THE PHYLOGENETIC Neil Snow^ LEMMATAL UTILITY OF MICROMORPHOLOGY IN LEPTOCHLOA AND S.L. RELATED GENERA IN SUBTRIBE ELEUSININAE (POACEAE, CHLORIDOIDEAE, ERAGROSTIDEAE)^ Abstract Micromorphological lemma features of the were investigated in the suhtribe Eleusininae (Poaceae) using scanning electron microscopy. Ninety-two taxa were studied, which included 48 genera and all currently recognized species of lA>ptochloa S.I. All species of Uptochloa and a majority of genera in Eleusininae have cork cells, hut silica cells are mostly ahsent in IjpptoMoa and most related genera Eleusininae. Enneapogonoid-type in microhairs are reported for Cladoraphis cyperoides and Psammagrostis wiseana, representing only the second report Eleusininae. Although in bicellular microhairs occur most in taxa. variations in microhair type, coupled with the occurrence of partitioning membranes, them give supra-specific phylogenelic utility. The occurrence of papillae on short and long cells varies within [j^piochloa and between genera. Prickles are ubiquitous in Leptothloa and most related genera, Macrohairs are present in all sj»ecies of Ij'ptochloa but a[>sent in some related genera. Clavicorniculate macrohairs are reported for one species eadi of Lt^ptochloa and Coelarhyninu and a corniculate macrohair was A observed Cyrwdon. in crispate macrohair occurring in three species of Coelachynim is described for the first time and probably represents a syn- apomorphy The in that genus. analysis oi Uptochloa suggests that micromorphological characters vary within a little genus, and thus have the potential to ser\'e as phylogcnctic markers at the generic level. Uncertainty concerning the homologous relationships of papillae, hooks, prickles, and macrohairs is discussed in light of this study and previous literature. Resumen Las caracterfsticas micromorfol6gicas de la lenm en la sublribu Eleusininae (Poaceae) se examinaron utilizando microscopfa electr6nica de barrido. Se estudiaron 92 taxa. 48 g^neros, y todas las especies de Leptochloa liasta ahora reconocidas. Todas las especies de Leptochloa y la mayorfa de los g^neros de Eleusininae tienen c^lulas suberificadas, pero las c^lulas silicfferas estdn ausentes en lA'ptochloa y en la mayorfa de los taxa de Eleusininae. Se reporta por segunda vez en Eleusininae la presencia de microjielos bicelulares de enneapogonoid tipo para Cladoraphis cyperoides Psammagrostis wiseana. Auntjue micropelos y los bicelulares se j)resentan en casi todos los taxa, variaciones en el de tipo micropclo, junto con la presencia de membranas dlvisorlas les dan valor diagn6stico para estudios filogen^ticos supraespecfficos. La presencia de papilas de c^lulas corlas y estrechas es variable en Leptochloa entre los otros y g^neros. Los aguijones se encuentran en todas especies de las Ijeptochloa mayorfa de g^neros y la los relacionados. Los macropelos estdn generalmente presentes en todas las especies de Ijeptochloa pero estdn ausentes en algunos taxa relacionados. Se reportan macropelos clavicorniculados para un especie de ambos generos Leptochloa Coelaehyrum, y un macropelo y corniculado se reporta para Cynodou. Un macropelo crispado en especies de Coelachyrum tres es descrilo por primera vez, y quizd represente una sinapomorffa de cierto nivel en ese g^nero. El andlisis de Leptochloa que indica las caracterfsticas micromorfolr)gicas varfan poco, por tanto tienen valor para estudios filogen^ticos su- y La praespecfficos. dificullad sobre las relaciones homologas de papilas. ganchos. aguijones, macrotricomas se discule y con base a lo halhulo en este estudio con datos anteriores de y la literatura. Cladistic eistirnationsof pliylogenctic history with easily defined character This tendency may states. morphological characters tend to favor the use of reflect the potential pitfalls discussed by Chappill distinct qualitative characters, or those possessing and Stevens (1989) (1991) in the uncriti(^al use of Mike ^ Veith facilitated the work with assistance at the Electron Microscopy Laboratory Washington University at I thank the curators of B, BM, BHI, CANB, MO, PRE, and US for sending loans and granting pemiission sample for lo directly from herbarium specimens. Hatch an S. sent original version of Morden's dissertation, and helpful discussions Ann. Missouri Bot. Card. 504^529. 83: 1996. Snow 505 Number Volume 4 83, Lemmatal Micromorphology 1996 Used documented thoroughly (Kaufman 1969; quantitative characters (but see Thiele, 1993). et al., Kaufman 1970) and are evident in grass fos- and due the perception of high levels of et al., alone, to morphology have of Mi 1984) homoplasy, characters of gross sils been considered inadequate as phylogenetic mark- elongation of long cells follows the initial differ- & (Kaufman and ers in the grass family (Thomasson, 1978; Hilu entiation of long cells short cells et & & Campbell, 1987; Davis 1969; Kaufman 1970). In addition to Wright, 1982; Kellogg et al., al., Soreng, 1993; Clark et al, 1995). Not surprisingly, their extended length, long cells (unlike short cells) revisionary studies of grasses are turning to micro- frequently have sinuous margins (Metcalfe, 1960; & Chen morphological characters for additional phyloge- Clifford Watson, 1977; Ellis, 1979; et al., and between long However, the distinction netic data. 1993). Thomasson Epidermal micromorphological characters of always absolute. short cells not is have systematic value between the ranks of long identifiable by grasses (1978: Id) illustrated cells, fig. subfamily and species (Prat, 1932; Tateoka et al., sinuous margins, having similar dimensions to {[^^ 1959; Metcalfe, 1960; Ellis, 1979). Descriptive short cells for a fossilized species of the genus A^a5- studies of micromorphological features in grasses (t^be Stipeae). The two types of short cells ^^//^ 1932; have focused on the surfaces of the leaf (Prat, recognized are cork cells (sometimes called suberin & summary 1976; see also Morden Hatch, (Kaufman in Ellis, which accumulate suberin et al., eells), & & Annable, 1987; Ddvila Clark, 1990; Peterson which accumulate ^^d silica ^979)^ silica cells, Chen 1990; Barker, 1993; Scholz, 1993; et al., observed bodies. j^^^ readily silica glumes (Lucas, 1979; Molina, 1993), the 1993), require Microhairs are bicellular structures that lemma Thomasson, 1986; Peterson, (Hsu, 1965; ^h magnification for detection (Tateoka et al., & 1990; 1989; Soderstrom Zuloaga, 1989; Kellogg, condition 1959). (For exceptions to the bicellular & & Valdes-Reyna Zuloaga Judziewicz, 1991; and Zu- Dahlgren 1985; Renvoize, 1985; see al, et Hatch, 1991; Naredo et al., 1993; Ball et al., 1993), j^^^^ ^^ ^^^ jggg ^j^^^ ^^^ present throughout the -^ & (summary Consaul Aiken, 1993: & and palea in ^^^^^ ^^^^p^ subfamily PooiJeae (Johnston j^^ micromor- The phylogenetic application of 1651). 1975) Microhairs have teen classified as ^^^g^j^^ phological characters has been more limited (Pe- "enneapogonoid," -^hloridoid," "panicoid," or & & Annable, 1992; Barker, 1993; Visser terson 1959; ^^^^^ ^^ typological variants (Tateoka et al., 1994; Guala, 1995). Spies, 1988 Only ontological status as a 1992). rarely their is An Overview Micromorphological of character uncertain. For example, as dis- distinct Characters cussed by Barker the "long slender papil- (1993), common & l^e" (Clayton Renvoize, 1986: 165) to micromorphological char- Several categories of genera Arundinoideae, given that they several in have been recognized. Short cells (which in- acters are reported to occasionally have the remains of a and long bicel- elude cork cells silica cells), cells, may be ^^^all apical cell (Renvoize, 1986: 328), and microhairs, papillae, hooks, prickles, lular icroh interpretable airs. examples micromorphological macrohairs are of all processes Papillae are short, undifferentiated characters (Metcalfe, 1960; Ellis, 1979). (See Ellis, summary terms hooks that arise from the outer cell wall. 1979, for a of alternative for Hooks have been considered processes having a and prickles.) rounded base and an apex that at least slightly is Short cells and long cells are readily distin- recognized as a and (when pointed (Ellis, 1979), but are not guished on the basis of relative size, may synapomorphy category In this study (see Discussion). Prickles are both are present) constitute a and Poaceae Doyle basally swollen processes having short, sharp api- uniting Joinvilleaceae (see also & The ces that typically point toward the apex of the struc- 1992; Kellogg Linder, 1995). differ- et al., ences between short cells and long cells have been ture (1 Columbus were provided by Filgueiras, D. Nicolson, G. L. Stebbins, aiiJ F. Zuloaga. The careful reviews of J. T. T. Spanish. and G. Davidse improved the manuscript significantly F. Lorea assisted with preparation of the abstract in Wash- Garden (Andrew W. Mellon Foundation); acknowledged Missouri Botanical Financial support gratefully from: is Louis (Evolutionary & Population Biology Program); and the American Society of Plant Taxono- ington University, St. and This paper represents a portion of a doctoral dissertation to be submitted to the Graduate School of Arts mists. Sciences, Washington University Louis). (St. , and Missouri Department PO. Box 1137, Louis» Missouri 63130, U.S.A., ^Washington University, of Biology, St. Garden, PO. Box 299, Louis, Missouri 63166, U.S.A. Botanical St. 506 Annals of the Missouri Botanical Garden Macrohairs arc generally unicellular structures with uncertain phylogenetic affinity (Dalilgren et & (but see Kabuye Wood, 1969) visible to the na- al., 1985; K(>]logg & Campbell, 1987; Doyle et al.. ked eye & 1992; Davis Soreng, 1993; Barker 1995; et al.. Among lemniatal niicromorphological & structures, Clark 1995; Kellogg The et al., Linder, 1995). only niicTohairs have been not reported to accu- subfamily Chloridoideae Rouy has about 13(X) spe- mulate The silica. presence of silica has been re- cies, which tend occupy warmer to dry, climates ported cork (Kaufman & for cells et al., 1972); long (Hartley Slater, 1960). (Taxonomic concepts fol- & & cells (Ball et al., 1993); papillae (Terrell Wergin, low Clayton Renvoize, The 1986.) largest tribe, & 1981; Consaul Aiken, 1993); prickles (Soni Eragrostideae 80 et Stapf, has about genera and 1000 Kaufman 1970; 1972 al., et al., (as "trichomes"); species. Within Eragrostideae, the most diverse & & Terrell Wergin, 1981; Valdes-Reytia Hatch, subtribe Eleusininae Dumort, and is only the high- 1991; Ball et 1993); and macrohairs (Consaul al., est latitudes arc excluded from the overall range of & Aiken, 1993). (Stomata occasionally occur on approximately 55 its genera. Relatively few genera lemmas. They are sufficiently infrequent in the taxa in Eleusininae occur outside of a single continent examined thus far to be considered an abnormal Coelachyrum, (e.g., Eragrostis, Leptochloa, Irichn- constituent lemma. of the In a study of the leaf of neura, disregarding minor exceptions and a few Kaufman Avena, et al. (1972) were unable to find cosmopolitan weeds). Excluding Eragrostis, with its silica in stomatal cells.) approximately 350 & species (Clayton Renvoize, & 1986; Van den Borre Watson, most 1994), of the Overview Lemma of MiCROMOFii'iii.ocY approximately 600 species in the subtribe occur in Up- The lemma '"^^^»'^''1>' ^'"^'^ g^'"'"-^ (10 species or fewer). is the lower of two bracts that (usu- worldwide genus ^^''" ^*'^"''- "^^'^^ subtend each The ^- '" ^ <,f ally) flower in grasses. upper ^f many ^^"'"* ^'thougH continue seg- to bract (the palea) not homologous lemma 'P*''^*^"' the is to '"^f the genus Dlplachne (McNeill, 1979; Phil- The lemma r^at*' (Clifford, 1987). a transformational is ^ ^^^^- ^'^''"" Simon, D^^'^i'^' 1992; 1993; homology serial (de Pinna, 1991) of the leaf by vir- ''I*'' ^•''""' l^^-"^)- tue of "ontogenetic its individualization" (sensu mlergenenc studies Eleusininae ^*-''''"* in dif- Wagner, 1989a, from b) the leaf (Pfiilipson, 1934; characters, and meth- '" '^'^'^''^' analytical Tran, 1973; Kellogg, 1990). transfonnational ^^ff Its ^^u-h hampers ^^^^"g'*'^' meaningful comparisons derivation from the leaf supported by rare is ata- & between the results (Phillips, 1982; Hilu Wrijrht, '^ • progenit,or hI aving ^ afl \ } ' ,^,^^ ' & . . 1982; Campbell, 1985; Hilu Esen, 1993; Duvall sheath, ligule, and blade (Philipsou, dim- 1934), its ""^ ^'•' ^'^^'^^- ^^'"^ 1"*"'*'"" '*'' '" »'»' g^'- inuitive size relative to regular leaves, restricted ^P^''^-'^^^'^ its "^"'' arrangement of Clayton and Renvoize (1986: occurrence in the inflorescence, the occasional non-mutually exclusive possi- presence of Icmmatal stomata (common j"^^' '"S^"^"*" ^^'''''' to the leaf), many ^7''''- leptochloa has close generic and ^^> '''^' the frequent infraspecific differences of micro- ""^^l^^f';^^', that U-ptochloa^ evolutionarily basal (2) is morphological features between and lemma the leaf & to related genera, and that Leptochloa may be (Thomasson, (3) 1978: 977; Terrell Wergin. 1981: paraphyletic The as currently recognized. mono- 706 ^^'* chloridoid groups, ineluding Upto- ^'^^^^ The most "extensive work on lenuna ''^ micrornor- '''^"'^'"' untested. ^^'^'''^ phology Eleusininae was in that of Vakles-Reyna '^'^Z'' The purpose of this paper create a more is to and Hatch (1991). Their sur\ey Eragrostideae in of ^'^^ensive data set for inferring phylogenetic- rela- 57 species (representing 30 genera) suggested such ^lonships in L-ptochloa and Chloridoideae (Snow, characters might be useful phylogenetic markers in ^^^itio"' terminological ^" inconsisten- the subtribe Eleusininae. This study extemled '" P'^f"'^' U.eir 'ies throughout the literature, coupled with re- generic survey of niicromorphoh)gy (Valdes-Reyna tl y' & '"^*' ^^^'^ q"*'^''"" "'c onto- Hatch, 1991) to include genera of Eleusininae '*'''^'"' I"""^'^*^ ^ '" ^''^*"' "*^ '^"^''^^ characters, and suggest the ^''^^ not previously examine<l, including a few other gen- "'^^^'^ ^"' ^ ^''"*'^' reevaluation of the characters, era and in Chlori.loideae, to characterize these fea- tures for all species in Leptochloa. Material and Methods An Overview of the Svstematics of The lemma adaxial surface of the was studied f ClILORrnOIDEAE 92 taxa using scanning electron microscopy (SEM) The composed grasses are of at least four mono- the Electron Microscopy at Lahorafor>' Washing- at phyletic suhfamilies and number contain a of tribes ton University (Apj)endix Samples were 1). re- Snow 507 Number Volume 4 83, Lemmatal Micromorphology 1996 was moved from her])anum specimens. To as- (except those with question marks, Table 1) directly with a photomicrograph, semaphorants 1981: verified sure the use of (Wiley, 119), I Cork Cork were abundant in the in- sampled lemmas from splkelets that con- cells. cells generally and mature caryopses. This precaution was Im- tercostal zones in all species of Leptochloa, al- lained As was of micro- portant, because spikelets in early stages of ontog- temated with long cells. true all when morphological characters studied, their distribution cny may be present even the inflorescence is lemma. There Three specimens could vary significantly on a single well exserted from the leaf sheath. decrease Some samples were was a general tendency for cork cells to most were observed. taxa of sonicated in xylene for 30 minutes to remove the in frequency near the apex. Among Desmos- OET, were absent cork in waxes can obscure surface fea- cells epicuticular that Ectro- treatment was not always effective. tachya (Fig. 14), Ectrosia gulliveri (Fig. 17), tures, but this Psam- and mucronata being placed on aluminum stubs the speci- siopsis, Odyssea (Fig. 60), After Odyssea E5000 Unlike congener, magrostis mens were coated with gold using a Polaron (Fig. 65). its The Apo- has cork cork cells of Samples were observed with 0*" paucinervis cells. sputter coater. tilt SEM Halopyrum and Neyraudia 20 kV on a Hitachi S-450 and photo- chiton (Fig. 2), (Fig. 23), at com- graphed using Polaroid 55 positive-negative film. (Fig. 58) were not noticeably darkened, as is SEM mon under observation mature cork Except Figures 10 and 42. photomicrographs for cells for all & However, since lemma toward (Valdes-Reyna Hatch, 1991). have the apex of the directed the Attempts standardize the scale of photo- their size and location suggested cork cells, they right. to were recorded present (Table micrographs proved unworkable, since various as 1). OERG except Cork cells were present in all magnifications were necessary. Valdes-Reyna and Hatch Spartina (not shown). With the exception of a The from format differs Tragus shown), cork cells microhairs and macroliairs are in- sparse distribution in (not (1991) in that common OERG. my on were in eluded, interpretation of papillae short cells Some had resembling cork cells different (see Results), and the degree of undu- taxa short cells is was and from which other structures could be lation of the long cells with papillae not re- in size, Valdes-Reyna and Hatch (1991) discussed seen developing, such as bicellular microhairs (e.g., corded. Diplachne gigantea, their findings based on four variations of cork cell Bewsia, Fig. 3), prickles (e.g., and macro- occurrence: cork cells adjacent to silica cells; cork Fig. 38, Leptochloa rupestrls. Fig. 48), L nee- Leptochloa fascicularis, Fig. 36, cells not adjacent to silica cells; cork cells papil- hairs (e.g., Many or cork cells not observed. This study merely sii, Fig. 45). taxa such as Eragrostiella (Fig. late; whose had and Indopoa short cells character (present 20) (Fig. 26) recorded t^ork cells as a binary^ and which were were rounded Eleusininae be- outer walls in profile, or absent). will refer to taxa in I become "Other Eleusininae Taxa" ontogenetically destined to cork cells. (Val- Leptochloa as sides & des-Reyna Hatch (1991: 536) referred to these (OET), whereas other taxa in Eragrostideae (fide & cork That they were destined to Clayton Renvoize, 1986) will be designated as papillate cells.) (OERG) become cork became evident by examining LL Genera" cells Other Eragrostoid lemma, development on a in This study focused on the presence or absence different stages of were of characters. No attempt was made to measure which cells at the apical and basal portions The development. stages of initially quantitative variation of the characters, but distri- in different and frequency on the lemma sometimes rounded outer walls of short cells appear to collapse bution is coincident with or prior to suberin deposition; this discussed. process was evident the shrunken tissue of the in coupled with darkened lumens of the outer wall, Results Leptochloa Leptocarydion, Fig. 28, short cells (e.g., A summary lemmatal microcharacters ob- ig. 52). of ifl were absent spe- Table Sillica ceUs. Silica cells in all monticola Valdes-Reyna and Hatch (1991) are in- cies o( Leptochloa except for L. (Fig. 41), results of' specimens Unless noted otherwise these re- which had abundant silica cells in all tercalated therein. L Valdes-Reyna and Hatch examined, and one specimen {Snow 581 1-A) of accord with those of suits which had a few silica cells (1991) for taxa examined in both stuthes. Figures fascicularis (Fig. 35), are presented at the end of the text, arranged in present in the intercostal regions. OET alphabetical order by genus name. Figures are not A majority of genera had silica cells. In Halopyrum they were poorly and characters observed or (Fig. 23, not visible) presented taxa for all obscured developed (minimal deposition?), every character recorded silica discussed, but virtually 508 Annals of the Missouri Botanical Garden Table Summary- of lemmatal micromorphological 1. characters for species oi Leptochloa, most genera in Eleusininae, and a few genera in Eragrostideae outside Eleusininae & (fide Clayton Renvoize, 1986). Genera not included in Eleusininae are indicated with a double asterisk. Taxa are arranged alphabetically, Taxa examined by Vald^s-Reyna and Hatch (1991) but not examined herein are followed by single asterisk; their data are intercalated convenience for by: relying on Watson and (1) Dallwitz (1992) for the type of microhair (as they determined from leaf blades); (2) examining herbarium material for the presence of prickles and macrohairs (Appendix and by leaving a question 1); (3) mark column in the for papillate short since cells, their interpretation of papillae differed (see Results). The generic abbreviation for species tentatively placed L-ptochloa in that apparently lack a valid combination in the latter D. is _* *« rx / j^ Di'pl'Mach1 ne)\ + ---fc + = BM (for specific epithet. Abbreviations: = present; absent; bicellular microhairs (C P - E = = = chloridoid, panicoid, enneapogonoid); Macrohairs (N normal, CC A CR clavicorniculate, apiculate, PLC = crispate); PSC papillate short cceelllls. Ta xon BM Cork PLC PSC Silica Prickles Macrohairs Acrachne racemosa C Apochiton burttii C N Beusia + hiflora C N Bouteloua curtipendula'^* C N Brachychloa schiemanniana C Chloris paniculata*'^ C + N Chloris verticillata'^'^ c N Cladoraphis + cyperoides E N Cladoraphis spinosa m m Coelachyru c brevifoliu CR Coelarhyrum + poiflorum c CR Coelachyrum stoloniferum c CR Coelachyrum yemenicum + c + CC Chondrosum gracile*^ c + N Cynodon + nlemfuerisis** c A/N * Dactylocteruum c Desmostachya + hipinnata C/F + Dinebra polycarpha C + N -I- H- Dinebra retroflexa C + N Drake- Rrockmania + haareri C + N Drake-Brockmania somalemis C N + Ectrosia gullireri p N Ectrosia leporinn + p -I- Ectrosiopsis hisioclada p 4- Eleusine indica + C -I- Eragrostiella bifaria C * Eragrostis spp. C/P Erioneuron spp.* C N Gouinia rirgata C N -I- Habrochloa bullockii P + N Halopyrum mucronatum c + + N Harpachne schiniperi c Heterachne abort iva P + Indopoa paupercula c N 4- Kengia serotiria c + N Ij'ptocarydion vulpiastrum + c + N -f- Ijeptochloa aquatica c + N D. caudata c N Leptochloa + chinerisis c + + N + Leptochloa chloridiformis c N Leptochloa ciliolata N Leptochloa coerulcscens c + N -i- D, ciispidata + N Leptochloa decipiem N Ijeptochloa digitata c N Leptochloa + divaricatissima c N Snow 509 Number Volume 4 83, Lemmatal Micromorphology 1996 Table Continued. 1. BM PLC PSC Macrohairs Prickles Cork Taxon Silica N C Leptochloa dubia cc + C D, eleusine + N + c Leptochloa fascicularis + + N + c Leptochloa fiisca I + c N/- gigantea D. N + c Leptochloa ligulata I N + c Leptochloa longa + N + c marquisemis Leptochloa + N + c Leptochloa monticola N c mucronata 4| Leptochloa N + + + c D, muelleri + N + c Leptochloa nealleyi N + c Leptochloa neesii N c Leptochloa ohtusifiora I N + c panicea Leptochloa + + + N c Leptochloa panicoides N c D, parviflora N c Leptochloa rupestris I N + c Leptochloa scabra ^- ^ Leptochloa sp. nov. + N c (Snow, in prep.) I + I N + c -h Leptochloa squarrosa I N c 4- leptochloa uniflora N + c + Leptochloa uninervia -f- N c Leptochloa virgata I N + + c Leptochloa viscida I I N c Leptochloa xerophila + N c Lintonia nutans**' -i- + N C/P Lophacme digitata N + I c ? Munroa squarrosa* I + c i- Myriostachya wightiana N + c Neesiochloa barbata N + p Neyraudia reynauldiana N + c Ochthochloa compressa + N + c + Odyssea mucronata N + + c Odyssea paucinervis + + N + c Orinus thoroldii -i- I N c Oropetium aristatum N c Pogonarthria Jieckii N c Pogoneura biflora -I- + E Psammagrostis wiseana + + c Psilolemma jaegeri c 9 Redfieldia flexuosa* 4^ N + c Richardsiella eruciformis I c macrostachyum Sclerodactylon + + N + c ? Scleropogon brevifolius* c Sohnsia filifolia* N r + c Spartina pectinata** I + + c barbata Steirachne + + N c pedunculatus** Tragus N + c grandiglumis Trichoneura I N + c ? Tridens spp.* N 7 + c spp.* Triplasis N + c Tripogon major + N + + p andropogonoides Triraphis + N c 9 Vaseyochloa multinervosa* H- + + + madagascariensis Viguierella 510 Annals of the Missouri Botanical Garden by epicuticular waxes, or hoth. AVng/V/ (Fig. 27, not chiton and Triraphis were th(7 cotnnion. Nirora and visible) Eleusine (not illustrated) had a few sil- (1962: 9) illustrated eniu^apogotioid inierohairs in ica cells that were localized over the c^oslal zones. Eleusininae for JSecragrostis, which Clayton and The proximity of silica cells to cork cells was a Renvoize (1986) included in Eragroslis, Tfie en- character noted by Vald^s-Reyna and Hatch neapogonoid microhairs observed here CUido- for (1991). Silica cells were not adjacent to cork cells raphis cyperoides and Psammagrosiis wUeami (Fig. Desmostachya in (Fig. 14), Ectrosia gulliveri (Fig. 65) represent only the second rc^port for this tyj)e 17), Ectrosiopsis (Fig. 19), Myriostachya (Fig. 56), in Eleusininae. All microhairs in Psammagrosiis Odyssea mucronata and (Fig. Steirachne had 60), (Fig. (Fig. 65) swt^llings <hstal to the base of the were 69). Silica cells adjacent to cork cells in Cla- microhair, a condition observed not clscwh(^r<\ OERG ioraphis (Fig. Ectrosia leporina Har- 7), (Fig. 18), All had chloridoid micro hairs. pachne (Fig. 24), Odyssea paucinenis (not shown). Papillae, Papillae occuned singly on short Psilolemma (Fig. Richardsiella and 66), (Fig. 67), cells or singly on the distal <Mids of long cells. Sclerodactylon (Fig. In Pogonarthria Three 68). (Fig. 63) clarifications are needed prior to presenting and Triraphis (Fig. 71) their location varied (cork observations of papillae. cells adjacent shown to silica cells not for Trira- First, the distal ends of the outer walls of <^pi- phi^). It was uncertain whether the dark, narrow dermal hmg cells can be noticcal)ly swollen, as in hands some adjacent to silica cells in Heterachne Eragrostiella Indopoa and (Fig. 20), (Fig. L'p- 26), (Fig. 25) w<*re artifacts of preparation (e.g., shrink- tochloa virgata (Fig. 52). These swellings were rcc- ing of cell walls adjacent to the silica bodies noted ognized by Valdcs-Reyna and Hatch (1991: 23) fig. by Kaufman & ct al, 1972: fig. 6, and Terrell Wer- in at least one instance as papillae. As inlcrpn^t I gin, 1981) or whether the narrow bands were cork them, outer walls that arc merely swollen are dis- cells that were partially obscured by overarching from The tinct true papillae. distinction between silica bodies; the uncertainty was enhanced be- long possessing cells a single, distal papilla (e.g.. cause not all silica cells fiad adjacent bands. Since Leptochloa coenilescens, Fig. 31) long anil cells that a few short cells with dark lumina were visible ad- are merely swollen can be seen with h^ptochJoa chi- some jacenl to silica cells, cork cells were scored nemis which (Fig. 29), in a single pa[)illa evident is as present Heterachm\ obsene for I did not silica atop the distal swelling of a long and ct^ll, willi /.. Lept /' by Valdes-Reyna and Hatch (1991: fig. 24, tab. 2), pillalc or miMely swollen. Second, the outer wall of they were recorded as present (Table 1). short cells prior to diffen^ntiation can a|)pear round- Among OERG, silica cells were observed in ed (papillate, fide Valdes-Reyna & Hatch, 1991) Spartina and Tragus, being more common when in the viewed in profile (e.g., Eragrostiella, Fig. 20; '^**^^- Indopoa, The rounded Fig. 26). outci* walls should Bicellular microhairs, Chloridoid microhairs be confused not with which papillae, are lo<-alized were observed in all species of Leptochloa except processes arising from the out(M- wall. Third, tlic L L cdiolata, Diplachne and cuspidata, decipicns. apical extensions of the lateral walls of loni^ cells Microhairs generally were scattered, but in L;pto- of Leptocarydwn and (Fig. 28) Steirachne (Fig. 69) L and c/t/oa cauf/a/a /on^^-a they were c(mmion near are distinct from Ime Palmer and papilhu-; Gcr- The the apex. basal cells of microhairs of L^pto- belh-Jones (1988: 94) have noticed a similar dis- L chloa panicoides and (Fig. 47) viscida (Fig. 53) tinction in Phacelurus (Panicoidea<% Andropogo- were alypically thick and nontapering lowanl their neae), as did Pt^t<'rson (1989) in Muhlenhcrgia basal inserticm. Additional study needed is to eval- (Chl(»ridoi<leae, S])(»robolInae). Thus, study for this uate whether shape the of the basal can be cell papillae arc taken to b<^ apically rounded, undif- hypothesized as a separate character. ferentiated processes from and that arise are local- Among GET, seven eleusinioid geru'ra had pan- ized on the outer cell wall, and which can b<u'ome & icoid microliairs {Ectrosia Habrochloa, & gullivieri, silicified (Clark Gould, 1975; Temdl Wergin, & Heterachne, Neyraudia, Steirachne, Triraphis, Vi- 1981; Consaul Aiken, 1993). (Ellis (1979) cited GET ^mere//a), whereas the remaining had Mdcalfe chlori- become (1960) that papillae can cutin- doid microhairs. Microhairs were observed not However, in ized. Metcalfe (1960: 668) indicated only Cladoraphis spinosa or Viguierella, the one latter of spcci<'s {Trikeraia hookeri (tribe Stipeac)) as which needs further study, since only one specimen having cutinized papilla<\ Furthermore, Mclcallc was available for study (Table 1). As with species (1960) did not indicate how he di-lermined the of Leptochloa, microhairs abundance varied in and presence of cutin.) It is imj)ortan1 to stress that oth- were generally scattered; only near the apex in Apo- er structures on grasses called papillae may not be Snow 511 Number Volume 4 83, Lemmatal Micromorphology 1996 L were Diplachne cuspidata, dig- homologous the papillae described here (Clark prickles rare in to L L L & occa- Gould, 1975; Zuloaga, 1987: 26.2c; Ddvila itata, mucronata, rupestrh, uniflora, fig. L common, cau- 1990 sional in most species, or as in I data. 1994). I A OET. were observed single on Papillae on short cells Prickles for all Papillae short cells. curred in several species of Leptochloa (L fas- prickle was ob serve d on Psilolemma\ however. one was observed on three specimens, since only L L and L most and because apex was directed toward the lem- panicoides, uninervia, viscida). In its toward cases they had a collapsed appearance, which may matal base (almost universally they point have been an artifact of preparation, or, since the the apex in Eleusininae), it was considered an ab- norm been documented has of papillae silicification & & fc- (Clark Gould, 1975; Terrell Wergin, 1981; most abundant in CWora;>/it5 spp. (Figs. 6, 7), & Gouinia perhaps was an indication Consaul Aiken, 1993), pyrum and Oropetium Unlike deposition had not occurred. Collapsed (Fig. 23), (Fig. 62). that silica Valdes-R^yna and Hatch (1991) papillae on short cells are frequent in the literature 0 microscope) on herbar- observed (with a dissecting OET had on ium specimens of Tridens spp., Vaseyochloa multi- Relatively few papillae short cells and purpurea, although ihey v^ere Drake-Brockmania nervosa, Triplasis {Dinebra spp.; spp., Fig. 16; present the apex. Only Hal- occasionally only infrequently at Halopyrum, Fig. 23; Orinus, Fig. 61). in OERG. opyrum were the papillae large. Unlike Prickles occurred on all Spartina (not (Fig. 23) Valdes-Reyna and Hatch (1991) did not detect shown) had prickles of widely different sizes. Prick- I awns were generally most abundarit on the of papillae on short cells of Eleusine (not shown). les Among OERG, only Tragus (not shown) had pa- species bearing these structures (e.g., Lintonia, Fig, Pogoneura, prickles restricted to the pillae on the short cells. 54); in Papillae on long In the 13 species oiLep- awn. cells. had Leptochloa tochloa having a single papilla on the long cells Macrohairs, All species of at were more frequently present least some macrohairs on the lemmas. They were (Table the papillae 1) SEM on Diplachne gigantea, not observed using /' 1551; but analysis of an isotype {Vesey-Fitzgerald ifl BM) microscope revealed the even over localized por- with a dissecting consistent in this feature ^ ^ M A « I tions of the lemma; some long cells had papill presence of short macrohairs along the edges of the _-r some lemmas. some species they are had only midrib of In others did not. Leptochloa viscida (Fig. 53) a few papillae near the lateral veins. The size of rare, such as the sparse basal occ on Lep- Most macrohairs ob- the papillae arising on long cells was more or less tochloa digitata (Fig. 32). were having smooth constant within Leptochloa, with the exception of L. served in Leptochloa typical in which the papillae were edges and a rounded or acute tip. However, a "clav- coerulescens (Fig. 31), in was found D. Tn some cases the papillae were icomiculate" type (see Discussion) in noticeably longer. was which the subapical portion no- weakly developed, as in Diplachne cuspidata (not eleusine, in and above which occurred a cor- shown) and Diplachne gigantea ticeably clavate, (Fig. 38). Although many OET had long cells with swollen niculate tip. This feature was noted eariier by Phil- Coelachyrum jemmicum, ends Coe/ac/zjram lips (1974, 1982) for D, eleusine, (e.g., BeM;5ia, Fig. 3; The compare with yemenicum, Lintonia, and Trichoneura. hairs Fig. 12; Trichoneura, Fig. 70, in part; and Coelachyrum whic^h the long shown for Tribolium obliterum by Visser Spies poiflorum. Fig. 10, in may have no swellings), relatively few had a single (1994: fig. 3b) be clavicomiculate, but this is cells shown. Drake- uncertain since only their apices are papilla on the long cells {Dinebra, Fig. 15; OET A lemma; had macrohairs on the Brockmania, Fig. 16; Kengia, Fig. 27; Odyssea, Fig. majority of Acrachne Cladoraphis and Oropetium, they were absent in (Fig. 60, not visible; Orinus, Fig. 61; Fig. 1), Desmostachya Ectrosia The greater basal diameter of pa- splrwsa (Fig. (Fig. 14), significantly 7), 62). Eleusine //a/opjrum from other taxa in Eleu- leporina (Fig. 18), Ectrosiopsis (Fig. 19), pillae in differs Heterachne smmae. (not shown), Eragrostiella (Fig. 20), Psilolemma Myriostachya (Fig. Long with papillae were absent in the (Fig. 25), (Fig. 56), cells OERG Steirachne Sclerodactylon (Fig. 68), (Fig. 69), 66), and Only two macrohairs were Prickles were observed in every spe- Viguierella (Fig. 72). Prickles. specimen (Faden base one observed the very of Leptochloa at quency toward the apex. Among Leptochloa species et ai 74/613, Appendix 1) oi Harpachne schimperi 512 Annals of the Garden Missouri Botanical MO The study of several herbarium sheets at re- phylogenetieally informative the generic at level, vealed that a few macrohairs occasionally occur on Cork were recorded when cells as present only the the edges near the hase of some lemmas. However, was short cell substantially darkened viewed by (as they are at most infrequent and are always rela- SEM), whereas Palmer and associates generally re- tively short. Moreover, given their length, one could corded cork cells as present when any undifferen- just as easily designate them as (relatively long) tialed short cell occurred adjacent to a silica cell prickles. Examinati(m of specimens of//, bogdanii Pahner 1985: and (e.g., et al., pi. le, 2e, 3e, 4c, (MO: Heady Bogdan 1466, many 4524), the other species others therein). agree with Columbus I T. J. in the genus, also revealed a few short hairs near (pers. comm., January 1995) that the extent to the base of some lemmas. Given their sporadic oc- which suberin deposition and actually occurs, the currence and their questionable status as hairs (vs. extent to which can be observed, needs further it relatively long prickles), have reeorded them I as investigation. If suberin deposition in short cells is present or absent (Table The abundant macro- eventually demonstrated be SEM, 1). to invisible to or on ApochiUm made hairs (Fig. 2) observation of oth- it is shown that "cork cells'' on lemmas are lacking The er features (Hfficult. clavicornieulate liairs of in suberin, then their use as phylogenetic markers Coehichyrum yemcnicum appear identical to those will need reccmsideration. These considerations A Diplachne of & eleiisine. crispate macrohair, iden- have aside, followed others Vald^s-Reyna I (e.g., by tifiable the irregular ("crisped") surface and ap- Hatch, 1991) by recording eork cells as either pres- parently reported here for the first time in grasses ent or absent. (Uphof, 1962; Metcalfe, 1960; Ellis 1976, 1979), Silica Cells, Lemmatal silica cells were mostly was observed Coelachyrum for hrevifolium (Fig. 8), absent from Leptochloa, and thus have minimal in- C poiflorum (Figs. and stohmiferum 9, 10), C, (Fig. frageneric phylogenetic value. Since only one spec- The L 11). crisping is expressed most thoroughly to- imcn of fascicularis (Fig. 35) had a few silica ward apex, and somewhat Some tlie less so basally. cells, their occurrence on that specimen probably is macrohairs of Trichoneura gmndiglumis (Fig. 70) similar to lemmatal stomata, which occasionally were swollen at the base. reappear as atavisms from the transformationally OERG Among the surveyed, only Spartina antecedent However, leaves. in Leptochloa monti' lacked macrohairs. Some specimens Cynodon of cola (Fig. 41) silica cells are common, which re- nlemfuensLs had macrohairs (Fig. 13) with distinct:llyy quires a reassessment of earlier research. Using ''apiculate'' (sensu Peterson, 1989) tips. several lines of evidence. Vails (1978) suggested L monticola was that generically misplaced. Based on leaf anatomy and and Discussion citing Clifford \\atson (1977), Vails (1978) suggested a possible alliance This study originated from morphologically with Chionochloa Zotov (Arundinoideae, Arundi- based cladistic studies of Leptochloa using various neae). To assess a possible alliance with Arundi- combinations of likely sister genera, as hypothc- neae, sampled lemmas from Chionochloa conspl- I sized by Clayton and Renvoize The use cua (1986). sole (Forst. Zotov subsp. cunninghamii (Hook, f.) f.) of gross morphology gave pooriy resolved consensus Z<»tov, C. flavesceru Zotov, Danthonia dominguemis & and trees elades with low support values (Snow, Hack. and Rytidosperma Pilg., pilosum (R. Br.) & unpublished). Moreover, the use of a single data Connor Edgar (Appendix set data not shown). Un- 1; may L have resulted in artificial groupings, a poten- like monticola, which had chloridoid microhairs, tial problem Hilu and Wright (1982) alluded to in the microhairs of C. conspicua van cunninghamii their phenetic study of chloridoid grasses. These and pilosum were R. panlcoid; fuilhermore, micro- suggested factors that additional characters miglit hairs were not seen domin- for C. flavescens or D. & more lead to "accurate" (sensu Hillis Bull, 1993) guensis. Of relevant note, Watson and Dallwitz estimations of the phylogenetic relationships. Given (1992) reported panicoid microhairs (for the abaxial that micromorphological characters have known leaf surface) for Danthonia, which does accord not systematic value in grasses (references in Introduc- with the chloridoid microhairs of Leptochloa mon- study was undertaken tion), this to enlarge extant ticola. In addition, the species of Chionochloa data sets for purposes of phylogenetic inference. lacked cork which were L mon- cells, present in Cork Given Cells. their universality in Lepto- ticola. Morever, the dumbbell-shaped bodies silica L chloa, cork cells are of no phylogeiK'tic value in in monticola differed from the saddle-shaped sil- the genus. ica bodies of C. flavescens and D. domingueiisis However, the variable presence of cork Whereas each cells in of these taxa has one at least dis- Eleusininae and L Eragrostideae suggests they are crepancy when compared monticola, lemmatal to Snow 513 Number Volume 4 83, Lemmatal Micromorphology 1996 Second, with few exceptions Eragroslis, micromorphology alone does not support an obvious taxa. (e.g., & L Watson microhairs can be readi- Arundineae. Dallwitz, 1992), monticola relationship of to & one morphological types: assigned of three Saddle-shaped Watson, 1977; Vails, ly to (Clifford enneapogonoid, based on dumbbell-shaped 1978), and cross- chloridoid, panicoid, or 1978), (Vails, and bodies have been morphological differences of the basal distal shaped (Metcalfe, 1960) silica blades Leptochloa. Vails (1978: cells and their length ratios (Tateoka et al., 1959; reported for leaf In 1988 82) has shown that both dumbell-shaped and sad- & below son Dallwitz, 1992; but see also). Lept Ovate and saddle-shaped silica bodies were ob- The elongated aspect and thin wall of the distal OET Habroch- seven Odyssea mucronata. The presence of two cells in the microhairs of (e.g., served for L them panicoid micro- shapes of silica bodies (m duhia and 0, mucron- loa. Fig. 22) characterize as which cases accords with the ata cautions against their uncritical use as diag- hairs (Table 1), in all Watson and Dallwitz (1992) for phylogenetic markers. Recognition of observations of nostic or shape differences of silica bodies as distinct char- 1 eaves. on seems premature The discovery of enneapogonoid microhairs acters (or character states) thus and Psam- lemmas Cladoraphis cyperpides can be demonstrated that infraspecific van- the of until it wiseana represents the third report of en- body shape minimal. This caveat magrostis ation in silica is & Pappo- Kim, neapogonoid microhairs outside the tribe Oryza supported by studies in (\^1iang is & and second phoreae (Chloridoideae) the in and Zizania WVrgin, 1981) (both 1994) (Terrell Amphi- which documented exten- Eleusininae. They were first reported for genera Oryzeae), in tribe Amarasinghe and pogon (Arundinoideae) by shape sive infraspecific variation in the of silica strictus (1988) bodies. moment body shape, and 232) were unable to find microhairs on leaf blades Ignoring for the silica cypewides from photographic material provid- focusing on mere presence or absence, silica cells of C. Of microhairs appear be important phylogenetic markers, since ed by R, P. Ellis. the bicellular il- to they were not observed in over half of the OET and lustrated by Tateoka et al. (1959), the microhair of OERG lemma seems most resemble those of S;>o- examined. Their presence on the C. cyperoides to is (Tateoka al, 1959: b,64) probably symplesiomorphic, given the near univer- robolus vaginiflorus et fig. and and Pappophorum elegaris (Tateoka et al, 1959: fig. sal occurrence of silica bodies in grass leaves, lemma homologous b,155). also somewhat resembles a microhair re- that the is transformationally to It however, ported by Peterson (1989: 2,d) for two species the Testing for parallel loss or gain, fig. leaf. approach (Snow, of Muhlenbergia, Tlie swelling distal to the base of require a cladistic in prep.). will was ab- Pscimmagroslis The lack of microhairs the microhairs in (Fig. 65) Bicellular microhairs. elsewhere Eleusininae. sent in Lept decipiens) not unequivocal evidence (or their ab- Although not identified as such, an enneapogo- is sence, since a single microhair was often all diat noid microhair was recently shown for the arundi- Pentameris distichophylla (Barker, was three or more specimens of a given noid species visible for 1993: The recognition of three mor- taxon. Microhairs were not restricted to certain por- fig. l,c). microhairs admittedly rep- when were usu- phological categories of lemma, infrequent, tions of the but, which can misguide found near the apex. In some taxa, such as resents typological thinking, ally (Mayr, they extend onto the awn. Given a sim{)le the understanding of biological reality Lintonia, would Thus, when making preliminary hypotheses presence or absence, microhairs probably ])e 1982). of phylogenetic value In Eleusininae. How- of homology (de Pinna, 1991), one might consider little whether "enneapogonoid" microhairs of Cla- two features of microhairs suggest their utility the ever, and Psammagrostis wiseana are Amarasinghe and doraphis cyperoides as systematic markers. First, occurring Watson demonstrated that ten chloridoid (1988) genera, including Leptochloa, have partitioning whether the relative elongation of the basal cell is been having More- a "character" the elongation membranes in the basal cell of the microhair. itself, enneapogonoid membranes were limited the subfamily achieved independently from the over, the to & morphologies of micro- Chloridoideae (Amarasinghe Watson, 1988: 307), type. Overall, the distinct Of the genera they surveyed in Eleusininae, Era- hairs in Eleusininae and their tendency to be re- and lacked the stricted generically together suggest their potential Pogonarthria, Triraphis grostis, membranes. The use of partitioning membranes as value as indicators of phylogenetic relationships. The and characterization interpreta- marker subfamily Papillae. a potential phylogenetic in the were confounded by several mcrits further investigation over a wider range of tion of papillae initially