PHYSIOLOGIA PLANTARUM 115: 87–92. 2002 Copyright C Physiologia Plantarum 2002

Printed in Denmark – all rights reserved ISSN 0031-9317

Silicification in sorghum (Sorghum bicolor) cultivars with differentdrought tolerance

Alexander Luxa,*, Miroslava Luxovab, Taiichiro Hattoric, Shinobu Inanagac and Yukihiro Sugimotoc

aDepartment of Plant Physiology, Faculty of Natural Sciences, Comenius University, Mlynska dolina B-2, SK-842 15 Bratislava, SlovakRepublicbInstitute of Botany, Slovak Academy of Sciences, Dubravska cesta 14, SK-842 23 Bratislava, Slovak RepubliccArid Land Research Center, Tottori University, 1390 Hamasaka, Tottori 680–001, Japan*Corresponding author, e-mail: lux/fns.uniba.sk

Received 10 July 2001; revised 31 October 2001

Sorghum belongs to a group of economically important, sili-con accumulating plants. X-ray microanalysis coupled withenvironmental scanning electron microscopy (ESEM) of freshroot endodermal and leaf epidermal samples confirms histo-logical and cultivar specificity of silicification. In sorghumroots, silicon is accumulated mostly in endodermal cells. Spe-cialized silica aggregates are formed predominantly in a singlerow in the form of wall outgrowths on the inner tangentialendodermal walls. The density of silica aggregates per squaremm of inner tangential endodermal cell wall is around 2700and there is no significant difference in the cultivars with dif-ferent content of silicon in roots. In the leaf epidermis, silicondeposits were present in the outer walls of all cells, with the

Introduction

The evidence is overwhelming that silicon (Si) should beincluded among the elements having a major bearing onplant life (Epstein 1999). Recently it has been necessaryto modify the traditional view that silicon deposition inthe cell walls was a purely physical process leading tomechanical strengthening of the tissue. Silicon depo-sition is under rather strict metabolic and temporal con-trol (Marschner 1995). The changes in cell wall metabo-lites interacting with silicic acid lead to bulk depositionof silicon into the mature cell wall structure (Perry et al.1986). A dynamic component of redistribution of siliconis also involved in the protective effects of silicon againstinsects and pathogens, in addition to a mechanical bar-rier (Heath and Stumpf 1986).

Takahashi and Miyake (1977) distinguished betweensilicon accumulators and silicon non-accumulators. Sor-ghum (Sorghum bicolor) is one of the important silicon

Physiol. Plant. 115, 2002 87

highest concentration in specialized idioblasts termed ‘silicacells’. These cells are dumb-bell shaped in sorghum. In boththe root endodermis and leaf epidermis, silicification washigher in a drought tolerant cultivar Gadambalia comparedwith drought sensitive cultivar Tabat. Silicon content per drymass was higher in leaves than in roots in both cultivars. Thevalues for cv. Gadambalia in roots and leaves are 3.5 and4.1% Si, respectively, and for cv. Tabat 2.2 and 3.3%. How-ever, based on X-ray microanalysis the amount of Si depositedin endodermal cell walls in drought tolerant cultivar (unlikethe drought susceptible cultivar) is higher than that depositedin the leaf epidermis. The high root endodermal silicificationmight be related to a higher drought resistance.

accumulators. Most of the silicon in sorghum, as well asin other graminaceous species, is deposited in the outerwalls of the epidermal cells of the leaves and in the in-florescence bracts (Hodson and Sangster 1989). The epi-dermal cell walls are impregnated with a layer of siliconand become an effective barrier against both water lossby cuticular transpiration and fungal infections. In sor-ghum, and many other grasses, a high proportion of sili-con in the leaf epidermis is also located intracellularlyin specialized idioblasts called silica cells (Sangster 1970,Esau 1977).

In accumulator species, silicon uptake is closely re-lated to root metabolism and not greatly affected by thetranspiration rate (Okuda and Takahashi 1965). Silic-ification of root tissues occurs in several species of grass-es. Three basic patterns of silicon distribution were rec-ognized in roots. It can be (1) restricted to the uniseriate

endodermis, the innermost cortical layer surroundingthe stele; (2) spread throughout all root tissues (as in theroots of Mollinia coerulea (L.) Moench); or (3) depositedinto intercellular spaces (Sangster and Parry 1981). Sor-ghum belongs to the first group, with the solid silicondeposits in the form of unique, regularly distributed,dome-shaped silica aggregates on the inner tangentialendodermal walls. These structures were first describedin the Andropogoneae by Borissow (1924, 1925, 1928),and called by him ‘Rasdorskys Körpchen’ (Rasdorskybodies). The high content of silicon and the importanceof this element for sorghum was demonstrated by severalstudies. Beneficial effects of endodermal silicification in-clude protection of the stele from pathogens and para-sites (Bennett 1982). The deposition of silicon in the en-dodermis of sorghum roots was the subject of an exten-sive survey by Sangster and Parry (1976a, 1976b, 1976c).X-ray analysis showed the development of silica aggre-gates with a clear acropetal linear gradient in seminalroots of sorghum (Sangster and Parry 1976a). Endoder-mal silicification of nodal roots of sorghum, using thescanning electron microscope and electron-probe micro-analyser for Si content determination, was investigatedby Sangster and Parry (1976b). In mature nodal rootsof sorghum no accumulation of Si occurred other thanin the endodermis, none was detected in the walls ofmetaxylem or in the sclerenchyma zone. In Sangster andParry (1976c) Si deposits in the endodermis were ob-served by transmission electron microscopy and a con-siderable involvement with the cellulosic structure of theinner tangential endodermal walls was indicated in theformation of silica aggregates.

The high tolerance of sorghum to drought has beenthe subject of investigation of many plant physiologistsand breeders. Several mechanisms, which permit sor-ghum to achieve economic yields under drought con-ditions, have been proposed. In our previous paper(Salih et al. 1999) the effects of soil moisture stress onthe rooting habits, transpiration rate and xylem ana-tomy of two sorghum cultivars, Tabat (drought suscep-tible) and Gadambalia (drought tolerant) were studied.Tabat had a higher root length density, higher numberof late metaxylem vessels per nodal root, higher leafarea, and higher transpiration rate than Gadambalia.Silicon has also been implicated in drought resistance(Doggett 1970). For example, greater endodermal silic-ification was found in the roots of upland rice withhigher drought tolerance, than in roots of lowland rice(Lux et al. 1999). The aim of this paper was to com-pare the silicification of the root endodermis and leafepidermis of two sorghum cultivars differing consider-ably in drought tolerance. Previously, the anatomicallyand physiologically characterized cultivars Tabat andGadambalia (Salih et al. 1999) were investigated. Inthe present work a novel technique of environmentalscanning electron microscopy coupled with X-raymicroanalysis for Si determination and quantificationwas employed and compared with chemical quantifi-cation.

Physiol. Plant. 115, 200288

Materials and methods

Plant material cultivation and preparation

The two sorghum cultivars (Sorghum bicolor (L.) Mo-ench) used for the experiments were Gadambalia, whichis grown widely in drought-prone areas of Sudan, andTabat, which is used for irrigated areas. Planting wasundertaken at the Arid Land Research Center, Tottori,Japan, during August–October 2000. Plants were grownto maturity under well-watered conditions. The field wascovered with transparent plastic, which permitted trans-mission of more than 95% of the incident solar radi-ation. The soil at the site was a loam containing 210 gkgª1 clay. Water was applied through perforated pipeslaid uniformly in the field at a rate of 50 l mª2 in eachirrigation. The irrigation interval was 5 days to providea soil water potential of about ª0.05 MPa in the rootzone (0–0.8 m) just prior to irrigation.

Samples were taken at flowering. The root systemswere washed free from the adhering soil and the longest,thickest nodal roots (with a diameter measured at theproximal end of about 3 mm) were chosen and excised.Roots originated from the basal node of the culm. Formeasurements, at least 3 roots from each of 3 plants ofeach cultivar were used. Silicon in roots is reportedlydeposited mainly in the walls of endodermal cells, pre-dominantly on the inner tangential face. In cross section,only a small part of this deposition site is exposed tothe beam and measurements. Because our intent was tocompare root endodermal and leaf epidermal silic-ification, we exposed the inner tangential endodermalwalls of roots and outer epidermal walls of leaves to X-ray analysis. The exposure of inner endodermal walls ina surface view is possible only if the peripheral roottissues are removed. Relatively thin radial and transver-sal cell walls of endodermal cells and thick, resistant in-ner endodermal walls enable mechanical removing of pe-ripheral tissues. Segments (about 5 mm long) of peeledroots were taken from the region at 20 mm from the rootbase. Samples of leaves were taken from the same plants,using the abaxial central part outside the middle vein ofthe third leaf below the inflorescence. The segments werefixed on the stubs using an adhesive tape, and immedi-ately introduced into the microscope without any treat-ment, desiccation or covering, and analysed.

Microscopy and X-ray analysis

A Nikon ESEM 2700 environmental scanning electronmicroscope was used for microscopy and X-ray analysis.The working conditions were: accelerating voltage 15 kV,pressure 520 Pa, working distance 8 mm, tilt 30æ. Theimaging gas used in the ESEM was water vapour. AnEDAX DX4 X-ray microanalyser (EDAX Inc., Mah-wah, NJ, USA) was utilized for elemental analysis witha live detector time of 100 s per analysis and a take-offangle of 43.2æ. For quantification of all samples, 2000¿magnification was employed. The number of measure-ments was more than 20 per sample for leaves (giving a

total number of measurements of .50 per cultivar) and.10 per sample for root (giving the total number ofmeasurements about 100 per cultivar). Si was expressedas weight percent compared to the sum of Si, C and O;this was calculated using the ZAF program of theEDAX system. The statistic of Sigma plot software ver-sion 1.02 with Student’s t-test was used.

Quantitative Si determination

Dried powdered plant sample (200 mg) was put into sil-ica-free beaker and ashed in a muffle oven at 500æC for5 h. Diluted HCl (1: 1; 10 ml) was added and the samplewas heated at 100æC. The process of dissolving in HCland evaporation to dryness was repeated three times.Then diluted HCl (1: 1; 15 ml) was added and the samplewas heated at 100æC for 2 min and filtered off throughthe quantitative ash-less filter paper no. 6. The filter waswashed three times with diluted HCl (1: 10; 5 ml), withdistilled water (100 ml), and put into a silica-free ceramiccrucible and ashed in the oven at 540æC for 5 h. Aftercooling, the weight of Si was determined.

Density of silica aggregates in root endodermis

For quantification of the density of silica aggregates onthe inner tangential endodermal walls, peeled roots wereprepared as described above. Silica aggregates exhibitblue autofluorescence in UV light, permitting their ob-servation, documentation and counting. Photographsfor quantification were taken by Nikon epifluorescencemicroscope Eclipse using objective 20¿ and projective2.5¿, the documented area was 0.0875 mm2. Samples of5 roots for each cultivar were used and 5 randomly se-lected fields per sample were counted, using root seg-ments approximately 15–20 mm from the base of theroot.

Results

The root morphology and anatomy and stem anatomyof two sorghum cultivars, Tabat and Gadambalia, havealready been described (Salih et al. 1999). Here only themost important facts will be summarized.

Root structure and silicification

The prolific root system of sorghum at flowering isformed mainly by well-developed and branched nodalroots. A single seminal root of sorghum with its numer-ous laterals constitutes the dominant absorptive systemonly up to about 4–5 weeks after germination. Subse-quently this is superseded by nodal roots (Sangster andParry 1976a). Nodal roots of both cultivars in wellwatered conditions were distributed to a depth of ap-proximately 0.7 m. The number of roots formed by thedrought susceptible cv. Tabat was of about 40% higherthan the drought tolerant cv. Gadambalia. The valueswere 35 ∫ 4 and 25 ∫ 4 roots per plant, respectively.

Physiol. Plant. 115, 2002 89

The mature nodal roots of both cultivars had rela-tively broad cortex with subepidermal sclerenchyma andwell-developed tertiary endodermis with intensivelythickened inner tangential and radial walls. The silicaaggregates are prominent on inner tangential walls andcan be observed even under the light microscope.

Si content of roots were 3.5% and 2.2% on a dry mat-ter basis for cv. Gadambalia and cv. Tabat, respectively.The difference is highly significant (P # 0.01). Silica insorghum roots is accumulated exclusively in endodermis(Sangster and Parry 1976a, 1976b). On the transversalsections of nodal roots it was possible to detect a highsignal of Si content in the endodermis using EDX map-ping. The aggregates can clearly be observed after re-moving the peripheral cortical tissues and the outer tan-gential endodermal walls (Fig. 1A,C). On the surface ofthe inner tangential endodermal walls Si aggregates arearranged in serial rows. They are formed by the neckand the irregularly shaped head is formed by a clusterof amorphous particles. In the majority of cells, onlyone row of aggregates is formed and aggregates are regu-larly spaced (Fig. 1C). Occasionally two rows of aggre-gates are formed or additional, smaller, irregularlyshaped outgrows can be present in the corners of radialand inner tangential walls (Fig. 1A). The analysis of den-sity of the aggregates has shown that there is no signifi-cant difference between the two cultivars (Fig. 2). Thedensity is 2.7 ∫ 0.1 and 2.7 ∫ 0.2 mmª2 ¿ 10ª3 for Gad-ambalia and Tabat, respectively.

However, the difference in silicification intensity of en-dodermis in both cultivars can be seen by a comparisonof the EDX map of the Si signals from both cultivars(Fig. 1B,D). Higher silicification in Gadambalia, ex-pressed by a higher Si signal, is therefore caused byhigher impregnation of inner tangential walls by silica.X-ray analysis showed C, O and Si as the only threeelements present in all samples. The peaks of Si signalin Gadambalia are higher than peaks of C and O. Thedata from EDX quantification are in Table 1. There wasa highly significant difference in Si content between thedrought tolerant Gadambalia on one hand and droughtsusceptible Tabat on the other hand. The values werehigher for drought tolerant sorghum.

The other form of data comparison is based on peakto background ratios of Si only. These values are sig-nificantly different between the cultivars, with highervalue for drought tolerant sorghum (Table 1).

Leaf structure and silicification

The difference between the two cultivars is evident inleaf morphology. Gadambalia has much narrower leavesthan Tabat, the leaf area of Gadambalia is 43% that ofTabat. (Salih et al. 1999). No difference was found be-tween the two cultivars in leaf anatomy. Sorghum has atypical Kranz anatomy of leaves characteristic of C4

plants with distinct bundle sheaths. Adaxial epidermalcells are bigger than abaxial cells. Silica cells (specializedsilicon accumulating idioblasts) are distributed regularly

in the epidermis in rows over the leaf veins, they aredumb-bell shaped (Fig. 1E,G). Their length along theleaf axis is between 12 and 20 mm, their width variesbetween 8 and 10 mm in both cultivars.

Si content in leaves is 4.1% and 3.3% on a dry matterbasis for cv. Gadambalia and cv. Tabat, respectively. Thedifference between cultivars is highly significant (P #0.01).

Silicification of leaves in the two sorghum cultivarswhen compared by X-ray microanalysis is similar. A Sisignal was detected in all epidermal cells with the highest

Fig. 1. Scanning electron micrographs (A,C, E, G) and corresponding X-raymicroanalysis mapping of Si signal (B, D,F, H) of root endodermis and leafepidermis of two sorghum cultivars. A, Band E, F sorghum cultivar Gadambalia,C, D and G, H cultivar Tabat. The innertangential walls of endodermal cells afterremoval of outer cortical layers are surfaceviewed in A–D. Note higher, moreuniform, endodermal silicification of cv.Gadambalia compared with cv. Tabat.Abaxial leaf epidermis is in E–H. arrow,silica aggregate; sc, silica cell. Scale bar Ω10 mm.

Physiol. Plant. 115, 200290

values in silica cells (Fig. 1F,H). X-ray analyses are sum-marized in Table 1. Silica content was expressed asweight percent of the sum of Si, C and O –– the onlytwo other elements detected –– and as peak to back-ground ratios. The values are higher for the drought tol-erant cv. Gadambalia. The difference is highly signifi-cant. When the extent of silicification of the root endo-dermis and leaf epidermis (based on EDX analysis) iscompared, the cultivars differ substantially. In thedrought susceptible cv. Tabat the leaf epidermal silic-ification intensity is higher. The values are close, al-

Fig. 2. Regularly distributed silicaaggregates on the inner tangential wallsof root endodermal cells after removing ofouter cortical layers in a surface view influorescence microscope. A, cultivarGadambalia, B, cultivar Tabat. Scalebar Ω 20 mm.

though the difference is significant (P # 0.05). In the caseof the drought tolerant cv. Gadambalia, the root endo-dermis silicification is more than 50% higher than theleaf epidermis silicification (Table 1).

Discussion

Greater root endodermal silicification is a characteristicfeature accompanying drought resistance. This phenom-enon was originally shown in rice (Lux et al. 1999) andin sorghum it is even more evident. Sorghum is adrought tolerant species. The ability of sorghum to growand bring yield in water limited conditions is caused bymultiple factors. Various metabolic and structural adap-tations of sorghum to drought have been shown in nu-merous studies (e.g. Doggett 1970). In the anatomy ofroots, drought resistance was accompanied by thickercell walls in sclerenchyma cells together with a fewer me-taxylem vessels (Salih et al. 1999). Another structuralfeature accompanying drought resistance is greater silic-ification of the inner tangential endodermal walls of sor-ghum. The curious nodular Si deposits, silica aggregates,caught the attention already of early anatomists (e.g.Borissow 1924). Their ultrastructure and developmentwere analysed by Sangster and Parry (1976c). The sameauthors also discussed the control of their formationand proposed two hypotheses. In the first case a physico-chemical control was assumed and in the other, somedegree of protoplasmic control (Sangster and Parry1976b, 1976c). The evaluation of these hypotheses in-

Table 1. X-ray microanalysis determination of weight percent Si (mean ∫) of the sum of Si, C and O and the strength of the Si peak tothe background ratio (P/B) for root endodermal cells and the leaf epidermis for two sorghum cultivars (upper rows). Percent Si on a drymatter basis from the whole roots and leaves of the same cultivars (lower rows). Values within a row followed by different letters (v,w,x,y)show significant varietal differences. Values within a column followed by different letters (a,b,c,d) are significantly different between organs.Student’s t-test, P # 0.05).

Gadambalia Tabat

% Si P/B % Si P/B

X-ray Root endodermis 37.3 ∫ 0.8av 126 ∫ 10av 20.5 ∫ 0.3aw 54 ∫ 2awLeaf epidermis 23.3 ∫ 0.5bv 93 ∫ 9bv 20.1 ∫ 0.2aw 59 ∫ 3bw

Gravimetry Roots 3.5 ∫ 0.3cx 2.2 ∫ 0.2cyLeaves 4.2 ∫ 0.2cx 3.3 ∫ 0.1dy

Physiol. Plant. 115, 2002 91

volves the correlation of root transport studies withpossible mechanisms of endodermal silicification. Todate, no common concept has been accepted for theproblem of root uptake and transport of silicic acid. Itis nor even clear if H4SiO4 first enters the endodermalsymplast and subsequently protoplasmic elimination viaa wall deposition pathway occurs or whether anotherxylem-endodermis pathway is responsible for the Si de-posits in endodermal walls. The mentioned xylem-endo-dermis pathway was supported by Sangster and Parry(1976c) using aerial, non-immersed roots of sorghum. Inthis case the silica deposition in endodermis may havereached its place only in an outward radial movementproceeding from the stele. Our results showed basicallythe same distribution of the silica aggregates in two sor-ghum cultivars differing substantially in Si content inendodermis. It seems to support the hypothesis thatphysico-chemical factors account for the characteristicserial arrangement and regular spacing of these struc-tures. The distribution of the aggregates might be re-garded as the result of the total net physical forces devel-oping over the entire inner tangential wall/cytoplasm in-terface (Sangster and Parry 1976b). The aggregateinitiation would occur at the lowest pressure points,evenly spaced along the interface.

The endodermal silicification is considered to act asan effective barrier against invasion of the stele bypathogens and parasites (Bennett 1982), for exampleStriga (Maiti et al. 1984). Drought resistance and mech-anical strengthening are other benefits (Doggett 1970).

Preventing of water movement from the stele by Si endo-dermal deposition might be responsible for increaseddrought resistance. Leaf epidermal silicification is wellknown in many grasses, but a direct comparison withroot silicification has not been performed. In this papera higher content of silicon was found by X-ray micro-analysis in root endodermis than in the leaf epidermisin drought tolerant sorghum cv. Gadambalia.

For drought susceptible cv. Tabat more Si was shownin leaf epidermis than root endodermis. In this cultivarX-ray data and gravimetric measurements of Si in rootsand leaves correspond well. However, for cv. Gadamb-alia gravimetric measurements showed more Si in leavesthan roots, while X-ray analysis showed the opposite.The gravimetric measurements of Si provide a rigorousmeasure of total Si, in this case of whole roots or leaves.X-ray microanalysis examines Si in a very thin layer ofone cell type. An important difference between X-rayand gravimetric data for the drought tolerant cv. Gad-ambalia indicate the difference of Si deposition to theendodermal cell walls between the cultivars. The Siquantity at the root endodermis and total amount ofSi in the roots is higher in cv. Gadambalia than Tabat.However, Si weight is 1.6-fold higher in root of cv. Gad-ambalia than Tabat, whereas percentage Si in root endo-dermis is 1.8-fold higher in cv. Gadambalia than Tabat.This then shows that cv. Gadambalia is putting a higherproportion of its total Si into the endodermis than Tab-at. In leaves however, Si weight and percentage are both1.2-fold higher in cv. Gadambalia than Tabat, indicatingthe same pattern of Si distribution through leaves. Wemay conclude that the difference in Si deposition of sor-ghum related with drought tolerance is located mainlyin roots.

When comparing the silicification intensity in droughttolerant and drought susceptible sorghum cultivars andtheir transpiration rates (Salih et al. 1999) it can be as-sumed that these two processes are independent. Thedrought susceptible Tabat had a higher transpirationrate and lower silicification of both root and leaf thanthe drought tolerant Gadambalia. This suggests the ac-tive silica uptake by roots of sorghum.

Acknowledgements – One of the authors (A.L.) acknowledges hospi-tality extended by Arid Land Research Center, Tottori University,Japan, during his stay as a Center of Excellence (COE) Visitor. Thework was partially supported by grants no. 1/7258/20 and 1067from Slovak Grant Agency VEGA and COST Action 837. Theauthors are thankful to Dr Philip J. White, Horticulture ResearchInternational, Warwick, UK for critical reading of the manuscript.

Edited by R. Munns

Physiol. Plant. 115, 200292

ReferencesBennett DM (1982) Silicon deposition in the roots of Hordeum sati-

vum Jess., Avena sativa L. and Triticum aestivum L. Ann Bot 50:239–245

Borissow G (1924) Über die eigenartigen Kieselkörper in der Wurz-elendodermis bei Andropogon-Arten. Ber Dtsch Bot Ges 42:366–380

Borissow G (1925) Rasdorskys Körpchen bei Ravenna-Gras. BerDtsch Bot Ges 43: 178–184

Borissow G (1928) Weiteres über die Rasdorskyschen Körpchen.Ber Dtsch Bot Ges 46: 463–480

Doggett H (1970) Sorghum. Tropical Agriculture Series, Rhind D(ed) Longmans Green and Co. Ltd, London and Harlow, 403pp, ISBN 0–582466474

Epstein E (1999) Silicon. Annu Rev Plant Physiol Plant Mol Biol50: 641–664.

Esau K (1977) Anatomy of Seed Plants, 2nd edn. John Wiley andSons, New York, NY, ISBN 0-471-24520-8

Heath MC, Stumpf MA (1986) Ultrastructural observations ofpenetrating sites of the cowpea rust fungus in untreated andsilicon-depleted French bean cells. Physiol Mol Plant Pathol 29:27–39

Hodson MJ, Sangster AG (1989) Silica deposition in the inflor-escence bracts of wheat (Triticum aestivum). II. X-ray micro-analysis and backscattered electron image. Can J Bot 67: 281–287

Lux A, Luxova M, Morita S, Abe J, Inanaga S (1999) Endodermalsilicification in developing seminal roots of lowland and uplandcultivars of rice (Oryza sativa L.). Can J Bot 77: 955–960

Maiti RK, Ramaiah KV, Bisen SS, Chidley VL (1984) A compara-tive study of the haustorial development of Striga asiatica (L.)Kuntze on Sorghum cultivars. Ann Bot 54: 447–457

Marschner H (1995) Mineral Nutrition in Higher Plants, 2nd edn.Academic Press, Harcourt Brace Publishers, London, 889 pp,ISBN 0-12-473543-6

Okuda A, Takahashi E (1965) The role of silicon. The Mineral Nu-trition of the Rice Plant. John Hopkins Press, Baltimore, MD,Manila, pp. 123–146

Perry CC, Williams RJP, Fry SC (1986) Cell wall biosynthesis dur-ing silicification of grass hairs. J Plant Physiol 126: 437–448

Salih AA, Ali IA, Lux A, Luxova M, Cohen Y, Sugimoto Y, Inana-ga S (1999) Rooting, water uptake, and xylem structure adap-tation to drought of two sorghum cultivars. Crop Sci 39: 168–173

Sangster AG (1970) Intracellular silica deposition in immatureleaves in three species of the Gramineae. Ann Bot 34: 245–257

Sangster AG, Parry DW (1976a) Endodermal silicon deposits andtheir linear distribution in developing roots of Sorghum bicolor(L.) Moench. Ann Bot 40: 361–371

Sangster AG, Parry DW (1976b) Endodermal silicification in ma-ture, nodal roots of Sorghum bicolor (L.) Moench. Ann Bot 40:373–379

Sangster AG, Parry DW (1976c) The ultrastructure and electronprobe microassay of silicon deposits in the endodermis of theseminal roots of Sorghum bicolor (L.) Moench. Ann Bot 40:447–459

Sangster AG, Parry DW (1981) Ultrastructure of silica deposits inhigher plants. In: Simpson TL, Volcani BE (eds) Silicon andSiliceous Structures in Biological Systems. Springer Verlag, NewYork, NY, pp 383–407. ISBN 0–387-90592-8

Takahashi E, Miyake Y (1977) Silica and plant growth. Proc IntSemin Soil Environ Fert Manage Intensive Agric, Soil Sci SoilManure, Tokyo, pp 603–611