type calcium oxalate and calcium oxalate on midrib. Results of this study provided support on the role of calcium oxalate crystal characteristics in plant taxonomy as has been reported in several taxa, such as in Amorphophallus muelleri⁴⁸ and Alocasia macrorrhizos⁴⁹. Taxonomic value of raphide-type calcium oxalate crystals and its important role as diagnostic characters for identification of plant microfossils residues has been mentioned by Crowther⁵⁰.

The use of leaf venation and micromorphological characters for plant identification has limitations in the preparation of microscopic slides that require adjustments from standard methods for different types of leaves. Since this is a crucial step, a sufficient knowledge and skill on of plant microtechnique is needed by researchers who are willing to apply leaf architecture characters in plant identification.

CONCLUSION

This study revealed taxonomic value of leaf architecture and micromorphology as supporting data for identification of Aquilaria malaccensis and Gyrinops versteegii at vegetative stage. The benefit of using leaf architecture in species identification at vegetative stage is particularly essential for plants with low capability in producing flowers and fruits.

SIGNIFICANCE STATEMENT

This study discover the role leaf architecture as supporting data for the identification of two agarwood-producing species, A. malaccensis and G. verstegii. The leaf architecture characters can be beneficial for plant species with low capability of producing flowers and fruits. This study will help the researcher to uncover the critical areas of plant leaf venation and micro-morphology that were rarely explored. Thus a new evidence on the role of leaf architecture for the identification of species at vegetative stage can be confirmed.

ACKNOWLEDGMENTS

The authors acknowledge financial support for this research provided by Universitas Gadjah Mada under the scheme of Thesis Recognition Program 2019 (RTA No. 3022/UNI/DITLIT/DIT-LIT/LT/2019) granted to second author. An expertise suggestion on microscopic observation from Dr. Maryani is highly appreciated.