Rhizosheath occurrence in South African grasses

Rhizosheaths are the sandy coatings which cover the entire length of each root of many of the indigenous grass species growing in South Africa. The resul ts of an extensive herbarium survey showed that rhlzosheaths occur on more than 80% of the grass species studied, irrespedive of the environmental cond itions to which the Individuals are exposed. Only 23 species did not have any sheath occurrence. The herbarium survey, together with growth experiments using Anthephora pubescens Nees, Oigitaria eriantha Steud and Eragrostis pal/ens Hack, revealed tha t the extent of the rhizosheaths (the th ickness and consolidation of the sheaths) varies not only between but also with in species. The within-species variation is a function of soH texture . The higher the sand content in the soil the greater the number of epidermal hairs produced and the greater the extent of the sheaths. A. pubescens, O. eriantha and E. palfens individuals in soil with 80% sand had 75, 11 and 100 root hairs per centimetre of roo t length respectively . In comparison, the individuals in soil with only 30% sand had 55, 5 and 45 root hairs per cm of root length respectively This relationsh ip indicates that wh ile species have a genetic predisposition to sheath development. the extent to which they develop is a facultative response to soil texture.


Introduction
Rhizos heaths are sa nd y coatings covering the ent ire length of every root in sheat h fo rmi ng grasses. Th ey have been described as thick so il cyl inders formed by modifi cat ions o f the rhi zosphere, hence the name ' rhizosheaths '. The desc ripti on of rhizosheaths dates back to the turn of the century when three authors noted their presence on grasses collected in Egypt and more widely in Northern Africa (Volkens 1887; Massart 1898; Price 1911). Further work in North America and Austra[ia in the 1980's confirmed the earlier descripti ons (Wull stein & Pratt 1981 ;Buck ley 1982). In itially th ese structures were defi ned as sandy sheaths if sand grains were held solely in place by the presence of root hairs. They were only defined as rhizosheaths if an adhesive agent was evident. However, as much of the earlier work was done using herbarium specimens, these specimens did not clearly exhibit mucilage which resu lted in th e definition of these sheaths being broadened to include a[1 sandy structures. It is now accepted that rhizos heaths consist ofa mass of sand particles matted together by a mesh-work of prolific ha ir-like epi dermal structures, mu cilage and other products released from the roots (Wullstein & Pratt 1981 ). Mycorrhizal associations are not present in rhizosheaths (Wu ll stein & Pratt 1981 ;Buckley 1982;Goodchild & Meyers 1987) but fungal hyphae may contribute to rhizosheath formation. Their role in the funct ion of the rhizosheath is unclear. Soil particles become entangled among the ep idermal hairs due to the intertwining growth of the hairs and may also be bound to the root by cell exudates from the hairs. The bonding between so il particles and epidermal hai rs is extremely strong. However, the strength of thi s bonding differs between species, si nce in certai n species the sheaths are easily removed from the root causing little damage to th e root.
Limited research has been done on rhizosheaths, particularly with respect to their occ urrence in relation to environmental conditions. Early recordings of the existence ofrhizosheaths resulted in the assumption that th ey were pecu li ar to xeromorphic species (Price 1911 ;Oppenheimer 1960 (Price 1911 ;Oppenheimer 1960;Leistner 1967;Wullstein el 01. 1979;Marneweck 1990). Duell and Peacock ( 1985) surveyed a number of mesophytic grass species growi ng in diverse soi l and mois ture conditions. Rh izosheaths were fOl llld on bo th cool and warm season perennial grasses occurring as either crop or weedy speci es. Rhizosheaths were also found to occur among species growing in both high and low fert ility conditions. There are a number of casual observations concerning th e extent o f rhizosheaths on South African grasses, not all of these species have di stribution ranges rest ricted to semi-arid nor sandy soil areas. This study a) described the occurrence of rhi zosheaths on 130 species of South African grasses and quantifi ed the extent of the sheaths and b) addressed the hypothesis that the extent of the sheaths increases as the sandiness of the soi l increases.

Material and Methods Herbarium Study
In order 10 determine whether r1u zosheath m;currcnce is excl usive 10 indi vidua ls growing in sand y soil and/or mid conditions. a herbariu m study was undc rtakcll. at the N ational Habarium (Pretor ia). Due to their hardy nature. rhizosheaths are nut eas ily rl.!Inoved from roots when the plant is extracted ["rom the soil nor when the roors arc lightly washed or shaken. They also press wel\. Therefore. herbarium specimens were ideal for slUdying individuals from many ditlcrent localities, soil types and rainfall regions. Sheaths of pressed specimens can be compared to those of fresh specimens as they are nor damaged by pressi ng. Th e nim of thi s study was not to detenn ine causc-and-effec t bu t rather to derenn inl.! whether a rdlltionship existed between sheath occurrence ,lIld thickness and env ironmenta l condi tio ns. The advantage or the herbarillm study was Ihar it allowed a wide spectrum of species 10 be samp led.
All specimens of grass sr ecies. indigenous tn So uth Afri ca, were studied provided they had intaci rhizosheaths showing no damage and recordings of soil texture cond itions and grid references. A total of I 260 specimens from 130 species was stud ied. ror eac h specimen , for each of the primary roots, th e exten t of the rhizosheaths. ie Ihe thic kness and conso lidat ion oflhe sheaths was rated lIsi ng a scale s. Afr. .l. Bol. 1997. 63{6) of 5 (maximum th ick ness and consolidation) down to 0 (no evidence of slll.:mhs): 5 -Sheathed roo t diamekr: greater than 3.5 mm. \vell consolidated sheath (soil particles adhere tightly to the sheath). Prolific e pidermal ha irs that arc no t easily visihle with the naked eyt! due to the large amoun t of sand ill the sheath. 4 -Sh.:athed rool diameter: 2.6-3.5 mm. shl.!<.u h otten less conso li~ dated \\ ith m ore obvious epiderma l hairs.
The annua l ra infa ll for the area in \vhich eilCh specimen grew was determined llsing the grid n::lerem:cs and rai nt~IiI maps of South A fricil .

Growth Experiment
Secds. collected from liel d sites. of Ant he ph ora pubescens, Eragrostis pal/ells and Digiraria erial/tha were used ill this expe riment. All indivlduals o f the ti r~il t wo sp~cies were fo und. during the herbariu lll study. to have d ist im:t rhizoshcaths. D . crianlha was J"ound to have gre,Uer vari ab ili ty in th e extent of the sheaths than the other species depending on th e soi l and rainfall cond itions under which the individuals grew. The ind iv iduals from whic h the seeds were co ll ected lwd thi ck rh izoshe<.lths.
Seeds we re gro wn in I L pots in a controll ed-cl im ate gnnvth chambt::r usi ng a 14 ho ur day (65-80 ~Lmohn -:!s-l quantum flux rate) at 25°C. followed by 10 hours of darkness at 16°C. The pot s we re randomi scd weekly to elimin ate differences due to pos iti on in the chamh.:r. FiVt: sets {)fpiants for each species were used. each Sl.!t \vas grown in a d ifkr en t sllil texture ( Table 1).
Each treatment consisted of five pots/rep licates, each containing hdwt::en ten and six tee ll ind ivid uals per rep licate. The ~m nd (80% sand. 20% d ay con tent) was co ll ected from a broad leaft:d savan na area wht: re E pallr.!1Is and D. er;amha grow. Kaolin powda was added tl) the san d to make up the clay fract ion in treatments 2-5. The texturc \\las accuratdy determined lIsing a O OliCOUS Hydrometer ( Day 19(5) The li eid capaci ty for each soi l treatmen t wa" ca lcu lated prior to the experiment and the plants were watered regul arly to keep the soi l at 70% tiekl capacity. It was necessary to water thl.! sandy soil pots more fr eqllt:lItly since this soi l drains fas ter th an clayey soil.
Watering was carr ied out by weigh ing each pot and adding the requ ired amoul1I o f waler. Th e nutrient conccntration in the different soil treatments \vas not controlled.  )0 After twe lve weds of growth tht! plal li s were uprooted and the presence and extent of rhiz() sheaths was scored for tl.!n individua ls per replicate using the same scale as ill the hernarium study. Using a dissec ting microscope. the number of I!pidermal hairs per unit length of root were counted on fi ve u f the pri mary roo Is 01" each ind iv idual.

Herbarium Study
Rhizosheaths occur o n ind ivid ual s grow in g in a ll soi l textures and in a ll rai nfa ll areas ( Figure I ). The presence ofrh izosheaths is therefore independent of soil texture and rai nfa ll conditio ns.
The presence is al so not restricted to ce rtain genera or even tribes ( Relat ionsh ips appear to exi s t between the extent of rh izos heaths and soil textu re and rainfall condi t io ns (F igure I ).
There is a h ig h representation o f spec ies in sa ndy. ari d (ra infall less than 400 I11m/annulll) areas. Among these species the rhizosheaths are ge nerally thi ck and we ll conso lidated. There is also a high representation of species in semi -arid a reas (rai nfal l 400-900 mm/a nnum ), es pec iall y on cl ay soi I. I n both these rainfa ll categories the indiv idua ls in the sandy so il region s genera lly have more extensive s heaths than th ose in clay soi l. These trends indicate a relation ship between th e extent of s heaths and so il texture.
T he average exte nt of sheaths in sa ndy so il is lower in high rainfa ll (m ore than 900 1111ll /annum ) areas than that in arid and semi-arid areas. W ith in high rainfall areas, Ih e sheath-forming species deve lop s heaths o f si m ilar cxtent irres pective o f the so il ty pe. There is a lso a hi g h nu mber o f s pec ies w ithout sheat hs in these areas. These tren ds show that a re lati onship ex ists between extent of sheaths (thi ckn ess and co nso lidati on) and rainfall co nditions, especially among indiv iduals found growi ng on sandy soi l.

Growth Experiment
The higher the sand content the greater the number of epidermal hairs produced and the greater the extent aflhe sheath, (Table 3).
Si mil ar trend s were ev ident w ith regard to the change in number of epiderma l hairs and ex tent of rhi zosheath s in respon se to the percentage s and content in the soi l.
The increase in rh izos heath ex tent ap pears to be a response to a n in crease in sand co ntent in th e soi l. T he thi cker, m ore conso lidated rh izos heath s occu rred on ind ividuals growing in soil with the highest sand content.

Discussion
The majority of SOllth African grass species have a genetic predispos ition to develop rhizosheaths. There is no re lationship between occurrence and thick ness of sh eaths and soi l texture or rainfall conditions. Sheaths occur on indi vi duals regardless of the soi l texture. Similru iy, sheaths occur on indi viduals fro m all rain fa ll regions studied. There is also no correlation between sheat h occurrence and whether the species is an ann ual or a pe ren ni a l.
A d isti nction m ust be made betw een the prese nce of sheaths and the extent of the sheaths. While spec ies have a genetic predispos ition to develop sheaths, in all except six spec ies, th e extent of th e sheath s is a fac ultative response to so il texture and rai nfa ll co nditi o ns which directly affect soil wa ter co nt ent. The genera l trend in all ra infall regions is tha t in sandy soil the individ u(lls have more extensive sheaths than ind ividua ls in more cl ayey soi l. T his is particularly marked w ithin the lower ra infall reg ions (900 mm/ann ulll o r less). Since diffe rent so il textures have di fferent c hemical as we ll as physical pro pert ies t hey present the pl ants with different envi ronmenta l cond itions, w hi ch resu lt in m ore or less favo urable growth conditi ons fo r the plants.
for exam pl e, sandy so il has a lower water ho lding capacity and nutrient content th an clay so il. Sh eath occurrence may be an in di rect response not o nly to th e di fferent textures but also to the d ifferent wate r ho ldi ng capacities and nutrient avai labilities of the soils. Th is hy pothes is is sup po rted by the obse rvation that, com pared with the si tuation in low rain fall a reas, in hig h mi nfa ll reg ions ( more than 900 111 m/annum ) the extent of the sheaths is similar in ind ivid uals on all so il types. Thi s sugges ts t hat the occurrence of thicker sheaths may be a res ponse to lo wer soil water co ntents.
The growth experiments confirm ed th e hypothes is that the ex tent of sheaths increase in response to increased sandi ness of the soil. It must however be remembered that th e nu trient co nce ntrat io n of til e so il used ill the exper iment was c ha nged by the additio n of cl ay. A dding vari able am ounts of nutri ents to each pot to com pensate fo r the add ition of nu trients in the cl ay fracti on would no t have bee n easy and wou ld probab ly ha ve co nfo unded th e expe rim ent. T he kao lin additi on may also have a ffected th e pH and thereby influ enced roo t growth, since higher amou nts of calcium and thus a higher pH may have in creased root prod uct io n.
A !tho ugh th e e xtent of the sheaths varied between th e three species, the trends were s imil ar withi n a spec ies. W ith in each species the most ex tens ive sheaths occurred in the ind iv idual s growing in sa ndy soil. T he hig her the sand conte nt in the soil the higher the density of epidermal hairs that were prod uced and the thicker, more conso li dated the sheath. This findin g is in accorda nce wi th previous work wh ich showed that the density of epiderm a l hairs is greater in sandy soi l (Kutchera 1960 in Leistner I <)67). It has been observed th at roots in sandy soil prod uce more ep idermal ha irs tha n roots of the sam e species in less sandy soi l ( Fey, pers eOl11m ). T he more ha irs there are th e g reater the numbe r of tigh tly bound soil pa rticles the re a re. T his bonding between soil particles and ep idermal hairs is extremely strong. especially when form ed in sandy soi l. T hi s work confirms that of Due ll and Peacock (1985), who showed that the occurre nce of rh izosheaths is not confined to a few species of grasses growing under xeric condit ions in sandy so ils. The sig ni fican ce of th ese rh izos heaths needs to be e luc id ated. T he re has bee n some suggestion that rhizos heat hs aid in seed li ng establishm ent un de r harsh environme nta l co nditi ons. T here is little data to su pport th is suggest io n. Due ll and Peacock ( 1985) reported that seedl ings had to attain a critical size before rhizosheaths were di scern ible on noda l roo ts. Rh izosheaths were never fo und o n semi na l roots. T he need fo r an improved mec hanism for overcomi ng d rought stress. as has been suggested in the case of desert species (Buckley 1982). would not seem appli cabl e to meso phytic grasses.
It is also suggested that these rh izos heaths are instru mentel! in the suppl y of ph osp ho ru s to plan t roo ts by creat ing an enviro nment favo urabl e for soil mi reo-organisms. Phosphorus. be ing relatively im mobil e, rel ies on d iffusion for uptake and rhi zoshe aths may aid this process. Studies to address this hypothesis are in progress.