Poa L. (Gramineae) in Malesia

49 Pages·1994·3.1 MB·

by  Veldkamp, J F

BLUMEA 38 (1994) 409-457
Poa L. (Gramineae) inMalesia
J.F.Veldkamp
Rijksherbarium/HortusBotanicus,Leiden,TheNetherlands
Summary
Arevision ofPoa L.(Gramineae) inMalesia ispresented.There are41 species and three varieties.
Ofthese 38 are native,oneis naturalized,two are casuals,tenaredescribed asnew, and fournew
combinations areproposed.
Introduction
Thelast general survey of theMalesianrepresentatives ofPoaL. (Gramineae) was
madeby Jansen (1953) who recognized 20 indigenous taxaand threeintroduced
ones,but only after this date didseriousbotanizing startinthemountains ofPapua
New Guinea, wheremost species occur. Imadeapreliminary survey (Veldkamp,
1979, mostspecies depicted in detail!)forNew Guineaandrecorded22taxa. Now,
severalothers havebeen addedagain. Atpresentthereare 38 indigenous species in
Malesiaandthreeintroducedones: one, P. annuaL., is nowwidely naturalized,one,
P. pratensis L., may possibly stillbepresent in thearea (Luzon, MountainProv.,
last collectedin 1953, and perhaps on otherislands), and one, P. trivialisL., has
beenfoundin Java (saddle of theGedeh) butwas last collected in 1932.The number
oftaxamakesthisessentially temperategenusthelargest oftheherbaceousgrasses
intropical Malesia!Ofthewhole family only thewoody bambooSchizostachyum
Neeswithmore than45species is larger.
Thisrevisionis againpreliminary butmay serve toidentify themore commonand
wide-spread taxa. Forabetterunderstanding ofthephylogeny oftheMalesianspecies
itwouldhavebeenvery useful ifthe Australianand NewZealandonescould have
beenstudied ascarefully. Theirnumber,however,andtheabsenceofmanyinL has
madethis impractical andwouldhaveenlarged this study farbeyond thescope with
which itwas setup.Time permitting aphylogenetic study willbeundertakeninthe
futureinwhichsomeselectednon-Malesiantaxa willhavetobeincluded.
As faras thecharactersusedare concernedIhadto limitmyself tothosethatap-
peared usefulinthedelimitationofthetaxa.
As thesubalpine area (above c. 3000m) is most extensive inIrianJaya, more
species maybeexpected fromthere.This area is still virtually unexplored. From the
Vogelkop nocollectionshavebeen madeonthemountainsprobably high enough to
supporta subalpine flora,e.g. Peg. Arfak (2940 m) and G.Lina(2870 m). For the
vastCentral Range (Peg. Maoke) theonly subalpine areas wheresignificant collec-
tionshavebeenmadeare theCarstensz [Kloss (1913), Wissel (1936),Hope (1971—
410 BLUMEA Vol. 38, No. 2, 1994
1972),Raynal (1973), and somecasualvisitors], LakeHabbema[Brass and Meijer
Drees (1938)], and theWilhelmina[Brass & MeijerDrees (1938), Mangen (1982,
1983, 1984)],andthePapua NewGuineapartoftheStarMts [Leiden-Lae Expedi-
tion(1975)]. Fromthe Doormanonly P. lamiiJansenis known,from theHubrecht
P. jansenii Veldk., certainly othersoccur thereas well,e.g. thewide-spread P. cras-
sicaulis Pilg., P. papuanaStapf, etc. For asurvey ofexploration see Van Royen
(1980: e.g. 307-315).
Having taken part inthe StarMts Expedition Iwas struckby theoccurrence of
manyspecies therepreviously only known fromIrianJaya,andIam convincedthat
thereis abiogeographical boundary betweenthemountainsofWestandEast atthe
TelefominGap, although aninitialanalysis couldnotprove this(H. Turner, L, un-
published).
As therearequite a fewtaxa knownfrom onelocality only, sometimesas asingle
collection, otherinstancesare tobeexpected. A curiousexample is P.borneensisJan-
senfromMt Kinabalu, whichhasbeenfoundsix timesby Ms.Clemensandonce by
Ms. MolesworthAllen,butnot by anyoftheothernumerousvisitorstothatmountain.
Evenwell-explored areasmay harbourspecies as yetundescribed:MtWilhelmin
Papua NewGuineaturnedup two:P. jeremiadisVeldk., only known fromthetype,
andP. muricataVeldk.insix collections.
Strikingisthe occurrenceofP.papuanaoverthewholerangeofthegenusinMale-
sia, moreremarkableevenis thatofP. epileuca (Stapf) Stapfwhichoccurs in Sabah,
Celebes, and thentakes afarjump to Centraland EastPapua NewGuinea. In the
field thisspecies is notlikely tohave beenoverlookedbecause ofthepeculiar grey-
ish-glaucous hueoftheleavesfrom whichitderivesitsname.
Seramappears tobea stepping stone betweenNewGuineaandCelebes forsub-
alpine genera.Compare in thegrassesthe distributionsofAnthoxanthumhorsfieldii
(Kunth ex Bennett) Mez(Y. Schouten& Veldkamp, 1985),Brachypodium sylvati-
cum(Huds.) Beauv.(Veldkamp & Van Scheindelen, 1989),and Bromus L.(Veld-
kamp etal., 1991).As yetthereare norecords forPoa. Iexpect severalof thewide-
spread species to be present, atleast P. epileuca and P. papuana.Examples from
otherfamiliesare presentas well, too numerous to cite. Othercurious disjunctions
are shownby P. lamiiJansen,P. multinodisChase,andP. pilata Chase(q.v.).
Noattempthasbeenmadeheretorecognize sections. Previousclassifications such
as thoseof Hermann(1939), Tzvelev(1976), Herndndez Cardona(1978), and Ed-
mondson(1978, 1980)apply totheEuropean (and FarEast) taxa, andnotto Malesia
(nor to Australia). Hitchcock(1935, 1951) recognized 7 sectionsforNorthAmerica,
briefly reviewedby Kellogg (1985). Nicora(1978) distinguished 3 subgenera for
Patagonia, whichinlinewiththe currentuseof infra-generic taxa inPoamightbetter
beregarded as sections,also. Ingeneral thecircumscriptions givenare polythetic as
was alsopointed out by Bor(1952a: 788):"combinationsofmoreorlessvariable(!)
charactersmustbeused." Soreng (1990) commentedthat"character statesused to
unitegroups ... mostly are continuousratherthan discrete, and oftenare inconsis-
tent."
Keng (1959) distinguished 7 sections with 10seriesfortheChinese species about
whichI can give no opinion, as thetextis in Chinese(and the newtaxa andcombi-
nationsproposed therehence areallinvalid).
J.F.Veldkamp: Poa in Malesia 411
Although Ihave,checkedrepeatedly Ihavenot foundanyAustralianor NewZealand
species thatseemedparticularly closeto aMalesianone.Vickery (1970) madeno at-
temptto distinguish groupsfortheAustralianspecies atall,probably finding differ-
ences between thespecies strenuous enough. Edgar (1986) mentionedsome groups
and subgroups fortheNew Zealandtaxa, withoutformally naming them.From the
latter's diagnoses itagain isclearhowpolythetic most groupsare. Italso becomes
apparent how much more diverse theAustralasian species are than suggested by
Soreng (1990), who based hisopinion ontheonly two taxa availabletohim, which
were placed in the same subgroup by Edgar. T.B. Hair (1968) distinguished only
threegroupsofNewZealandPoa basedon caryotypes. Cytotypes ofthe largest and
mostwidespread groupwere mostlyindistinguishable fromoneanother.
Becauseofitssize nooverallcladisticanalysis hasbeenattempted forPoa.Soreng
(1990) madeoneforchloroplast DNArestriction sites of46taxa representing 19sec-
tionsor groups.Five groupscouldbeidentifiedandbasedonthishemadeanassess-
mentofbiogeographic events. The DNA tree suggested that Poaoriginated in Eur-
asiafrom whereNorthAmerica was independently colonizedsix times,two groups
invadedSouth America,andfrom thereone,(provisionally named ‘Australopoa’),
closely relatedamongitself(butsee preceding paragraph!) andtotheAndinaeNicora
(as asubgenus), migrated toNewZealandandAustralia.Unfortunately, noMalesian
species couldbetakenintoaccount.
TheAndinaeare circumscribed as being gynodioecious, the femaleflowerswith
large staminodia.The ‘Punapoa’ groupofSoreng is gynomonoecious. Poalabillar-
dieriSteud. andP. sieberiana Spreng. andmostotherAustralasianand Malesianspe-
cies are hermaphrodite (cf. Edgar, 1986:427; Connor& Edgar, 1987). There are
only 3 gynomonoecious taxaand3 dioeciousones out of35 nativesinNew Zealand
andnoneinAustraliaandMalesia.
Ifitturns out thatSoreng was correct inassuming thatatleastthemajorityofthe
Australasianspecies are as closely related as is suggested by theDNAoftheonly
two (related) species studied, and by Hair's (1968) andConnor& Edgar's (1987)
cytological analyses oftheNewZealandspecies, then itseems tomethatonly after
migration gynomonoecy,gynodioecy, andfinally complete dioecy musthave devel-
opedindependently, as itsurely didinthegynomonoecious P. annua(and therelated
P. infirma H.B.K., P. maroccana Nannf., andP. supina Schrad.) ofsection Ochlo-
poa Aschers. & Gr. Thisseems aneasier explanation than accepting a reversal to
bisexuality. Morelikely is thatinvasioninAustralasiahastakenplace atleasttwiceto
accountforthemanyhermaphroditespecies presentnow.
In viewof thedistributionofthe genusin Malesia, this migratory route agrees
withmyintuitivefeeling thatthe species havean Australianorigin: 25 species in
Papua NewGuinea(excl. theStarMtsarea),20inIrianJaya(incl. theStarMts area;
butmore to be expected), three inCelebes, and threein Sabah. Noneare known
fromSumatra, Malaysia, Java,the LesserSundaIslands, orthePhilippines (P. luzon-
iensisMerr.is P.pratensis, andseemsintroduced). Soreng's remarkthatthe"species
(from SoutheastAsia) canbe accommodatedwithinsections foundin the USSR,"
referredonly to thecontinentalones,andnot theMalesian(Soreng, inlit.).
Somedirectionsintheevolution ofmorphological charactersintheMalesianspe-
cies seemapparentandwill bementionedinthe nextchapter.
412 BLUMEA Vol. 38,No. 2, 1994
MORPHOLOGY
Whenchecking charactersitisadvisable to check afew specimens toknow theap-
proximate rangeofvariation.
In Poa shoots may be intra- or extra-vaginal. Inlra-vaginal is theusual way of
branching both at thebaseand at thehigher nodes oftheculm: the axillary shoot
elongates andpushes the sheathaway. Inthis case the prophylls are usually elon-
gated. In somespecies, however,someofthebasalshoots (at substrate levelorjust
below) pierce thesheath,andtheprophylls are less than6mmlong.Thefirstleaves
ofsuch an 'extra-vaginal' brancharereducedtocataphylls, modifiedsheathslacking
ligules andblades,moreinduratedthannormalsheaths, withwhichthey mightother-
wisebe confused whenthe bladeshavewitheredaway. Theshootprotrudes atfirst
horizontally, latergeniculately so. Itmaybenotedthatextra-vaginal shootsmayeas-
ily belost whenthe tussock is too roughly torn fromthe substrate, as is the usual
methodofcollecting grasses. InP.pratensis and P. trivialisstolonsmaybe formed
from extra-vaginal shoots, butIhave notobservedthese in herbariumspecimens,
possibly duetothiskind ofcollecting.
Extra-vaginal branching hasbeenusedas a characteratthesectionallevelinEur-
asia, butinMalesiaitseemsoflittleuse.
Cushion-forming plants are not very common in New Guinea. In grasses this
growth-form ispresentin atleast
Danthoniacraigii Veldk. PoahentyiVeldk. (sometimes)
D.montis-wilhelmiVeldk. &Fortuin P. inconspicua Veldk.
D.oreoboloides(F. Muell.) Stapf P. lunataChase(perhaps sometimes)
PoacallosaStapf P.papuanaStapf(sometimes)
P. crassicaulisPilg. P.pilataChase
P. dozyi Veldk. P.pulviniformis (Veldk.) Veldk.
Theligules are scarious, theuppercaulineusually thelongest, except inP. borneen-
sis wherethey are theshortest. Theshape rangesfrom rim-likeand truncateto tri-
angular and acuminate, butit seems not possible to use shape andlength to distin-
guishgroups by, as is donein Eurasia, e.g.sect. OreinosAschers. & Gr. ex Nannf.
withlongcaulineligules, andthesectionsFlaccidulaeHernandez(Leptopoa F. Her-
mann, nom. inval.)and Macropoa F. Hermannex Tzvel. withcollar-shaped ones.
Thebladesareusually involuteprobably as a defenseagainsttheharsh climate at
highaltitudes: UV-light andcold.
Mostspecies havesmoothblades,butsomeare moreor less strongly scaberulous:
P.erectifolia Hitchc., P. muricata, andP.papuana.
Two species have deciduousleaf blades, P. hentyi and P. hypsinephes Veldk.,
wheretheblade disarticulates withthesheathabovetheinsertionofthe ligule. The
baseoftheplant thenhasajagged outlinebecauseofthe densely imbricate, distich-
ous sheaths.Intheotherspecies thebladesjustrotaway leaving thetougher vascular
bundles as an inconspicuous network. Species with similar disarticulating blades
areknown fromNewZealand(Edgar, 1986): P. acicularifolia J.Buchan.,P. astonii
Petrie, P. colensoi Hook,f.,P. hesperia Edgar, andP. maniototoPetrie, butother-
wisetheseseemtohavelittleto dowiththeNewGuineaones.
J.F.Veldkamp: Poa in Malesia 413
Theshape oftheinflorescencehastobetaken atorafteranthesis, as immatureones
are contractedtoexpand laterornot,asthecase may be.
LikeBor(1952a: 794), andunlikeVickery (1970), Ihave observed thatthenum-
berof lowermostbranches ofthe inflorescence isof diagnostic value,but I would
hesitatetoregard itas a characteratthesectionallevel, however, as was doneto dis-
tinguish sect.Coenopoa Hyl. fromsect.Flaccidulae.
InPoa, as inmany othergenera,thelongest lowermostinflorescencebranchoffers
additionalcharacters, usually neglected or imprecisely defined: length, numberof
spikelets, and 'nakedness',i.e. absenceof spikelets inthelowerpart. Thelatterhas
roughly been measured fromthe attachment to the mainaxis to the glume ofthe
lowermostspikelets.
Thescabridity ofthepanicle branches hasbeenused to distinguish sections, e.g.
insect. Oreinos they are smooth. In Malesiathe conditionis variously expressed,
severalspecies havingboth smoothandscaberulousbranches.
NineMalesianrepresentatives areexceptional forhaving strictly 1-floweredspike-
lets,andanotherseven 1- or2-floweredones, whereby thefirst cannot beidentified
to genus with mostkeys and the second only with difficulty. Hencesome have
atonetimebeen attributedtoAulacolepis Hack, (nonEttingsh.), a genusmorerelat-
ed toAgrostis L., asA. epileuca (Stapf) Hitchc. The most extreme oneis P. incon-
spicua, where theinflorescence isreduced to a single 1-flowered spikelet hidden
among thefoliage. Poapilata sometimeshas 1-spikeled, 2- or 3-floweredinflores-
cences, too.
There is awhole range ofspecies with 3-5-floweredspikelets through some
with either 1 or2 floretsto strictly uniflorousones. Itis tempting to see thisreduc-
tionas anautapomorphy forthelast groupofspecies.
One-to four-flowered spikelets arealso present in P. takasagomontana Ohwi
fromTaiwan.Inotherspecies 1-floweredspikelets exceptionally occur in depauper-
ateindividuals.
Cleistogamy wasconcludedto bepresentwhenwell-developed fruits werefound
withthethreestamens stillentangled amongthestyle branches. Itoccurs inseveral
species, apparently exclusively soinP. callosaandP. crassicaulis. Itwas observed
intheonly specimen ofP. celebicaOhwiex Veldk. Bothcleistogamy and chasmo-
gamyoccur inP. aurigae Veldk., P. keysseri Pilg. var.brassii (Hitchc.) Veldk. and
var. saruwagetica (Pilg.) Veldk., P. languidior, P. lunata, and P. nivicola Ridley.
Apomixis andparthenogenesis isknownfor manyspecies ofPoa, contributing to
thetaxonomicproblems in thegenus. In Malesia itmay beexpected tooccur inP.
pratensis.
Poaannua(and relatives) are gynomonoecious, thatis, withbisexual(lower) and
female(upper) floretsinthesamespikelet, arareconditionin grasses.
Hybridization, such a powerful means ofspeciation and very common inPoa,
hasnotbeenobservedin Malesia.Thatis,no clearcases ofintermediatecollections
havebeenobserved.Aboutkaryology nothing canbesaid, asthechromosomenum-
berofonly 4nativetaxa isknown: 2n=28 forP. crassicaulis, P.papuana, and2n=
56forP.keysseri var.keysseri and var. saruwagetica (Borgmann, 1964).
Thelength ofthefirstglume relativeto the firstlemmais measuredin entire
spikelets.
414 BLUMEA Vol. 38, No. 2, 1994
Measurementsofthelemmahave always beentaken fromthefirst, or lowermost
one.Thisoneis thelargest, best-developed, andlongestretained atmaturity. Higher
onesare smaller, maybeless scaberulous andhaveless nerves. Theirantherstendto
beshorter, too.Severaltypes ofhairscanberecognized on thelemma,which have
beenusedinsectionaldivision.
1.Crinkly hairsattheabaxialsideofthecallusofespecially thefirstlemmaand
sometimesalso along thelowerpartof themidrib, theso-calledweb. Whereitoc-
curs itis usually easyto see,butoccasionally theweb is reducedto asingle hair, or
evenabsentinthesameinflorescence, e.g.inP.keysseri Pilg. var. saruwagetica, the
formdescribedasP. archboldiiHitchc., P. languidior Hitchc., P. longirameaHitchc.,
P. lunataChase,andP. nivicola(e.g. theformdescribed as P. egregiaChase).
Inthenative species thereis no othertrue indument,although thespicules present
especially onthenerves maybecomeelongated, 'ciliolate' (as in P. crassicaulis) or
'setosely scaberulous'(as inP. nivicola).
2.Crests of'true' hairs adherentatthetopbecoming especially prominent when
thelemmais moistened.Thisoccurs intheintroducedspecies only.
Likewisethekeels ofthepalea maybesmooth, scaberulous(atleastintheupper
half), ciliolate, or provided withcrestsof hairs, e.g. P. annua. The latterhasbeen
placed inthesection Ochlopoa because of this,but in-andoutsideMalesiamaterial
withcompletely glabrouskeels wereregularly observed, whichthrowsdoubts onthe
valueofthisfeatureas asectionalcharacter.
REFERENCES
Bor,N.L. 1952a.Thegenus Poa in India,Part 1.J.BombayNat. Hist. Soc. 50:787-838.
Bor,N.L. 1952b. The genus Poa in India,Part2.J. BombayNat. Hist. Soc. 51: 61-103.
Borgmann,E. 1964.Anted derPolyploidenin derFloradesBismarcksgebirges vonOstneuguinea.
Zeitschr.Bot. 52: 118-172.
Connor,H.E.,&E.Edgar. 1987.Australasian alpinegrasses: diversification andspecialization. In:
B.A.Barlow,Flora and faunaofalpineAustralia: 413-434.CSIRO, Australia&Leiden.
Edgar,E. 1986.Poa L.in New Zealand. NewZeal. J.Bot. 24:425-503.
Edmondson,J.R. 1978.Infragenerictaxain Poa. Bot. J.Linn. Soc. 76: 329-334.
Edmondson,J.R. 1980.Poa.Floraeuropaea 5: 159-167.Cambridge,etc.
Hair,T.B. 1968.Contributions to achromosome atlasofthe New Zealand flora. 12. Poa(Grami-
neae). NewZeal.J. Bot.6: 267-276.
Hermann,F. 1939.Zur AbgrenzungderGattungPoa und zurGliederungihrereuropaischenArten.
Hercynia 1:451-461.
Herndndez Cardona, A.M. 1978. Estudiomonogrdficode los gdnerosPoa yBellardiochloa enla
PeninsulaIbdricaeislasBaleares. Diss.Bot. 46: 1-365.
Hitchcock, A.S. 1935.Poa. Manual ofthe grasses ofthe UnitedStates:100. NewYork.
Hitchcock, A.S. 1951.Poa. Manual ofthe grasses ofthe UnitedStates,ed. 2: 100.New York.
Jansen,P. 1953.NotesonMalesian grasses. I.Reinwardtia 2: 322-333.
Kellogg,E.A. 1985.Abiosystematic studyofthePoa secundacomplex. J.ArnoldArbor. 66: 201-
242.
Keng, Y.-L. 1959. Poa. Flora illustrada plantarumprimarum sinicarum. Gramineae: 134-224.
Peking.
Nicora,E.G. 1978.Poa. Fl.Patagonica3.Gramineae: 138-207. Buenos Aires.
J.F.Veldkamp: Poa in Malesia 415
Royen, P. van. 1980;ThealpinefloraofNew Guinea 1:1-317.Vaduz.
Schouten,Y., &J.F.Veldkamp. 1985.Arevision ofAnthoxanthumincludingHierochloe (Grami-
neae) inMalesia andThailand.Blumea 30: 319-351.
Sorcng, R.J. 1990.Chloroplast-DNAphylogeneticsandbiogeographyin areticulating group: study
in Poa(Poaceae). Am. J. Bot.77: 1383-1400.
Tzvelev, N.N. 1976.Grasses ofthe SovietUnion 2:649-722. Washington,New Delhi (Engl.ed.
ofZlaki SSSR 2:433-478.Leningrad).
Veldkamp, J.F. 1979. Poa. In: P. vanRoyen, The alpineflora ofNew Guinea 2: 1056-1111.
Vaduz.
Veldkamp, J.F.,M.Eriks& S.S.Smit. 1991.Bromus (Gramineae)in Malesia. Blumea 35: 483-
497.
Veldkamp,J.F.,& H.J. vanScheindelen. 1989.Australopyrum,Brachypodium, and Elymus(Gra-
mineae) in Malesia. Blumea 34:61-76.
Vickery, J.W. 1970. Ataxonomic study ofthe genus Poa L.in Australia. Contr. N.S.W. Nat. Hb.
4: 145-243.
POA
Poa L.,Sp. PI. 1 (1753) 67; Veldk. in P.Royen, Alp. Fl.New Guinea 2 (1980) 1056.—Lecto-
type:Poapratensis L.
Perennials(P. annuaannual). Ligule membranous.Paniclelaxtocontracted; branch-
es solitary tofascicled. Spikelets pedicelled, not in glomerulus, 1-several-flowered,
laterally compressed, elliptic to lanceolate. Glumesunequal, 1-3-nerved.Lower
glumepresent. Lemmasherbaceous, 3-7(-ll)-nerved, usually muticous, exception-
ally shortly aristate, dorsallykeeled.Rachillaglabrous to hairy-scaberulous. Lemma
more orless imbricatewith thenextone, base moreor lesscuneate, glabrousorad-
axially with1-numerouscrispy hairs('web'), margins not broadly membranous,at
mostscabridon thekeels. Ovary glabrous. Stamens3. Hilum punctiform,subbasal.
x =7,only 4counts inMalesia, elsewherewithmanyaberrationsfromthis.
Distribution About500species, 41 inMalesia,2formerly introduced, perhaps
-
still present, 1 naturalized.The Malesianspecies are clearly ofasouthern(Austral-
asian) origin, mostly occurringin NewGuinea,3 inCelebes,3in Sabah.
Note- 16 species have1-floweredspikelets, and ingeneral cannotbe identified
with most generickeys!
KEY TO THE TAXA
Note- When devising thiskey, itbecame necessary toomitexceptional cases,e.g.
where spikelets usually were2-3 mmlong theobservationofoneof3.25 mmwas
disregarded. Inthe descriptions such exceptions have beenplaced betweenbrackets.
Itisrecommendedthatseveral specimensor spikelets bechecked.
la. Callusoffirstlemmaadaxially withcrinkly hairs('web') 2
b. Callusoffirstlemmasnot webbed 13
2a. Panicleserect 3
b. Paniclessecundand nodding 8
416 BLUMEA Vol. 38,No. 2, 1994
3a. Caulineligules triangular, acute torounded 4
b. Allligulescollar-shaped, truncate. — Tuftedrhizomatous plants. Firstlemmas
midribandouternerves withacrestofhairs 31. Poa pratensis
4a. Basalligules triangular, acute 5
b. Basal ligulescollar-shaped, truncate. — Culmsloosely tuftedto mat-forming,
weak. Bladesflaccid, flatto involute.Spikelets 3.5-5.7 mmlong. First lem-
mas 2.9-3.5mm long.First palea halfaslong as to distinctlyshorterthanthe
lemma,keels scaberulous 20. Poa languidior
5a. Culms solitary or few together. Blades flaccid, flat. Spikelets 2.5-3.8 mm
long. Rachilla process 0.4-1 mm long.First lemmas2.1-3.2mm long. First
paleakeels smooth 6
b. Culmstufted. Blades stiff,involute.Spikelets 4.2-7.5mmlong. Rachillapro-
cess 1.25-3.25 mm long. First lemmas 3.7-5 mm long. First palea keels
scaberulous 7
6a. Culms weak, 13-25cm long. Blades 1-1.5 mm wide. Paniclescontracted,
3.5-4.5cm long, branches appressed. Spikelets 1-flowered.Lower glumes
2.5-2.8 mm long, c. 0.7timesas long as thelemma. Upper glumes 2.9-3.5
mmlong. First lemmasmidribwithoutacrestofhairs. First palea subequal to
thelemma.NewGuinea 21. Poa leptalea
b. Culms stiff, 35-60cm long. Blades 1.75-4.5 mmwide.Panicles lax, 9-22
cm long, branches erecto-patentto patent. Spikelets 2- or 3-flowered.Lower
glumes 1.35-1.8mmlong,0.5-0.67timesas long as thelemma, upperglumes
1.8-2.6 mm long. First lemmasmidribwith acrest ofhairs. First palea dis-
tinctly shorterthan thelemma.Java 39. Poa trivialis
7a. Culms 17-38cm long.Paniclescontracted,branches appressed. Glumesscaber-
ulous on the nerves. First lemmas midriband nerves smooth to scaberulous
23. Poa lunata
b. Culms 48-63cm long. Panicleslax, brancheserecto-patenttopatent.Glumes
and atleastthefirstlemmasscaberulousallover 33. Poa quadrata
8a. Caulineligules triangular (the basalmaybeaccolade-tocollar-shaped!). Blades
erect 9
b. Allligules collar-shaped. Bladeserecto-patenttopatent 11
9a. Basal ligules triangular, more than2.75 mm long, acute to acuminate. First
lemmascuneate toapiculate 10
b. Basalligulesaccolade-orcollar-shaped, upto 1.5mmlong, truncate,erase. First
lemmasgradually acute.—Culmsbranching intra-andextra-vaginally atbase.
Blades loosely folded.Anthers0.4-0.9 mm long.... 22. Poa longiramea
10a. Culms branching intra-andextra-vaginally atbase. Basalligules acute.Blades
loosely foldedor flat.Anthers 1-1.7mmlong 27. Poa nivicola
b. Culms branching intra-vaginally at base. Basal ligules acuminate. Blades
involute. Anthers 2.6-3mmlong 37. Poa telata
11a. First lemmas 2.25-4 mm long. — Lowermost, longest panicle branches
usually 22-tovery manyspikeled. Rachillanodesusually scaberulous ... 12
b. First lemmas 4-5.5 mm long. —Lowermost, longest panicle branches 10-
22-spikeled. Lower glumes 3-5.25 mm long, upper glumes 3.5-5.3 mm
long. Rachillanodesusually smooth.... 17b. Poa keysseri var. keysseri
J.F.Veldkamp: Poa in Malesia 417
12a.First lemmas2.25-3mmlong. —Lowermostpanicle branches 3-12or more
together. Lowerglumes 1.7-3mmlong, upperglumes 2-3mmlong
17a. Poa keysseri var. brassii
b. First lemmas 3-4 mm long. — Lowermost panicle branches 3-6 together.
Lower glumes 2.4-3.5 mm long, upperglumes 2.75-3.8mm long
17c. Poa keysseri var. saruwagetica
13a. Sheathsdisarticulating withthe blades.—Lowermostpanicle branches solitary
orpaired, nakedinthelower0.7—1th 14
b. Sheathsnot disarticulatingwiththeblades 15
14a. Plants greenish yellow. Lowermostlongest branches naked inthelower0.85-
0.88th, 1- or 2-spikeled. — Leafblades 0.5-0.85mm wide. Basal ligules
0.4-0.9mm long. Panicleswith3-13 spikelets 12. Poa hentyi
b. Plants moreor less glaucous. Lowermost longestbranches nakedinthelower
0.7-0.8th, 3-5-spikeled. — Leafblades 0.75-1.5mm wide. Basal ligules
0.8-2 mmlong. Panicles with 10-27spikelets . . 13. Poa hypsinephes
15a. Bladesrough, muricate(lens x25!) 16
b. Bladessmooth 18
16a. Basalligules triangular, acute.Spikelets usually 2-flowered.Glumesscaberu-
lous allover (lens x25!). First lemmas scaberulous atleaston themidriband
nerves. Firstpalea shorterto subequal tothelemma 17
b. Basal ligules collar-shaped, truncate,erose. Spikelets 1-flowered.Glumes and
lemmasscaberulous onthemidrib.Paleausually distinctly longer thanthelem-
ma 28. Poapapuana
17a. Blades filiform, 8-18 cm long. Panicles rather lax, 4.5-8.5 by 1.5-5 cm.
First lemmas 3-4.15 mm long, faintly5-nerved. Anthers 1.1-2.25mmlong
10. Poa erectifolia
b. Blades subulate, 2.7-8 cm long. Panicles compact, 2.5-4.5by 0.4-1.1 cm
long. First lemmas2.4-2.8 mm long, faintly 3-nerved. Anthers 0.6-1 mm
long 26. Poamuricata
18a. Caulineligules collar-shaped 19
b. Caulineligulestriangular 24
19a. Culmsstiff. Caulineligules 0.15-0.7mm long. Paniclebranches appressed to
erecto-patent 20
b. Culmsweak. Caulineligules 1-1.7mmlong. Paniclebranchespatenttoreflex-
ed.— Ligules subequal, truncate to acutish. Blades green
25. Poamultinodis
20a. Ligules truncate, erose, to acute,subequal. Bladesgreento glaucous. Panicles
contracted.— Paniclebranchesusually appressed 21
b. Ligules acuminate, unequal, the uppercaulinelongest. Blades greyish-glaucous
above. Panicleslax, brancheserecto-patent.—Bladesflatfishtoinvolute,0.5-
1.5mmwide.Lowerglumes faintly 3-nerved 9. Poa epileuca
21a. Blades involute. Longest lowermost panicle branches 1- or 7-10-spikeled.
Lowerglumes faintly to distinctly 3-5-nervcd 22
b. Bladesloosely foldedtoflat.Longest lowermostpanicle branches 2-6-spikel-
ed.Lowerglumes 1-nerved 23
418 BLUMEA Vol. 38, No. 2, 1994
22a. Plants densely cushion-forming, culms 1-6 cm long. Blades erecto-patentto
patent. Panicles 1.4-3.7cm by 2-3 mm, 1-6-spikeled, the longest lower-
mostbranches 1-spikeled. Spikelets 3-3.75mmlong 30. Poa pilata
b. Plantstufted,culms 17-40cm long. Bladeserect. Panicles5.5-8cm by 10-
15mm,many-spikeled, thelongestlowermostbranches 7-10-spikeled. Spike-
lets5-5.5 mmlong 35. Poa stellaris
23a. Plantscushion-forming. Blades apicallyrounded.First lemmasacute,5-nerved,
midribscaberulous, distallyciliolate 7. Poa crassicaulis
b. Culmssolitary orfewtogether, or tufted.Bladesacute. First lemmasacuminate
to mucronate, 3-nerved, midribsmooth to distally scaberulous
41. Poawisselii
24a. Bladeslooselyfoldedtoflat 25
b. Bladesinvolute 32
25a. Panicleserect.Upper glumes 1-3-nerved(sometimes 5-nervedatbase) .. 26
b. Paniclessecund. Upper glumes 5-nerved.— Plantsperennial. Spikelets 3.8-
6.75mmlong.First palea keelsscaberulous 22. Poa longiramea
26a. Plantsperennial. —Lemmaandpaleakeels smoothtoscaberulous, never with
crestsofhairs 27
b. Plants annual.— Paniclebrancheserecto-patenttoreflexedin fruit.Spikelets
3.75-6mm long, 3-5-flowered.Lower glumes 1-2.5mm long. Lemmaand
palea keelsglabrous or withcrests ofhairsonthe nerves .... 1. Poa annua
27a. Spikelets 3.1-7.6mmlong. Lowerglumes 1.5-5.25mm long, upper glumes
1.8-5.1mm long.First palea keels scaberulous 28
b. Spikelets 2-3mm long. Lowerglumes 1-1.5 mmlong, upperglumes 1.25-
1.75 mm long. First paleakeels smooth. — Paniclebranches erecto-patent.
Spikelets 1-flowered 2. Poaaurigae
28a. Paniclebranches appressed 29
b. Paniclebranchespatenttoreflexed 30
29a. Culms 2.5-8 cm long. Paniclescontracted, 2.5-3.5 cm by 3-6 mm, the
lowermostbranchessolitary. Spikelets usually 3-flowered.Lower glumes
4.15-5.25mm long and0.87-1 timesas long as thelemma,theupper glumes
4.5-5.1 mm long. First lemmas 4.6-5.35 mm long
18. Poa kuborensis
b. Culms 12-30cm long. Panicles lax, 5.5-8 cm by 37-40mm, lowermost
branches 2-4 together. Spikelets 1- or2-flowered.Lower glumes 2.25-2.5
mm long, c. 0.6 times as long as the lemma, upper glumes 2.65-2.9 mm
long. First lemmas3.5-4mmlong 40. Poa wilhelminae
30a. Glumes smooth. First lemmas 2.6-3.5 mm long, midrib smooth to scab-
erulous 31
b. Glumes scaberulousonthenerves. First lemmas 3.6-4mmlong, midriband
nerves scaberulous. —Culms stiff. Basal ligules collar-shaped, 0.5-1.5 mm
long, truncate. Blades stiff, flat toloosely folded. First lemmasnerves scab-
erulous 15. Poa jansenii
31a. Culms weak. Basal ligules collar-shaped, truncate. Blades flaccid. Lower
glumes 1.5-2.5mmlong. Paleahalfas long to distinctlyshorterthanthelem-
ma.— Caulineligules 1-4.5mmlong 20. Poa languidior