American Journal of Botany 85(4): 461466. 1998.

VARIATION

OF SILICA BODIES IN LEAF EPIDERMAL OF

LONG CELLS WITHIN AND AMONG SEVENTEEN SPECIES

ORYZA (POACEAE)1
AND

SUNG SOO WHANG, KYUNGSIK KIM,3

WILFORD M. HESS2

Faculty of Biological Sciences, Chonbuk National University, Chonju 560-756, Republic of Korea; 2Department of Botany and Range Science, Brigham Young University, Provo, Utah 84602 Morphometric procedures were used with scanning electron microscopy backscattered images to study silica bodies in epidermal long cells of four different leaf veins of 17 of the 20 species of Oryza. The veins studied were midrib, large vein, small vein, and marginal vein. Image analysis was used to study morphological variations among the silica bodies. Statistical analyses were based on 11 variables. Even within a single leaf, silica bodies were not uniform. However, the degree of morphological variation normally showed a distribution of morpholgical types around one modal shape. The most signicant differences observed were between silica bodies of the midrib and those of other veins. Bodies varied with respect to both size and shape. Computer-assisted image analysis is an effective tool for categorizing basic data and for statistical analysis of variation among silica bodies. Morphological variation among silica bodies of a single leaf may be related to waterconducting systems and their inuence on silica availability and phytolith formation. Key words: image analysis; Oryza; Poaceae; silica bodies; variation.

In order to use phytoliths as a taxonomic tool, it is necessary to understand morphological variation in phytoliths within an individual plant and among taxa. Variation can be introduced by both botanical and environmental factors (e.g., Mulholland, Rapp, and Ollendorf, 1988; Mulholland et al., 1990; Ball, Brotherson, and Gardner, 1993). Although phytolith analysts have used a variety of approaches to overcome these problems (Blackman, 1971; Norgren, 1973; Terrell and Wergin, 1981; Rosen, 1983; Brown, 1984; Hodson, Sangster, and Parry, 1985; Piperno, 1988; Mulholland, Rapp, and Ollendorf, 1988; Mulholland et al., 1990; Ball, Brotherson, and Gardner, 1993), many difculties remain unresolved. Until better experimental procedures are developed, phytolith investigators must exercise caution with sampling procedures, particularly with reference materials for phytolith systematics. Therefore, studies in phytolith variation are needed before investigators can develop more accurate and more convenient procedures. Past studies of morphological variation in leaf phytoliths have largely focused on a particular portion of leaves, such as left and right halves, bases, midsections, or tips of leaves (Blackman, 1971; Takeoka et al., 1984; Mulholland, Rapp, and Ollendorf, 1988). Here we investigate the possibility that silica bodies formed in epidermal long cells over different veins of a single leaf may provide more reliable samples of phytolith variation with individual grass species. Monosilicic acid in plants, absorbed by roots from ground water, moves through the plant in water-conducting xylem (Frey-Wyssling, 1930; Parry and Smithson, 1964, 1966; Blackman and Parry,
1 Manuscript received 23 August 1996; revision accepted 21 July 1997. The authors thank Dr. D. A. Vaughan (IRRI) for providing materials in this study. This study was partly supported by a grant from the Basic Science Research Institute Program (BSRI 95-4427) of Ministry of Education, Republic of Korea. 3 Author for correspondence: Faculty of Biological Sciences, Chonbuk National University, Chonju 560-756, Republic of Korea.

1968). Differences among leaf vein systems may affect silica availability and phytolith formation. Studies to test this hypothesis have not been conducted previously. Species of the grass Oryza are ideal for these types of studies, since they are silicon-accumulating plants. Furthermore, their parallel leaf veins can be clearly grouped into four categories (midribs and large, small, and marginal veins) (Metcalfe, 1960). Recently, studies of phytoliths as tools in systematics have successfully employed computer analyses of morphological variation in phytoliths. Attempts have been made to test statistical variation in phytoliths from different plant parts (Pearsall, 1978; Piperno, 1984; Russ and Rovner, 1987; Ball, Brotherson, and Gardner, 1993). It is recognized that this approach should be tested by different investigators to develop condence in the methodology and to justify acceptance of the procedures for phytolith systematics. The aims of this study were to use scanning electron microscopy backscattered images of phytoliths as a morphometric tool to study Oryza leaves and to assess morphological variation in silica bodies, from epidermal long cells of leaves from 17 economically important Oryza species. MATERIALS AND METHODS
There are more than 20 species of Oryza (Tateoka, 1963), but some species are now either extinct or are very difcult to collect (Vaughan, 1989). Therefore, 17 species of Oryza that were available are included in this study. Oryza sativa L. was collected from near Chonju-city and also obtained from Honam Crop Experiment Station in Korea. Leaves were freeze dried for later use. Freeze-dried leaves of other species (O. alta Swallen, O. australiensis Domin, O. eichingeri Peter, O. grandiglumis Prod., O. barthii Chev., O. minuta J. S. Presl, O. ofcinalis Wall, O. punctata Kotschy, O. rhizomatis Vaughan, O. rupogon Griff., O. rupogon [intermediate type], O. granulata Nees, O. meyeriana Baill, O. longiglumis Jansen, O. ridleyi Hook, and O. brachyantha Chev.), were obtained from the International Rice Germplasm Center (Herbarium, IRGC) in the Philippines.

461

462

AMERICAN JOURNAL

OF

BOTANY

[Vol. 85

Samples (6 mm2 samples) were excised from the middle portion of mature leaves. The leaf squares were sonicated in hexane-chloroform (1:5) for 5 min to remove epicuticular wax and mounted on aluminum stubs using double-sided adhesive tape. After sputter coating with gold, samples were observed by scanning electron microscopy (SEM), using backscattered electron imaging (BSI) at 30 kV with a Robinson detector. Backscattered SEM images of leaf surfaces were used for the morphometric studies. For each of the 17 species studied, 50150 phytoliths were analyzed in each of the four types of leaf cells (epidermal long cells from midrib and large, small, and marginal veins) from the abaxial side of leaves (Figs. 14). The images were copied onto tracing paper using a sharp lettering pen. Eleven morphometric parameters were used to evaluate silica bodies (Table 1). Morphometric studies were carried out with an IBAS 2000, Kontron Bildanalse image analyzer. Factor analysis was used to interpret relationships among the 11 variables, and to simplify the immense amount of raw data obtained through image analysis. Factors over one in eigenvalue were selected and scored. An analysis of variance using the factor score ( factor score coefcients for each variable the standard score for each variable) was performed to determine the signicance of morphological shapes of silica bodies formed in long cells of the different vein types. The Tukey multiple comparison test was used to test the signicance of differences among the silica bodies in the four different veins studied for each of the 17 species. All signicance levels were determined at P 0.05. The mathematical analysis was performed using an SPSS-X Release 2.2 Program and CYBER 93231 computer at Chonbuk National University.

RESULTS Morphology of silica bodies was statistically analyzed using 11 variables measured with image analysis (Table 1). After factor analysis, the three factors that had eigenvalues that summed to over 1.0 were extracted (Table 2). Each of these factors was characterized with respect to several characters that had high absolute values for factor loading. Factor loadings indicate the overall importance of each character, while eigenvalues for each factor indicate the portion of total variance accounted for by that factor (Table 2). In distinguishing high factor loading, loadings with values 0.5 of the absolute value were arbitrarily adopted as the standard. Factor 1 was characterized by high loadings for area, perimeter, maximum diameter, minimum diameter, feret x, feret y, form ell, and dicircle (Table 2). Except for the form ell, which showed a low communality below 0.5, these parameters had apparent relationships with the size of the silica bodies. Bigger values were correlated with larger silica bodies. Thus factor 1 represented the size relationship factor. Factor 2 was characterized by high loadings for form pe and aspect ratio (Table 2). These variables were correlated with the shape of silica bodies. The maximum value of form pe was one and smaller values indicated increasing irregularity. The values for aspect ratio were obtained by dividing the maximum by minimum diameter value for the silica body. If the value approached one, the shape tended to be round and/or square. Therefore factor 2 represented shape relationships. Factor 3 was characterized by high loadings of feret x and feret x/y. The value of feret x was based on the maximum dimension of a silica body in the x direction. These variables were also related to the size of silica bodies.

They were correlated in that with bigger values for feret x, values were also bigger for feret x/y. To access tendency and extent of variation, silica bodies formed in epidermal long cells over midribs, large veins, small veins, or marginal veins (Figs. 14) were evaluated using the following procedure. Factors 1 and 2 had high eigenvalues: they were used for analysis of variance among factor scores (Table 3). Except for factor 2 for O. australiensis, factors 1 and 2 differed signicantly for each of the species of Oryza considered. That indicates that silica bodies produced by epidermal long cells in the four different veins of a single Oryza leaf did not have uniform morphology. To determine which silica bodies differed signicantly among the four different types of veins under study, Tukeys multiple comparison test was performed on average scores for factors 1 and 2 (Table 4). With respect to the size factor, silica bodies formed in epidermal long cells above the midrib differed most from those of other veins. It was observed that the silica bodies from the midrib and marginal veins of O. ofcinalis and from the midrib and small veins of O. meyeriana and O. brachantha had nearly identical morphologies. Silica bodies formed in epidermal long cells along each of the four veins of O. minuta, O. sativa, O. granulata, and O. longiglumis differed signicantly from each other. On the other hand, only silica bodies from the midrib differed from those of other veins for O. grandiglumis, O. barthii, and O. punctata. Silica bodies developed in epidermal long cells along the midrib were signicantly different from those formed along other veins considered. However, silica bodies formed over large veins and small veins of O. australiensis were not different from each other. Except for those formed over the midrib, silica bodies formed over veins of O. punctata, O. meyeriana, and O. brachyanta were not signicantly different from each other. Silica bodies formed in epidermal long cells above marginal veins of O. barthii and O. ofcinalis and small veins of O. ridleyi differed signicantly from those from along other veins of those species. DISCUSSION AND CONCLUSIONS Because phytoliths assume characteristic and often morphologically distinctive shapes in many plant taxa, they can be used to identify microfossils from specic plants once present at prehistorical sites. During the last 40 yr, various methods have been used to identify phytoliths from a variety of plants (e.g., Rovner, 1971; Rapp and Mulholland, 1992; Whang, 1993). It is likely that phytolith morphology will become an excellent taxonomic tool in phytolith studies. Phytolith morphology should make it possible to help workers rene species identications from prehistoric strata and dene specic tissues from which the phytoliths came. However, we must recognize that phytolith systematics is still far from complete for many reasons. One of the main obstacles in phytolith systematics is variation in silica body morphology from tissue to tissue and within specic tissues. Many studies have been conducted to better understand these variations. Several factors may inuence within-individual variations, including the stage of plant maturity (Hodson, Sangster, and Parry,

April 1998]

WHANG

ET AL.VARIATION OF SILICA BODIES OF

ORYZA

LEAVES

463

Figs. 14. BSI (backscattered electron imaging) photographs of silica bodies on abaxial side of leaf blades of different Oryza species 300 (bars 30 m). Arrows indicate silica body. 1. Midrib of O. grandiglumis. 2. Large vein of O. brachyantha. 3. Small vein of O. ofcinalis. 4. Marginal vein of O. australiensis.

464

AMERICAN JOURNAL

OF

BOTANY

[Vol. 85

TABLE 1. Variables measured using an image analyzer for statistical description of phytoliths in Oryza.
Parameter Measurement

TABLE 2. Factor loadings, communalities and eigenvalues obtained by factor analysis of 11 variables based on characteristics of silica bodies from Oryza.
Parametera Factor 1 Factor 2
b

Area Perimeter Maximum diameter Minimum diameter Feret x Feret y Form ell Form pe Dicircle Aspect ratio Feret xy

Total area of silica body (m2). Distance around the edges of the silica body (m). Length of the maximum diameter of silica body (m). Length of the minimum diameter of silica body (m). Maximum dimensions of silica body in x direction (m). Maximum dimensions of silica body in y direction (m). Shape factor for elliptical silica body based on y/x. A measure of the shape of silica body, calculated as 4 area/(perimeter)2. Diameter of an equivalent circle for a silica body area, calculated as 2(area/). Ratio of maximum diameter to minimum diameter. Ratio of feret x/feret y.

Factor 3

Communality

Area Perimeter Max. diameter Min. diameter Feret x Feret y Form pe Form ell Dicircle Aspect ratio Feret xy Eigenvalue
a b

0.97603 0.98862b 0.94479b 0.94203b 0.78231b 0.89404b 0.10563 0.61928b 0.98632b 0.15954 0.46337 6.72901

0.02711 0.01163 0.29736 0.31111 0.04618 0.22962 0.97276b 0.09282 0.00465 0.95945b 0.00264 2.11638

0.00820 0.09045 0.03958 0.03710 0.58508b 0.36526 0.07036 0.23871 0.02742 0.16000 0.87107b 0.33398

0.95343 0.98568 0.98262 0.98558 0.95647 0.98544 0.96236 0.44910 0.97360 0.97159 0.97348

Variables correspond to those in Table 1. Factor loadings considered to be important to each factor.

1985), intraspecic variation within the plant taxa (Piperno, 1984, 1988; Mulholland, Rapp, and Ollendorf, 1988), the amount of soluble silica in local ground water (Norgen, 1973), the rate of leaf transpiration (Rosen, 1983), the tissue within which the phytoliths form (Blackman, 1971; Terrell and Wergin, 1981; Mulholland, Rapp, and Ollendorf, 1988; Ball, Brotherson, and Gardner, 1993), the location of phytoliths in leaf blades (Blackman, 1971; Takeoka et al., 1984; Mulholland, Rapp, and Ollendorf, 1988), genetic variation among plants, and geographic location where the plants grew (Mulholland, Rapp, and Ollendorf, 1988; Mulholland et al., 1990). Additional phytolith studies, particularly those that concentrate on variation with a single species, will help to provide more efcient and more accurate methods. The avenues and rates of movement of silica in plant leaves are related primarily to vein location and structure within leaves. Since vein systems are structurally different from each other, silica availability and phytolith formation could reasonably be expected to differ. Unfortunately, studies on phytolith variation specic to different vein systems are not yet available. In the genus Oryza, most silica bodies are formed in epidermal long cells above leaf veins. They often occur in rows (Whang and Kim, 1994). This is one of the reasons why Oryza was chosen for evaluation of morphological variation among silica bodies formed in epidermal long cells of a single leaf. Although we expected the leaves to have relatively similar silica bodies throughout, the bodies of at least midribs differed signicantly from phytoliths of the other three vein types considered. Differences were based on both size and shape relationships. In many cases, silica bodies did not differ among veins other than the midrib. Since several studies have been reported on the variation of phytolith assemblages within leaves (Blackman, 1971; Takeoka et al., 1984; Mulholland, Rapp, and Ollendorf, 1988), these are not the rst descriptions of phytolith variation within leaves. However, this is the rst report of the variation of silica bodies specic to different leaf

vein systems. Our ndings might be useful to workers attempting to develop a known phytolith assemblage for comparison with unknown samples from prehistoric sites. Unfortunately, none of the systematic studies conducted in the past have compared silica bodies from different leaf veins. With earlier studies, phytolith assemblages from known species were obtained from either entire leaf blades or whole plants, to homogenize variation in the sample (Mulholland, Rapp, and Ollendorf, 1988; Ollendorf, Mulholland, and Rapp, 1988). Such procedures have probably contributed to the confusion in establishment of diagnostic characteristics of phytoliths. At least such procedures did not help to resolve taxonomic confusion. Russ and Rovner (1987) statistically tested variables that seemed to distinguish morphological categories (crossbody vs. bilobate) that had not previously been shown to be useful in identication of plant taxa. Our study has shown that certain species, for example O. brachyantha, had only cross-shaped silica bodies, but most species had either bilobate only or both bilobate and
TABLE 3. Results of ANOVA test of silica bodies formed along four kinds of abaxial veins in Oryza. The test was based on factor score coefcients (see Table 2).
Species Factor f prob.a Species Factor f prob.

O. alta O. eighingeri O. barthii O. ofcinalis O. rhizomatis O. rupogon (intermediate) O. granulata O. longiglumis O. brachyantha
a

1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2

0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0002 0.0000 0.0000 0.0000 0.0000 0.0000 0.0001

O. australiensis O. grandiglumis O. minuta O. punctata O. rupogon O. sativa O. meyeriana O. ridleyi

1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2

0.0000 0.1236 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0001 0.0000 0.0001

f prob. probability of signicant difference.

April 1998]

WHANG

ET AL.VARIATION OF SILICA BODIES OF

ORYZA

LEAVES

465

TABLE 4. Results of multiple comparison tests using factor scores for silica bodies from the four leaf veins of abaxial leaf surfaces of species in Oryza (*: signicantly different at 0.05; f1: factor 1; f2: factor 2; vein 1: midvein; vein 2: large vein; vein 3: small vein; vein 4: leaf margin).
Species Mean (f1) Leaf Vein Mean (f2) Leaf Vein

TABLE 4.
Species

Continued.
Mean (f1) Leaf Vein Mean (f2) Leaf Vein

O. granulata

O. alta 0.6844 0.1130 0.1187 0.5744 O. australiensis 1.4328 0.2101 0.6016 0.6174 0.8203 0.2442 0.0028 1.0626 1.8398 0.5748 0.5867 0.6807 1.1069 0.2595 0.3444 0.5093 1.5689 0.0182 0.4842 1.0660 0.6097 0.3302 0.4316 0.5098 1.7740 0.5295 0.6130 0.6311 1.4929 0.0940 0.6104 0.7888 0.9619 0.1116 0.4382 0.6414 1 3 4 2 1 3 4 2 1 4 2 3 1 4 3 2 1 3 2 4 1 3 4 2 4 1 2 3 1 2 3 4 1 4 3 2 1 3 2 4

1342 * * *** 1342 * ** ** 1423 * * *** 1432 * * * 1324 * * * 1342 * ** *** 4123 ** ** 1234 * * * 1432 * ** ** 1324 * ** ** 1423 1 4 2 3 1 3 2 4 * * *** 1324 * ** *** 0.8191 0.1270 0.4699 0.6347 0.1611 0.0972 0.0249 0.2301 1.2648 0.3056 0.3946 0.5633 1.0849 0.0860 0.3860 0.6074 1.0217 0.2739 0.2205 1.0806 0.5469 0.4098 0.3058 0.6560 0.4250 0.0867 0.0127 0.5096 1.3754 0.4383 0.4528 0.4787 1.2025 0.1767 0.3845 0.6371 0.4597 0.1681 0.1453 0.4876 1 4 2 3 2 4 1 3 1 4 3 2 1 3 4 2 4 3 2 1 3 4 1 2 4 3 2 1 1 4 2 3 1 3 2 4 3 2 1 4

1423 * *t* *t* 2413 O. longiglumis * 1432 * * ** 1342 * ** ** 4321 * ** *** 3412 ** *** 4321 * * *** 1423 * * * 1324 * * *** 3214 * *** 1324 1 3 2 4 1 3 4 2 ** ** 1342 * * *** O. brachyantha O. ridleyi O. meyeriana

1.0203 0.1889 0.4364 0.7735 0.8557 0.6401 0.3843 1.1063 1.0979 0.2819 0.4652 0.9110 1.0734 0.0210 0.2506 0.7739 0.7042 0.6096 0.3361 0.9826

1342 1 3 4 2 1 3 2 4 1 4 3 2 1 4 3 2 1 3 2 4 * ** *** 1324 ** *** 1432 * ** *** 1432 * * *** 1324 ** ***

0.5928 0.2545 0.3922 0.4499 0.4421 0.0364 0.2366 0.2442 1.1460 0.0951 0.4136 0.6339 0.4074 0.0494 0.0831 0.3736 0.4610 0.0637 0.1517 0.2402

1243 1 2 4 3 1 4 3 2 1 3 2 4 3 2 1 4 1 3 4 2 * ** ** 1432 * * * 1324 * ** ** 3214 * * ** 1342 * * *

O. eichingeri

O. grandiglumis

O. barthii

O. minuta

O. ofcinalis

O. punctata

O. rhizomatis

cross bodies or a continuum of phytolith shapes between the cross and bilobate bodies. However, size and shape relationships were not affected by types of silica bodies (i.e., cross and/or bilobate bodies). Thus, computer-assisted image analysis was shown to be an ideal tool for development of an effective and reproducible system for phytolith classication. Conclusions for this study are summarized as follows: (1) silica bodies that formed in epidermal long cells differ signicantly depending on whether they formed along midribs or lateral veins (in some cases silica bodies differ among lateral veins, as well); (2) silica bodies of midrib cells constitute a distinct type in comparison with phytoliths associated with other veins; (3) silica body morphological variations might be related to different waterconducting tissue systems that inuence silica availability and phytolith size and shape; and (4) computer-assisted image analysis has proven to be a useful technique for collection of basic data and testing that data for statistical signicance. LITERATURE CITED
BALL, T. B., J. D. BROTHERSON, AND J. S. GARDNER. 1993. A typologic and morphometric study of variation in phytoliths from einkorn wheat (Triticum monococcum). Canadian Journal of Botany 71: 11821192. BLACKMAN, E. 1971. Opaline silica bodies in the range grasses of southern Alberta. Canadian Journal of Botany 49: 769781. , AND D. W. PARRY. 1968. Opaline silica deposition in rye (Secale cereale L.). Annals of Botany 32: 199206. BROWN, D. A. 1984. Prospects and limits of a phytolith key for grasses in the central United States. Journal of Archaeological Science 11: 221243. FREY-WYSSLING, A. 1930. Uber die Ausscheidung der Kieselsaure in der Panze. Berichte der Deutschen Botanischen Gesellschaft 48: 179183.

O. rupogon

O. rupogon (intermediate)

1.3279 0.1756 0.2966 0.7524 1.5986 0.2265 0.3290 0.6358

0.3203 0.2239 0.1813 0.3293 1.3251 0.1600 0.0446 0.7229

O. sativa

466

AMERICAN JOURNAL

OF

BOTANY

[Vol. 85

HODSON, M. G., A. G. SANGSTER, AND D. W. PARRY. 1985. An ultrastructural study on the developmental phases and silicication of the glumes of Phalaris canariensis L. Annals of Botany 55: 649 665. METCALFE, C. R. 1960. Anatomy of the Monocotyledons. I. Gramineae. Oxford University Press, London. MULHOLLAND, S. C., G. RAPP, JR., AND A. L. OLLENDORF. 1988. Variation in phytoliths from corn leaves. Canadian Journal of Botany 66: 20012008. , , , AND R. REGAL. 1990. Variation in phytoliths within a population of corn (Mandan Yellow Flour). Canadian Journal of Botany 68: 16381645. NORGREN, J. A. 1973. Distribution, form, and signicance of plant opal in Oregon soils. Ph.D. dissertation, Oregon State University, Corvallis, OR. OLLENDORF, A. L., S. C. MULHOLLAND, AND G. RAPP, JR. 1988. Phytolith analysis as a means of plant identication: Arundo donax and Phragmites communis. Annals of Botany 61: 209214. PARRY, D. W., AND F. SMITHSON. 1964. Types of opaline silica depositions in the grass leaves. Annals of Botany 22: 543549. , AND . 1966. Opaline silica in the inorescences of some British grasses and cereals. Annals of Botany 36: 781783. PEARSALL, D. M. 1978. Phytolith analysis of archaeological soils: evidence for maize cultivation in formative Ecuador. Science 199: 177178. PIPERNO, D. R. 1984. A comparison and differentiation of phytoliths from maize (Zea mays L.) and wild grasses: use of morphological criteria. American Antiquity 49: 361383.

. 1988. Phytolith analysis: an archaeological and geological perspective. Academic Press, San Diego, CA. RAPP, G., JR., AND S. C. MULHOLLAND. 1992. Phytolith systematics, emerging issues. Plenum, New York, NY. ROSEN, A. M. 1983. Phytoliths and marginal agriculture in the Chalcolithic of southern Israel. Society for American Archaeology, 48th Annual Meeting, Pittsburgh, PA. Program and Abstracts, p. 92. Society for Amercan Archaeology, Washington, DC. ROVNER, I. 1971. Potential of opal phytoliths for use in paleoecological reconstruction. Quaternary Research 1: 345359. RUSS, J. C., AND I. ROVNER. 1987. Stereological verication of Zea phytolith taxonomy. Phytolitharien Newsletter 4: 1018. TAKEOKA, Y., T. WADA, K. NAITO, AND P. B. KAUFMAN. 1984. Studies on silicication of epidermal tissues of grasses as investigated by soft X-ray image analysis. II. Differences in frequency of silica bodies in bulliform cells at different positions in the leaves of rice plants. Japanese Journal of Crop Science 53: 197203. TATEOKA, T. 1963. Taxonomic studies of Oryza. III. Key to the species and their enumeration. Botanical Magazine, Tokyo 76: 165173. TERRELL, E. E., AND W. P. WERGIN. 1981. Epidermal features and silica deposition in lemmas and awns of Zizania (Gramineae). American Journal of Botany 68: 697707. VAUGHAN, D. A. 1989. The genus Oryza L. Current status of taxonomy. IRRI Research Paper Series Number 138: 121. WHANG, S. S. 1993. Phytolith morphology and its taxonomic signicance in the genus Oryza. Ph.D. dissertation, Chonbuk National University, Korea. , AND K. KIM. 1994. Opal phytolith morphology in rice. Journal of Plant Biology 37: 5367.