Skip to main content

Full text of "Annals of the Missouri Botanical Garden"

See other formats

of the 


Volume 90 
Number 1 

rs. Authors should write the Managing Editor for informa- 

ie back oft 

:rg/mbgpre £ 

the A.N 


Editorial Committee 

Victoria C. Holhmell 


Missouri Botanical Garden 

Amy Seheuler McPherson 

Managing Editor. 

Missouri Botanical Garden 

Gerrit Davidse 

Missouri Botanical Garden 

Roy E. Gereau 

Missouri Botanical Garden 

Peter Goldblatt 

Missouri Botanical Garden 

Gordon McPherson 
Missouri Botanical Garden 
P. Mick Richardson 
Missouri Botanical Garden 
Charlotte Taylor 
Missouri Botanical Garden 
Missouri Botanical Garden 

s the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 

Volume 90 
Number 1 

of the 







Nigel P. Barker, 2 H. Peter 

Cynthia M. Morton,' and ! 

Mark Lvl<" 

Missouri Botanical Garden 

The grass genus C.aitadfna (including the in- and nearhv islands, and lh«- anotiKi lon> 

lameutal "pampas grass."* ('.. sclloana (Schull.) dii is lr New ( 'uinea. 

\scli. «.*< Craeb) lias a < . Cortaderia is presently delimited I 

ng Ion nd in temperate anil \ in lean South \merica. nice (if a g\ iio.lioe.ious breeding sysle 

and Islands). New Zealand (including Chatham (sects. Hijida. Cortaderia. Mutira. am 

iml I'itl Mauds!, ami New Cuinea. lata) on t It. ■ basis of sexual dimornhisii 

the plaslid rpo(.2 gene iHarkei el al.. I 009) found nrhloa Znlo\. with the New Cluneal, Cortaderia 

tin support lor tli«- n-.-omiilinii oldie ('....lad- arehholdii uilennediale between die two genera 

erieae. as Cortaderia is included within a large (Clayton K benvoi/e. 1980). Hitchcock (1030) 

group .4 genera dial appro\imatel\ agreed with llie originalb described Cortaderia arehholdii as a spe- 

tril.e Danthonieae >ensu Watson and Dallwil/ cies of Danthonia. During the M(.Os the l.roa.lh 

(1002). These genera ha\e now l.een ele\ated to circumscribed genus Danthonia was broken up into 

subfamily status as I >anlhonioideae (( ,V\\ ( ,. 2001 ). numerous segregate genera. In M'riea. all species 

This siililaniiK is defined l>\ the presence of hails- formerly in Danthonia were placed in Merxmuellera 

lorial sMiergid cells (\erl.oom el al.. 1004: CI'W C. Conert. karrooi Idoa Conert i\ liirpe. Dre^eoi hloa 

2001). and all species ,4 Cortaderia examined to Conert. and Dseudojienti minis Conert (Conert. 

dale possess swiergi.l cells (Philips..,,. 1000. |070. 1 « »7 I : Conert ,\ hirpe. MOO). In New 

1077: Philipson & Connor. M8I). Zealand. Chionoehloa. \olodanthonia Zotov. Kr\- 
thranlhera /oto\. and Pxrrantltera Zoto\ were 

TaXONOMI erected (Zot..\. M<>5). In \ustralia blake (1072) 
resurrected Plinthanthesis Steiul. and Monaehather. 

Although no recent species-l.-\cl tax »m\ exists ;lll( | A'w/,/o.v/;e/7//</ Steud. was recogni/ed in South 

lor the entire genus, (.onnor and Kdgai ( IO, |) and \meriia (Nicoia. 1073). Rytidosperma and \oto- 
Connor ( 1 083a) ha\e re\iewed the nomenclature. danthonia ha\e subsequently l.een synonymi/ed 
The New Zealand species have l.een well studied (Connor K II MM iltl I tl has l.een 
laxonomicallv (Connor ^ K.lgar. M71. I O.w: hdgar some dispute al.out the most appropriate name 
vK (.onnor. 2(KM)). hut the South Vmerican species (Yeldkamp. M80: Jacobs. 1982). the \ali.litv of rec- 
have only l.een studied on a regional basis (e.g.. ogm/uig Ruulospenna as a genus distinct from 
Davidse X bold. lOOl; \st.-giano .-I al.. 100.",. Danthonia (Conert. 1075: Jacobs. 1082). ami the 
1000). Hie number of species in the genus is at circumscription ol Rrtidosperma (Cla\ton & Hen- 
present uncertain: Clayton and Renvoi/.- (M80) voi/e. M80). More recently, binder and Yerboom 
and Watson and Dallw it/ ( M)2) considered the ge- (|000) and binder (1007) have once again revised 
mis to comprise 21 species, but byle (1000) ,on- ||„. \ U slralasian taxa. Danthonia arehholdii has not 
sidere.l there to be onb 22 species, and Astegiano escaped the fragmentation of Danthonia as it is 
et al. (1995) recognized 25 species. However, at clearly not Danthonia seiisii Zol<>\. Connor and Kd- 
least four South Vmerican species may not be .lis- gar (M71) transferred the s|..-cies to Cortaderia. 
tinct. and one is considered to be of hybrid origin while virlualb smiultaneousb Conert (1075) trans- 
It.', seneantha X mtula. I ,\ le. unpublished). Most ferred it to Chionoehloa. The mallei .4 where this 



i,n t;„n 

nor& Edgar (1974 

Beuzenherg, per*, c 

■ited l»\ Connor & 

ioiuin.ini, (]'), 

Connor & Edgar (1 

!,mr (J9n 7) «•! 

t«-.i i 

,5 Conn 

or & Edgar (1974) 



n & 

Edgar ll<>7ll 

hi. E. C.iiitinr. 

(1 <>(,(,) 



!;;;;," I,;;,;'; 

nor ,\ 1 

'Agar (1974) 



: ;::;,m,<;;^ 


riled In Connor M 

_ Edgar 


1 explain the une 
ear (ITS) genes f 

mal groups of dantlu 
er el al. (2000). Re] 
boldii are also claril 

Tile Simlh 

American species was sister to hun- 

en informal lineages identified hv Barker el al. 

pmthysus a. 

i(j the New Zealand species was allied 

(2000). Table 2 details the lava sampled in this 

tolhe Auslr 

alian genera \olorhloe Domin and Plin- 


\ subsequent investigation using Inter- 

Leaf material was dried in silica gel (Chase & Hills. 

nal Transcri 

l.ed Spacer (ITS) sequence data from a 

1901). DlNA was extracted from these samples us- 

number of > 

-pecies of CorlaJrrin supported the re- 

ing the hot CTAB method of l)o\le an<l Doyle 

suit ohtaiue 

el from the rpoC'2 study (Barker et al.. 

1 l'»;;7i. [I,, .-Mine 1 IS region n;h amphlied l>\ tin 

2000). A 1 

ombined analysis was conducted hy 

Polymerase Chain Reaction (TCR) using the prim- 

Barker et a 

1. (2000). hut this analysis had iiumei- 

ers 17SK and 26SK published by Sun et al. (1994). 

i\\ llmsc nl ' Corttulcria. \ul<- dial la\a lot which lioth ITS anil ,/»>< !J dala haw ahvaiU lirrn | > 1 1 1 .li-hi-il aiv mil 
lliosr s[M-cics listed as "r\ Cull. Vu Zealand" were |.iu\ iilnl l>\ Mr. I«. Uu\|i.n from the Holanic Cardrn ; 
\ Zealand (C = Carden key: MWC = l)N A extract numher |.ruvided liy M. Chase (Jodrell L.lxir; 
v): NI'H = collected l.\ Y I'. IWker. HI'l. collected In II. I'. I.mder. CHI! Christchurch herharium. 

IaI,-X Canllo'L'lh \eue/„cla \R5.V. 


C :>(»!!!!: ex Cull. Vu /.-aland Barker , 

cincta (Nees) Co 

\ I ■■•',.-,.-,«»').-, 


Volume 90, Number 1 

(BOF) AF355983 

Barker el al. (1999) \F267 

ent) using an annealing temperature of 52°(: 
I. extension at 72°C (.'J mill.) and denatur- 
I 97°C (I min.). A l()-.ninute extension pe- 
72°C followed tlie :i() cwlcs. prior to cooling jr^ " Tow|l) r [) APSA was also ~ use( ? 

■d as luo separate products. II S 1 and ITS2. 

Coipoiah.Hi). (.ompleled se(|iiences were then ii 
ported into the existing data set of danthonioid [' 
and i/mirl sequences, and aliened u-ing D\I K 
(l)\A And Protein Sequence Alignment: K. 
Ilarley. Dept. Chemical Pathology. I nivcrsih 

ubsequent phylogeneti 

The/poC2and ITS data \ 

pi of Tris-FDTA buffer or water using '"'''•«' l > 
•A (,)l \(h,ick PCU Product Pi 

version FOh.ia (Swofford. 2000). Ccnlmfindia -/ 
were carried out using '" (Nees) '■ A - (: °l )e a,ul Mcxiiutellera mi. 
PRISM or Alii PB.ISM PIG in K cycle H'i's-) < : «>"<->< ««-«' u*«l a * outgroups , all ,, 
,.,- kit :.e.-.mlin. f to tl.e man, , fact i, rorV in- s, ' s ' Tn, ' sr *P''<"- ll; >^ pl'e\ mush boo,, sl.ovu 
he related to the danthonioid lineage (Parker. PJ 
Parker et al.. 1090. 2000). Two hundred ran. 
input order searches were conducted (TKKKP - 
MULPARS OFF). The most parsimonious Ii 
lound using this search were suhjecled to a 
heuristic search (TBR option. MUTPAHS ()\). I 

us. Th«- ITS PCK product was sequenced 

I7SF. .xf'.S. r>.»H. and 26SK published In 

ml internal primers were used to sequence 
(Hsiao et al.. L998), 

I" ,.v'-c\<': ict cgg c\ \ <".g(, I! "" MM| ' ««»-•'>-- _»-»«w 10.000 . 

i.psir (.v-riT gcc gtg acg ccc- < s ' ( ° | , ' ,l " < ' 1 on lr ,m,,vl,,lial and 

(12 PCK product was sequenced using 

primers ", 7 ,oC2-l" and "r /M >( 12-2" Combined molecular data set. 

I.. 1000). Sequencing was carried out and against when and how to con 

iL IW(>: \i\nn K Carpenter. I<><><». cies combined lien- are both North American, they 

mil Kellogg | !')<)()) discussed unions vvavs in are considered In be -ullici. ill I \ < li>-.l\ related t.> 

.vliich conllicl between data sets can lie assessed. allow data ■ ■oinliiiuli.ui. I'he llmd lictive taxon rep- 

niil advised on when it i- siiilalile to conduct -neh iv-cnl> a -ingle species in a nioimly pit genus. \s 

ests. Following their suggestion, combination of suggested bv their rank, the two subspecies of 

lata sets is possible if the trees do not conflict, or < '.hat -lohi, >mw. iniolm rains arc ohv ioush closely re- 

I conflict receives low bootstrap support. Thus each lale.l. and were once considered as distinct species 

lode on each of the two trees is tested for congru- (see Yerboom iX Finder. 1998. (or a summary). The 

•nee against the other tree. Where the nodes are two subspecies share a unique morphological char- 

ongnieii! (thai i-. d t contain conflicting inlor- acler ol a lull d hair- below each spikelet. corrob- 

nalion. e\en if they are not identical). thev are ob- orating ill.' monophvlv of the species. The cytology, 
iotisly combinable. Where thev are incongruenl. ecology, and taxonomy of this genus has been ex- 
he bootstrap support for each of the conflicted lensivelv studied (Spies el aL !«>*><>: \erboom & 
lodes is examined. If the sup|>ort is less than 709L Finder. I99K). and these two entities can be readily 
he incongruence is interpreted as being due to combined. Triholinm has been recently re\ ised 
■hance. and that there is no hard conflict. This log- (Finder & Davidse. |W7) and now includes the 
c is derived from Mason-Gamer and Kellogg prey iously inoiiolv pic Irochlucna Nees (= T. pus- 
1996). ami ultimately from Rodrigo et al. (IW{) ilium (Nees) Findei ■ & Dayidse). Although data for 
uid was recently applied by Fldenas and Finder this latter species are available for both data sets 
2(KK)). Tests for incongruence (such as that pro- and thus could be used to represent this genus, a 
>osed by Karris et al.. |99J. 1995) were thus not broad,'.- sampling of Triholinm was considered ap- 
•onducled. as the- trees oblained here show no well- propriale. hence the combination of the two species 
iupported conllicl. of Triholinm. 

lo reduce the amount ol missing data in tin- With the use of fictive taxa. the combined data 

oinbiiied analysis, a combined data set was ere- set comprised (."> taxa (including oulgroups). and 

ited from only those taxa lor which both ITS ami included representatives of all the seven informal 

poC2 data were available. I nlorlunalelv. four lineages recognized by Barker et al. (2(MH)). "Total 

;enera are not represented by a species common Kv idence" (IK: see de Oueiro/ et al.. \W7y. Huel- 

•fictive taxa" I Kellogg K Finder. I99.",) or "hy- set was carried out as described above for the in- 
n-id taxa" (Wiens K Reeder. 1995) that repre- dividual data sets, 
ented the four genera for which different species 

ir subspecies bad been sequenced for each gene. Supcrtrccs. \ supertree approach to combining 

'hese lictive taxa thus represent genera, rather die results of the ITS and rpoC2 analyses was taken 

hun species. This approach is acceptable when <<» include all species that were sampled for one of 

he monophvlv of the higher group is known or dies,' two regions. The supertree was calculated 

ceepled (Wiens vK Reeder. I«)%). These fictive from strict consensus trees of the two molecular 

axa represent Dtinthonia la combination of IK data sets, using matrix representation parsimony 

picata (F.) beauv. ex Roe.n Ox' Selmlt. \rpoC2\ (MRP: Bauni. 1992: Ragan. I 992). Kach consensus 

ml I). California Bol. |ITS|). Pcnlamcris Beauv. tree was described as a matrix, with each node con- 

a combination of /'. llmarii Beauv. |/yx>< !2 1 and stituling a character. Species described by that 

'. macrocalycina (Steud.) Sehweiek. |ITS|). node were scored as I . and species not included in 

'.havlohromns \ees (a combination of ('.. iniolu- the node as 0. Species absent from the tree were 

ratus (Schrad.) Nees subsp. imolucralus [ITS] scored as unknown \'f) for these nodes. In order to 

ml (.'. inrolucralus. (Schrad.) Nees subsp. drc- reflect the support each node had within its data 

canus (Nees) \erbooin |///oC2|) and Triholinm (a set. we weighted the characters (nodes) relative to 

oinbiiialion of T. uniolac (F. f.) Renvoize \rpoC2\ their I Islrap support lev, -Is: bootstrap support of 

ml T. hispidum (Tlu.nb.) Desv. [ITS]). 0.7-0.79 was weighted X M 0.B-0.B9 weighted X 

Pcntameris has been recently revised (Barker. and 0.9-1 weighted X 9 (Ronquist, 1996). Since 

99.? | and is clearly a moiiophv lelic lineage. How- the supertrees search for resolution, and not lor 

ver. the monophvlv ol Ihmlhonia has not been ex- nodes supported by all possible trees of a given 

mined critically and. lor the purposes ol this study. level ol parsimony, we used a semi-strict consensus 

- Centropodia glauca 

^"•"d Merxmui 

gg | Pentameris i\ 

CLQOl Pentaschistis 

Cortadena toetoe 

100, Tribol 


Joycea pallida 
Rytidosperma pumilum 
Rytidosperma nudlfloru 

- Pseudopentame 

- Chaetobromus ir 

- Lamprothyrsus p 


Cortadena h 
Cortadena rucliuscula 

H62 | Merxmuellera cm 

Im.iIi Kilch optimization ami opt uni/aliori pro- edited by removing iminforiiialive 

ng rc\ersals. lo |irc\ cut shared absences from la\a fur which -e(|iienee data were u 

>.ng. as suggested l.v Bmmda-Kmonds and data were compiled l.v HIM, from 

it (1998). both herharium and fresh eolleetioi 
published revisions and anatomic, 

'HOLOGICAL DATA were consulted. These characters ; 

pari of a larger study (Barker et al.. 2000). are listed in Appendix 2. which also 

Annals of the 

vond search, iii which nil most paisirnon kmis I 
ere kepi). This search was limited to 10.000 li 

lg the data by the it-sealed consistency index, 
■peating the above search protocols. 

these nodes ranges from UV '( {Wrrximtcllcni riijti. 
M. dccoro. and »/. lupulinu clade in the r/)oC2 tree) 
to 1007. (numerous small clades in hoth trees). Of 
these nine common nodes, live comprise small 
clades of two to three taxa. and (with the exception 
of the lull lade) noi I tl I . 

common correspond to the lineago recognized h\ 
Karker el al. (2000). There is thus some congruence 
between the two data set>. and. more importauti). 
then- are no conflicting nodes where I l-trap -up- 
port is greater than 0.V7,. The data sets were thus 
considered to lie comhiuahlc. The TK analysis in- 
cluded 2(>2 ph\loi;cuclicall\ iulormative sites, and 
analysis of these data resulted in 32 equally par- 
simonious trees of length 713 steps (C.I. - 0.534. 
R.I. = 0.737). The sir 
is shown in Figure 1. 

hut these do not affect the final alignment 
. This alignment contained SO taxa and re- 
in <)H phvlogenelically informative sites. The 
mm parsimony analysis found 0710 equally 
.onions lives of length 105 sleps (C.I. = 
R.I. = <>.«<*>>. \ randomly selected most 

i lie i i .-i sequent . - i ( in;, ■lohnmiii.s (and lo a 

lesser extent I'scudopciihiinciis) were difficult lo 
align in two regions: one at the heginniug ol the 
ITS I region (a deletion in the Churl, ihnmiiis se- 
quences), the other al the beginning ol the I I S2 
region (an insertion in the Chuclohromits scquenc- 

rohlciiialie. hut ineorporaled mam small gaps, 
loslly one lo Cue hases in si/.e. The alignment ol 
I sequences produced l0<5 phv logenelieally inhn 
u.liye sites. I'Lvlogenelic analysis resulted!.. 10 
lost parsimonious trees will, a length of .">.V> steps 

ex of 0.201. and retention index of 0.0 I I. 

hili/ed allei three roiiml- of re-weighling. and this 
weighted data sel resulted in 080 lives, with a con- 
sistency index of 0.503 and retention index of 
0.850. The strict consensus of the most parsimo- 
nious trees obtained following re-weighling is 
shown in figure 5. Uoolslrap support \ allies were 
not calculated owing to the high degree of liomo- 
plasv. poor resolution, and the yen high number ol 
. lound in the abo\e 

ombllleil with I 

■ inoiphologieal data were i 

a single clade for Corladeria. including C. 
haltlii and l.amprotliMsus. C.orhulcria is sist 
Cliionixli/oti. and both are nested within a 

group ol I 

the molecular analyses, the relationships 
ig these groups arc quite different. Mlhough 
ipport analyses were conducted, the low rcso- 

solution than the >lrict consensus 
odes according lo their bootstrap 
• trees added only a single node 
elude. Treating all characters as 

ible characters located a further node that link 
and Notochloe. Including C. archboldi 
the collapse of the three included Danthonit 
species. C. arrhbohlii. and the PUnlluinllu ■>/> clad< 

•rsible for the unweighted matrix prod need the , () ,-„„„ ., |M) | vt((im wit |, t | lr , un 
result as obtained from the weighted matrix with ,)„. Hylulosperma clade. and the ( 
parsimony, while the weighted matrix with ir- (|)()( | es marked j n K ig. (,). 


Chionochloa rigida 
Chionochloa macra 
Chionochloa pallens 
Cortaderia archboldii 
Cortaderia fulvida 

Cortaderia rudiuscula 
Cortaderia splendens 
Cortaderia richardii 
Cortaderia toetoe 
Cortaderia bifida 

Pseudopentameris macrantha 

inhumation. However, the use of mdels hi phvlo- 
oted above, the alignment used line was genel it- anab ses is slill somewhat eontenlioiis. and 
il hi dial presented b\ Harker et al. ( l { > ( )<>). mini. toiis . . m lin- iiicllidds haw been pmpo-ed (>ee. 
additional ind.'ls had to he ineorpoitated lor example, (dribet K Wheeler. IW); Simmons & 

I heplameric repeal- following insertions) being of species and I'luilliuiitlir-i, and \oloihlor share 

equenl (Barker. 1005). |{eeause of the possible a two-base pair insertion in ITS I. 

on-random nature of indel ^ain and loss, as well Of (lie South American taxa. C. rolombiana and 

- the finding dial inultiple I,,— ,■- (deletion-) fo|- (.. bap,doln'rba -hare a one-ba-e pair insertion in 

)\v gains (insertions), indels were not eoded (or ITSI. and ha\e a se(|iienee similarity of OoVP/r . 

iclusion in these analyses. These two speeies are consistently resolved as sis- 

Tlie rjuA'rl sequences lor the New Zealand spe- ler la\a in all analyses. Field observation- indicate 

ies are identical, with the exception of ('.. turharia. thai these two species overlap morphologically and 

Inch has a deletion corresponding to two he|)ta- geographically : <-'. rolombiana is functionally gyn- 

icric repeat motifs. This deletion is in the same odioecious and occupies lower altitudes, while (], 

■gion as a deletion in /.</////</o//nrw/\. emphasizing hup<dolriiha is aponuctic. and is found at higher 

le honioplasv of 1 1 1< l< I particularly deletions. altitude-. It i- po— ihle that llie-c two "species" are 

ir this region. The sequence similarity between extremes of a single species that shows altitudinal 

amprollnrsus and C. turharia is 07. V/(. variation (M. I .vie. unpuhli-hedl. 

An examination of //»<*( !2 sequences from species Sequence similarity between /.amprollnrsus and 

i the South \meriean clade (including l.amprotli- the "south \itiericau Cortadcria species ranges from 

rsus) reveals a similar lack of sequence variation ( H).iV7( ((.'. srllouna sequence from Hsiao et al.. 

mong the species, with similarity ranging from 1008) to 0».0<7r ((,'. bifida). The least similar se- 

( ).'2'7( to 1 ( M I9r . C.artatlcria <<>/omhiana (Pilg.l Pilg. quences are those of I], sclloaua (Hsiao el al.. P)OH) 

n.l «:. hapalolri.ha (Phil.) Conert are retrieved as and C. rolombiana (04.2<7r ). ami the most similar 

chide (in all analyses) within the South American sequences are l *KlV'/( similar (('. sclloaua (this 

roup, but these sequences are not identical study )('.. juhala ( L.ui.i Slapl. and ('.. juhala-C. ar- 

) < )..'i'/r similarity), \inong the South American aurana Slapf comparisons). 
>ecies. C. scricantha (Steud.) Hitchc. is uni(|iie in 

ay u.g a deletion of 12 bp (two heptameric repeats). ITS alignment— problems and implira- 

ul the remaining sequence (i.e.. excluding the re- """<- Tll «" ' ' S sequences of Cbarlobromus. and to 

ion with the deletion) is identi .1 t .. f the a l ^ s «-'' «^«"»» I'srudopentanuris. were difficult to 

[her species, including Ijunprotbvrsus. '>»>»• «'»• *ere unexpectedly divergent from the 

The lack of resolution of relationships among the remaining data set. This high divergence might in- 

iccies of Cortadrria by these data is disappoint- ,li< ^«- «»" accelerated mutation rale in this lineage 

lg. and is caused by high sequence similarity of ol " paralogous copies of the gene. Data for C. in- 

iccies in each of the chides. This may be as a rolurratus subsp. dregeunus show at least two dif- 

■sull of recent speciation and radiation within the b'rent copies of the ITS region in the genome, 

ade. or thai the grass-specific rpoV.2 insert mil- ^ lli(l ' ""'> relied the polyploid nalure of lln- -uh- 

ites more readily by slipped strand mis-pairing species (Spies el al.. 1000: \erl m K binder. 

)roducing insertions and deletions) than by sub- I'^l- This second subspecies was excluded from 
itulions. and that the actual sequence is con- the analyses discussed here because of this v aria- 
rained by the as-vet-iiiikiiown function of this re- tion. as the sequences were not usable for a portion 

.n. However, other r/;oC2 data from species of the <>' lh<* ITS region. However, in regions where I 

rican genu- I'cntasrbislis (\ees) Spach (Parker se,|iieiices from these two subspecies are alignable. 

(dlbert. unpublished) suggest that this region diere is nonetheless a substantial amount ol se- 

y undergo both substitutions and insertion-de- quence divergence. This suggests thai, despite their 

ion events. Thus the lack of sequence diversity present status as subspecies (and even their past 

both the South American and New Zealand status as species, for that matter), the ITS region is 

des of Cortadrria may be the result of a recent presently undergoing substantial re-arrangements 

gin and diversification. in this taxon. possibly as a result of ongoing poly- 

ploidi/ation .vents. These events may have been 

SKOI i:\ci: Dl\ i:ii(,K\(K delected here due to the incomplete and/or ongoing 
process of concerted evolution in these lineages 

Willi the exception of C. spicules Connor and (Arnheui, el al.. I OHO; et al.. I OHO: Arn- 

turhariu, sequences of all the species of Cortad- hoim. MMtt), the results of which will ultimately be 

/ from New Zealaml are identical. The former a uniform copy of rl)N A loci within the taxon. This 

■cies is 00..,</r similar to the remaining taxa. and has been shown to be possible in other polyploid 

id la\a sampled, and addili 
ipertree. The remarkable « 

tradictory evidence, n-sultinjj; in suhslantial loss of 
■■■■solution when C. archiho/dii is included. Of more 
interest arc nodes dial conflict helween the TK tree 
and the siipertree. I or example, tlic siipertree finds 

il'cnliiM ///.*//> c\mU-. \Chiono, h/aa. [Corhnlciia. I!\ 
lidos/icrma))). while tlie TK tree ohlains (Chimin, h- 
/<«/. \Pcnlaschislis cladc. [Corladcria. R\lidos/>cr- 
ma))). Curiously, i lei I her analysis of iudiv idual data 
sets located the node found in ll.e TK live, which 
must llms |>e regarded as tlie novel node. suggesl- 


markahly simila 

• TK I 

. The , 

•oinhiniiig A1//V07 i / / 

l (Kirk) H. P. 

iinler. A', nudtjlorum \\\ Morris) Connor t\ Kdgar. 
arroo<hh>a jtnqmmi (K. f.) Concrl tK Ti'irpe. W 
ulanlhonia »m,-ilis I kirk) /o|o\. .m,l \„^ ri „l„n 
\ania ftirs/iilnsa (Candied. | II. V. Kinder: the proh- 

lis elade. However, this agrees with the result of 
ic ITS analysis, and again the TK analysis located 
>dcs not found in the analyses. There 
e. however, some positions of uncertainty: 
The presence of a "Corladcria cladc" with Dan- 
lhonia. riintliantlirsis. and Xolochloc included. 
This is retrieved holh hv ITS and the comhiiied 
analysis, hut always with less than 507r hoot- 
strap support. However, the conflict of /■/«>( 12 
and ITS, and the low level of support, indicate 
that this grouping cannot as yet he taken seri- 
ously. Kiirlhcrmore. the siipertree is unahle to 
resolve the relationships of any of the species of 
Danlhonia. I'/inlhanlhcsis. and Xolachloc. 
The other fictivc laxa survived well: Pcnlamcris 
remained firmly in the Pcnlaschislis cladc in the 
siipertree: Chactohromu.s retains its position as 

morphological data set lacks <le- 
ic main groups extracted are re- 

to the molecular trees, although 
there are numerous differences in detail. There ap- 
pears to he iid morphological support for a hiphv- 
Iclic Ci triad I i i i | | the morpho- 

logical data set retrieved a moiiopliv lelic Corladcria 
cladc. which includes Corladci m. I.anij>ii>lh\rsus. 
and Corladcria arcliboldii. This cladc is hascd on 

■ the lack of re-olulion on 

■ of support for nodes. Kor example, the 
places Cfiioiiochloa in a hasal position. 
• the I'cnlaschislis node. There is prohahly 
acter support for this (and it conllicts with 

Ml molecular analyses clearly show that Corlad- 
cria is not monophvletic. In the TK analysis. Cor- 
ladcria is retrieved as one cladc. hut with less than 
:■)()'/( l.ootstrap support, and with Danlhonia. Xo- 
lochloc. I'linihanlhcsis. and I .anijnoih\rsus emhed- 
<lcd in it. This cladc is divided into three suhcla- 
des: the Australasian clatle. the S(.uth American 
cladc. and the C. archholdii-UcUw Danlhonia 
cladc. The Australasian clad.' includes I'linlhan- 
ihcsis and \olochloc (HiJ'/r hootstrap support in the 
comhined analysis: 7 \</< in ITS analysis: \\V/< in 
r\»o(;l analysis, hut also includes Danlhonia). The 
South \mericaii cladc includes l.anijiiolh\iMi\ and 
receives (>'<\7< hootstrap support in the TK (slightly 
up from (*)<Y ( in the ,poC2 analysis. 7A c /< in the ITS 
analysis). The C. archholdii-\'u[\\c Danlhonia 
■ lade receives <.«>'v support in the TK analysis. Re- 
lationships helween these chides and the Hylidos- 
pcrma lineage {IA) ( A hootstrap support in the TK 
analysis) and the I'scndo/tcnlanicris cladc (I <)()'# 
support) are not well supported. There are thus crit- 
ical nodes in the topology that are p Iv supported. 

and additional data, holh molecular and morpho- 
logical, are required hefore the relationships of 
these lineages can he resolved with some certainly. 
However, il is interesting to note that the resolution 
of Corladcria into the South American and New 
/calami lineages is supported (at least in part) hv 
chemotaxonomic dala. Connor and Kurdie (I<W1) 
reported thai Cylindrin, a triterpenc methyl ether, 
was restricted lo some species of section Corladcria 
(excluding C. sclloana). and ahsenl from the South 
American C. bifida (sect, lii/ida) and all the New 
/calami species tested. Kurtherniore. the New Zea- 
land species also possessed a range of other triter- 
penc methyl ethers that were not found in anv 
South vmericaii species, hut which are known from 

alK ||„- adaxi; 

>ed midrib. Furthe 

stoloniferous Haiti 

\otoehloe).,- ITS data 

III,- lie 

. Dan- 

npling ul' Danl/ionia species in this study. The 

ition l.,-lu,-,-n III,- two cla.lcs ul" Cortarieria and 
• species nl' Danlhonia when (.'. arehholriii is in- 
i(l,-,l. Although bootstrap support for th,- position 
li the Australian species of Cortarieria is stron- 
in the r/>oC2 tic- than the support lor a position 
h th,- South American species as suggested by 
■ ITS tree, node support in different parts of the 
i- indicates that it would he simplistic to surest 
I the rpoC2 results arc more trustworthy. Cur- 
itly there is therefore strong evidence that C. 
hboldu belongs in the hroad Cortarieria. rather 

i- it is generally possible to distinguish th, 
s of southern Uriean Mer.xnuiellera from New 
mil Chionorliloa. and these from Cortarieria. 
■maining similarities are still quite nun hieing 
<ecnnd character group involves the evolution 
• gvnodioeeious breeding system. Possibly the 

me lealurcs are lound largely or only in , 
■ Cortarieria segregates. In many South An 
species of Cortarieria there is a variously 
ied layer of collenchv malous cells below 

( iiioth>,-lrioa. clade. but 

furl he 


in be made in determining where in this clad,- i 

It is tempting to argue for the recognition of Cor- 
delia m, id,,, Idu a- a mouotypif genus, related to 
e Australasian species of Cortarieria and to Ihm- 
onia. However, the morphological evidence is not 
■ong enough, nor is the molecular evidence cur- 
nllv convincing, liven il Cortarieria archboldii 


there ■ 

eparate genus for this 

everal species of softly herbaceous species o\ Dan- 

hoiria associated with the ■ lineage, and the 

mall. I.ealhlai.,1 genera \oloelrioe and Plinlhan- 
hesis associated with the other, means that several 
typical" character of Cortarieria must have either 
volved Iwice. or been lost twice. These characters 
■ inflorescences, spikelels with 
with abundant 

cies it may be much narrower. It is sometimes 
linked lo the outer bundle sheath via well-devel- 
oped extension cells. However, this is also found in 
one species. C. Julriria (buehan.) Zotov. from \ew 
Zealand. The South American species also have 
characteristically yiHous inflorescence branches, a 
character that is remarkably homoplasioiis in the 
grasses. A divided phloem is typical of the C. sel- 
loana clade. but is absent from most other species 
of South American Cortarieria. \ha\ial ridging of 
the leaves has a similar distribution. 

Most of the New Zealand Cortarieria species pos- 
sess curious islands of clear cells in the chloren- 
chvina. a feature also found in l'hnthanthesis and 
Xotoehloe. However, this character is missing in C. 
splenriens. as well as ,.'. anhholriii. but this could 
be interpreted as a secondary reversal. In addition, 
all New Zealand species of 'Cortarieria have a well- 
developed and prominent leaf midrib, a feature rare 

'" ,M< ' danlhoi Is. but also lound m the Ninth 

\merican C. ruriiuseula Stapf. 

Over and above these features, it is worth revis- 
iting the cytological data in light of the above re- 
sults. The New Zealand species all have 2n = <)0 
<'■<■•■ -"«• decaploid. if v = *)). while the South 
American species have variable chromosome num- 
bers, ranging from 2n = M, (tetraploid). through 2/; 
- 72 (ocloploid). to 2/, = 108 (duodecaploid). As- 

the ba.c , 

The lirsl involves the formal ion of lough, long-lived 
plants, with tough leaves. This has happened in 

having 2// = l.v = 36, 2n = H.v = 72. and 2n = 
I2.v = |()8. The chromosome count of the New Zea- 
land species of 2n = <W) may be a decaploid inter- 
medial,- between that of 2// = 72 and 2/; = 108. 
suggesting these species could have arisen 
bom an allopolyploid ancestor, which was a hybrid 
between 2n = 72 and 2/1 = 108 parents. While 
resorting to an ancient hybridization event as an 
explanation may be viewed ; 

eh a history could explain 
the conflict found here between the nuclear ITS and 
pl.i-hd rpo(';l data sets, with morphology reflecting 
a mixed or mosaic history, further confusing tin- 
phvlogeiietic signal in the analyses of multiple data 

,, :i •' ■:,!,,.'< ',,' i ; in i - < - I \> > I, cages; one \Us- 
tralasian. the other South American. The relation- 
ships of the New (iuinea species C. arclthnh/ii arc 
not salisfactoriK resolved, other than to allow us to 
state that it is not a member of either C.orlndrrin 
lineage. These relationships are not supported in 
analyses of morphological data, hut the available 
morphological data are still poor, and it is therefore 
possible that there arc undiscovered morphological 
markers for these molecular clades. Morphology i- 
thus m in r that needs further research. For ex- 
ample, aspects such as canopsis morphologv nee. I 
investigation. In addition. morphological -Iii.Im- 
are complicated bv the presence of both bisexual 
and leinali-oub spikelets. creating sampling prob- 
lems. Having saiil this, we realize that the molec- 
ulai -ampling is also incomplete, since most ol the 
(jnltti/iTHt spciies sampled here are \ndean laxa. 
and the lowland species have not been included. 
Only once more laxa and samples have been in- 
cluded in the molecular data sets, and an exhaus- 
tive field-based laxonomic study of the species (and 
their breeding systems) in both lineages has been 

be made. 

Literature Cited 

Arnlleim, IN. 1983. Concerted evolution of mull. gene fam- 
ilies. Pp. 38-01 "• M. Nei & R. K. Koehn (editors). 
Isolation of Cedes and Proteins. Sinauor. Sun. I. ilai'd. 

. VI. Krvstal. H. S< l.mickel. C. Wilson. 0. Uvder 

exchanges anion- rihoso.nal ^eiies on non-homologous 
chromosomes in man and apes. IVoc. Natl. Acad. Sci. 
U.S.A. 77: 7323-7327. 
•Vstegiano. VI. K.. A. VI. Anion & H. K. Connor. 1995. 

rn.i (I'oaeeae) en Vlgeiilina. 

, & . 1090. Flora laiierogamica 

Wgentina. Fascicule 22. 10. I'oaeeae. part.- 2. Tnhu 
13. \rundineae. Programa PB< >FI .( >B \ (COMCKTl. 
Cordoba. Vrgentina. 

Barker. \. I'. 10').'',. \ |,i,,s\ s|,. m .il ic ol I'mianu-ns 
Beam. (Arundiuoi.leae. I'oaeeae!. Bothaha 23: 23-17. 

. 1001. Kxlernal fruit morphologv of southern \f- 

rican Aruiidineae I Vrun.linoideae: I'oaeeae). Bothalia 

21: 55-66. 

. 1005. A Vloleenla. ■ Pliv logenv ol ll„- Nihlainil v 

Arundinoideae (I'oaeeae). I'll. I). Thesis. I nnersitv of 
Cape Town. 

. II. P. Finder c\ K. II. HaHev. 1007,. Polvphvlv 

of Arundinoideae (I'oaeeae): Fv idenee from rhc\. se- 
quence dala. Svsl. licit. 20: 123-135. 

grass-specific insert in the ehloroplast r/i,A:2 gene elu- 
cidate generic relationships ol the \ ruii.lmoi. leae 1 I'o- 
aeeae). Svsl. Hot. 23: 327-350. 
. X C. Morton. 2000. The Danthonieae: 

(a-nerie eomposil ion and relal loiish 1 , ,s. I'p. 221-231 in 

aties and' Kvolulion. CSIBO Publishing. Colli, igwood. 

Baum. B. IC 1002. Combining trees as a ua\ ol ', ■on. Lin- 
ing dala sets for phv logenetie inference, and the desir- 
ability of combining gene trees. Taxon II: 3-10. 

I'.iniiida-I monds. ( ). R. P. & H. N. Bryant. 1 998. Prop- 

Syst. Biol. 47: 497-508. 
Blake. S. T. 1972. 

review of the \1ist1alian spei-ies ol Ir/iloflilna (Crami- 
neae). Contr. Queensland Herb. 14: 1-19. 
Borgmann. K. 190 |. \nleil der Polvploid.-n in der flora 
■ les Bisinareks^ebiiges win (Miiril^iin.';i. /oil. F Bot. 

52: 118-172. 

Bowden. \\. VI. <.N II. A. Sen... 1902. Chromosome num- 

l!oi. Id: I 115-1124. 
Brower. V. V. /.. B. I), Sailed \ Vogl. , I99(, („-,„• 

speclive. Annual Hev! Fcoh Svsl. 27: 423-150. 

Chase. VI. \\. yN II. IT Hills. 1991. Silica gel: \n ideal 
material lor field preservation ol leaf samples for l)\\ 
studies. Taxon 40: 215-220. 

Clark. I.. (;.. W. Zhang & J. F. Wendel. 1995. A phvlog- 
etiv of the grass familv (Poaceae) based on nilhV se- 
quence data. Syst. Bot. 20: 43o-W>0. 

Clay. K. 1983. Variation in the degree ol eleislogamv 
within and among species of the grass Danthoniu. 

urn. kew \ddilional Series Mil. Il.r Majesty's Nalio- 

\ruiidineae. J. Cramer. W ienheitl). 
. 1900. Dre^-orhlou. erne neue Callung der Cra- 

mineeii iCianiineae. Viuii.l ideae. Danthonieae). 

Senckenberg. Biol. 47: 335-343. 
. 1970. Mer.xmuellem. eine neue ( ,altiing der < ,ra- 

mineen (I'oaeeae: Vrunduioideae). Senckenberg. Biol. 

51: 129-133. 

. 1971. The yeans Danlhunii, ill Vfriea. Vlitt. Bot. Vlunehe.l 10: 299-308. 
. 1975. Die Chionovhloa- Arten von \11slralie11 

and Neuguinea. Senckenberg. Biol. 50: 153-104. 

Arundinoideae! Pp. 239-250 in T. B. Sodersirom. k. 
\V Hilu. C. S. Campbell K VI. F. Barkvvorlli (editors). 
<a-jss Svs|,.|i.aiics and Fvolulion. Smilhsoiiian Institu- 
tion Pre-s. Washington. D.C. 

& A. VI. Tiirpe. 1900. Karranrhioa. eine neue 

(ialtung der Cramineen I Poaceae. \r„„. I, noideac Dan- 
thonieae). Senckenberg. Biol. 50: 289-318. 

Connor. II. F. 1903. Breeding systems in New Zealand 
grasses IV Cv iio.lioeeis.u in ( Ki latin h,. V» /.calami 
J. Bot. 1: 258-264. 

. 1905a. Breeding systems in New Zealand grass- 

li.iM-r M.imiLM'in. 10: 292-290. 

— & J. L Doyle. 1987. A rap 
lure lor small quantities „f In- 

I .1. Hoi. 9: 519-525. 

•I. Taxon 23: 595-605. 

c, J. P., J. J. Bull & C. ' 

1:52- 1 5,' J. 

A. Meinke, I. Dory & 

>rla<L>na Stapl. Pp. 

Kell,,j.». K. \.K II. I'. Under. 199.-,. IMix lo-em ol I'oak 

Pp. 511-542 in P J. Kudall. \\ .1. Crihl,. 1). K Cull 

& C. J. Humphries (editors). MonoeoU ledous: Svste. 

tonoma de Mexico. 

Louis; The Natural 

Oogenetic structure 


(erred from ehloro- 

& G. Davidse. 1997. The syslc.nalics of Tribi, 

ium Desv. (Danlhonicae: I'oaceae). Hot. Jalirb. Sys 

interlire survival strale^ies in ihe Cape r'vulios grassc 

35. Separate versus 

Bothalia 20: 91-103. 

leiroz, A. 1993. For o 

I. 42: 368-372. 

-. M.J. Donofrhueft J. 

,. Kcol. Sx si. 29: 057-081. R\ti(hps,-rm„ Daiillioni ■ I'oaccacl complex I', 

et. J. M. J. 1951. The ocums Ihimhnnia in -rass (y. 597-628. 

lo-env. Amct. J. Hoi. 11: 204-2 1 1 . , J. K Thompson. H. I>. Kllis ,\ S. M. IVrold. 

r ; /,,i ,',, ,-■,,,, ,1,,-ns. eiu hemeikungswerter Kail chromoso- 

maler ami oekologisclior I )i\ergen/. Oeslerr. Hoi. Z. 

Ill: 193-207. 
Simmons. VI. P. ik II. Oeholorena. 20(H). Gaps as chur- 

aclrrs m sequence-based |ili\ Ingenel ic analyses. S\st. 

Biol. 49: 369-381. 
Spies. J. .).. II. I)u Plessis. N. I'. Barker & S. M. C. van 

Wyk. 1990. Cytogenetic studios in ilie genus Chaeto- 

hrnrnns (Poaceae: \ rundinoao). Cr ne 33: 646-658. 

Slobbms. (,. \..& B. VI. l.oe\e. PHI. A evlological study 

of California forage grasses. Amcr. J. Hoi. 2!'.: :{7 I 383. 

Sun. Y.. I). Z. Skinner. G. H 
1991. I'hylogenelie anabsi 

Nieora. E. C. P>73. Novedades agiustologieas patagoni- 

cas. Darwiniana 18: 80-106. 
Nixon. K. C. ex .1. VI. Carpenter. I90(>. On simultaneous 

analysis. Cladislies 12: 221-241. 
Philipson. VI. Y 1 97-. Hauslorial svnergids in CorUid- 

eria (Cramineae). New Zealand J. Bot. 15: 777-778. 
. 1978. \pomi\is in Cmldilrriu jiilxilo iCi.iin! 

neae). New Zealand J. Bol. 16: 45-59. 
& H. K. Connor. 1981, Hauslorial sxuergi.K in 

danthonioid gra— . Bot. (,;,/.. 1 15: 78-82. 
Pagan. VI. \. \ ( >'>2 h 1 •■ n i< mi i n i hu-< d on ma- 

: X S. 


: 97- 

Veldkamp, J. F. 1980. Conservation of Notodunllumiu 

lm <Cramineae>. Taxon 29: 293-298. 
Verboom, A. G. & H. P. Linder. 1998. A re-evaluati( 

Nordic J. Bot. 18: 57-77. 

■ .;< . i 

53 . 

Kodrigo. A. (;.. VI. Kelb-lio.ges. P. I!. Berg,,| k \ P I. 
lie <...(] ii .1 <»<»;> \ randomisaliou lesl of die null h\ - 

| .i 1 * i ill i I ii c i mi n H I 

parameti c phylogenetic tree. New Zealand ,|. P.oi. .'51: 

Bonquist. K 19%. Matrix representation of trees, redun- 
dancy, and weighting. Syst. Biol. 45: 247-253. 

Sang. f.. I). J. Crawford ex T. K SIucssn. 1995. Docu- 

icer sequences of nuclear 
"or biogeography and con- 
Acad. Sci. U.S.A. 92: 

)64. Danthonia (Sieglin- 

riliosomal l)N\: 
(,i; I. '5 -(),'! I 7. 

Wiens. J. J. (.x I. V\ I 1,1' I m I ii i 

S\st. Biol. 14: 548-558. 
Zimmer. K. A.. S. I.. Martin. S. M. Beverlev. V. W. Kan 

& A. C. Wilson. 1980. I!a|i,.l dupliealion and loss of 

genes coding for the a chains of hemoglobin. Proc. Natl. 

\cad. Sci. U.S.A. 77:2158-2162. 
Zotov. V. I). 1963. Synopsis of the grass subfamily Arun- 

dinoideae in New Zealand. New Zealand J. Bot. 1: 78- 


inetn is described in cN-lail !»> Kllis (I 
2'». \,l,i\i,il riilfiing: (0) absent; I I ) |ii«-sciil 
.'50. t/«i.v,W//^/,^:|(M : ,l,s,,,l:(l||m.s,-Ml 
.'51. C.lriu cells III chin, nirh \ ma: |()| ah-enl 

sparse (1-2 law,.): |2) llmk , +■ .'5 lav. 
.■',.".. Ilullijoim cells: (01 absent: III iml> flat 

/r,,,,:<0):i:|l,.-,-7:(2l<)o, more. 12. Mullicrllula, uiamh 

ape: «») enlne; (I I olwiuvK l.ilol.r.1: |2) cell. liar -lands we 

1 bed. (1990).] 

i basal huH'. 14. \dii.\iul sclcrenclivma cup: (0) anchor-shaped: 

lens-shaped; (2) p\ramidal. 
'lit. 45. (>///(-/ /x/w/A- .vAra/// of primary vascular bundles 

16. Scallrrnl large cells in chlorenchxma: (<>| absent 

17. Chlorriichuna spaces I leaf mesic/xeric): (0) absent 

2H. /W/r«/e fcr/.s//,".s: (0) absent; ( I ) , 
21, hulicule microhairs: (0) absent; ( 
2. r >. Carvopsis shape: (0) linear: (I) <>l 




i values are n< 

it equivalent among th 

e VI,' 

xican Musci. Poacea 

e. and Astera< 

■eae. The numl 

>er of endemic 

■s 1mm nut' y.n 

"" 'M| III: 

IL' [:><•■ nil,! 

life cvcle. di-pcrsal al 

onsc K, select 

plant -roups 

include 2373 endemic 


Cluster analvsi 

ren UCghhonn 

,g slates and an 

'"■ - '-'I' 11 

• Hell 

ays have a geographical 1 

,asis. From the 

, standpoint of 

key words 

: Asteraceae 

, endemism, Mexico. Musci. 

, Poaceae. 

Although Mexico, witli 1.072.511 km', is I In ( I'm iim|iii ;i . 1903). IVrlim-nl literature for vascular 

third in biological diversity (Mitlermeier. 1988). It 1991a. h). in which the endemic taxa and their geo- 

harhors approximately 3().<)<)(> >prrie> ol vascular giapliical ranges are identified, 

plants, including more than 21.600 in about 2.500 Preliminary observations indicate that the num- 

genera of flowering plants (Kzedowski, 1993). ber of Mexican endemics is associated with eli- 

Among these, more than 300 genera and between mate- and geography -dependent factors. Thus, for 

SO and (AYY< of the species arc endemic to this instance, in the lowland moist areas of southern 

country (Ramainoorlhv cv Lorenee. 1987). There Mexico llie percentage of endemic vascular plant 

are 49 Mexican specie- of pine-,, represeiiling more genera is I lie lowest in the count rv. while their nu m- 

than 50% of the total for the world (Nvlc. 1993). bers increase toward the drier (R/.edowski, 1978) 

and 900 to 1000 fern species (liiba. 1903). The and cooler areas. On the highest mountains, the 

bryophytes include about 1700 species (el. Sharp extreme clunale mav have caused many species to 

et al., 1994; Fullord i\ Sharp. I ( )90). and among become narrow Iv adapted to the environment of the 

them, the mosses compose about 25 r /r of the Meo- alpine meadows and siibalpine elevations. Beaman 

tropical moss flora. and Andresen (1900). in a survey of the vascular 

High plant diversity and the large endemic clc- flora of the summit ol Certo Polosi in northern Mex- 

menl are features that set apart the flora of Mexico. ico. delected 27 ol <>l species (12.2%) endemic to 

Information on the number, origin, and distribution the sierra Mad re Oriental: 13 ol them were restrict- 

of endemic- (e.g.. l!/edow-ki. 1078: sharp. 1953) ed to Cerro PotoM. High endemism values have 

is still imprecise, but current data suggest their been detected in the dry lands of the Tehuacan 

concentration in certain areas such as the Neovol- Vallev (Snutli. I0(>5) where endemism approaches 

canic Belt, a mountain range bisecting the country 17% (Villasenor, 1993). 

between 19 and 20°N. and the Sierra Madre del The significance of these observations cannot be 

Sur, along the southern Pacific coast, which arc fully evaluated for the entire flora, flic main lim- 

considered centers of endemism for manv groups iling factors are the lack of complete data on the 

I I'll'l ' .,,1 .1 ' ' ll' " I' ' < "'I lion Vhm, \„ ion lid. Inv < s|,gado.« s 

lo ||„. ||, S | author, which pro\ ided Mipporl for Mrs. liodnguc/. IL 1 1 r i< 1 1 1 < - Orliz llci miido/ prepared the map ol Mexico 

'l V,'-,' ' n I .im. ■', I . ,' ',' l! " 'l iV.l'.'in i'.' . . . InMiIuIo d, |:.ol„ , . I- Postal . I' -•',', 1 1 1 ". I ( I Mexico. D.F. moya@ 

I „ ,,li ,. I ,|, I ., „ , -, , , \\ d« los Barrios no I. Los Keyes l/tacala. 51000 Tlalnepantla. Mexico. 

pda\ ilaa(«sci\ 

Ann. Missouri Bot. Card. 90: 25-34. 2003. 


.ra. Although the. II i I I it I lor 1 998). The number of >pooie> known from Mexi 

ears yet. the use <»i' an alternative strategy in this family was expected to rise |o ahout 30 

II permit reasonable estimate: I I and. with this, an increase in ll number I 

illation in Mexico, ami how en.lemism re- deinie taxa in certain areas: Villasenor (1993) si 

the geography of the country. In this con- geste.l a trend toward higher endemism \alues 

n. we make Holistic comparisons among laxa stales located in the drier northern and southe 

•h preliminary lists and geographic ranges areas or in the mountain region of Mexico. Unpu 

lahle. As specialists, we have produced anil lished data for the Yallev of Tehuacan recognize 

ir 358 Aste 

'It of cen- 1 restricted taxa out of the total 188 species. The 

i. flora ol \a\arit comprises 117 species of Astera- 

lata in a ceae. 15 of which are restricted to the slate (Ortiz- 

nber and Berimidez et al.. 1998). while in the Yucatan Pen- 

flora of the Vicalan Peninsula (Delgadillo. I « »: i 1 1. 
In the dr\ lands of Zacalecas (Delgadillo c\ Car- 
denas. 1987) and the Tehuacan Valley (Delgadillo 
& Zander. 1081). the proportion of endemics is low. 
but appears comparalivelv higher than in the trop- 

are recognized as endemic in Zacatecas (4.3% of 
the moss flora), and lour in the lehiiacan \alle\ 
(7 c /c of the moss flora). In the alpine areas there 
are 19 endemic species that account for \~7( of the 
moss flora there (Delgadillo. 1971. 1987). 

The studies on the Poaceae of Mexico are mostly 

on the distribution ol species in the country (e.g.. 
Hernandez V. 1939. 1964: Johnston. 19 Id; Miran- 
da. I960: Rzedowski. 1992. 1963. 1973. 1978. 
1993: Sharp. P>53). A valuable discussion on grass 
endemism was contributed bv Yaldes and Cabral 
(1993). who indicated thai a total of 272 species 
(3()'r of the grass flora) are endemic to Mexic... The 
Chloridoideae have the high,-! number of endem- 
ics, with 73 species, followed In the Panicoideae 
with |0. and the Pooidcao with 13. Ycording to 

' Jalisco. Mexico. Ye. 

• correlated with edapluc am 

ci. \sleraceae. ami Poaceae restricted In lb- polit- 
ical limits of Mexico was compiled from biblio- 
graphic sources and support from herbarium 
specimens. Sharp el al. ) 1994) and Delgadillo et al. 
(1995) were the main sources for mosses. In addi- 

compiled from publications such as Davidse el al. 
11901). McViugh (1083). and Yaldes-Revna and 
Davila (1995) for the Poaceae. and \h\augh 
(1081). Rzedowski and Calderon (I Wo). Strother 
(1099). and Turner I 1907). as examples, for the As- 
teraceae. The main herbarium sources include 
MKXl fo, the mosses, and MKXl . KXCB. IBl 0. 
MICH, and IS for the Poaceae and the Asteraceae. 

'alabase infoi 

determine the Hoi 

i was used to compute J 

lo Microsoft l'\< I I tables as the first step to use 
an NTSYSpc version 2.02 software package iRohlf. 
1008). A |)resence-absence OOPs (Operational 

(geographical I nil-, i.e.. -tale>) matrix served to 

group arithmetic averages method) dendrogram- 
(Figs. 2-1) were generated by the SAHN-cluslering 
command in \ TS^ S-pc. Similar procedures were 
used to review the relationships of individual 
groups oi sin ill. i u< is m Mexico « g.. the Neo- 

ic Belt states. As a whole, the present anal- 
mcerns 2373 endemic species, including 15 
•ries and 339 varieties, in the three plant 
studied. The database and the similarity 
es are available on request from the authors. 


•e Navaril. Jalisco. Colima. Michoaean. Gue- 
. Hidalgo. Mexico. Distrito Federal. Morelos. 
da, Puebla, and Veracruz. The density values 
in Table 2 represent the computation of a 
> density index of ecology textbooks, i.e.. 

species In Mexico. 1 of which arc represented In 
siibspecific la\a. Most endemic species are known 
from below 2800 in. but there is a group of about 

! i i .i ii i I i ' 

ever, most states along the Neovolcauic Belt are 
grouped together, and harbor. ahuii: with Oaxaca 
and Tamaulipas. more than 10 endemic la\a liable 
1). Despite the disparity in group si/e. the Poaceae 
and Asteraceae show similar beha\ ior. i.e.. they are 
best represented in certain adjacent slates, in the 
Neovolcauic Belt staler, ami in Oaxaca. The \alues 
for all three groups seemed to roiilirm this trend 
(Table 1). 

With respect to the Poaceae. a total of 257 en- 

eties— out of 950 grasses, have been registered for 
Mexico (Tables 1.3) for 27% endemism. Some spe- 
cies, such as Festuca hintoniana Alexeev. are 
known only from one or a few localities, while oth- 

is the case olSchajfnerella gracilis (Benth.) Nash. 
By contrast, many endemic species, including Bou- 
lelona scorpioides Lag.. Xlnhlenbergia gigantea 
(Fourn.) Hitchc. 17. jirma Beal. Bothriochloa hir- 
lifolia (J. Presl) Henrard. I'anicum decolorant 

Kunth. and / mcliloa me.-iatm lllilclie.) Morrone «\ 
Zuloaga, are wide-pread in Mexico. Fxcept for Ta- 
basco, there are endemic grasses known from every 
Mexican state, mostly distributed at intermediate 
elevations (ca. 1500-2800 in). The highest number 
of endemic specie- is found in Jalisco. Mexico, and 
Michoaean, with 55 or more species, but the slates 
ol Chiapas. Chihuahua. Durango. Guanajuato, Nue- 
vo Peon. Oaxaca, Puebla. San Puis Potosi. and Ve- 
in contrast. Baja < California. Gampeehe. Quintana 
Boo. Tlaxcala. and Yucatan have less than 10 en- 
demic species. In contrast to the results reported 
by Valdes and Cabral (1993). the present study in- 
cludes species with a strictly Mexican range only. 

Stales (California. Wizona. New Mexico, and Texas) 

species would increase to about 300. with the high- 
est number of them occurring in the semiarid hab- 
itats and the alpine grasslands. Fndemic Poaceae 
are present in low numbers in the slates of Cam- 
peche and Quintana Boo. and are unknown from 

The flora of Mexico includes about 3003 Aster- 
aceae: 1972 of them, or 65.7%, are endemic to the 
country. However, tor the analysis. 2030 -peeies. 
subspecies, and \arielies of endemic Asteraceae 
were accepted, i.e.. incorporating 58 taxa not ('till \ 
documented and increasing the percentage to 67.0 
(Table 3). The endemic laxa include 10 subspecies 
and 328. varieties. The known altiludinal interval 
for the Mexican Asteraceae places many of the en- 
demic taxa in the intermediate elevations (ca. 
1500-2800 m), and their individual ranges are fre- 
quently broader than those of mosses and grasses. 
Some species of Asteraceae have narrow ranges 
that depend on the presence of special habitat-. 
e.g.. (,cissolepi.s suaedijolia B. P. Bob. or Slephan- 
odoria tomeiilella i B. P. Bob.) Greene that are en- 
demic to gvpsophilous grasslands in San Puis Po- 
tosf; other species, such as Psacalium peliaiuin 
(kunth) Cass., range extends from Chihua- 
hua and Durango south to Puebla and Oaxaca. 
demonstrate a comparatively bmad distribution in 
Mexico. In terms of the states, those with the largest 
number of endemic species are Jalisco. Mexico. Mi- 
choaean. Oaxaca. and Durango. with 3P.5 to 520 
species in each slate. A second group, formed b\ 
Guerrero and Puebla. contains between 317 and 
365 species (Table 1). 

Cluster analysis of similarity .lata for the Aster- 

Plant Endemism in Mexico 

I j< >r plant tiiunp in\ ot ilMleil in \lc\icu. Number ul -perir- 

., , 0) 

■>75 (:.«'!) 

11 (ii) 

:',2i u.o) 

■>->:\ i loi 

205 (571 

1 13 (10) 


12 (.5) 

IKO (1) 

25 (0) 



213 (2) 

251 (2) 

300 (55) 

271 (15) 

2» (0) 

300 (15) 

.520 u>li 

no ( ir>) 

10 (12) 



100 (25) 

55 ( 1) 



230 (.">) 

22 (O, 

25S (5) 


277 ll.-,) 

27 (5) 

3,11 ,22. 


257 (11) 

.".2 H) 




10 (3) 

55(1 il25i 

23 (3) 

:si7 ( i.-it 

10 < 1 ) 

3«0 (17) 

1 (0) 

2 IK (0) 


237 (0) 

regional le\el. Vmia-calienlo. /aeatecas. Guana- 
juato. Ouerelaro. Hidalgo, and San Luis Potosf form 
the first block of neighboring stairs thai -hate nu- 
nirrous rndrinir taxa. The stairs in the peninsulas 
of I'.aja California and Yucatan sta\ together in ihr 
endemism dendrogram (Fig. 2) as do groups of 
states in northeastern (Coahuila. Nuevo Leon, and 
Tamaiili|)as). uorlhurslrrn (Chihuahua and Duran- 
go). and central Mexico (l)islrito Federal. Tlaxcala. 
Puehla, Veracruz, Guerrero, Mexico. Michoacan, 
and Morelos). 'Die position of certain slates does 
not conform to geographical vicinity as, for in- 
stance, in the case of Oaxaca. which is closer to 
Morelos and Michoacan than to Puehla and Guer- 

demic flora of Chiapas remotely links that state to 
the rest of the country. The data set for mosses and 
grasses modifies the \aluc of the similarilv eoeffi- 

dative position of many states in the 
! (Fig. 3). Such states as Aguascalienles. 
Guanajuato. San Luis Potosf. /acateca-. Chihua- 
hua, and Durango from the first Mock in Figure 2 
ha\e a dilierenl ( >;ii r i ti- arrangnnrnt in Figure 3. 
Also, individual analyses for mosses and grasses 
fail to produce reliable dendrograms, as indicated 
by the lack of similarity among neighboring stales. 
perhaps induced by the low number of records and. 
in mosses, by the absence of endemic records for 
about six states. 


indirectly related to the size of each 
lor instance. Aguuscalientes. Colima. 
i are among the smallest states in Mex- 
■ some of the lower numbers (Table 1 ). 
. the low numbers exhibited by the 
\iicalan Peninsula (Campeche. >uca- 

02 >hared endemic in<.>^|i..<-i,> in ihr Bell iJ!,V;i. 

country (8.87.). as shown in Table 3. A distinction 
must be made between "shared" and "reslricled" 
endemics: in lhi> contribution the former refers to 

latter are known from a single state. 

The Asteraeeae are represented by 1640 species 
and infraspeeilie ta\a along the Belt, or ne.irb 5.V, f 
of the Mexican Asteraeeae. About 1095 of them an- 
gle Neovolcanic Belt stale (Table 3): the percentage 
of endemism nationwide (07.69f ) is nearly the same 
as that for the Belt (60.8%). By contrast', there are 
222 species of Boaceae (23% of all Mexican grass- 
es) along the Belt states. 162 of which are shared 
with other states (73</r of the \eo\olcanic Belt Po- 

range (Table Ml. The Neovolca.uc Bell mas be eon- 
phml ,,n„p.| and higircndeiin.ind.-ilO endemics. 
be recognized, by these criteria, as a separate flo- 

.. I i-. 2) -iiuiiesl areas ol eiideiiusm . , . , , ' 

uther analvs.s ° ^ • S,HI,ll, ' m Ganges province (see Fig. 1). 

The dendrogram in figure 1 shows the overall 
)NG THE NEOVOLCANIC BELT relationship of eiidemism among the slates along 

the Neovolcanic Belt with a general trend in a west- 
resent an important element in the east direction. Jalisco and \a\arit. on the western 
eovolcanic Belt. The Belt occupies coast, are ver\ similar to each other, with about 240 





133 (10) 


2.30 Kl.31) 

oil, (12) 

: ;i ).;;;n, 

311 (221 


1.13 10.031 

Dislrilc. Federal 



1 1.07 «>.->7) 

i:;:; kh 


3.31 (0) 


.".no iir,i 


1.13 (0.07) 

237 id) 

2.07 (0.03) 



2 1 .333 

2.21 (0.1 1) 


332 chii 


0,3O (((.or,) 


258 (5) 

3.21 (0.1 Oi 

."»::•) (in 


1.13 (0.03) 


:i3 1 (30) 



) among states along the Neovolcan- 

portions of the slates of Colima. Jalisco. Mavaril, 

shared endemic taxa; the st; 

ites of Mexico and Mi- 

D.slrito Federal. Tlaxcala. Hidalgo. Querela.,,. 

ehoacan also share niunero 

lls taxa (300) and to- 

Mexico. Michoacan. Morelos. Puebla. and Veracruz. 

getber constitute a separate 

area of endemism de- 

thus extending the width of the country (Fig. 1). 

spile the geographical vicii 

litv with states on the 

About ,28 moss species are known from this area. 

western coast. Colinia and Tlaxcala have the lowest 

By virtue ol this number, the states along the Bell 

endemism numbers along the Belt (Table 2). and 

may be considered br\ologicall\ diverse, for thev 

this is attributed in pari to tl 

include about 7 1% ol the mosses known from Mex- 

The latter state, however, is 

florislicalb more sim- 

ico. The Belt stales are easib accessible, and llieir 

dar to Distrito Federal (Fig. 

4) than to Puebla or 

coUectmg record is better than that of other Mexi- 

Veracruz that surround it. 

iherwise. the close flo- 

aiidt 1 ;;::,!;;::!:;,;; 1 ^ percentageof 

— ^usci.Asteracea 


Mexican species 082" 

3003 n) 


Mexican endemics 80 

2030 257 


67.6 27 


1040 222 

Percentage from total 74.1 

54.6 23.4 

Restricted to one Belt state 17 

190 11 


i.ii-. .inr- iini— .- '' ^ii iii-nitv 

',„„■,.,„■. .„..! Mu- 

tate- along the \co\olcanic licit, (geologic history, 
ical pographic <lilTcrcmcs. and climatic regimes n 

Jaccarcfs Index of Similarity 

llic most important areas <il endemism 

The studv of en.lemism h\ resorting to asse, 

.•.volcanic Hell, Together these states 

Mages of species from widely different plant ■■mil 

ic largest portion ol the I >• -1 1 . and most 

niav he advantageous hccause ihe, inrj Mu, 

a in the three major groups are repre- 

le or the oilier states. Nevertheless, the 

entire flora to en\ factor- thai oper ile o\ 

of the numherol emlemie la\a per 100 

hroail geographical areas. The „|ui„ lb dki.l\anln 

■s that Dislnlo Federal (uilh a «lensit\ 

"f thi- approach i- that the la\a under imoti-alii 

07). \lorelos(o.2l). amlTlaxeala (.'}.:} 1. 

difler in -i/.-. evolutionary his|or\. and hi. .loi-ical . 

) aie. In lar. Ihe most important areas 

Inhules and. thus, in llieir res|)ouse lo selective pre 

. in the hell. \ more accurate measure 

sures. For these reasons, it could not he assiiin. 

ui and their luimhcrs per unit area. In this 
■ density iude\ calculated lor the endemic 

those of Asleraceae or Poaceae in the same r 
Kneh laxon. In \ iilue of a differing life cycle n 
logical |)references. does not operate under the 

The use of a similarity coefficient and cluster due to relatively rapid spread of taxa fol 

analysis has shown thai the Holistic relationships <-ialmn. Mosses are usualK considered 

of slates nun he eslahlished with certain degree of <'vnlvuig organisms, hul ihe elfeci of soi 

accuracy. The resolution of ihe analysis, however, lm,ls <> 

depends on ihe amount of field or herhariimi inlor- populai 

million and an adequate laxuiiomic hackgro.ind. !■''«' <vcle. there are theoretical considerations I 

The input of data from oilier major plant groups which al least some populations nun undergo rap 

should assist in refining the scheme of relationships evolutionary change. Kor instance, if polyploids a 

among such stales, particularly among those thai produced In diplospory or aposporv. or if a somiit 

show little similarity Willi llieir neighhors. In this mutation is retained in an otherwise haploid orga 

• tin- unusual position ol a stale may ism. with the aid of asexual rcprodiiclion these pr 

a. old, 

Iron, , 

t. Rapid speciation, 
I disperse gradually. 

Thus, endemic mossr-, should he comparative!) obvious pnl t i-rns. In fad, there are few endemic 
scarce in Mexico and elsewhere. This is supported species in die desert areas of Zacateeas (e.g., Cur- 
by current phytogeographic and geologic informal ion liramca me.xicana iTIid.) Crum and ,lajjuclihr\um 
suggesting that the moss Mora of Mexico has not arsrnri (Ther.) flier.: I >elgadilln i\ Cardenas, 1987), 
evolved in isolation. In addition to the examples giv- hut in the alpine areas where mosses are dominant 
en in the introduction, a recent stud) in the lowland 
areas of Chiapas (Delgadillo K Cardenas, 2002) re- 
ports a single endemic specie-.. Pylaisimii 
pii Crum. for the I aeandon fore-i and more lli.ui 
130 species shared with other continental areas. The 
broad geographical patterns exhibited l>\ Mexican representation. High UV radiation, daily tempera- 
moss species and the age and North-South onenta- tine fluctuation, low organic- nitrogen and phospho- 
tion of mountain ranges agree with the hypothesis of rus in the substrate ol alpine and subalpine areas 
rapid dispersal of newly evolved species in Mexico. seem strong seledivc loives lor all plants, including 
The exploration of poorh known areas is not ex- mosses, grasses, and composites, 
pected to produce a sharp increase in the number There are other differences that are evident in 
of endemic species, but rather the decrease in per- the present analysis. Degree of endemism varies 
centage endemism values as the distribution of de- among groups, and (here are disparities in their 
scribed species is better known or as modern taxo- altiludinal and latitudinal gradients and in the 
nomic evaluations result in svnonvmv. A few years types of vegetation they occupy. Contributions by 
ago, Delgadillo (1994) calculated nearly I V/t moss Delgadillo ( 1979. 1984) and Delgadillo and Zander 
endemism in Mexico: this contribution records onlv (1081) attest to the uneven distribution of endemic 

creased from 943 in 1994 to about 982 in 2(X)2. in the Yucatan Peninsula, and in the Tehuacan Val- 

Compared to mosses, grasses and composite: 
represent heterozygote systems where sexual repro 
duction, the length of the life cycle, and dispersal among geographical areas, even among those of 
retard evolutionary events. Assuming similar rales similar surface area. On a national scale, these pre- 
of speciation, but differences in dispersal abililv liminary findings may be the basis for the identi- 
and age of taxa. vascular plants wmld be expected ficalion of areas of high endemism and the selective 
to be geograpliicallv limited, genetically stable, and forces in operation. They will also assist in unveil- 
narrowly distributed, more so than mosses. Long- ing the history of the Mora in Mexico and its rela- 
lived moss species have been documented in the tionships to other floras in the American tropics, 
fossil record (Frahm, 2000: Miller. 1984), and The Neovolcanic Kelt, as an example, has been 
broad continental and intercontinental ranges arc shown to be one such region where portions of the 
common among mosses (cf. Sharp el al.. 1004). mountain range » I >i — I ■ i • < > federal. Jalisco, Mexico, 
This may not be the case in vascular plants where. and Michoacan) haw higher endemism concentra- 
in addition, selection does not immediately elimi- lions than the rest. However, a detailed floristic 
nate mutant genolv , ' I ■ pressions rec- knowledge ol less known in uiidin olleeled parts of 
ognized as endemic taxa may remain for a longtime the country, or even of adjacenl areas in other coun- 

negalive selective pressures. In Mexico, although eniial dist i ibnl ion ol endemics is not an artifact 

the number of moss and grass species are similar. derived from historical collecting preferences. 

ortion of endemic 

lifferenee may be ( i|( , ri|un , ( iu . (] 
sought imonii lh< I la li In it I I 

The present study illustrates how endemism values ^utum, J. II. & I. W. An-lrese,,. I'M. I lie vegetation, 

t , ■ , . . , • -i • ll<.n-i i« .in. ple.l »■ ■ ■ -; |.le. . I ill- ■ ii' i eim 

may not be equivalent between sumla. laxononnc ^^ ^.^ w Mi() |. Natura | isI 7 5: 1-33. 

categories, but din i i , i i Ii hllerences in |) ;IU( j sr . (;.. \1. Sousa <X \. Clial.-r (editors). IW4. IV 

the taxonomic hierarchy. aceae. Pp. 476-^84 in Flora Mcsoaniericaiia. Vol. 6. 

Local climate's certainly act as strong selective Alis.nataeeae a ( M ,n, „,;„■. I nivn -i.h.l Yici.mal \u- 

forces for every plant grou|). Mosses, grasses, and (■''J.n s'i Lmis''' 1 .! 1 | , ''T 
composites, however, show differential responses to 
climate. The distribution of the first group in the 
drier areas of Mexico does not apparent Iv follow 

.uis: ami The Natural History Museum. 

A. Lot & J. Fa (editors). Biological Diversity of Mo 

Origins and Distribution. Oxford I ni\. Press. New \ 

Rohlf, P. J. 1998. NTSYSpc. Numerical Taxonomy 

Multivariate Analysis System. Vers. 2.0. Exeter J 

Kiolmpi, a 2(r. 12-16. 
& A. Cardenas S. I ( ),'!7. Musgos de /acatecas. 

& R. H. Zander. 1984. The , 

Imacan Vallev. Mexico, and notes on 

Ursoiogisi ;::': hi 9-322. 

. B. BelloB. * \. Cardenas S. 1995. I.ATMOSS. 

a Catalogue of Neotropical Mosses. Monogr. Svsl. Bot. 
Missouri Bot. (iard. Vol. 56. 

enuM|uia V. I. I 093. („■,,!,, U x ,,! Mexico: A synopsis. 

l-aledilm.i. Biological Din,, -iK ol Mexico: Origins and 

1959 Palrones < 

lexicol \± :\<ri 
Pp. 97-135 in 

i- Yiluiale- lie- 

common to North and South \mcrican deserts. J. \r- 
nold Arbor. 21: 356-363. 
.1c\augh. B. 19B3. Oraniineae. /„ \\. H. Anderson (ed- 
ilor). flora \o\, ,-< ,al iciaua. \ Descriptive Account of 
the Vascular Plants of Western Mexico. 14: 1-436. 
Univ. Michigan Press, Ann Arbor. 

. 1975. An ecological and phy togengiaphical 

anaKsis ol the grasslands ol Mexico. Taxon 21: 67-450. 
. 1978. \egetacionde Mexico. Pedieion. p. \:\± 

f.ditorial LIMUSA, Mexico, D.F. 
. 1991a. Diwrsidad \ urigcncs de la lloia laiie- 

rogamica de Mexico. Ada Bot. Mex. 14: 3-21. 
. I991h. fl eudemismo i-n la Mora lam n.nannca 

mexicana: Una apreciacion analilica preliminar. Acta 

Bot. Mex. 15: 47-64. 
. 1993. Diversity and origins ol the ph. in. lo-amir 

Mora of Mexico. Pp. 129-1 14 in T. P. Bainamoorlhv. B. 

Bye. \. Lot ,x .1. fa (editors). Biological Diversity of 

Mexico. Origins and Distribution. Oxford Univ. Press. 

i \ernonieae. floi 
il del Bajio. Pat/.cu 

,l, i;. i<)«»5. 

ihm()s Smith. C. (;.. Jr. I%5. flora. Tehu; 

brother. .1. I . \'>'>'K < omposilac- I hliaullieae -.1. I 
232 /,, T. f. Darnel (editor), flora of Chiapas. I 

■Ivies. B. T. 1993. Oenus/W: \ \lexi< an Punic 
397-120 in T. P. Bama.noorthy. B. Bye \. Lot & 



: diversity and the trop- 


\,. .1. 1 . {')',;•< Manual para la idcul ifi.-acinu d 

las Compositae de la Peninsula de Yucatan v Tabascc 

!tor& 0. Telle/. 1998. 

Technical Report No. 4. Rancho Santa Ana Botanica 

ido de Nayarit (Mexico). 

Carden. Claremont. California. 

. 1993. La lamilia Asteraceae en \lr\ioi. Keust. 

nee. 1987. Speciea vi- 

Soc. Mex. Hist. Nat.. Vol. fsp. 44: 117-124. 

-. 0. Ibarra ,\ 1). 0, ana. 1998. Strategies lor 1 1 1. 

.dansonia 2: 167-175. 

conservation ol Asteraceae in Mexico. Conserv. Biol 




yuccas (Agavaceae) and yucca moths (Lepidoptera. I'rodoy i.lar). in win 

r progeny I 

■ classically < 

■ 1872, our limit-islanding ol 1 1 1< - ecology and evolution of lliis ass( 
ha- increased dramatically in the past decade. Here I rc\ic\v current information on organiamal diversity and 
genetic relationship-, ecological i olat ionshi|)s. origin and reversal ol the mutualism, and the |io|enlial lor an 

patterns of co-speciation and the historical role of coe\oluli n -pcilic trails in «lfi\ inii diversification in tin 

action. Majoi uo\e| de\elopmenls in recent voar- include the recognition ol a large species complej ol polli 
previously thought t 

■cular clock to phylogenetic data s 

other pollinator -pocic-. This appeal- to have happened nol llnnMgli -election for cheating. I. lit rather 
of a phonological -lull to an iinexplnilod -eed resource, in which case pollination behavior became redui 
ol' parallel di\ ei si Ileal ion and charade) eoe\ olnl ion are 1 1a m p<ivd hv incomplete phylogenetic informal io 
level, especially for the plants, but also for the pollinators, \\ailable data indicate considerable devia 
co-speeiation. and no evident examples of this process. \iialv-o- of the role of coev olulionarv proeesse 
diversification of yuccas and yucca moths will be possible once fully resolved phvlogenies become ava 
Key words: coevolulion. flesperoyucca. mutualism. I'aratc^eliiiihi. I'lodoxidae. Tegeticula, Yucca. \i 

(devolution, in the sense of reciprocally induced plant-feeding insects often have increased rates o 

evolution, is one o| I be major proeesse- driv ing di- div ei-ilieat ion ■ on i pared to sister groups with <lif 

versification and speciation (Farrell & Mitter. 1993; Cerent life habits: thus one or more life history as 

Thompson. 1991. 1999a. I>). Since first applied in pects of these groups appear to he important it 

plant-animal interactions as a hypothesis to explain driving diversification and speciation. This migh 

diversification among butterflies and flowering involve, for example, chemical, physiological. an< 

plants (Khrlich & Raven. 1961). it has been ap- morphological arms races between the interactin 

plied successfully in comparative analyses to test organisms. 

rates of diversification in ecologically I Hod Our understanding of (revolutionary processes a 

groups, such as plant-feeding insects and parasil- population il ami species levels is still in its infanc 

oids (Mitter el al.. I ( >i!«'»: \\ ieginann el al.. IWi; because identification of proximal factors of diver 

Becerra, 1997; Farrell, 1998; Becerra & Venable. sification relies on strong phylogenetic hypothese 

1999). Several such studies show that plants and for the intera< 

>ars. They include 
Groman, Beau Crabb. Mary Ann feist, Mark Bi 

Out/. James Ooldmovor. \nslev Grimes, and Kri 

Long provided helpful information on George K 

memory ol t,| ,bc Schnndl \ lelse,,. who revolul, 
- Department of Biological Sciences. Univi 

Ann. Missouri Hot. Card. 90: 35-55. 2003. 

R>9 I' I I! >•>;; l«i •■», I.).. vieusiye hi. hisinn 
<l;il;i. find 1 1 1 1 1 1 1 1 . 1 1 < - 1 \ . \| x-ii imiiial testing ol' emeig 
ing candidal. • trails ( Armbruster et al., 1997). Oflrn 

llic phvlogeuelie Irnineworks are missing, and there 

mel. A recent exception is I he study of liircrra 
(1997), who used data lor members of the plant 

'dies (Hlcpliarida; I llm sonn ■- 
chemical d. 'lenses and detox- 

Kn^'lm.mii uii lli. first ni«lit lliat lie ohsened n 
). jilanu-ntos,, ll.meis. I ah.-l likely written l.y C. 

One of the most often cited cases of convolution 
is the obligate mutualism between yuccas (Yuan 
and Hcspcro) uvea. \ga\aceae) and yucca moths 
(Tegeticula and I'anitegeticuta, Rrodoxnlae. I .< j.i 

adult moths for then [ •« ■ 1 1 1 r i i. while the moth Since the lalesl reviews ol this interaction, by Bak- 

larvae require developing seeds to complete their er (1986) and Powell (1992), information on sys- 

development. This association was first recognized lematics. plis In-, ti.lie relationships, and life his- 

over a century ago (Anonymous. 1872; Riley. lory has increased dramatically, especially in the 

1872), and then served not oidy as an example of moths, and the complexity of the association at dif- 

rciiiaikable |...lhn, niuliialiMii. hut also as one Cerent hierarchical levels is now <|iule different. 

ol ihe lirsl and si routes I examples ol evolulion bv 

means of natural selection. Together will, a few oth- Ka|{| y HlsioiO (». St. in Ol TIIK Rl.ANT-MoTII 

er models of obligate mutualism that involve seed- Interaction 

ealing pollinators, the yuccas and y ueea moths form 

a class ol associations lliat are excellent systems The first observation of the yucca moths was 

Cor studies of convolution, as well as of evolutionary made by Oeorge Kngelniami in St. Louis in 1872 

and ecological dy n i, i I i , isked Charles Ril- 

solution. This stems in part from the relative sim- ey. then stale , ■iilonioloiusl <>l Missouri, to explore 

plicity of measuring fitness costs and benefits in the relationship between the moths and the plants. 

these interactions; in most instances both plant cost Raker I 1 980) pn.v idol a pas-age from Kngelmamfs 

and bennfit nan bn measured in seeds. Second, in notes written on 13 June 1872 about the initial ob- 

wilh a single pollinator species per plant species. pears to have survived. Riley donated his very hug. 

making it easier to measure reciprocal effects than insect collection to the United States National Mu 

when webs of many, simultaneously interacting taxa scum of Natural Hislorv (Smithsonian Institution) 

have to be analyzed. where it became the nucleus lor the creation of tin 

Considerable progress has been made in under- Department ol Knlomology. Among his yucca motl 

standing this unusual type ol obligate pollination specimens is one female Tc^clirulu yiicrasclla (Ril 

mutualisms in the past 15 years. This is certainly ey) specimen labeled "Cound in Yucca flower — Kn 

true of the long-recognized yueea-vueea moth and gelm. June I 2/72"" ( Fig. 1 ). This dale coincides \y ill 

fig-fig wasp associations (Riley. 1872; Weiblen. that ol Kngelma.ins original observations al tli< 

2(K)2). The two other documenle.l examples of such Missouri Rolanical Car.len. and is obviously a motl 

obligate associations (Trollius nimpiicus L | Ran given to Riley by Kiigelmanii. Although this wouh 

uiiculaceae| and (.!; • Ihplera: \iilho- have been an obvious candidate for holotype, it n 

myidaej, and L>phoc<'reus schollii (Kngelm.) Iliilloii not. Riley, a driven and opinionated worker, nnvni 

& Rose |Caclaceac| and the moth lipiga rirrsn-ns bothered to designate or label type material for any 

(Hulst) [Uepidoplera: Rv ralidae|) were actually lirsl of the many species that he described, but insteac 

documented during this |)eriod (Pcllmyr. 1989. would mention in his descriptions the numbnr ol 

1992; Fleming K Holland. |9'>8: I & Jaeger, specimens used for the description (Davis, 1967) 

1999; Jaeger el al., 2001). Here I will review our A leetolype having already been designated lor 7. 

current understanding ol' the association between yiircasclla. the surviving moth from Kiigelmanii 

original observations lias now been labeled l«» in- 1875: :i 10-3 12). « >lhcs < ha.ged not only that Riley 

(Urate its historical significance, was ineorreel l>nl 1 1 1 « - very |)henomenon of in- 

Cliarles Kilev was In .luminal.- llie Held ol vucea seel pol linalion was a dubious notion in the first 

moth studies up until his sudden death in 18%. plaee (Boll, 1876; Meehan, 1876); Boll went on to 

despite this being a sideline in his job as the first state that aetive pollination "belongs in the land of 

federal entomologist (Sorensen, 1995). One of his fables." Yet other erities eliallenged that his argu- 

most important contributions was Ins involvement merits about exchisivilv <>l mollis in pollinating yuc- 

in the salvation of the French wine industrv (Smith. cas were overstated (e.g.. Ilulsl, 1886). Riley re- 

1992). I mention it here because it indicates Riley's sponded to his critics with experimental results, not 

general understanding of the process of 'plant-insect always published in lull, ollen with a singularly 

eoevolution. Bv the early 1870s. North American sharp pen (Davis. 1007). A prolific writer, with 

grape phylloxera aphids" (l)aklulosphaira rilifoliar some 2 KM) entries in his bibliography (Ho & Yuille, 

(Fitch)) accidentally introduced in central Kurope 1999). Kilev used the empirical data as he knew 

caused massive mortalitv ol Kuiopean grape culti- them to rebut and often scold Ins critics (e.g., Riley, 

vars by attacking their roots. Kilev (1871) reasoned 1877. 1881, 1887), and occasionally even stooped 

phylloxera, and thus might tolerate them better. A Following Riley's death, a hiatus arose in the em- 
grafting program with Furopcan cullivars and pirical study of the association. Trelease worked 
American roots proved highly successful in reduc- with Kilev on behavioral and botanical aspects, 
ing phylloxera impact, and carried the industrv to performing extensive fieldwork, and published de- 
financial survival: Charles Valentine Kilev may be tailed observations on pollinator behavior as well 
the only individual to have received the French be- as plant morphology and systematics in his works 
gion of Honor for contributions to eoevolution. (Trelease, 1893. 1902). Considerable collections of 

An extraordinary observer and able thinker. Rib both moths and plants were made by Susan Mc- 
ev unfolded the hash natural historv ol the plant- Kebev lor her monographs on southwestern Yucca 
moth mutualism and documented the life histories (McKelvey, 1938, 1947). Kusck (1947) attempted a 
of the pollinator Te^elicula \ih casd/n and lire bo- reassessment of inolh-planl associations based on 
gus yucca moth Prodoxus decipiens Kilev within a McKelvey 's insect material; his conclusions when 
decade of the initial discovery (Kilev. 1880. 1881). correct generally followed those of Riley, but Busck 
In contrast to the records of most of his eonlcm- misinterpreted morphological variation that he was 
poraries. there are yen few inaccuracies in his ae- the first to document among pollinator yucca moths, 
counts, simply because of his reliance on empirical cheater yucca moths, and bogus yucca moths, 
observation, hi this, he arguably belonged in the Since the 1960s, information about the assoeia- 
exclusive group of exceptional naturalists with lion has accrued at an accelerating pace from sev- 
whom he regularly corresponded, such as Charles eral lines of investigation. This includes systematic 
Darwin, Alfred Russell Wallace. Henry Walter and phylogenelic studies of the organisms, as well 
Bates, Thomas Belt. Fritz and Hermann M tiller, and as the ecological and evolutionary studies of the 
Asa Gray. As one of the earlv protagonists of evo- interactions between the moths and the plants, 
lution by natural selection in the I nited States. Ril- 
ey went beyond natural historv to use the relation- Natural History 
ship between the yuccas and the moths in oR(;ANISMAL DIVERSITY 
discussing more general issues such as mimicry 
and aninw nation (Kiley. 1871, 1892). T ™ ?™ ' ,,( ' I™'***™ P"" <»' '"«' N( > ' 

The relationship between yuccas and yucca and Central American famdy Agavaccae < I ,,. 2). 

Recent data suggest that the sister group ol \ga 
vaceae may be the small family Camassiaeeae. eon- 
fined primarily to mesic habitats of western North 
America with the exception being one species in 
eastern North America (Pfosser & Speta, 1999). 

11 letter from Charles 
Darwin as "the most remarkable example of fertil- 
isation ever published" (Burkhardt & Smith, 1994), 

first 15 years alter the discover). Kilev was chal- 
lenged on numerous occasions regarding the ac- 
curacy of his observations. This included the ar- l V. T. Chambers, an amateui 
gument from P. C. Zeller. a German entomological 

authority whose experience with yucca moths was j^^^ i'^). In a n^'uttal. Ililey" 
limited to three pinned specimens given to him. | u -.i<>n ami used Chambers's moth 
that it was simply too improbable to be true (Zeller, Prodoxus (Gr., "judging of a thing 

used the first non |>. Ill in; t>.yu- \ ucra moth to chal- 
lenge Uilevs description of pollinator yucca mollis (Cham- 

< americ-leyel relationship- are partly iiuresolv c<l 

involved in the obligate mutualism willi yucca 
moths (Bogler & Simpson, 1996; Clary, 1997). The 

ii i I i i i I It x, rilicd a- a 

distinct subgenus based mi such features as a cap- 
itate stigma (Kiigelmaiin. 1871). was long consid- 
ered a section within Yucca (Baker, 1986). Recent 
analyses show llcspem\ net ti to he the sister group 
til Hcsperaloc. a small genu- ..I the Sonoran and 
Chihuahuan deserts (Bogler & Simpson, 1996). Im- 
portantly. Ilcsperaloc taxa are not associated with 

the yucca moth- nil i » i 'n n i ! 1 1, 

(Pellmyr & Augenslein. 1997) and hats 
(Engard, 1980) lor their pollination. Jointly. Hes- 
pentvueca and Hespcraloc constitute the sister group 
of all remaining yuccas (Bogler & Simpson. 1990). 
Yucca is divided into three sections: spongy - 
fruiled section Clistocarpa. the fleshy-fruited sec- 
lion Sarcncarpa. and the capsular-fruited section 
Chaenocarpa. Section Clistocarpa consist- solely ol 
Yucca hrerij'olia I aigehn.. whereas the two other 
sections consist of no more than 20 to 25 species 
each (Clary. 1007). Section Clistocarpa is charac- 
terized h\ the -ingle aiilapomorphy ol a thickened 
exocarp, as observed by 'I release "(1893). Its posi- 
tion relative to the other yuccas is uncertain Iml 
possibly tied lo the -cries Ihipicolar of capsular- 
fruited species (Clary. 1997). The longstanding in- 
terest in yucca- ami their importance in many bi- 
ological communities notwithstanding. )u<ru 
taxonomy ami syslemalics remain in a state of flux. 
yyith much need tor a modern revision, h'ey i-ionai y 
work i- complicated by the rtdaliyc -can -ily ol her- 
barium material, caused In part by the logistic 
problems of preparing specimens from these large. 
succulent plants. Horticultural interests in the 

group ilso In . I I Intra ol name-. 

with many taxa narrowly delineated using in effect 
a typological species concept (sensu Mayr, 1963). 
Observed variation frequently has been attributed 
to assumed hybridization and introgression (e.g.. 
McKelvey, 1938, 1947; Webber, 1953), but this 

should be t i.n- . I ill i a- there is only 

one example where genetic e\ idence lor introgres- 
sion between two \ u< v.\ -p< . ies is pn id i (I la:i 
-on. | < >< >l2 > . I'h\ logenelit analyses are limited thus 
far. but appear not to \ iolale a — uniplion- ol mono- 
phyly of both section Sarcocarpa and section 
Chaenocarpa (Clary, 1997). The use of horticultural 
material or yucca cultivars of unknown origin in 
some studies may contribute to historical confusion 

The Hesperoyucca-Hesperaloe-Yucca clade i- na- 
liye to \orlh \merica I fig. 3). and ils contiguous 

n-ni; ■■ ha- rn < \ i '1 i ii" " • itral America and 
northern South America through the cultiyation ol 

» . , <;/ - ■ log. I lol then edible Mower- (Tre- 

lease. 1902; Matuda & Pina l.ujan. 1980). Several 

nenls. including in Europe since the late 1500s 
(C.-ranle. 1033), but yucca moths have never been 
I. ill i id either soul h of Mexico or on oilier conl incuts. 
Hiley (1881) allcmpled |., establish them by send- 
ing halche- of polliuatoi larvae In their cocoons to 
Darwin and Stainton in England, Planchon in 
fiance. 11. Midler in Germany, and Asa (hay in 
Massacluisells. for eslahlishnicnt on cultivated yuc- 
cas. Miiller (1871) reported that moths hatched, but 
no local ornamental plant- were in llowci. Darwin 

coons lo Joseph Hooker at Kew. where their sub- 
sequent fate is unknown. 

The two larger Yucca sections, section Sarcocar- 
pa and section Chaenocarpa, have wide range- that 
overlap in area- north and -outli ol the border of 
Mexico and the United Stales (Fig. 3). The fleshy- 
Iriiiled set lion Sarcocarpa i- primarily southern, 
ranging throughout the Megamexieo-1 biogeograph- 
ic region ol B/e.l.twski MOO;',), ami extending in 
one species northward to southern Colorado, flic 
aberrant Y. aloifolia L. occurs in the northern Ca- 
ribbean and along the U.S. Mexican Gulf and 
southern Atlantic coasts: it reproduces yegetatively 
but is not known to have a native pollinator. Il- 

capsular-fruited yuccas are more northern in dis- 
tribution, ranging from the northern edge of the 
(heal Plains in southern Canada -oiilliward lo the 

latcl • 

Whereas yuccas generally , 
shriih desert, chaparral, or grasslands, many Mex- 
ican species often grow in pine-oak woodland (Ma- 
tuda & Pina Lujan, 1980; Gentry, 1982). Packrat 
midden data from the Wisconsin glacial show that 
species such as Y. rostrata Engelm. ex Trel. that 
currently inhabit shrub desert grew in pine-oak 
woodland- in area- such as the Big Bend region of 
Texas during wetter periods (Van Devender, 1990). 
The most unusual habitats are those of the south- 
ernniosl y in < i- 

G. Pompa & Valdes. Both occur in rainforest, with 
the lonner haying a Icrrc-irial habit whereas the 
latter is epiphytic or epilithic (Matuda & Pina Lu- 
jan, 1980; C. Beutelspacher. pens, coram.). 

The yucca moths. The yucca moths belong to 

the I'rodoxidue. a basal family within l.epidopiera 
of 78 described species (Davis, 1998; Pellmyr, 
2002) and at least 15 additional uiidescribed spe- 

i. IH7 t Muhi 1001: In I, l<)()2 I iU,s|, a conservative n 

s give actual .sites). For Y. aloifolia (black squares), the same sow 
for records from the \ulilles jTrelease. |<>()2). specific locations a 
Kind areas south of Mexico have been excluded. 

ties (Fraek, 1982: Nielsen. 1982; Pellmyr & Bal- lecular clala (Brown el al.. 190k Pellmyr & Fee- 
eazar-Fara, in prep.). The sister Family Cecidosidae hens-Mack. 1999) lo-«-tli.-i surest that the mono- 
consists of gall-makers feeding mostly on Anaear- basic Prodoxoiilcs. the nnl\ soulhern hemisphere 
(liaceae (Nielsen. I')?'.:")), and il shows a typical prodoxid moth, is ihe basal genus in ihe family 
Gondwanan dislribiition. The presence of sister (Fig. 4). Creya is a diverse genus of boreal and 
genera in Africa and South America of these moths. temperate humid to semiarid areas of western North 
which are highh sedetilarv. -IioiilJs indicales an \merica (Da\ is et al.. 1002). with the exception of 
origin of I Ins lamiK. and h\ inference tin- Prodox- a few basal members reeeiilb doeiimenled from 
idae, before the South Atlantic breakup 95-KM) easternmost Asia (Kozlov. 1006). Telmgnm is coti- 
million years ago (PellmynN, Fecbens-Mack. 1000). fined to North America, whereas the large genus 
Morphological (Nielsen i\ Davis, 1085) and mo- himpnmia is holarclic in distribution. These gen- 

ion. Estimated minimum 
a molecular clock, calibrated based on bio^eojjiapliic dala from the sis 
dates from IVllimrand I.eebens-Maek (IOW). Numbers in parentheses . 


Morphological data, v 

era use a remarkable variety of host plants, includ- sues other than the seeds. They are not involved in 
ing species of the Myrtaceae, Apiaeeae, Hosaeeae. |)ollination. Virtually all yuccas host Prodoxus spe- 
(irossulariaccac. and Sa\ilra^accae. i.e.. represen- cies that f ■ ■ 1 1 uisid. lli< inflorescence scape, and 
talives from four plant orders (A PC 1998). Incases most fleshy-fruited and spongy-fruiled yuccas also 
where immature stages are known, the hu\a leeds host species thai Iced inside hardening -alls in the 
inside plant tissue dining early inslars. and then exo- or mesoearp portion of the fruit. The recently 
from the outside while concealed inside folded described Prodoxus phylloryctm Wagner & Powell 
leaves or cases during the final instars of (level- is so far unique within the genus in feeding as a 
communal gall-maker in lleshy yucca leaves (Wag- 
observed ner & Powell, 1988). In addition, the peduncles of 
i of arid at least six Agave species are used (Frack, 1982) 
habit to by some Prodoxus species. I will not deal with them 
having larvae that W-vA inside host tissue until feed- further hen-, as ihey are not directly involved in 
ing is complete (Davis. 1907: r'rack. 1982). Mese- the pollination mutualism. 

piola feed on members of Nolinaceae. whereas die The pollinating yucca moths belong in the genera 

three yucca moth l < << i Pmoli gcln ula, with 

and Tt'grlirula bed on members of Agnsaceac. Pro four described species (Pellmyr cv Halcazar-Lara, 

doxus (the "bogus yucca moths" of Riley (1880)) 2000), is unique in having lost the linear cutting 

coexist with the two other genera, but feed on tis- ovipositor of prodoxid moths used for inserting eggs 

where eggs are laid (Davis, 19<>7: Powell. 1WI). In 
species with known biology, thev ;iIm> differ in 
the larva bores into the young f i nil. where il causes 
(lie formation of a gall-like structure ("cyst" of Pow- 
ell. 1984) formed from modified placental tissue- 
arid a few immature seeds that in effect fuse and 
are consumed from within. Tegeticula was until re- 
cently held to consist ,.| llnce -pecies (T. maculata 
(Riley). T. synthetica (Riley), and T. yuccasella) with 
broadly similar lile histories (Raker. 1986). Mor- 
phological variation had long been reported within 
'/.' \uccasella lull considered as ml i aspecidc varia- 
tion (Busek. 1917; Davis. 1967); Davis (1967: 53) 

staled dial more thai <■ "biological entity" may 

exist, but refrained from delimitation on the 
grounds of in>ullicienl inlormatiou. Mile- (F>8.!| 
used iiioi-phoiuelric dala to demonstrate the pres- 
ence of at least three unnamed host-spccilic euli- 
ties. Further studio using mor|)liological and mo- 
lecular tools have so far led to the description of 
13 species (Pellmyr. 1999). and several additional 
taxa remain to be described (Pellmyr & Balcazar- 
l.ara. in [irep.). Tegeticula maculata is morpholog- 
ically and moleciilarlv highly di\ergenl and may 
well consist of several biological specie- I Ton. II t \ 
Mackie, 1966; Segraves & Pellmyr. 2001). and T. 
synthetica as currently cir< uniscribed contains two 
species (Pellmyr. in prep.). All species consume 

timing and location. Pollinators oviposit at the time 
ol llowering. ImiI Tey< ■ticula species, sometimes re- 
ferred to as "cheater yucca mollis." delay oviposi- 
tion to the fruit stage and have independently lost 
the behavioral and morphological trails of active 
pollination (Pellmyr et al.. 1996a; Pellmyr & 
Krenn. 2002). Intrageneric phv Intend ie inlorma- 
tion for I'aralegelicula and Tegeticula is relatively 
well established (Pellmyr & Feebens-Maek. 2000). 
with the major remaining uncertainties revolving 
around a rapid burst of radiation creating most lin- 
eages within the 7. \uccasella complex and the in- 
clusion of remaining undescribed species primarily 
from the southern portion of the range. A note of 

moths of the T. yuccasella complex is indit 
because of the historical lumping, many st 
must be interpreted very cautiously and are s 
times of little value, as studied species an 
identifiable and because as many as three cm 
ing species may have been treated as one. 

'-pollinator int 


and the moths, and here I only outline major sh 
elements. The female yucca moth of both pollii 
genera is equipped with unique tentacular mc 
parts that she uses for pollen handling (Kiley. 1 
Davis, 1967; Fig. 5A). She collects pollen 
yucca flowers by dragging her tentacles across 
anthers. The pollen is embedded in copious p< 

using the tentacles, and then stores it as a bal 
underneath her head (Fig. 5A. B). The pollen 
kept in place by adhesion alone, and the tentac 
play no part in holding it in place. This load c 
be substantial, reaching nearly 10.000 grains in 
geticulu yuccasella females, and constituting nea 
10 percent of the moths body weight (Pelhr 
1997). Pollen collection can recur on an occasioi 
basis during the active life of the female, so I 
pollen load may consist of multiple pollen gei 
types. Following pollen collection, the female set 

, be ac- 

may also be subject to ov iposition (Riley. 188')). In 
Tegeticula, the female first walks around the ovary, 
and her decision whether to oviposit is influenced 
not only by the flower itself but at least in some 

Tyre. 1995; Hull, K Pellmyr. 1999). In T. yucca- 
sella, females deposit a host-marking pheromone 
during oviposition, and subsequent visitors perform 
a crude estimation of pheromone quantity (llulh M 
Pellmyr, 1999). Visitors become increasingly un- 
likely to accept a llovverwith increasing number of 
prior visits. In one case of two cnexisliug poilina 

whereas the other made ovipositii 

lions herself in a species-specific location on the 
ovary and cuts into il (Fig. 5C). Mosl species pen- 
elrale the ovary wall and lav eggs inside the locale. 
bul a few species oviposit very superficially under 
ihe epidermis. The female then n-es the tips of her 
tentacles to sera pi' oil a small amount o| pollen 
lioin hei I ..ill h. walks up |,, til.- stigma, and places 
ihe pollen on the papillose internal surfaces of die 
style using a series of 10-20 bobbing 
s (Fig. 5C, E). The only exception in this 
regard is /.' maculata. which pollinates the capitate 
stigma ol llcsperoyucca uhipplei 'loir, using the 
same scraping behav : "- — : - "-<"' ''—' ""II-— -•"' 


r as is used for pollen 

Yuccas and Yucca Moths 

,. —A. Head i.f V'frrlinthi , «mn, 
[I proboscis indicated by black and v 
; pollen just collected from a Y. fil 
into (right) a V. ovarv. - 

below ihc head. 

Carriere flower. Moth win 
length 73 mm. For a set c 

nols I! I 10 I 

lection. A female may repeat oviposilion and pol- 
lination main times on a Mower, especially if -die 
started on a virgin (lower. In T. yucc(is<-'i,i. \». i- 
nation almost invariably happens following llie lir-l 
oviposilion on a (lower, but females then become 
increasingly likeK to skip pollination during sub- 
sequent oviposilion bouts, and they also deposit 
less pollen per pollination event (Hulli iS I VI I my r. 
1999). Females of I he species that encounter a 
flower visited by one oilier female first typically 
perform about half as many ovipositions and pol- 
linations as the first female (Huth & Pellmyr, 
1999). and a smaller yet significant reduction was 
observed in T. altiplanella Pellmyr (Addicott & 
Tyre, 1995, referred to as "deeps"). Once a female 
moves on, she usually walk- to adjacent flowers and 
inspects them for suitability, (hen visits other side 
branches, and eventually she Hies off to other in- 
florescences. Consequently, females perform both 
geitonogamous and xenogamous pollinations (Uilev. 
1H92: Fuller. 1990: Dodd & Finhart, 1994; Pellmyr 
et al., 1997; Marr el al.. 2000); there is no exper- 
imental evidence of plant self-incompatibility and 

fruit set readily occurs following both type> of pol- 
lination, but selfed fruits are highly susceptible to 

outcrossed fruits (Pellmyr et al., 1997; Richter & 
Weis, 1998; Huth & Pellmyr. 2000). 

Fggs of Tegeticula hatch within a few days, and 
larvae of species thai lay eggs inside the locule start 
feeding on seeds immediately. In species thai ovi- 
posit superb, lallv. llie larva lust burrows in the 
ovary wall before entering the locule to feed on 
seeds (Wilson & Addicott. 1998; Pellmyr & Fee- 
bens-Mack, 2000). Larvae consume a variable 
number of seeds (Fig. 5F), depending on the spe- 
cies and factors such as the presence of abortive 
seeds that can reduce per capita consumption 
(Powell, 1984; Ziv t \ Bronsteiu. 1 996; Bronstein & 
Ziv, 1997). Upon completion of feeding, the larva 
creates an exit path. It preferentially exits during 
rain, either night or day (Whitten, 1894), but per- 
haps more commonly at night (Groman & Pellmyr. 
unpublished data), and can spend extended time 
waiting inside the fruit for optimal conditions ||',,\\- 
ell K \lackio. |0M... The larva burrows into the 

civil Willi soil mi s;iiii) particles. Tin- exact I . .. : i > 

in the ground has never been reported, hut from 
lab trials Riley (187.'*) reported depths of 7.5-10 
em and Kan (1945) 2.5-7.5 em for T. yuceasella 
and perhaps also /.' intermedia IVllmyr. Powell 
(1984) reported depths of 1-3 em in shallow con- 
tainers for T. maderae IVllmyr. The larvae of five 
Tegeticula specie- i /.' \m (usclla. T. intermedia. T. 
cassandra Pelhnvr. /.' Ircculeanella IVllmyr. '/.' air 
nerosaneila IVllmyrl reared in my lah commonly 
created their cocoons at a depth of 20 cm where 

thev reached llic n i . h.ilil in nl die renins 

canisters. The var 
Tegeticula species 

The larva enters diapause inside the cocoon and 

pupates a lew week- lie fore emergence. This may 

happen alter , c-veai diapause, hut the larvae 

can remain in diapause in lah conditions lor at lca-1 
four years (Riley, 1892). Very high fruit set during 
mass flowering episodes in \ ■., I ■ .■ 

then effectively cease llowering almost completely 
for several years (IVllmyr, unpublished data) mi,- 
gests that tin- modi larvae are capable of diapausing 
for several years in the field as well, and dial there 
are unidenlificd cue- dial trigger coinplclioii of d.- 
\elopun nl and adult moth emergence. This is not 
lo suggest lhal moth emergence is perfectly -\n 
elironizcd with host llowering we know it is not 
(Frack, 1982)— hut rather that a sufficient number 
has remained in diapause to emerge at the time of 
mass flowering lo cause high levels of pollination. 
The life history of Parategeticula is known in 
less detail than thai ol Tegeticula. but o\ iposilion 
and larval biology ol one species. /! pall, I •. 

vis, has been described in detail bv Davis (1907) 
and especially Powell (1984). The most obvious dif- 
ference is that Parategeticula females oviposit on 
pedicel- and in pel. a Is, rather than into the ovary. 
In this ease, the larva chews ils wav into the ovary. 
and then proceeds to h'ed on partly modified seeds 
as described above, larvae of P. pollcnifera pupat- 
ed at 1-3 em in shallow containers (Powell, 1981). 
and P. elephant ipel la IVllmyr & Italenzai Lata 
formed their cocoons at 2-4 cm depth in 15 cm of 
loose soil (IVllmyr K Ralcazar-hara, 2000). Para- 
tegeticula pollcnifera from southern \rizona invan 
ably required Iwo years to complete development 
(Powell, 1984). whereas the tropical P. elephanli- 
pella emerged in the lab without a diapause (IVll- 
myr & Balcazar-Lara, 2000). 


af host spccijicit\. In the traditional 

relatively high levels of host specificity (Khrlieh o* 
Raven. 1901; Price. 1980; Farrell & Milter. 1993; 
Thompson. 199 I), especially when the phenologieal 

very narrow, \ucca moths, which only live for a lew 
days (Kingsober. 1981; Powell. 1981). must access 
the plant during the short llowering period, so moth 
populations would have lo be locally adapted for 
the flowering periods of different hosts, for exam- 

Chihualuian desert, lour yucca species coexist and 
have largely non-overlapping flowering periods 
spread out from February to early June. II a single 

lliis would rcipiire intraspecific poly morphisni in 
emergence phenology with lour distinct peaks in 
the moths. Busek (1947) and Davis (1967) specu- 
lated thai T. Mierasella may be a complex, hut suf- 
fered from a dearth ol material available for study. 
The first solid data supporting the hv poll esized 
complex were provided by Miles (1983). who 
showed lhal the pollinators of three sympalric yuc- 
cas in southern New Mexico differed greatly in 
morphology. She described the enlilies hut did not 
formally name them. Addieott (1996) likewise pro- 
vided morphometric data suggesting the existence 
of several more hosl-specilii species, and Pelhnvr 
e| al. ( |99();i) provided moh-cular phylngenelic data 
indicating the presence of a large complex. Thir- 
teen species, including eleven pollinator species. 
have since been described (IVllmyr. 1999). 

Given the revised moth species delineation, diet 
breadth among the pollinators is now more uniform 
I 1'ig. I)). I -lllg the yucca specie- delineations Used 
in IVllmyr (1999).' members of the T yuceasella 
complex have been recorded from 17 host species. 
Seven of the eleven pollinator species within the 
complex are iiionophagous. one has Iwo hosts, two 
have three hosts, and one has six recorded ho-l-. 
Thus more than 707. of all pollinator laxa are mo- 
nophagous. and the most ohgophagous species uses 
six host species. The reason for this level of spec- 
ificity remains to be explored, but certainly involves 
phenologieal specialization on hosts and probably 
also selection lor specialization on plant- with crit- 
ical differences in ovary morphology. Interestingly, 
the two derived non-pollinating yucca moth species 
are known lo use four and six hosts, respectiv eh. 
giving 1 1 1 < in a -iginlieanlb broader host range than 

Wallis lesl. x 5-M. P <>•<>•")• Proximal rea- 

Volume 90, Number ' 


□ „.,» 


■| pollinators 2000 


g cheaters 2000 





however, no fruil set occurred. I lulsl ( 18! !<» wa !,, 

tion v-\ I i < - 1 ■ stating that honev bees i \p 
L.) can he copollinators. Frustrated with such un- 
tested hvpolhcscs presumed lo l)f tine. Riley used 
a range of experimental and observational ap- 
proaches lo gather data to lest them. Whereas he 
presented Ins conclusions in print (Idles. 1887. 
1889, 1892), the original data were never pub- 
lished. Tabulated result sheets found in the ar- 
chives of the Missouri linlanical (iardeii showed 
5 I » 11 _ 13 15 1/ that inclusion experiments using Iwo of the most common llower visitors. 1. mcllifera (25 bees, 72 

umber of recorded hosts of described 7c- | lr _) JIM( | ( | lr soldier beetle Cliaiilingnalhiis pet,s\l 
m.legetinda species. The two open bars ;wmm (|( . (] ^ r (Canthari(lae ) (36 bee tles, 24 hr.), 
ie lour -peeies recognized lielorc IW. . , , . , 

number of hosts for described pollinator IM s «'l» i "' ill< ' K«» z «" » a £ s coiilai M i ng; single )uccafi- 

lamenlosa L inflorescences failed t 

of 2(KK». Cheater speeiev ha\e siiiinlieanlb ruon- host- per fruit development, whereas control inclusion ex- 

specics than pollinator species (kruskal-Wallis test, X" = periments with yucca moths caused fruit produc- 

5.68, p = 0.017). tion Ri|ey ()889 ^ lg92) and Trelease (i 8 93) f ur . 

ther argued against copollinators on the basis of 

extensive visitor behavior observations. For exam- 

i for the wider diet of non-pollir 

,,..-) , 1 1 

explored Because the non-polhnator larvae feed ^^ ^ ^ ^ ^^ ^ 

seeds side In s,de with pollinators, diet ,s an ^ .^ ^ ^^ { ^ ^ ^ 

hkelv exphina.ion. In.l hvpo,hes,s „, ^ ^ ^ _ ^.^ ^ ^ ^^ 

.include a broader phenologieal window ^^ ^^ ^^ ^ ^ ^^ R ^ (J 

s that oviposit into , modes thai ^^ ^ ,., .j,^,,.,, |||;|| |)|; , M|s m areas wi| 

■ less likely to select for specialization, or higher 
cnlial for establishment on novel hosts through 

, for example where plants receutlv had I 

m/ ''""" 7 " . ; ' fruit, even though a wide array of other insects were 

tl< " 1 VV " 11 "^ l "" '"' ,""" '', 'V ^ lr, ' l ;" S found on the 11,,,,, ll, also noted that in areas 

Mack, 2000; Marr e, al. 2001). Alternatively, ., ^ (< , ^ ^.^ ^ 

may reflect species age; because the non-polhna- ^ ^^ ,„„„,_,„,,,,,,„,.,, s|| „ |( . s with different 

tors are among the vou,,,e, -,,-s „,.,.■ compl... ,,_ ^ |ilii||o|i ^_ ,-„,;, s< , was neyer observe(Ji 

they have simply had less tune available lor pole..- ^^ ^.^ ||()W( ; ri||g nmi( , ( | ( , l( , 1 of individual 

tial diversification through host specialization. ^^ wjih a ^^ yucca hmJ been knQwn U) resu ,, 

The role of copollinators of yuccas. Suggestions in fruit set. 
of pollinators other than yucca moths appeared Speculation about eopollinalors was raised anew 
shortly after the original description of the plant- by Oodd and Linhart (1994). A lauxaniid fly {Pseu- 
moth interaction, and this arguiiieiil ha- n--uil;i<-ed ...!-,:..■ nee on Yucca glauca 
in the last decade, leading one monographer to the flowers, with some individuals found to carry mod- 
unfounded conclusion that "when moth populations est quantities of pollen on their bodies, was sug- 
are low, the fly Pseudocalliope may be an important gested as a possible vector. No attempt was made 
alternate pollinator |of yuccas)" "(Verhoek. 1998). to test experimentally whether the flies cause pol- 
For this reason, it is worth revisiting the support lination. There is reason for skepticism, because, 
for this untested hypothesis. as Riley (1892) first pointed out, lack of fruit set is 

I will discuss first all species other lhati Yucca common in vueea populations for a variety of rea- 

aloifolia, which is a special case in this regard. The sons, yet llower visitors other than moths are often 

earliest claim of copollinators was made by Median common in those same imputations. Further, even 

(1879). who presented fruits resulting from geiton- if occasional modest pollen transfer were to take 

ogamous hand-polliiuhon on a cultivated ). glauca place through visitors other than the moths, it 

Nutt. as evidence that other animals could serve as would likely be of little or no ecological and evo- 

pollinators. In the absence of hand-pollination, I utionary significance. This follows because flowers 

& Weis. 1008; Hull, c\ Pellmyr. 20(10); thus a poor 
vector is expected to contribute niiiiimallv to pi. ml 
fitness. The ceiilurv -nlil hvpnlhe-is about existence 
of collimators could readily be settled by the 
proper experiments. A sini|)le experimental ap- 
proach could exploit the size differences between 
yucca mollis and piopos.d , opollinalors bv using 
selective screens that permit enlrv to smaller visi- 
tors (such as the lauxaniid llv ) but exclude the larg- 
er Te^eticula mollis; llus approach worked well in 

m<] bumble 


vely screened on Trollim 
curopueiis ( I '. 1 1 rnx t. I').'!''). Civen high levels 
ol 'genetic diversity in yuccas (Feisl. |Oo:>; Musm-v 
& Hamrick. 1008). routine genetic analyses of am 
resulting serd progenies could al-o provide infor- 
mation about selfing and outcrossing rales. 

)ll<T,l uloijulia |- |||, ,111-1. ■ exception lo the lack 
of evidence for pollinators other than the mollis. 
Introduced as a garden plant in Kurope no later 
than 1596, in Australia by 1883. and in Melanesia 
bv 1880. il has been reported on several occasion, 
to set fruit in locations outside North America even 
though there never have been any coincident moth 
reports. For example. Fngehnami I b".7:i) saw Iriul- 
ing plants in Italy, l.avard (1880) in gardens of .New 
Caledonia on what undoubted!) was >. aloifoliu 
(MacKee, 1994). Riley (1891) conveyed a report 
from Australia, and Calil (1900) reported fruit set 
m a cultivated plain in l-raol. The plant historii allv 
was scattered along the southeastern North Amer- 
ican Atlantic and Cull coast, especially along sundv 

isiana(H». 3). Occasionally planls sel fruit in parts 
of thai range. Ivpicallv as a result of colonization 
liv Te^-lieulu \m-casellu anil '/.' russutulra from co- 
existing ami siniultaneoiislv (lowering Y. (ilamcn- 
tosa (Kngehnann, 1873; Riley. 1873; Pellmyr, 
1999). but fruiting plants without oviposilion scars 
or larval damage have also been reported ih'ilev. 
IR02; (.Ionian. I''')''!. Rilev h> pothesized from flo- 
ral structure il mav have resulted from autog- 
amy, but Trelease (1803) found that he could pre- 
vent fruit sel by excluding all floral visitors with a 
gauze bag in a plant thai previously had produced 
fruit. This is the onlv reported experiment for am 
yucca that provides even moderate support for other 
visitors as copollinalors. Il needs to be replicated 

The situation is more complex as Yucca aloifoliu 
can have fruiting and iion-lruitmg iiillorescences 
within populations and even individual plant- 

ublished obs.). In sites with moths on 
it i> general I > explained bv mollis 

Y. aloifoliu inflorescences that coin- 
cide with those of the native host (Rilev. 1802). 
Klsewhere llii- cannot be the case. (Conceivable ex- 
planations include inlrapopulational variation in 
v isilor guild-, anil po—ilulilv ol " aulogaiiiv or gei- 
loiiogamv. ImiI 1 1 lev arc lelaliveK uulikelv expla- 

ex pen an iila I da I a exi-l lo explore these or any oili- 
er hypotheses. Il is also po-- ible that ). uloi/oliu is 
iindci liiiiiled -election loi ma i ill a i in ng sexual re- 
pi-odiK lion, a- il reproduce- verv \igoiou-lv bv veg- 
etative propagation, flu- happen- both through rap- 
id clonal extension and establishment by broken-off 
plant parts (Rrown. 1050); in coastal North Caro- 
lina, local residents disseminate the planl bv lul- 
ling slenis in l()-eni pieces thai are tossed on the 
ground in disturbed sandv sites (J. Oroman. pers. 
conun.). This habit of elevated vegetative propaga- 
tion, absence of an endemic pollinator, core loss in 
the fruit, and poorly synchronized llowei ing spread 
across many months suggest that ). aloifoliu may 
be an escaped cullivar. Described from Kuropeaii 
gardens. Trelease (1803) referred to it as a species 
without a known ideographical origin, and this is 
-lill ihe ca.-e. Il i- nio-l clnselv related to Y. e/c- 
phunlipcs and ). lacum/o/iica. which are tropical 
forest dwellers along ihe Cull" side of Mexico from 
Veracruz to Yucatan and into northern Belize (Ma- 
tilda & Pifia Lujan. 1080; Davidse et ah, 1094); 
within this range, they sel fruit through ihe actions 
of a specific vueca moth (Relhnv r *K Ralca/ar-Lara. 
2000. unpublished data) whereas plants are sterile 
elsewhere. This is most evident in ). elephuntipes, 
which is widely cultivated throughout Mexico and 
southward at least to Panama lor its comestible 
flowers. )ucca aloifoliu has been reported from 
Mexico ( & Pina I.ujan. 1980). but exami- 
nation of available lo 1 1 ..n i collections al I NAM 

for their records indicates that these refer to cul- 
tivated specimens and lo ). elephuntipes (Pellmyr 
Ox Ralcazar-I.ara. unpublished dalal. In addition lo 

Us disinl, ii alom; ihe shoreline ol southeastern 

North America. ). aloifoliu is reported I > also estab- 
lished on Cuba. Jamaica, the Bahamas and Ber- 
muda (Trelease. 1902). where pre-Hispanic cul- 
tures are suggested to have used its roots for soap 
(Kngelmann. 1873). This use, together with a dis- 
junel geoL'iaphic range from ihe remainder of the 
genus and trails characteristic of cultivated plants, 
make- plausible a hypothesis that Y. aloifoliu orig- 
inated from ). clephuutipes as a cullivar selected 
for its high vegetative propagation. If correct, phy- 
logeographic studies are predicted to show a ge- 

,i. ill. ill I i ' ,'»■ . •/■ • n. -led within ). 

:>cs. I'his would be an important analvsis 
In perform from the perspective ..| f 1 1< - p i :> -II 

imply that occasional fruit 

moths in >. aloijnlia i- irrelevant to understanding 

diversification and eoovohit ol tile plaut-molli 

The origin ol the ■ -111 1 1 ; II -i > II •: n r i-ie; m 

known, and was subject to little speculation for 
more than a eentun after its initial di-eoverv. Two 
limiting factors loom important in this context. 
Kirst. life hislon difference- and variation in out- 
comes of vucca-yucca moth interactions had vet to 

historically recognized yucca moth species (Davis. 
1967; Powell. 1992) and yuccas were held to be 
obligate mutualisms, so there was no apparent 
transformation series to anab/e. Second, and per- 
haps more important in retrospect, the phv logenetie 
Iramework ol the yucca moth- at the laimlv and 
genus level was not determined until the 1980s 
(Frack. 1982: Nielsen & Davis. 1985). At that time, 
life hislon dala al-o -tailed to appear for the close- 
Iv related genera of prodoxid moths (hack. 1982: 
Davis et al.. 1992). 

> for establish* 

moth mutualism. The fossil rec..rd is quite pooi 
for these plants and eflectivelv absent for the 
moth-, providing little a — i-lanoc in dating the e— 
tablislmieul and div . r-ilnalion of the planl-polli- 
nator association. The onlv pre Pleistocene yucca 
maerofossil is a 14- My old trunk segment described 
as Protoyucca shadishii Tidwell & Parker from Ne- 
\ada. most resembling the extant )urca t,iv> ifoli,; 
(Tidwell & Parker, 1990). Fossil pollen described 
as Agarc has been described from the mid Miocene 
(Axelrod, 1979; Palacios & Rzedowski, 1993). Bre- 
mer (2000) used clocklike behavior in rbcL to es- 
timate the minimum age of the Funkiaceae, which 
is the sister family of Agavaceae + Camassiaceae 
(Pfosser & Speta, 1999), at 21 My, whereas Eguiar- 
te (1995) provided an independent rbcL-based es- 
timate for the Agavaceae of 14 My. For the moths, 
a mitochondrial DN A sequence data set was used 
to estimate age of their diversification, using bio- 
geographic events for calibration (Fig. I: Pedum 
& Leebens-Mack, 1999). Colonization of yuccas as 
a host was estimated at having occurred about 1 1 .7 
Mva. with the diveisificalion of the three genera 
that inhabit vin ca- b« in." so iapid that their dates 

.•tunes ill'- -p 1 iicto ,-( i j ii (.in, Hitlirami ; <-ni a 

(between T. mac u lata and all othei species), show- 
ing thai the pollination • i I wa- established in a 
comin.,11 ancestor very close in lime to the coloni- 
zation of ihe vueoa- b\ prodoxid moths, (oven this 
rapid diversification ol tin- moth lineages, we can 
infer that a basal radiation of yuccas was in exis- 
tence bv llns mid Koeene dale, pre-dating current 
independent estimates for the plants. For compar- 

y molecular data from s 

wasp lineages suggesting t 

the obligate mutiu 
ps had originated I 
Mya (Machado et a 

required to switc 

mutualism. This would happen, for example, bv 

variation in outcome based on ecological context. 

N - nl I'M inning-. ..i 'iion ba-a piodoxid i olhs 

that do not feed on yuccas lend support for this 
model of pre-adaplations. and also reveal two sep- 

three members of the genus Grcya and their saxi- 
fragaceous hosts (Pellmyr et al., 1996a). 

The first studies were made of Greya politella 

nl a -| alisl ol several spec ies of 

- ii _ i i I llmv i vK I honip- 

son. 1992; Thompson & Pellmyr, 1992). The female 
moth oviposits into the ovary through the tubular 
hypanlhium. and poll. -ii often i- transported on an 
. oi galed ibdi mil I -c; n i at I \pei uneiits showed 
that oviposilion was a highlv effective pollination 
behavior. At study sites in Washington, an exten- 
sive guild of copollinators of mostly bombyliid flies 
ami solilarv bee- al-o provided ero>--pnli r,i< ma. 
Although none of the co|tollinator species was as 
effective per visit as the ovipositing moths, their 
relative abundance and far higher rate ol visitation 
made I hem impoi 'ai ' ■<>-\\\ b ilnr- i. [ >< • 1 1 nation n 
the study population. In two years of study, G. pol- 
itella was estimated to have contributed 0.8-2*7, of 
all seed set in the study population. I lieu positive 
effect wa- eflectivelv masked, a- there was no sig- 
nificant e flee I of moth oviposilion on net seed set. 
Their negative effect through larval seed consump- 

lion was also masked by other sources of variation ly abscised will perish. Kloral abscission is highly 

in seed production. The outcome of 'this molh-pl, ml selective, with fertilized Mowers resulting from 

interaction is thus strongly dependent mi copolli small pollen loads or self pollen having a much 

moths for improved pollmall 'Ilicicncy unless it Kichter & Weis. |<W8: lluth cK IVllmyr, 2000). For 

found for Greya cnchr\ s,/ Davis o, IVllmyr. a highly \ided by female yucca moths can result in differ- 
effective pollinator of its hosts in Hciichrra, where enlial abscission ol Mowers containing moth eggs, 
abundant hiimhh I » . I I ■ m In lal el is I. males providing huge amounts of pollen de- 
fects on seed set (Pellmyr et al., 1996b). In these crease the risk of abortion. Importantly, this trait 

across I lie ranges of the species (Thompson iK Fell- of relative!) ineflicient. nectar- and pollen-coiisum- 
myr, 1992; Gomulkiewiez et al.. 2000). potentially ing floral visitors. In a second step, reciprocal spe- 
leading to sustained selection for a stronger mutu- cializalion in the plants on the increasingly eflec- 
alistie e(|uilibriuni between the moths ami plants. tive yucca moths is expected as the net fitness 
The third ease and second origin of pollination in contributions attributable to the ancestral neelar- 
Greya involves C. mitcllac Davis Ox I VI liny r. a spe- consuming visitors relative to energetic investments 
eies whose larvae led inside the flowering stalk in the nectar reward became negative. Both selec- 
and in leaf peduncles of Milclla stauro/wlala Fiper. t 1 I il I I I i ind high cost 
Moths pollinate while drinking nectar from the of nectar production were novel traits to the yucca- 
flowers. Whereas virluallv all pollination was pro- yucca moth association in the sense that they are 
vided by the moths in study populations, no selec- not present in the plant-moth interactions inunc- 
tion on increased pollination efficiency is expected dialcly basal to il, and they may point to factors 
in this interaction as larval fitness is unaffected by that could facilitate similar transitions in other as- 
ide incidental seed production during adult nectar- socialions. Consistent with this prediction, much 

direct link between female pollination efficiency lio are characteristic ol the recentlv described ob- 

and progeny fitness to cause selection toward in- ligale n alism in the Sonoran desert between the 

creased pollination efficiency and potentially obli- columnar cactus Ln/tharrrciis schottii and its polli- 
nate mutualism (Pellmyr et al., 1996a). naling moth. I i»»a rin>sirn< (Holland & Fleming, 

Mapping of several life history trails that were 1999). 
necessary prerequisites for lb. origin of the mutu- 

alistic behavior by yucca mollis indicated that most |{| \Kus\l OK Ml THAI ISM 

traits were basal to 1 id an lh >i at least had 

evolved before the lineage leading to the common Mulualislie interactions contain an underlying 

ancestor of the pollinator genera. Hence this sup- evolutionary eonllicl in that the interacting partners 

ported a scenario in which the life habits of pro- are under selection for increased exploitation of 

doxid moths commonly have stales that make pol- each other (Tri vers, 1971; Hull ov bice. 1991; IVll- 

linator function easy to acquire. At the same time, myr & Huth. 1994). In a phmt-pollinator relalion- 

obligale muliiallsm llial requires novel traits lor ship, this mighl manifest as selection for higher ef- 

the family, in the true yucca mollis. Why did this smaller or more inaccessible rewards in the plants. 

happen in the yucca-yucca moth association, but In laciiltalivc relationships, such conflicts may re- 

not in Ihe others? IVllmyr el al. (1996a) used an- sull in arms races that shut out excessive exploit- 

cestor reconslriicliori of the yuccas to erect a hy- ers. For example, a decreasingly rewarding plant 

polhesis in which h i I I mil ion in the species may he abandoned bv Mower visitors that 

moths evolved first, followed by exclusion of an- have a choice, while plant trails that reduce losses 

eeslral eopolliiialors through elleclive cessation ol to poor pollinators in theory can evolve to complete 

nectar production in the plants. A general feature exclusion. In obligate mutualisms lhat involve a 

ol the \ga\aceae is resource-limited fruit set (Suth- single pollinator and plant, this conflict has a po- 

crland, 1982), where only a minor fraction of all tentially different dynamic. In such instances, the 

flowers give rise In mature fruit. \s pn.dnxid mollis evolution ol a cheating mutant with a fitness ad- 

eolonized yucca ovaries, they thus encountered a vantage over mulualisl individuals is expected to 

major new mortality factor for their progeny, be- lead to reciprocal extinction of the mutualists. at 

cause all eggs inside pollinated flowers subsequent- least at the population level and possibly on a spe- 

depending on : > ; ) 1 1 * Us <>l gei e How I'm 
i, obligate mutualisms such as those be- 
cas and yucca mollis were long roii-id- 
itionars dead ends (Soheron Maincro vx 
lei Rio. I9!!5: Hull K Rice, 1991). This 

is clearly not the case unde 

two distinct specie- !>l nor |>ol hating cheater y uc- 
ca moths derived from pollinating ancestors have 
been identified (Pellmyr et al., 1996a). The two de- 
scribed species. Te^clieula intermedia and T. cor- 
niplri.\ IVllinvr. oviposil directly into fruits at dif- 
ferent stages of development, and the larvae 
consume seeds in coc\istcnee with larvae of the 
pollinator species I fig. ■")!)). I'heir presence can be 
very costlv for host seed production; in one study 
of >. Jilamenlasa. seed destruction was tripled in 
populations where cheater moths coexisted with 
pollinator moths (Pellmyr et al.. 1996a). Kcological 
data did not reveal any competition between co- 
existing larvae of the pollinator T. vueeasella and 
the cheater T. intermedia (Marr et al., 2(K)1). so 
coexistence is evidently no| a problem, but the sep- 

r I-, .sialjlc . 

cheater life habit remains to be explained. 

Phylogcnetic analyses based on mitochondrial 
DIN A sequence data suggest that the two species 
originated separately around 1 .26 ± 0.9 \1va: thus 
these are not ephemeral lineages (Pellmyr et al., 
I'KlOa: IVllmvi & I .eebeiis-Uack. 1999). A simple 
solution to the problem of escaping the evolutionary 
dead end of obligate mutualism is coexistence of 
two or more mulualisls on a shared partner, for 
exam | lie. if two v ueea mollis were to coexist on one 

ciil extinction is only expected if both mutualists 
independently abandon the pollinator habit. With 
the recent recognition of a large number of polli- 
nator species, it has become apparent that coexis- 

linalors sharing a lio-l in all or part of its range 
(Davis. 1967: Powell. 19}? 1; Tyre & Addicott, 1993; 
Pellmyr, 1999; Pellmvr & Baleazar-Lara, 2000; 
Pellmyr & Leebens-Maek. 2000). One of those 
sympatry zones is implicated in the origin of T. in- 
termedia. This species is most closely related to the 
pollinator T. cassandra. and available data suggest 
thill it mav have evolved where T. cassandra came 
into coexistence with '/.' \ueeasella in part of its 
range. The pollinating sister species of both T. in- 
termedia and the other cheater species oviposit in 
a way that distinguishes them from all other polli- 
nator species, and they have a characteristic ovi- 

positor thai allows them to oviposit into either a 
flower or a young fruit. Thus, these pollinator- mav 
be preadapted for a switch to oviposition into fruit 
once a sympalric pollinator species is available to 
perpetuate pollination. Because of a selective ab- 
scission mechanism in the yuccas causes (low- 
ers with many moth eggs of most pollinator species 
to be abscised within a lew days ol pollination, a 
huge proportion ol the seeds are simply not acces- 
sible for larval consumption by these pollinator 
species. Hence, a pollinator species can delay 
oviposition by a few days and oviposits directly into 
voung fruits can bypass the plant's abscission pe- 
riod and exploit a rich seed resource. In this sce- 
nario, the phonological shift can be an adaptive 
step into a novel niche llial precedes the loss of 

fruits become the target ol oviposition. \vailable 
data thus suggest that the origin of cheater yucca 
moths Iroin pollinators did not result from selection 
for cheating per se. but rather as a byproduct of 
selection for exploitation ol a pieviously untapped 
seed source (Pellmyr & Leebens-Mack, 2000). 

By analogy to evolution of non-cooperative pol- 
linators, it is in theory possible that ehealm plant- 
could arise in an obligate mutualism. In the case 
of yucca plants, that would entail the evolution of 
mechanisms that maintain pollination but prevent 
seed destruction by pollinator larvae. This could 
happen through mechanisms such ;h prevention ol 
successful oviposition. or killing of the eggs or lar- 
vae. If an alternative, cooperative host species ex- 
ists in the area, such cheating by plants could be 

plants in a single plant-single pollinator scenario is 
predicted to lead to extinction (Bull & Rice, 1991). 

The only proposed case thus far involves a popu- 
lation of \iieea haecala Ton., where Bao and Ad- 
dicott (1998) reported that the fruits of a substantial 
proportion of all plants lacked evidence of larval 
damage, and speculated that this might be evidence 
ol a cheating mechanism in these plants. They did 
not speculate regarding a mechanistic basis, but 
mentioned that fruits without larvae had a distinc- 
tive shape. Further studies will be needed to de- 

should he emphasized that the strongest. 
:t evidence for selection for cheating in a 
sm would be direct evidence of individual 
fitness gains. Such data are wanting for 

I, io'Ji lie 

Inn • < lore 

8 .1 

i I i i j f ! f 1 1 i « 1 S 



ami ISiilra/a.-LiM |2(KMI). 

suitable technological tools will he available to 
solve this problem. In the yuccas, longevitv of de- 
cades or centuries (McKelvey. lO.W; Webber. 1953; 
Matuda & Pina huja.i. 1980: Webb. 19%; Comanor 
& Clark, 2000). with iteroparily in all but one spe- 
cies, and also different magnitude and po>-ible 
pla>licil\ in vegetative propagation, makes it diffi- 

ijiience. surrogate 


When species are tightly < 
prohabilih that thev mav co-speciate (Kichler. 
1948; Huelsenbeek et ah. 2000). Such parallel ,li- 

bascd divergence mi (rum eoev ohil iouarv | r.-srs 

between the species (I'age. 1991). For this reason, 
obligate pollination iniiliiali-ni- between seed-par- 
asitic pollinators and their hosts should be good 
candidates lor parallel di\ ersili. atiou. as potential 
divergence may derive from linked host speciali- 

zation in the pollinators and pollen-mediated gene 
How in the plants (Bogler et ah. 1995; Pellmyr et 
ah, 1996a). The lig-fig wasp associations and yuc- 

\nal\ses of the association between figs anil fig 
wasps have indeed indicated a high level of parallel 
diversification at the level ,,(' fig genera and sub- 
genera (Herre et ah. 1990). while there is emerging 
evidence that this pattern breaks down to a fair 

degree at lower lavonomic levels (hop.v \aan le 

et ah, 2001; Machado et ah, 2001). Analysis for the 
yuccas and yucca moths is still rudimentary as the 
vucca phvlogenv is incomplete!) resolved, and the 
nin. solved polvtomv in the moth phvlogenv a bo 
limits analysis (Kig. 7). (dven the current ' unre- 
solved plant relationships, there are no -trong can- 
didates for parallel diversification, although tin- 
tnav change with increasing phylogeneli. i nl. >i m.i- there are iiiiinerou> instances w h. -re coloniza- 
tion has occurred. The most obvious instance in- 
volves recent colonization by Tcgt'ticu/a vucnisella 
of Yucca ulni/nlia. Similarly. 7.' hnccalcll,, IVllmvr. 

which feeds on a fleshy-fruited hosl. i- nested amid CONCLUSION 

-pe< es lli;il leed iv : " | «■- i; i dialed '-ijhj- 1 I \<> , .. . ... 

1 l ■ I'n :-.. -. <-n a -» mi -/.I . : - ibdnc I i. I. .. i v.l i. I. 

7), and thus supports a pasl shift assuming thai (<) ^ [|ia( ^ ( . ;mm)t v<i| |)rrj . )mi ^^^ ^ <>f 

monophyly of fleshy-fruited yuccas is upheld. 'I he ^ m l e of coevolulion in the diversification of vuc- 

cheater T. corrupt n.x. also arming Iron, an ancestor ^ ^ ^^ m()[ ^ ^ ^ reasf)n {& simply that 

on a capsular-fruited yucca (rig. , ). uo« utilizes ., ,. i| . f , < , ammml (|f inforinalion aboul morphology, 

both fleshy-fruited and capsular-fruited species. V( . \ (> ^ natm ,,| | lislorv . an( | nhylogeny is required 

Second, the coexistence on a host of non-sister taxa (■„. an> ()He ass()( .i a ,i () „ | M .f ore analyses of the his- 

of Tegeticula pollinators cannot he explained by torical impact of coevolution can be explored. Most, 

parallel diversification; in principle, coexistence of | )Ut MO , a ||, f , nrsr requirements are now largely 

a Parategetieula and a Tegeliculu pollinator on a „,<>,. The last 15 years have seen a dramatic in- 

host could reflect two independent parallel diver- crease in our understanding of organismal diversitv. 

sifications with the hosts, but there is very little especially among the insects, although much infor- 

support from published host data for this explana- mation irom the Mexican range of the yucca-yucca 

tion (rig. 7). Third, instances where a pollinator moth associations remains to be published. Ecolog- 

species utilizes more than one host species (Fig. o) ioal and evolutionary dynamics have also become 

cannot reflect parallel diversification, although they far heller understood in the last decade, including 

may possibly reflect an ancestral association with the expansion into the- realm of reversal of mutu- 

subsequent unilateral diversification in a monophv- alism. Phv logenelic information is now arguably the 

letic group of hosts. primary limiting factor for analyses of coevolution 

Co-speciation does not require coevoh.tionarv '»"• «**veral other major questions, but there is rea- 

processes. and coevolution can act on organism.. *»» "> »"P<" »"» «'«>»«• '"fonnation soon will be 

regardless of their history of association: thus the available lor both groups. Ongoing parallel projects 
on subsets of fig-fig wasp associations (e.g., Lopez- 

Naamondeel al.. 2001: Uachado e| al.. 2001: We, 

bleu & Bush, 2002) as well as other mutualism 

' ., .. . involving seed-parasitic polhnalor- 1 1 ic-pi'c- el al 

that vary among species nrav a.,s.- ether from the- ** ' ,.,..,. , _„,,,. __„. 

iweel iili.' plants and ie (...I i m;iI. m - i- a i n i •• i l n 
separate matter. Selecti 

interacting part 

interaction, lor example, trails likely In affect ninth 

.. '.. ... 20021 also offer possibilities for grander compari- 

or from factors extrinsic to the _ ^ ' s ^ 

Whatever generalizations aboul factors mediating 

opposition success, such as Moral ova,, morphol- , m ollilialor rmi|lia||sms rmrrge from these 

ogy and moth ovipositor morphology, may be strong |n „ ||U ^.^ ; ^ (| _ m M)OI1 be used as a 

candidates for reciprocal selr i ,s il„ y , lb u ^ ^ m m ^ /n ^ ^^ fmw vomp{ex plant . 

affect plant and polluial..r illness, Meanwhile, hail- |l(1 ||j M .,,,,,. mM | lia |i sms . 

such as petal shape and color mav be more likely 

to be under selection based on a wide range of I iterature Cited 

"ier herbivores, as . 

... . . Adclicott. J. I . !'" Ii i. in Micca/molli iiiulualisni. 

line the historical Nature 380: 1 14-1 15. 

role of coevolution in the diversification of an in- & \. j. Tyre. I«W5. Cheating in an obligate mu- 

titioned to remove exlrinsic components, i.e.. in es- , k(,s l2: :ili2 ~?l"- , . .. , , ca 

, \<u n . .;, ! i . , » . in m.i i iiu- 

" '"" o ri-ino i ! .hi I oaition in the ( , rs " \. {[uw 5. 442.444. 

inli tac ling iMniip- ■ factors I'his \l'(, | \ngiosperm Phytogeny Group]. 1998. An ordinal 

obviously requires groups of plants and pollinators ' l.--'l" •'""" '" l'" '•"' dowering plants. Ann. 

that have coexisted during much or all of their di- a^™"^ g^j" ^ Hovvafd^T. P. Clausen. K. M. 

versification, as is the case for yuccas and yucca |>el.ew-c. I. C. I,.',,, nam. M. Matsuki. H. Cerendolo & 

moths. It also requires well-resolved phylogenies. F. Andel. 1997. Do biochemical exaptatimis link evo- 

well as abiotic factoi 

I hypotheses anil e\p< a iincnlal tests with Diih-rhii. 

preferably with estimates of internal bran. ",' "| '/^^"^ ^ 

lengths. This criterion is not yet met for the yucca- ^ vi ^ AmleVL^ist'uQr^^.' 

yucca moth association, nor for any other similar 
plant-pollinator association. Such analyses will be 
mi -lib iis.-liil : l i-\aluaiiug lac hisiurical role ol co- 
in plants and pollinators. Bao, T. & J. F. Adclicott. 1998. 

rvev. I 998. 

Wraelough. T. G.. A. I'. Vogle. ■& P. II. 

Revealing the faclors that promote speciation. Philos. 

Trans.. Ser. B 353: 241-249. 
iecerra. J. X. IW,. Insects on plants: Maeroev ohilionar v 

cllemieal trends in host use. Science 270: 253-256. 
& I). I.. VnaUc. |9<><). Vlacrocvolntionol'inscct- 

planl ;i>-.i i.ilion-: The relevance ol ho-l biogeographv 

to host affiliation. Proc. Natl. Acad. Sci. U.S.A. 96: 

1996. Phylogen) of \ga- 

.1. Hot. . 

. ei B. B. Simpson 
wed or. ITS rDNA 
I: 1225-1235. 
L. Neff& B. B. Simpsoi 

Acad. Sci. U.S.A. 92: r.8<> G(.8t>7. 
toll, J. 1876. Ueherdie Befruchtung der nordame 
ischen Yucca- \rten. Knlomol. Zeitung 37: 4 ( ) 1 — 1 1 

flowering plants. Proc. Nail. Acad. Sci. U.S. A 

Iritton. N. U. & C. K Millspaugh. 1962. The \k 

Flora, llafner. New York, 
.ronstein. J. U. & Y. Ziv. 1997. Costs of two non- 

ologia ||- .-{79 385. 
'..own. C. A. 1959. Vegetation of the Outer Par 
North Carolina. Uouisiana Stale Univ. Press. 

mis (l.epidoplei 
524: 1-88. 
Despres, U. & N. 

ii (editor). I land! k of Zoology. I\ 

la. I.epidoptera. Moths and Butler- 
r de Gruyter. Perlin, New York. 

I "99. fuillltln, VipoMtloll 

rr in lh« -lobeflouer fir*'- Cluas- 
iidae). J. K\ol. Biol. II 1 : 822 

hi iii the globcllower llv Cliiastachcla 
thomyiidae) in relation to host [)lan 

: 258^268. 

I. & Y. B. Linhart. 1994. Reprodu. 

J. M., (). Pellmyr, J. N. Thompson & B. G. I 

1991. Mitochondrial |)\\ p|,\|„g.. m <>f lliePro- Kichler. W. 1948. 
ac (Uepidoplera: Incurv at ioideal indicates a rapid 
deal *liveisili*ali..ii ..I ||,e mi,-, a moths. Ann. En- 

Eg.liaile. I . I-. P)9.V Mulehinsoll ( \ga\al.-l v *. llllhci 
\ HahUieii I Vsparagales): \nalisis molecular*- miIhc la 

Hutchinson dentro de las monocotiledoneas. Bol. Soc. 
Bui. Mexico 56: 45-56. 

. V. Sou/.a iK A. Silva-Monlellaiio. 2909 Iwolu- 

cion de la familia Agavaceao: Filogenia. biologia re- 
productive v genelica de polilaeiones. Bol. Soc. Bot. 
M.'-vic uM>: 131-150. 

Khrlich, P. R. & P. H. Raven. 1964. Bulterllies and 
its: A study in (Devolution. Evolution 18: 586-608. 

Bn.kl I A >. Smill. (Editor). I9'U. Calendar of ihe 

C..m-| deuce of Charles Darwin. 1809-1882. with 

Supplement. Camliridgc I uiv. Press. Cambridge. I .k. 

Busck. \. 19 17. The vuca moth. Pp. 180-185 in S. I). 
McKelvey (editor). Yuccas of the Southwestern United 
Slates. Vol. 2. Arnold WLorelum. Jamaica Plain. 

Chambers. V. T. 1877. The Tineina of Colorado. Bull. 
U.S. Ceol. Ceogr. Survey Terr. 3: 121. 147-149. 

Clary. K. H. 1997. Phvlogenv. Charade. F.volulion. and 
Biogeography of Yucca U. (Agavaeeae) as Inferred from 
Plant Morphologv and Se(|iicnces of the Internal I'ran- 
-eril.ed Spacer (ITS) Begin., «.f lli.- Nuclear Bibnsomal 
DNA. Ph.D. Thesis. I mversih of lc\as. \usim. 

Comanor. P. U. <K V\. II. Clark. 2000. Preliminary gmvvlh 

Fngelmami. C. 1871. )licca and Hcspcraloc. Watson's 
Botany of the Fortieth Parallel. Report of the United 
Stales Geological Exploration of the Fortieth Parallel 



I foe. 

. 1872. The Howe 

Bull. Torres Bot. Club 2 
. 1873. Notes on 

Sci. St. Uouis3: 17-54. 
Farrell. B. I). 1998. "Inordinate>" explained: 

Why are there so many beetles? Science 281: 555-559. 
*X C. Milter. 1993. |'| lx logenetic determinants of 

insect/plant community diversity. Pp. 253-266 in R. E. 

Ecological Communities. I niv. Chicago Pie--. Chicago. 
Feist. M. A. 1995. Population Genetics of Yuccas: Their 
System and its Consequ* 

(). ( hat,-, (Edit, 

>ati. Cincinnati. 
•0. 1991. Fleming. T. H. & J. . 
i Cypera- obligate pollination mut 

Thesis, University of Ci. 
Holland. 1998. The evoluti* 

Incurvariidae). Bull. 

1987. Incurvarioidea. Pp. 354-362 in E W. 

Fuller. 0. S. I99C 

(editor). Immalure Insects. KendalUHunl. Du- 

s, Iowa. 

ticula yuccasella 

1998. The Monolrysian Heteroneura. Pp. 65-90 

Thesis, Universi 

Oecologia 111: 30.8-375. 
ck. I). C. 1982. \ Systematic of Prod,,- 
noths (Adelidae: Prodoxinae) and their Hosts (Ag; 
•eae), with Descriptions of the Subfamilies of Adeli 
s. lat.). M.S. Thesis. California State Polytechnic I 
ersilv. Pomona, California. 

Affecting III*' Balance nl 

Galil, J I 909 Intentional pollination in yucca flowers. Machado. C. A.. K. Jousselin. F. Kjellberg, S. G. Compton 

leva Va'arelz I I: 1-12 |in Hebrew]. & K. A. Herre. 2001. Phylogenetie relationships, his- 

Garefa-Mendoza, A. 1000. Una nueva espeeie de Yucca torical biogeographv ami character evolution of fig-pol- 

(Vnavaceael ilc ( )a\aea v Pliehla. Mexico. Acta Hot. limiting wasps. I'roc. liov. Soc. I .omlon 2()i'.: I -10. 

Mex. 12. I ... MacKee, H. S. 1901. Catalogue ,les I'lantes Introduites 

Gentry, H. S. 1982. The \gave family in Sonora. el Cultivees en \ouvolle-( .alcdomo. Museum National 

b.S.D.A. Sen. Agric. Handbook 309. dTlistoire Naturelle, Paris. 

Gerarde, J. 1633. The llcrball or Generall Historic of Marr. I). I... J. Leebens-Mack. I.. Kims & 0. Pellmyr. 

Plantes. Adam Norton ami Uichard V\ hilakcrs. London. 2000. Pollen dispersal in )„cca Jilamcnlosa (Agava- 

Gomulkiovuc/. IS... |. Y Thompson. I{. I). Holt. S. I, Nuis- eeae): The paradox of self-pollination behaviorjn Te- 

mer & M. K. Ilochben:. 21100. Hot -pot-, mid spots. nclicula utccasclla | Prodoxidae). \mor. J. Bot. 87: 670- 

and the geographic mosaic thmrv ol .revolution. Amer. 677. 

Naturalist 156: 156-174. . M. T. Brock & 0. Pellmyr. 2001. Coexistence of 

(ireat Plains flora Association. 1977. Atlas ol the flora mutuali-ts and antagonists: Kxploriiig the impact of 

of the (ileal Plains. Iowa State Inn. Press. Ames. cheaters on the wieca-vueca moth mutualism. Oeeologia 

(irnman. J. I). I0<><). | ». ,p,ilal ion Geuetio and life His- 128: 154-463. 

ton Consequences of a Becenl llo-l Shift in I he Bogu- Massev. L. K. & J. I.. Hamrick. 1998. Genetic diversity 

Yucca Moth I'm,/. > I'odoxi.ln I and population stuuluie ol )ii<(<i lilniiu nlnsa 

M.S. Thesis, \amlcrbilt University. Nashville. eeae). Amer. J. Bot. 85: 340-345. 

see. Matuda, K. v\ I. Cola l.ii.j.m. I9JM). las |-|antas Mexican- 
Hanson, M. A. 1902. Dispersed I lnlr,,grcs as del Cenero )ti,,a. Scrie Fernando de Alva Ixtlilxo- 

sion from Yucca sclmii-cta oilo ) ticca hairala (Agava- chill. Colleeion miscclauea eslado de Mexico, Toluca. 

eeae). Ph.D. Thesis. Clarcmonl Graduate Sch , Po- Mexico. 

inona. California. Mayr. K. 1903. \nimal Species and Involution. Harvard 

Herre. K. A., C. A. Machado. K. Bermingham. .1. I). Na- Univ. Press. Cambridge. 

son. I). M. Windsor. S. S. VlcCalJeilv. V\. \a.illoulen X Mokclvev. S. I). 1930. 1917. ...ccas of the Southwestern 

K. Baehmaiin. 1990. Molecular phvlogenies ol figs and I nited States. Vols. 1. 2. Arnold Arboretum, Jamaica 

their pollinator wasps. J. Biogeogr. 4: 521-530. Plain. 

Ho, J. & W. Yuille. 1990. The Papers of Charles Yalci- McVaugh. I!. 1003. )ticca. Pp. 201 283 in W. B. An- 

tine Biley: A Begister of his Papers in the National de.-M.n (general editor), flora Novo-Galiciana. A l)e- 

Agricullural Library. I .S.D.A. Natl. Agric. Libr. Bibli- scriplivc \ecounl ol the \a>eular Plants of Western 

ographies and Literature of Agriculture No. 92. Mexico. Vol. 15. Bromeliaceae to Dioscoreaeeae. Univ. 

Holland..). N. v\ T. II. I lommg. 1000. Mutualislic inter- Michigan Herbariui ' 

actions between I />/>« rircscens (Pyralidae), a 

eeae). Kcologv 00: 2071 2001. 

Huclsenbeck. J. P, B. Kannala cK B. Larget. 2000. A Kr.lomol. 1:33-30. 

Bavesian framework lor the analysis ..I cospccialion. Miles. N. J. 1983. Variation in host spccihcitv in the 

Involution 51: 352-364. yucca moth, Tcgclicttla yuccasclla (liicurvariidaci: \ 

Hulst, G. D. 1880. ISocielv news.] Kntomol. Americana morphometric approach. J. Lepidopterist's Soc. 37: 

2: 184. 207-216. 

Huth, C. J. & <>. Pclhnvr. 1990. Yucca moll, opposition Miller. C. B. Farrell & B. Wiegmann. 1988. The phylo- 

and pollination behavior is affected bv past flower vis- genetic study of adaptive / -: I Li- phvlophagy pm- 

itors: Kvidence for a host-marking pheromone. Oecolo- moled plant diversification? Vi.ier. Naturalist 132: 107- 

gia 119: 593-599. 128. 
& . 2000. Pollen-mediated selective Mohr. C. 1901. Plant Life of Alabama. Contr. U.S. Natl. 

abortion promotes evolutionary stability of mutualism Herb. 0. Washington. D.C. 

between yuccas and yucca moths. Ecology 81: I 1(H)- Moss. K. H. 1959. Flora of Alberta. Univ. Toronto Press, 

1 107. Toronto. 

Jaeger. N.. F. Pompanon & L. Despres. 2(H) I . Variation Muiawski. I). A. 1007. Mollis .-mm- to light. Natl. Geogr. 

in predation costs with Chiastnchcla egg number on Mag. March: 40-57. 

Tmllius ctirapacus: How many seeds to pay for pollina- Miillci II 10,1 ( ., . n- Vbhangigkeil von Bhimen 

lion? Fcol. Entomol. 26: 56^62. und sie befruclilc.iden lns«-c|eii. /oologische Garten 

Kingsolver, B. W. 1984. Population Biology of a Mutu- (Frankfurt) 15: 377-382. 

alislic Association: Yucca glauca and Tegelicula vac- Nielsen. K. S. 1982. lncurvariidae and Prodoxidae from 

casella. Ph.D. Dissertation. University ol Kansas. Lawr- the Himalayan area (Lcpidoptera: Incurvarioidea). ln- 

ence. secta Matsumurana N.S. 26: 187-200. 

Kozlov, M. L. 1996. lncurvariidae and Prodoxidae (Lep- . 1985. Primitive (non-ditrysian) Lcpidoptera of 

idoptera) from Siberia and the Bussian Far Fast, with the Andes: Diversity, distribution, biology and phylo- 

descriptions of two new species. Entomol. Fenn. 7: 55- genetic relationships. J. Bes. lcpidoptera Suppl. 1: 1- 

62. 16. 

Layard. F. F. 1880. Yuccas under cultivation. Nature 22: c\ I). B. Davi-. I ««■,;,. The first southern hemi- 

606 607. sphere prodoxid and the phvlogenv of the Incurvarioi- 

Lopez-Vaamonde. C.. J. V Basph.s. (;. I). Weiblen ,K J. dea. Svst. Entomol. 10: 307-322. 

M. Cook. 2001. Molecular phvlogenies of l,g wasps: Page. B. 1). M. 1994. Maps between trees and cladis 

Partial cocladogencsis of pollinators and parasiles. Mo- analysis ol In- • il ■ i r, h ■ genes, organ- 

lec. Phvlogenel. Kvol. 21: 55-71. isms, and areas. Svst. Biol. 43: 58-77. 

logical .'volution. Nature 101: 877-884. 
Palacios. I{. A J. Rzednwski. 1993. Kstudio palinold 
de las Moras fosiles del mioceno inferior y princi 
i) medio de la region de Piehucaleo, Chia 

Vllmvr. 0. I 

Bot. Mex. 24: 1-96. 

i: 53-59. 

The plnlogenv ol'a mutualism: F.volnlion 

3iol. J. Linn. Soc. 47: 337-365. 

Polluialing seed eaters: Win is active pol- 
..<•? Kcologv 78: 1655-1660. 

42: 182-187. 

K M. Balca/ar-l.ara. 20(H). Svslemalics ol the 

vucca modi genus I'.iiulcgcliculu ( I ,e| ,ldop|, la: Prodox- 
idae). with description of three Mexican species. Ann. 
Knlnmol. S»c. \nier. 93: 132-139. 

A C.J. Huth. I'WI. Kvolulionarv -labile ..I mu- 
tualism between yuccas and vucca moths. Nature 372: 

v . 2000. Reversal . 

Price. P. \\. 1980. Kvolulionarv Biologv of Parasites. 

Ban. P. 1915. Ihe vucca plant. Yuccu Jilamcnlosu. and 
the yucca moth. Tegeticula ( Pronuba I vuccasella Bilev: 
An ecologico-I.ehavior studv. Ann. Missouri Bot. Card. 
52: 373-394. 

Kichler, K. S. K A. K. V\eis. 1998. Inbreeding ami .un- 
crossing in Yucca whipplei: Consequences for the repro- 
ductive success o! plant and pollinator. Ecol. Lett. I: 

Bilev. C. V. 1871. Third Annual Report on the Noxious. 
Beneficial and Other Insects, of the Stale of Missouri. 
II. Wilcox. Jefferson (atv. Missouri. 

. 1872. Ihe fertilization of the vucca plant l.v 

I'ronuha uncasctla. Canad. I'.ulomol. 1: 182. 

. 1875. On the oviposilion of the vucca moth. 

Amer. Naturalist 7: 619-623. 

. 1880. The true and bogus yucca moth, will 

. 1881. Further notes on the pollination of y; 

and on I'rotiulia and I'mdo.xus. Proc. Amer. Assoc. 

Sci. 29: 617-639. 
. 1887. Mr. Hulst s observations on Pronuba 

casclla. Knlnmol. Americana 2: 253-236. 
. 1889. Notes on I'ronuha and Yucca pollina 

Proc. Kntomol. Soc. Wash. I: 150-155. 
. 1891. Fertilization of Yucca in Australia. In 

Life 4: 74. 
. 1892. The vucca moths and Yucca | K >lliiiai 

Annual Rep. Missouri Bot. Card. 3: 99-158. 
Rowlands, P. G. 1978. The Vegetation Dvnamics of 

Joshua Tree (Yucca hrcri/olia Kugelm.) in the : 

Rzedowski, J. 1993. 

erogamic flora ol > 
moorlhv. B. Bve, 
Diversity of Mexic 

hi). Th.Ms. 

■xico. Pp. 129-1 14 in T P. Kama- 

. Lot & J. Fa (edilorsi. Biological 
Origins and Distribution. Oxford 

B. (;. Harris, 
148: 827-847. 

sofmu- Univ. Press. New York. 

I. Acad. Segraves. K. K (). Pelhnvr. 2901. Phv logeographv oft 

yucca moth Tegeticula maculala: The role of histori. 

tualistic hiogeographv in reconciling high genetic structure w 

limited specialion. Molec Kcol. II): 12 17 1253. 

Smith, E. H. 1992. The grape phvlloxera— A celebrati 

of its own. Amer. Fnlomol. 58: 212-221. 
Soberon Mainero. J. & C. Martinez del Bio. 1985. Che 

havior and sell-poll 
Pfosser. M. v\. F. Spela. 1999. Phv logeneti 

n yuccas. Oecologia 109: 

i Bol. Card. 86: 852-875. 

,J. A. 1984. Biological interrelationships of motl 
)ucca schollii. I niv. Calif. I'ubl. Kntomol. 100: I 

P.I t :.!.,- idae. Prodoxidae). 

Pp. 192-210 in I). H. Boucher (editor). The Biology 
Mutualism. Oxforil f niv. Press. New York, 
orensen. W. C. 1995. Brethren of the Net: Amerie 
Knlomologv. 1810 1880. I niv. Alabama Press, Tusc 

v.lhe.Lmd. >. 1982. The Pollination Biology ol Panicu- 
late Agaves: Documenting the Importance of Male Fit- 
ness in Plants. Ph.D. Dissertation, University of Arizo- 

'homp.-on. .1. V 1991. The < revolutionary Process. Univ. 

Chicago Pic--.. Chicago. 
. 1999a. Specific hypotheses on the geographic 

mosaic of coevolution. Amer. Naturalist 153: SI-SI 1, 
. 19991, \\|,ai u, know and do not knov about 

Pp. 7-30 in H. Olff, V. K. Brown & Ft. H. Drent (edi- from Yucca baccata: First rc|iort of a leaf-mining pro- 

lors). Herbivores: Between Plants and Predators. Black- doxine ([.epdioptera: Prodoxidae). Ann. Entomol. Soe. 

well, Oxford. Arner. 81: 547-553. 
& 0. Pellmvi. 1002. Mutualism in pollinating Webb. K. II. 1096. Grand Canvon: A Century of Change: 

seed parasites amid co-pollinators: Constraints on spe- Bephotographv of the 1889-1890 Stanton Exp. 

cialization. Ecology 73: 1780-1791. Univ. Arizona Press, Tucson. 

Tidwell, W. D. & F. K. Parker. 1900. I'roKnncm shmfi- Wdil.ei. J. \1. 1053. hi.rih of the Southwest. U.S.D.A. 

shii gen. et sp. nov.. in A > m <i. u, I! in iih \l ^i. Mo. 17. Washington, D.C. 

ndar\ g. vvth li . the middle Miocene of north- VXcihlcn. (,. 1 1. 2002. Mow I,, he ., |',g wasp. Annual Pes. 

western Nevada. ISA. Hex. Palaeobot. Palvnol. 02: 70- Kntomol. 47: 299-330. 

95. & C. I.. Bush. 2002. Speciation in fig pollinators 

parasites. Molec. Ecol. 11: 1573-1578. 

1, J. B. 1891. The emergence ol I'nmitha from 

220. yucca capsules. Ann. Rep. Missouri Bot. Gard. 5: 137- 
. 1902. The Yuccaeae. Annual Rep. Missouri Bot. 138. 

Gard. 13: 27-133 + pis. 1-99. Wiegmann, B. M., C. Mitter & B. Farrell. 1993. Diver- 

Trivers. R. L 1071 lh.- . v..lnli..n ol i. . ipm. .il jIimiimii. -ideation of.aniivon.u-. para-ilic insects— Extraordi- 

Quart. Rev. Biol. 46: 35-57. narv radiation or specialized dead end. Amer. Naturalist 

Turner. 11. M.. J. E. Bowers & T. L Burgess. 1005. Son- 142: 737-7.54. 

oran Desert Plants: An Ecological Atlas. L iiiv. Arizona Wilson. B. I). & J. F. Addicolt. 1008. Bogiilalinn „| ', mi- 
Tyre, A.' J. & J. F. Addicott. 1993. Facultative non-mu- behavior responsive to -eh-dive abscission of flowers? 

in.,!,., hohaviour l.v .,„ obligate mitualist-cheating bv <>ikos 81: 109-118. 

yucca moths. Oeeologia 04: 173-175. ' Zeller, P. C. 1875. Beitrage zur Kenntniss der nordamer- 

Van Devender. T. B. IO')(). Fate Oualeniary vegetation icanischen Nachtfalter besonders der Microlepidoplcr- 

aud climate" ol the ' liihiiahiiau d. ■ ... I I - , - ■•"". "'00 M.lheihing. Verb. Zool.-Bot. Ges. Wien 25: 

and Pp. !<«., I'l'. m J. I . F.-l.incom.. T. B. Nan 2"7 MA), pis. 8-10. 

Devender & P. S. Martin (editors), Packrat Middens: Ziv. V A J. L Bronstein. 1996. Iniertile seeds of )ucco 

The Fast 10.000 Near- ol I'.iolic Change. I niv. Arizona schottii: A beneficial role for the plant in the yucca- 
Press. Tucson. vucca moth mutualism? Evol. Ecol. 10: 63-76. 
Yerhoek. S. 1998. Agavaceae. Pp. 60-70 in K. Kubitzki Internet Bcsource 

(editor). The Families and Genera of Vascular Plants Pellmvr. <). 2002. I Yodoxi. lac. The Tree of Fife \\ob Project. 

IN I lo\V< • ■ 1 1 P III % 'l« I I I II- I ill III I (< I I illl|> I ll ,!,,!..) 

Orehidaceae). Springer. Berlin. Meolepidoptera) 


types. Tetraploidv [2n In) is documented for ill.- first tim.- in -oinr | illations of //. , 

ic Smith American representatives of ill.- -.-mis an- all' on x = 1 in contrast |„ more o 
ices anion" Kuropean taxa (2n - (>. », 10. 12). This pattern <>l -rcaler cvtological unilori 
h<- New World (.-a. :><> known. .'12 mm .•onnt.-.l chromosomallv I suggests thai th.- group w 

eciesdevel biogeography in ill 

al.. 1953: Cant. 1981). little is Chromosome counts as well as .letail.-.l karvo- 

licsc same processes in llie flora logical studies have been investigated for Kuropean 

To help reveal these dynamics specie-, ol llypochaeris (Parker, 1976; Mugnier & 

cgions of southern South Ameri- Siljak-Yakovlev. 1987: Hargl.i ct al.. 1989; Siljak- 

ca. we have begun a series of investigations on //>- Yakovlev et al.. 1994; Cerbah et al.. 1995. 1998a). 

pochaeris L. (Asteraceae. I.acluceac; l/ypoi hoc/is which appear to possess symmetrical karyotypes. 

L., orthogr. var.). This is a genus of approximately and which show general correlations of different 

60 species, with 9 in Kurope (DeFillips. 1976) and chromosome numbers (2n = 6. 8. 10. 12) with tax- 

ca. 50 confined to South America, many of which onomic sections. Molecular phvlogenetic studies 

are localized in the Southern Cone (Bortiri. 1999). using ITS regions among these Kuropean taxa (Cer- 

Hypochacris is a good genus to extract information bah et al.. 1998b). and including four species from 

on modes ol spoliation and patterns of biogeogra- South America, have also revealed useful insights 

phv because ol lli. ■ large si/e and small number of on broad evolutionary patterns within the genus 

chromosomes (Stebbins. 1971). the diverse mor- that correlate, to some considerable degree, with 

phologies of the taxa varying from small acaules- previous sectional limits. 

cent alpine forms |<> broad-leaved large he.i. led Despite (It. • large number of species of Hypo- 

speeies. and distributions in very different ecolog- cluicris from South America, thev have been poorly 

ical zones ranging from sea level to over 3000 m studied kary ..logically. Chromosome numbers. 

(Lack. 1979). Apparent rapid and recent special ion sometimes accompanied by photographs, have been 

1 The stndv was support.-.! l.v I \\ V project No. PI mVvliK ) to T. K Sluessv. 

2 Deparlment of Higher I'lanl Svsle malies and Kvolulion. Inslilule of liolanv. I niversilv of Vienna, Rennvveg If. 
\ ll).;i) \,r,in;,. Vuslna. I.mI.sI liessv (."univ 

1 Institute for Systematic Bolanv and Botanical Carden. I niversilv of Munich. Vleii/inger Slrasse (.7. i',1 )(>:'»!; Munich. 

■* Departamento de ISotamca. llniversidad de Concepcion. Casilla lo()-C. Concepcion, Chile. 

\NN. Ml— 'I HI Uni. (,\l!l>. "0: :»(, <>/.. L>( K).",. 

Volume 90, Number 

; Chromosome Reports 

//. rlnllrnsK X / 
//. clullemis X / 



risel al. (195! 

i); Parke 

.(1971): 1 

. |l«>7«>) 


«-l al. « 1 ■>— Jilz 


kovlev el 

et al. (1995): Cerbah et al 

. (1998a) 


11992. 1998) 








sicunrphala (A. Gray ex Wedd.) 

. taraxacoides (Walp.) Benth. & H 

. /wwrftei (Hook. f. & Am.) Cabre 
variegata (Lam.) Baker 

Siehbi.isel al. (1953) 

Sefmack * Cnvas (1947) 

Moore ( 19» I) 

Saez (1949); Siljak-Yakovlev et al. (1994); Cerbah et ; 

(1995); Kuas et al. (1995); Cerbah et al. (1998a) 
Wulff (1992, 1998) 
Diers (1961) 
Saez (1949); Siljak-Yakovlev et al. (1994); Cerbah et t 

(1995); Kuas el al. (1995); Cerbah et al. (1998a); 

Wulff (1998) 
Wulff (1992) 

Cerbah et al. (1995); R 

uaset al. (1995); C 

erhah et 


Diers (1961) 

Turner et al. (1979) 

Ruasetal. (1995) 

Olsen (1980) 

Turner et al. (1967) 

Stebbinset al. (1953); 

Diers (1961) 

Parker (1971) 

Saez (1949) 

Ruas et al. (1995) 

a* by Bortiri (1999) 

ies on Hypochaer 

is, therefore, the ] 

i of this 

- paper are to: (1) 

summarize all av, 

ailable } 

its for South Ame 

rican species of 

r Hrpocliaeris; (2) report original con 

nts for t 

a\a col- 

f let-ted in Chile a 

ru\ other Andean 


( Ar-e.i- 

. tina. Bolivia. Kcuador, Peru, and Venezue 

(3) comment on 

Ihe chromosomal 



y merited within South America, parli 

cularlv i 


erence to the already well known 

eWolojrical pic- 

ture of European 

i species. Several 


(lotls ol 

- each taxon have , 

>ftcn been analyzed so tli 

al infra- 

1 Treated a 

unifoim Ictrjploi.U known (//. incana, Moore. 
1981; H. stenocephalia Slebbins et al., 1953; Diers, 
1961), and two cases of infraspecific telraploid\ 
(i.e., H. chrondrilloides, Wulff. 1998; H. sessiliflora, 
Olsen, 1980, both 2n and 4n reports). 

To set the sta,L« l< ddi n;a 'In In i . n 

Table 2. New chromosome counts of Hvpochiieris from South America [all plant 
iomaii numerals) unless indicated otlierw ise|. Counts are mos||\ from mitotic prepaia 
md I- 1 pollen mitosis) are indicated In *: populations for which holh diploid and haploi 
vith **. Abbreviations of collectors: CI? = C. Bae/a; DC = I). Crawford: JC = J. C 
ituessy; HF - H. Forlher: V1VV = VI. Wcgend. Vouchers on deposit in Wl and CONC (, 
md Lope/) and \l (all other collections), t t == fii-l icpoillsl lor la\on: + = new chr 

Ta\on. voucher, anil ehromosoim iiiicil,. i 

+ +//. apargioides Hook. f. & Am. 2n = 8 

VIII. Parque Naeional Liguna Fa l.aja. 25 Jan. 1W1. TS & DC 1.1170: Trancas. 30 Jan. 1998. TS 15481. 15485. 
13 Jan. 1999. 7'.S rt „/. /55-W*. Nail,' llermoso. /.T.^/. /55.5/. I\. Kesena Nac.onal Vlalalcahuello. 12 Id, !«)«>». 
77S rt «/. 15509**, 3 km VV of Fonquimay. 13 Feb. 1998. /;>.5/.?; 10 km F of Puente Fonquimav. '/'.S e/ «/. /5.W5*. 
17 km F of I'uente Fonquimay, /5;>/6. 29 km F of Puente Fon,|u,maN. 15517. 8 km F of Chilean Aduana. 13 Feb. 
1998. /.55/«; 3 km VV of Fnnquimav. 19 Jan. 1999. TS & CH 15570*. 8 km F of Chilean Aduana. 20 Jan. 1999. 
15588. 3 km W of Chilean \duana. 1550 1**. 25 km F of lonquinu.N. 15505**. 17 Id, 2(HH). 15500. Volean IJaima. 
21 Jan. 1999. 75602*. Volean Villarrica. Piedra de Aquila. 22 Jan. 1900. 15012*. to I, .-una lluinhuea, 23 Jan. 1999. 
15021*. id to Fonquimay. 10 Feb. 2000, I5809A. 
+ +//. vlarionoides (Bemy) Keiehe 2h = 8 

He-ion Melropolitana de Santiago. 6 km VV of l.a Parva. 18 Feb. 1998. TS & PS 15527. 2 km VV of U I'arva. 
/5529*. /555/*. 
//. elata (Wedd.) Griseb. 2n = 8 

BOLIVIA. Fa Paz. Puerto Perez. Fake Titicaca. s.d.. Kams Hyp-48. 
+ /7. meyeniana (Walp.) Criseb. 2n = 8 

PERU. Dept. Cajamarea. Prov. Conluma/a. Bosque de Cachil. s.d.. MW 98/554. 
+ 11. meyvniaiM (Walp.) Criseb. +2« = 16 

PERU. Depl. Tacna. Prov. Taranta. S of Volean Tutupaca. 20 km above Candarave. sal.. MW & HF 97/688. 
+ +77. palustns (Phil.) De Wild. 2/1 = 8 

VIII. Valle de Las Nieblas. Termas de Chilian. 15 Jan. 1999. TS el al. 15500*. I\. 20 km F of Chilean Aduana. 20 
Jan. 1999. TS & CH 15588*. Volean Vdlarrica. 22 Jan. 1000. 15000. 15007. rd to lacuna lluinliuea, 23 Jan. 1999. 
15622, Volean Casablanca. 21 Jan. 1999. 15628, 15629. 
11. radicata L 2/; = 8 

VIII. Cerro Ponpon. 21 Jan. 1998. 7.S I5I50H. I5I50J. Parque llualpen. Coneepcion. 15451; 2 km VV of Angol. 24 
Jan. 1998, TS & DC 15151, 2 km into Parque Naeional \ahuelbula. 15158*. Pan,ue Naeional Nahuelbuta, 75467, 
Piedra de Aguila. 15405. 15466, 17 km F of Anlucn. 25 Jan. 1998. 15108. 15100*. Parque \aconal Faguna Fa l.aja. 
15471*, 15478: Fola. S. huapa. 27 Jan. 1008. TS n al. 15477. Termas de Chilian. 31 Jan. 1998. 15487. IX. 12 km 
F of Curacaulfn. 12 Feb. 1008. TS el al. 15500. 5 km S of Funquunav. 12 Id, !<><);;. 1^08. 8 km F of Chilean 
Aduana. 13 Feb. 1998. 15519; 14 km F of Cherquenco. 21 Jan. 1900. TS & CH 15597*. 1 km VV of Pucon toward 
1999. 15004. 11 km SF of Agnus Calientes. 21 Jan. 1000. I5020-. licgion Vlctiopohtana 

KCIADOR. I'irhincha. Norn,. NAY of Ouiio. s.d., JG Hvp-36. 
+ +H. spalhulata (Remv) Reiche 2n = 8 

IX. N of Pucatrihue. 25 Jan. 1999. TS & CB 1 5633*. 
'/. stenocephala (A. Gray ex Wedd.l kuritze 2n = 16 

PERU. Dept. Puno. Prov. Puno. Ruins of Sillistani, s.d., MW 
+ +//. tenuifolia (Hook. f. & Arn.) Griseb. 2n = 8 

VII. Prov. Talea, Laguna del Maule. s.d., JG s.n. VIII. Prov. 

1 In-I.n.. 

31 Jan. 1998, TS et al. I54H6. 1549 
, TS & CB 15577-3*: Volean, : 

- (!«■ Chilian, Catania del Diablo, s 
eb. 1998, /5 792a 75498, /5505. 
1999. 756WA Volean Villarrica. 22 J 

. 7:S el al. 15510. 

I thrincioides (Remy) I 

, TS & DC 15456**. 

specific euploidy or dysploidy might be revealed 
and better interpreted. 

omy and nnnif I // ■ liaeris in South 

America needs comprehensive revision. Despite 
ibr \er\ lielpfnl I real merits on \ i gent inean species 
by Cabrera (1971. 1974. 1976. 1978) and more re- 
cently Bortiri (1999). there still remains confusion 
in proper limits and correct names for many ta\a. 
! these important 

l ; it n in. I' 

!!! mi 

Keulgen staining with Scbiff's reagent following 
standard methods (Fukui & Nakayama, 1996). Root 
tip- and Howcr bud- were washed in distilled water 
to remove the fixative, hvdrolvzed in 5N HC1 for 30 
min at 20°C. and washed and stained with Schiffs 
reageiil in darkiio>- lor |-_ } hr. Sijiiasli preparaliotis 
were made in a drop ol 15' y acetic acid. After cov- 
erslip removal on drv ice. preparations were dried 
for 24 hr. at 37C and rnonnted in DPX. Chromo- 

Bud and seed materials of Hypochacris were coi- 
ned in South America (Table 2). Collections were 
f populational sample- m all in-laiiee-. \nuchers 
re on deposit at CONC. M. and WU. 

Root tips were prctrcaled with 0.1% cold 

2 hrs at room temperature in darkness, f'r 
I (ethanol : acetic acid) for 24 hr.. and storei 
- al -2()°C. Kor mciotic chromosome c 
wer buds were fixed in modified < atno\'- -o 
3:1; chloroform : absolute el ha i rol : glacial 

ed from 85 populations of 15 species of Hypochaer 
'- from South Vnierica (Table 2) including f i r — I re- 
ports for nine taxa: //. apargioide: 
H. palustris, H. 

ilr-.. II. -.palhnlala, fl. tenuifolia. and //. thriiu ioi- 
(Irs. Ml counts give 2/r = I!. with iiilraspccilic let 
raploidy (2n = 4x = 16) al-o documented in // 
s,;>r:»ntr<ir and //. tenuifolia. A new let rap loid pop 

These new data, in correlation with pre 
published reports for Hypochaeris in South A 
(Table 1). yield 32 species (of ca. 50) now e 
from 138 populations. A brief combined ai 






6 Jit_ 

7 "— 





12 "* ' " _\ 



9 ' 




shows 2.") uniformly diploid (2// - 8) species, two 
species uniformly lelraploid Via 10; //. incutm 

;md //. slenocephala). and five species wild both 
diploid and leh i| li - I 

//. nieseniana. II. seorzonerae. II. sessiliflora. and 
i) Out spe< ies. //. gardneri. has been 
reported previously as n = 5 (Coleman. 1908), and 
another. //. sessiliflora. has been documented as // 
= (Turner et al, 1907). 

because of the strong uniformity in chromosome 
levels anions South American species of llypo- 
chaeris (all n = 4 or n = 8). further eonunenls on 
these two deviating reports are in order. The n = 
report (Turner et al., 1907) for H. sessilifolia con- 
trasts with more tvpical /; = 1 (Jansen o\ Slne-s\. 
1980) or n = 8 (Olsen. 1980). Kxamination of the 
voucher of //. sessilifolia reported by Turner et al. 
(1907; Wurdack 437. TKX) with camera lucitla 
drawing of meiotie bivalents attached. suggest-, llial 
n — 4 is probable, with some homologous ehro- 
niiNiiiifs being pulled apart in nieluphase l/carb 
anaphase- I earlier than ihe oilier-. Turner, in fact. 

"In hindsight and with more thought this could be 
n = 411!" The voucher for the count of n = 5 for 

//. gardneri ha- not \el been located despite an 
herbarium search. It is worth mentioning thai die 
Kuropean H. glabra with n = 5 is also known to 
be adventive in South America (e.g.. Matthei. 1995: 

loilui I')*' 1 - 1 ! ! . i i phological vari- 

baeris, and heru e difficulties 
nub iilicalion, an examination of the voucher 

this represents a new aneuploid level in nalbe spe- 

Karyotypes of all newly analy/.ed South \meri- 
can species of Hypoehaeris are bimodal anil asym- 
metric and similar in overall morphology to kar- 
yotypes of species analyzed prey iouslv (f'igs. I -1.1; 
see also references in Table 1). In general, they 
consist of two large and two small i 
pairs. Two of these pairs, one large suhleloceiilrie 
and one smaller acrocentric, appear to bear satel- 

Dcspile this gem ra a I i ,. karyotype and a 

reasonably stable . Iiromosoiiii i amber (2/; = 8). 
karyologieal differences among taxa do 

exist. The main differentiation of karyotype con- 
cerns chromosome size changes anil the presence 
of satellites (Figs. 1^0. 8-15). A detailetl analysis 
of karyotypes of South American Hypoehaeris and 

their evolutionary importance will be reported else- 

Although the <'inphasis in this paper is on new 
cytological reports of native South \ 1 1 1< -r i< an spe- 
cies of ll\po< hoars reports for the in- 
hodueed //. radicala are also included. This taxon 
is abundant in Chile (Matthei. 1995: pers. obs.), 
growing from sea level to over 2000 m and often 
found intermixed with native species, because of 
the possibility of hybridization between //. radicala 
and native congeners, which could eniiliise inter- 
prelalions of pattern-, and processes of evolution, 
sampling of this taxon was also included (Table 2). 
b'csidts show the typical ehroino.soiiie level (2n = 
8: Fig. 7) for this species and symmetric karyotype 
and no irregularities, effectively excluding hybrid- 
weedy species of the genus. //. glabra, (distinctive 
with 2n = 10. Stebbins et al.. 1955). is much less 
common in South America, al least win i n 

examined ev lologically during this slmb. 

Tetraploidy (2// = 4v = 10) has been previously 
i. 'ported for lour South \meriean 1 1 \ pu< Itaeris spe- 
cies: H. chondrilloides (Wulff, 1998), H. incana 
(Moore. 1981). //. sessiliflora (Olsen. 1980). and //. 
slenocephala (Stebbins et al.. 1955; Diets. 1901). 
for //. chondrilloides and H. sessiliflora. both ploidv 
levels (2.v and h) were reported. The present paper 
atlds three more species. H. meyeniaaa (big. 5|. //. 
seorzonerae. and //. temdfolia 'Figs. 15, 14), in 
which lulraspet ilic polyploid cytotypes are known. 
Karyologieal dala provide some suggestions on the 
mode of origin of these lelraploid cvtolv pes. The l.v 
races in //. leniuj • id// ■■ > \ • 

karyotypes consisting of lour e<|iial-sized sets of 
. probable, therefore, thai these 
polyploids are of aulopoly ploid origin. We suspect 
this to be the case also in H. sessiliflora (Fig. 8) 
and perhaps also foi //. clmndrilloiiles thai contain 
both 2.v and \\ cyloly pes. although no detailed kar- 
yotypes arc available for ihe filler. I riderslanding 
the origin of the apparently uniformly lelraploid 


] 15 Mitotic chromoso. 

tic- ,.i Soiilli \nierican species ol 

IH haeris Scale bar = 

5 |JLItl 

1 H < 


TS e 

t al. 15565.— 2. H. apargioides 

, TS 15485.— 3. 

II. clarionoides. 

TS & PS 15527.— 4. H. , 

Ilk II,,, 

wyeniana. M\\ Wl/551.- <>. II. paluslris. IS r, „ 

1. 15566.— 7. 11. 

radicala. TS & PS 15533. 


■ra. .11. 


38.-9. H. setosa, JGa 1048.- 

des. JG Hyp-36, 

— II. //. spalhidala, TS & (.li 

12. //. 

11 A /// <>7 ll'> 

13. H. tenia folia 

,TS&CB 15605.— 14. H. lenufr-b.. '> 

el al. 


-15. //. 

ilmmioides. TS 154501. 

SlCIIOII'plldld .111(1 pOss||>|\ 

Soulh \nieriean species nl Ihpochiieris. special at- 
tention was given to searching loi ineiolic irn-gu- 
larities that might signal hv hi idizalion. VIeiotic 
counts -houeil cm 'ln-i\i-l\ lour regular bivalents in 
all examined populations. IYe\ ions reports on mei- 
otic chromosome numbers ha\ < ■ shown similar re- 
sults (Wnlff. 1992. 199}!). Only experimentalK ob- 
tained hybrids between //. chillensis and H. 
mri!<>j>t>'">:::-, n showed si. me irregularities in ho- 
mologous chromos, ■ pairing with III ami o.-ca 

sionally 311 + 21 (Wulfl. 1992). ['resumptive nat- 
ural In buds ol //. eiiilletisis and //. ;;;,, ;>!;;•!>) 

var. dlhijlnni. however, showed regular bivalent lor- 
malion sii-i;es|j M M 1 h. p..ssibilil\ ol recent origin of 

these species ; 1 1 1< I high degree ol karvotvpe- simi- 
larity (Wulfl", 1992). Two instances in our'own field 
collections suggested possible hv bridizalion based 
upon morphological lealiues: //. aptir^ioides X //. 
irnm/olia \TS el al. 15554) and //. palustris X H. 
tenuifolm (TS & CH 15607). Because some differ- 
ences in chromosome si/e and presence of satellites 
are obvious in the case of//, tipurgioidrs I big. 2) 
mil H. lei // lit, (Kig. 13). hybrid individuals be- 
tween these taxa should be detectable. These po- 
tential hvb rids were loiind to be diploid, however, 
with no differences in size of chromosomes of the 
two haploi.l se|s. bi-ed on k . 1 1 \ • .morphology, pos- 
sible hybrid origin ol these populations is unsup- 

In contrast to broad cytological diversity among 

nine Kuropean species of llypoclutcris \2n = <>. o. 
10, 12; Mugnier cK Siljak-Yakovlev, 1987: Cerhah 
el al.. I'»''.".ai. . vlologieal uniformity of the iNevv 
World members of die genus suggests several as- 
pects icgai. hug evolution of the group. First, be- 
cause the karvotvpe of Soulh American taxa rep- 
resents only one general pattern, in contrast to 
several found anion- biiiop.Mii species, it can be 
hvpolhes|/ed that the former evolved from out of 
the latter. This hypothesis is also corroboialed bv 

(ITS: Cerbab el al.. I9D ( ;|>: IT Samuel et al.. in 
prep.) and chloroplast {lrn\.: K. Samuel et al.. in 
prep.) genes. Second, the abundance of South 
American taxa representing so main different 

alions. such as eiiplnidv. ,\\u\ marked 
karvotvpic change. 1 1 \ pneluieris represents a genus. 
therefore, in which perhaps more minor amounts of 
karvotvpic. and eerlaiiilv genetic, change lias ac- 
companied speeialion. because ol this mIu.iIioii and 
due to the large and few chromosomes, the genus 
in South \merica provides an excellenl opporlunilv 
to map tin genome and determine minoi kaivolvpie- 
changes during speeialion. within context of the 
relativelv stable 2/; <". karvotvpe. as has been 

done successfullv for Helianlhus In North \merica 
(Rieseberget al.. 1995; Hieseberg. 2001). 

1971. llvpo, horns. \y:\<>7-min M. Y 
Correa (editor). Flora I'alagi'.nica. Parte- VII. Composi- 
tae. IYT\. Mucins Aires. 
. 1971. Hypochoeris. Pp. ">I2 Si?, in \. Burkarl 

leilil.irl. Mm. i i .|e Knlie Kins i \igentiiia). I'arle 

Huhiales. Cueiirbitales, Campanulales (lneluso Com- 
puestas). IYTA. Buenos Aires. 

. 1976. Vlaleriales para una revision del uenem 

Hypochoeris I. Hypochoeris chillensis (If. H. k.) Hien.n. 
Darwinian.. 20: .'512-322. 

1978. Hvpochoens. I',.. r>7I-686 in A. I.. (la- 
's Jujuy, Parte X. 

, Flora de I* Provi 
. INTA, 

Cerhah. \1.. J. l.niilau.l. I'.. I.o.l.-ll.- ,X S. Sil jak-Yakuvlev 
1995. Genome size, fluoroehrome handing, and karvo 
tvpe evolution in some Hypochoeris species. Genome 
38: 689-695. 

..I. Coulau.l K S. Mhak-Naknvlev. 1998a. rl)N/> 

Ihpo h, m l\s|,i „, „ i | ||, ii ditv 89 512-518 

, T Semza-( lilies. VI. F. Jihier. 15. Fejeune i\ S 

Siljak-Yakovlev. 1998b. Molecular plivle.geMiv of the 
genus Hypochoeris using nil. < transcribed spacers n 


15: .' 

Charrier. G. (1. (Felitor). 1997. Fl Alliplano: Ciene 
Cheruhini. (1. 1982. N.i.ii.-m ele- .rnmose.mas ele algi 

War. I no. Y,c <:„ v , 22: 215-25. " 

zilian Ciinipositae. Blmde.ra 70: 228-21(1. 
I)<-Fillips. I!. \. |'»7o. Ih/„ lt Imrns. I'p. :it»!J-3l() i 
C. Tutin. V fl. Ilevvvooel. \. A. ({urges. I). M. Vl„ 

ats is slronglv indicative of rapid and recent 
ive radiation. Third, the' mechanisms of spee 
ccoinpaming ihis explosive evolution have' < 
eit been driven by gross maero-cvtologieal 

Flora Fureipaea. ^ 
)iers, L. L961. Dei 

. Press. 

;. & S. Nukavania (Fditors). 1996. Hani Cliro- 
les. Laboratory Methods. CBC Press. Boca Kalon. 
1981. Planl Speciation. 2nd ed. Columbia L niv. 

■ Hvpochoeridinae (As- 
and the- Southern 
here. Pp. 26.5-270 in K. Larsen & L.-B. Holm- 
(editors). Tropical Botany. Academic Press. 
lalthei. O. I')'),"). Manual cle Las Male/as ((lie Creeen en 

Chile. Published l>\ the author. Santiago. 
1oore. D. M. 1981. Chromosome number .,| I u.-ian 

aumo-penn-. P.ol. Sue. lirot.. Ser. 2. 53: 995-1012. 
lugnier. C. & S. Siljak-V.kovlex. I9JS7. Kan , .logical 

(Composite). Cam.lo-ia 10: 319-323. 
Ike,.. |. 1980. /„: Chromosome numbers reports L\\ II. 

Taxon 29: 347-367. 
'arker. J. S. 1971. Ihe Control of Becombination. Ph.D. 

Thesis. I niversity of Oxford. 

i of Hypochaeris 

1976. Th. 

Kuas. C. F.. P. M. Una-. \. I. Mal/mbad,,,. ( ., lio^. < , 
Bernini & \. I.. I.. Yanzela. 1995. CMo^-nelic Mudie- 
of some Hypochoeris species (Compositae) from Brazil. 
\mer. J. Bot. 82: 369-375. 

F. A. 1949. Estudio citologico comparative tie al- 

\meiiea <l.| Snr." Lilloa 19: 97-101, 

ick, B. & G. Covas. 1947. Fstudios eariologicos en 

,.lilas. Haumania 1: 32-41. 

J. Bot. 72: 1496-1502. 

Plants. Ldward Arnold. London. 

. J. A. Jenkins X M. Walters. 1953. 

and phvlogenv in the Compositae. tril 
I'niv. Calif. Puhl. Bot. 26: 401-430. 

omb. VS.. K. I.. Chambers. I). \\. k>lms 

. I!. I... 

I X .1. 

1967. Chrt 

nctics I II: 1163-1 

Aim. Missouri Bot. Card. 54: 172-177. 
. J. Bacon. L. Urhatsch & B. Sim 

Chromosome numbers in South American 

Amer. J. Bot. 66: 173-178. 
Wulff. A. F. 1992. Hihridizacion natural entre especi.s 

Sudamerieanas de Hypochoeris ( Asleraeeaei. Iliiuiin- 

ana 31: 167-171. 

. 1998. Lstudios cariolojzicos en \steraceae. \ 1 1 1. 

I)arwiniana35: 37-13. 


\l!STI! \CT nun-ion- .in- uicrra-inglv recogui/.eil a- .1 kev problem lor ill. conservation ol biological iliversitv. Ilimcir 
the scientific recognition that some species, when inl roilnceil outside thru native range, rails. ' a decline in indigeiiou 
species, goes hack to at least the writings of Charles Darwin. In the l ( ). r >()s another British biologist. Charles Klloi 

Over the last 20 \ears biological invasions have ,. , • .. . /lt ,<,-,, .,,.,, ... 

. " . . . ^ . . . lerenl slimes, lie wrote Darwin, \iW): .5 1.5 : "In 

•' in d ! ■"" 1 i" itti 1 1 fri 1 oli gists him the . . . , .. . .. . ,. ... 

a|)id 11 1 I tl lentihc and popular arti- ... \. . . . , ,■ f . . , 

.' ... . . ' ' . , llir lace ..I the ulobe. llie ln-l ureal I. in uhicli 

I I i I 1 1 hi 1 I 11 1' 11 

. . -iMrvC- I - .- ll ,1 ,i -1)1 . 1; ic - II II. 1 m no the ill- 

onclude that invasions are a rela.ivelv new phe- ^.^ „,. ^ mha|)il;mls ( , f va| . |((Us |()ns ,. ;m 

■omenon In la,,, as humans his. began .0 move |m . mu iH ,. M . A ,,, |, v ( ., im , l;l | ;1114 , „„„., ,„.._ 

'T" 1 I 1 "" ™rtlMlH-> l«H,k Ian, r plants a.,.1 an- , ( . al ( , )M(jit|(ms . Th( .,. ( . |s ,,„,„, a ( .| imat( / ( , 

mals w.ll, them lor use as food, nieiluuie. or tech- ( . (m(|i|i(m m „„. ()|(| NW|(| M] ^^ ^ 

.ology (rntz, 1994). We know that ina.xe was found a||rM „, „ ie v . w ,„.,, |)()u „,,,., ,,,,.,.'„,„, 

uropean explorers l|lf ,. organi(i 1)( . ()( , u( 

, \orlh ' 

1 the New World. 

Today we understand fully that the diversity of s 

IN ' U " '" lM> , '" ■'■ |N I" 1 '- eies on on. plane, i.. in part, the result „l eonti- 

uitive Ameiieaiis 111 easteiii Ninth \1ner1ea suggest || ];|t S() m t r igued Darwin and all liiologists since. 

uit it was likely traded and earned by indigenous However, this geogra|)h\ sets up one ol the most 

eople to the region prior to the arrival ol hum- |)m f m .„d threats to the'Kurth's diversitv. Because 

cans. And some species are known to --hitchhike" llu . continents differ more in species than climates, 

long with human travelers. The Polynesian rat each continent has the potential to provide invaders 

Rattus exulans) was brought to islands bv to ( , t | ier places. Humans are a potent force for as- 

le early Polynesians (Merlin K Juvik. 1992). most sisting that invasion. 

Charles Darwin provided the necessary context friends and enemies. Darwin realized that this eco- 

>r understanding the biological invasion problem logical context provided the ultimate cheek on 1111- 

lmost 150 years ago. One of the observations ilia! fettered population growth. Thus, there is another 

liinulal.-d Darwin In conceive of evolution bv nal- important perspective to tin invasion problem: vv<' 

for Urban Horticulture. Iniversitv ol Washington. Bov .V> 1 1 I fT. Seattle. Washington 9B195. U.S. A 
of North Carolina at Chapel Hill. Chapel Hill 

uo>. <>o: (.1 (.*.. 2011.T 

Reichard & White 

often transport species without their revolutionary llie knowled i wering these ques- 

eonlext. This may result in poor performance (e.g.. lions? Kaeh symposium consisted of distinguished 

if a plant lacks a pollinator), hut il ma\ als suit biologists, mosl ..I whom worked in related areas of 

in the potential for rapid population growth (e.g.. if study, though not spccificallv invasions, attempting 

the species is free from natural enemies or other to answer these question- from their understanding 

constraints). Darwin (f859: 370) also recognized of the biola of their continent. These volumes (e.g., 

that some introduced species may threaten native MacDonald el al.. 1986; Mooney & Drake, 1986) 

species. In his chapter on geographical distrihu- serve as "stale of the knowledge" documentation 

tions he reported, ". . . many Kuropean productions and are notable for their general lack of specific 

cover the ground in La Plata, and in a lesser degree dala on invasions. This was I.e. a use there were few 

in Australia, and have to a certain extent beaten studies specifically on invasions from which to draw 

the natives " This may be the first scientific conclusions. Reading the chapters, il appears that 

Although there were studies ol invasions in the the course of the symposia and book production. 

years since Darwin, biologists and ecologists did These books fueled the newly emerging field of in- 

not focus on the magnitude of the growing problem vasion biology, as a generation of graduate students 

until f958 when Charles Klton (( |95«| 2000) pub- read the chapters and seized the opportunity to at- 

lished a book that has come to be regarded as the tempt lo answer the questions derived during the 

seminal volume in this field of study. The Ecology SCOPK process. 

of Invasions In Animals anil Plants is a slim book In the approximately 15 years since the publi- 
borne from his observations as an animal ecologisl. cation of the SCOPK books the field of invasion 
Three 1957 BBC radio broadcasts he developed on biology has not only influenced the content of many 
the subject of "Balance and Barrier" apparent!) biological journals, but it has also developed its 
stimulated Klton to write a book on invasions. own journal (Biological Invasion*, Kluwer Press), 
aimed at a lay audience, that laid out what are still its own set of terminology, and its own set of raging 
the fundamental issues in invasion biology: thai scientific debates. \\ liile the SCOPK volumes pro- 
each continent has its own unique Nora and fauna. \ ided the initial fuel for the scientific field of in- 
that human migration and trade were breaking vasion biology, it was the mounting evidence of se- 
down the barriers that had led lo the uniqueness of vere environmental degradation that focused the 
the biota, and thai this breakdown ol the barriers attention ol so many people on invasions. h>r in- 
could have severe consequences for the mainte- stance, invasive species were identified as signifi- 
nance of diversity. The book is amazing in its pre- cant threats to biodiversity at more Nature Conser- 
science and yet was in large part underappreciated vanev preserves with completed conservation plans 
until the 1980s, when it became widely recognized than any other type of threat, including develop- 
that the invasion of non-native species was one of ment. lire suppression, and altered hydrology. In 
the biggest threats to naturally occurring species fact. 94% of those sites responding as of the sum- 
and ecosystems. rncr of 2000 listed invasive species as a serious 

In 1982 the Scientific Committee on Problems of problem (J. Randall. The Nature Conservancy, pers. 

the Knvironment (SCOPK). a committee of the In- comm.). In 1998, a study found that invasive spe- 

ternational Council of Scientific Unions, met in a cies were second only to habitat destruction and 

general assembly in Ottawa. At that meeting the fragmentation in threatening endangered species in 

invasive spread of plants, animals, and micro-or- the United States (Wilcove et al., 1998). This study 

ganisms introduced by humans outside their native reviewed listing information for species proposed as 

ranges was identified as a problem of global con- endangered or threatened under the Kndangered 

cern, amenable to interdisciplinary synthesis. This Species Act and found that 49% of the imperiled 

around the world in the mid 1980s and resulted in cause of invasive species, 
two important books (MacDonald et al.. 1986; Moo- 

ney & Drake, 1986) that both address three mi- KlTUKK CllAI.I.KN<;KS 
portant questions: (1) What are the factors that de- 
termine whether a species will be an invader or With the recognition that invasive species are 
not?; (2) What are the site properties that determine one of the most serious conservation concerns to- 
prone to. or resistant to." invasion?; and (.T) How the problem is getting worse very rapidly. Global- 
should management systems be developed using izalion of trade and advances in technology mean 

Missouri Botanical Garden 

thai species arc moving around Ilic earth more ire- ducing llie entry and impact of such species. As 
quently and are coming from some places, such as scientists, we will need to ensure that our science 
China, that have been closed oil from most ol the is not only viewed |>v our peer* a* is traditional 
rest ol the world during the last several decades. but that agencies managing invasive species are 
For instance, the Asian long-horned beetle {A no- also aware of our lindings. As scientists we must 
plophom glabripennis). which is native to China learn Irom die past, examine the present, and plan 
and considered to b«- a pest there, was first detected for the future. As the field of invasion biology 
ill the (ireeiipoiiil neighborhood of l!r....klyn. New moves ml,, i|> adolescence, it will continue to test 
York, in 1W0. The I uilcd Stales Department of theories basic to ecology and to form new hvpoth- 
Agriculture subsequently determined that the in- cses to address the novel situations that arise fol- 
sect arrived in solid wood packing material on lowing the introduction of new' species. These dis- 
goods imported from China. Not only arc new path- eoveries. if implemented in management and policy 
ways opening, but the |ourneys that used to take practices, ma\ play a substantial role in lessening 
weeks by ship may now lake hours by plane, allow- environmental degradation through invasions, 
ing more organisms to survive die trip. Trade agree- 
ment* in I i i| , v .1 \ biterature Cited 

can Free Trade \o„v....-nl (,ee N \ I T \ website. "T'"" \" ■ ."^ !''"' *!""'" "' 1 S ' ), ' ,lrs l,v Mt ' ;ms ,,f 
(w\o-* i i i ■ i w^. i i ,v ,, Natural Sell • ti i i n rnel li i ■ lelov 

2002). launched .,, January 1091. and the World ,,; ,. (; s ,,„-,,, , (MM) .,.,„, ,... o|( ^ n| |nui> . ^ ,„ 

Trade Organ, /alio., (see WTO webs,|e. 2()(>2|. Xni.nak : ...<l Planl>. |{e.„-i„t. uiih a'new f«,.-.-«.„-,l I,; 

I Wo. limit the restrictions I). Sind.e.loff. I ni\. Chic 

ies can pla, 


Fritz. G. L. 1994. Preco 
ssp. a 

goods themsebes. in the case of YlaeDonald. I. \. W 
id animals, or as contamiiiaiils of 'he Keolngv ami 

•h as insects found in shipping dun- 

■n. Africa. Oxford I nix. Press. Cape Tow ,,. 
.•Hit.. M. I).K JO. .lunik. \*m. Relationships am., 

Pp. -W7-OJ I . i P s .„„ . , .' \y slniil, 'j. T . T un ' is 

)f the U.S. economy, will, current figures of about ^""^ Alit "» ,>,u »« Rasmus in Native Keosvstems 

BI 10 bi n a vea. (up from about $2U billion in J^'"' 

I OHO). 0\er l(>.5 million international visitors en- Unit. 

cred the I niled States in I WO. with a projected Moonev. I 

mniial growth of \\-\ c /< (Dogged. 1007). although [ n * a ^ 

ourism may decline over the next several years a 

i result of fears regarding safety in traveling. Trav 

ders often inadvertently carry invasive species as OTA-F-565. U.S. Coven 

litchhikers on their person or property, but they iv . ,on < l)( • 

. ■ , ,. n , ..'.,,. • Wileove. I). S.. l.'.-llisiei... |. Dul.n.w. \. Phillips t M 

t I 1 ill I I „ . | *. For m- Uosos. PW8. Onantilying llneats U, imperiled spc< 

-lance. K.,99, international passenger* cheeked j„ ,',. I ,,,!.•<) SiaL. !!„,*, „„,-, -18: ni7 i.|' .Y 

luring one week in May 1990 at the Los Angeles , ntemH Krsoui ,,. s 

nternational Airport were found to be carrying Darwin. C. 1859. Online version of "On the Origir 

357 lots of fruits and vegetables and 32.") lots of Species." Outp://pa»es.hrilislilihrarv.nel/charles.dan 

u.inial products, for a total of 2035 kg of contra- n t«ls/ori S i..l85«)Am«i,ui,n.l,.ml) 

- '<- < — -» ■» , ^: f 7;v ,, M\ , i,, 1 :;::; 1 ,o^::;;;:v , ;;:; 1 : , ; 1 ;;, , ;::x:Vd: 

>nng species with them lor personal use or as gifts index. html) as of Fehmarv 28. 2(KK). 
or friends and family. 

To address lli< problem* caused b\ invasive sp 
ies and the many pathways by which thev ent( 


ALIEN SPECIES TO John M ' Randal1 ' 


1m1.iI dot rue will likely i.mlli|il> these effects. Many 

"Xowadays we lire in a very explosive world, and ,.,,.,,,. „„,,, - ;l |,. ( | up || lr Welland Canal aflei \'.V2 ( > In 

„■/,//,. „e nun no, l.non where o, when ,he next on!- (|e( .; iml ,e ,| 1( . <; reat Lakes" trout fishery. 

('""' '"" '" "' " ' '""!""" I- hI " >■* , ,., I. ,. I, ■ ,. bolstered by hun- 

doiin Us /one. Its not pisl > 

nuclear homhs and wars thai threaten us. though these dreds (.1 mere recent invaders, like common cru- 

rank very high on the list at the moment: there are pjna {Critfiina vulgaris Cass.). Asian tiger mosquito 

other sorts of explosnms ... eeologual explosions. .. . (\ e dvs albo/uetus Skuse). and iht* zebra mussel 

/"■/'"■"<"/ ' '/"''-' " ' ''", " s/ '" (Ihnssena ,,nhmn,phn Pallas) ( 1993: Mac- 

;;;;;;;: ;::;; ~ T,',',: "— - i ( "><>- »»«i»- « n»n- ■**>': ^-*>- <* **•> 

effects...." 1999). However, although the quote above may be 

—Charles Elton (| 1958] 1972: 15) j ust as true as it was in 1958. the context of bio- 

Thus did the Irtish ecologis, Charh-s Klton. in '^ "' -—.'-. « >"'"^' ^'uheantly 

hi. i9.-, !{ i k //„■/,,/„, „//, „ ,,,,,/„ i™„/, K ""- 1 -'7 1 - Im "~ : h ; r 7 <al,,l,(l - a,1< 

„,„/ Plants, foieel nake ,,, that s,,n, »»■" P"'"""'* ' ^ >-"-'-« -l« >.,,, ,.>n hax e cnerged. 

. - * For example, the end of the Cold War. globalization 

translocations iliu h ! - are trans- , r i u , i 

...... ,. ill- a of trade, and free trade agreements have connected 

tornuiig the biological world and help inspire the . n 

, h • . r- • i prey mu-dy i-nlnled parts of the world and increased 

n " W '-^-'"f 7" K "' "'7- I"' -7"' 1 ; 1 '„,. . 1 ,,,. c among them (e.g., 

impacts. E ton iMuh a. enco ,,y eahng ea>, .>.,„! b]mw | (ransp(lrt 

■ es. escribing, for example how c icstim. big. . ^ __, ^^J.^ (|f jn 

,-m/ Mumll Hair, virtually ^ u • u- 1 

I I lc \mericin chestmil from extern " U(lr M "" ;1 " l,,a ' """' , ' s l"' ( "' s are l^ng shipped 

', ""'!' : ' 1 " ',"'■ ""'" ,a ", ( o7i"" 3m Tif and >hipped more rapidly around the globe and. 

U.S. forests during the early 20th centun and how " _, (1 ] ■ ._ .^ f^^., ^ 

parasitic sea lampreys. Petromyzon inunniis il 
native to the Saint Lawrence Uiver and bake 

, () " * becoming established in new areas. Accelerating 
rates of habitat destruction, climate change, and 

1 We appreciate the help of Bum Hire ami l.i/ Hamilton in developing ami analyzing <lala on The Nature Conser- 
vancy's experience with imasive species. Several ideas in this paper were developed during discussions with TNC 

colleagues including Alan Holt. Maggie C< Handy Curtis, krisly Ciruiia. and MarU Our thanks also goto 

Peter Haven and the Missouri Botanical (,arden for focusing the 2()<tl Annual Systematic. Symposium on the important 

'" '\nll,ll!'l!r<^Tesponden(-e. P.O. Box 22665. Cannel. California 93922. U.S.A. 

; wildland In\asi\e Specie- Team. I lie \atiiie 0oii-erv uuev. I l.parlmeul of \egelable Crops & \Xeed Science. Uni- 
versity of California. DaOs. Calilonn,, <)5(>16. 

Ann. Missouri Bot. Card. 90: 67-76. 2003. 

other global change phenomena are disrupting na- 
tive species assemblages and creating new invasion see OTA, I'W.'i; |\\illiams & Mefle. 1996]; Mack 
opportunities (e.g., Mooney & llobhs, 2000. and el al., 2000; Fwel el al.. 1999; Bandall. 2000). 
papers therein; Chapin et al., 1997). Perhaps the most frequently doeumented effects of 
At the same time, there is reason lor optimism. invasive speries are their suppression of native spe- 
Worldwide efforts to reduce international transfers of cics populations through predation. competition, par- 
invasive species. |>;ii!i< ular K llics. harm agrieul- asitism. or disease. A recent example is the Asian 
ture and other economic interests have grown signif- swamp eel. Uonoptents ulhus (/uicvv). a fish originally 
ieantly over the past half century (e.g., Mr IN eel v. from tropical to wann temperate portions of east Asia 
2001). A broad to the conservation of that was first detected in Florida and Georgia in 1996, 
biodiversity has emerged as manifested in interna- possibly following escape from aqiiaculture facilities 
tional treaties such as the Convention on Biological or intentional releases of aquarium specimens | Hen- 
Diversity |see Internet l{esources| and in national son et al.. 2001 1. Populations ol native fishes appar- 
laws, such as the U.S. Kndangere.l S|>ecies Act of ently have declined in areas where this geiieralisl 
1973 |see Inlernel Hcsoiirccs|. Meducing the ecolog- predator has become established [Benson et al.. 
ical threat that invasive species |>ose to biodiversity 2001 |. 'fhe eventual lull impacts of the Asian swamp 
is an important part ol these efforts, and our technical eel in the U.S. are as yet uncertain, but the fish's 
capacity to do so has grown. In comparison to the capacity to tolerate extended periods out of water and 
1950s, todays richer -cientific understanding of in- at low temperatures surest its range and associated 

ger bams for targeting efforts and lor designing ellec- Certain in\asi\e species hybridize with congener- 

tive prevention and control strategies. ic native species. At a minimum, this alters the gene 

For those engaged in the conservation of biolog- pool „f the native species, sometimes in maladaptive 

ical diversity, then, the picture is one of urgency ways. Al a maximum, ititrogression can eventually 

■eked; Hope, pletely displace the native species. A recent well- 

s that would documented example i> llie invasion of San Francis- 

comain mis tnreal Have never been greater. co Bay sail marshes by smooth cordgrass (Spartina 

alterniflora boisel.). originally from the Atlantic and 

IMPACTS <>!< Invasivk Spkciks on Bh>I,<)<;k;ai. Culf coasts of ihe United States (Anttila el al.. 2000; 

DIVERSITY [San Francisco Fstuarv Invashe Spartina Project, 

2(K)2|: Vila el al.. 2(H)(1). Spartina alhrnijhra hy- 

"Invasive species" here refers to non-native spe- bridizes with the native Spartina foliosa Trin.. al- 

cies that become established in new localions. though rarely, and the resulting hybrids outcompele 

spread, and then cause ecological or economic- ll.e nalive cordgrass. spread clonally. and interbreed 

harm or threaten human health. This definition is with both parental species (Anttila et al.. 1998, 

consistent with recent uses in I S pobev. such as 2000. \yres et al . 1999; Vila el al . 2000). Where 

Fxecutive Order \A I 12— Inrasirc Sprrirs— signed ihey have displaced the native cordgrass, the non- 

by President Clinton in IW9 (Clinton. 1999). Note, native species and the hybrids alter marsh architec- 

however, that it differs from strictly ecological del- ture. because, in comparisoi he native species. 

ones provided In h'icliardson el al. (2000) and Be- denser rhizome mats that trap sediments and raise 

jmanek et al. (2002), which locus on the biological marsh elevation. These allribules enable- the invael- 

attributes of a species that enable il to spread and ing species and hybrids to expand into large areas 

become established in new locations. that previously were open mudflats, displacing cum- 

Bough estimates are thai invasive species, as de- nmnities of algae and invertebrates associated with 

fined here, comprise about I0-20<X of the i.on-na- mudflats and reducing foraging resources for certain 

live species that have become- established outside- migratory waterfowl and other animals (Callaway & 

and plant pathogens u, the United Slates (ef. | \\ il- ,-jes are indirect- mediated b\ the altered behavior 

hams Ox Meflee. |09o|; (»TA, 199.5; Williamson & or other characteristics of intermediary native spe- 

Filter, 1996). Their ecological impacts range from cies that are not themselves harmed. Invasive 

loc^al suppression of single native species to species plants that have showy Moral displays, like purple 

extinction and wholesale changes in llie funclioning loosestrife [Lyihrum saliraria F.), have the potential 

Volume 90, Number 1 

Chomesky & Randall 69 
Threat of Invasive Alien Species 

to alter the abundance or behavior of insects lliat 
pollinate native plants and thereby indirectly affect 
native plant reproduction and fitness (Crabas K 
Laverty, 1999: Brown et al.. 2002). Certain invasive 
shrubs may indirectly reduce the reproductive suc- 
cess of native songbirds, because their architecture 
and lack ol sharp thorns allow predators greater 
access to songbird nests than do native shrubs 

meltdown"— accelerating impacts on native ecosvs 

anion- repealed species invasions. 

Bv suppressing native species populations and 
altering habitats and ecosystems, invasive species 
have contributed to the imperilment of nearly half 
of the plants and animals now considered rare, 
threatened, endangered, or extinct in the United 

(Schmidt & Whelan, 1999). 

The most harmful invasive species are those that 
have system-level impact.-, fundamentally altering 
the ecological processes that structure communities 
and ecosystems (Vitousek, 1990). A well-known ex- 
ample is m< I il ii ! • >,">< '" '"" l ( <>\ ' 
S. T. lilake. an Australian tree that has spread and 
become locally abundant in Florida's Fverglades. 
The tree forms dense thickets that trap sediments 
and debris and elevate the topography. Because of 
the plant's high rates of evapotranspiration. mela- 
leuca appears to draw down water levels and alter 
the hydrology of areas where it becomes abundant 
(Sclimilz. el al.. 1997: llolstetler. unpublished data). 
Melaleuca is pyrogenic and increases the temper- 
ature and duration of fire. Overall, the tree is fun- 

rida's "Biv 

wetland habitats of i 

plants and their 
iunas (e.g., Gordon, 1998). 

zebra mussel. Ih< i^-< ■■■, 

other invasive species that is causing broad Ostein 
level effects (Maclsaac. 1996: Stray er et al.. 1999). 
The mussel directly suppresses native freshwater 
bivalves bv overgrowing and smothering them. 
Dense accretions of zebra mussels on hard and soft 
lake bottoms create new habitat space for benthic 
invertebrates. Because the mussels' efficient filler 
feeding clears the water of algae thai previously 
supported pelagic predators, zebra mussels have 

luiidaineiilallv transformed the I I web dv mimics 

of places like the (beat bakes. Most production and 
biomass now occur in benthic parts of these sys- 

dominated (Maclsaac, 1996). 

Certain invasive species facilitate the establish- 
ment of other non-native species 
harmful effects. The laurel fig. Finis 
{.. for example, was introduced into Florida in 1012 

natural areas, including the Big Cypress National 
Preserve and the I .oxahatchee National \\ ildlile Bef- 
uge (Fangeland & Burks. 20(H)). Simberloff and Von 
Nolle ( !<)<)<)) have hypothesized that "invasional 

Stales (\\ ilcovc et al., 1998). According to an anal- 
vsis by Wileove et al. (1998). non-native species 
are the second most frequent cause of species im- 
perilment: the most frequent cause is habitat deg- 
radation and loss. 

In addition to the nearer-term ecological impacts 
discussed above, species invasions also may be po- 
tent drivers of evolutionary change (e.g.. Grosholz. 
2002: Mooney & Cleland. 2001; Sakai et al., 2001). 
Some invading species in.iv pass through an evo- 
lutionary "bottleneck." for example, if certain phe- 
nolvpes survive the pathway of species transit or if 
the receiving environment imposes strong selective 
pressure on invading populations. Conversely, sue 
ces-fiil invaders may cause genetic changes m pop- 
ulations of native species by hybridization or by 
qualitatively or quantitatively altering selection 
pressures through ecological interactions or through 
ehanges in important habitat qualities or ecosystem 
processes. Many examples exist of hybridization 
between invasive non-native plants and closely re- 
lated native species yielding fertile, and in some 
cases, highly competitive hybrids or new species 
that are reproductively isolated from both parents 
bv mechanisms such as polyploidy (Vila et al., 
2000: Mooney & Cleland, 200 1). 


species unpads on native sp< 
is. Synthesizing this informatio 
ive picture has proven challer 

portance ( 

biodiversity, and one whose unpads can be signif- 
icantly reduced by effective action. 

Here we describe how one major conservation 
organi/alion. The Nature Conservancy (TNC), has 
developed an overarching assessment of the threat 
that invasive species pose to biological diversity. 
TNC is one of the world's largest place-based con- 
servalion organizations, working al thousands ol lo 

•alH.ii> around the world. Mm organization^ melh- Table I. Examples of plant speei. 

ods lot -fl.M luiii .11 1. J d. -i-miiil c on-. i\ ihoii -lies ''<>us ihreal to many sites or aeross ; 
have evolved significantly over the past 12 years to im ' a ''> TN,; managers'. 
include the identification of conservation threats _ 
and their solutions al multiple spatial and ecolog- 
ical scales. From this new methodological rigor, a 
new picture ha- cm. ig.-d ,,| how pervasive and sig- 

uork an. I lor hiodi 

The Conservancy selects s 
tion through the process of "ecoregional planning"— '^Z7\hlUKbV.cktho,' 
a stepwise process thai uses ccoregiotis as the basic Euphorbia e.sula L (leafv spurge) 
planning unit (Groves el al.. 2(X)2: (TNC. I»|). Each ~ 
ecoregional plan identifies a collection of "conser- 
vation areas" required to fully protect the ecorcgions 
biological diversity. Invasive species were identified 
as among the top threats to TNC's conservation tar- 
gets in 15 of I,", completed ecoregional plans for 

North American ecorcgions in a summary completed cnt conservation areas. Other highly 
in 2000 (Groves & Valutis. unpublished data). Hab- ' 

ital destruction was the onlv other threat thai was ""'" " ri "" l """'<' a '■•). ani1 spotted knapweed {Centaur 
cited more frequently in the plans. 

A site plan is prepared for each of the conser- 
vation areas where TNC works <„• intends to work. cerns. with 49% of the respondents identifying in- 
Each plan formally ;,>,»,, the relative severity of vasive species as one of their top 5 management 
known threats to the areas biodiversity — such as concerns. 

unsustainable timber harvest, altered hydrology. Garlic mustard. Allium peliolata (M. Kiel).) Ca- 

second home development, or invasive plants— and vara cK Grande, is one example of the kind of in- 
sets priorities foj 

areas from A<) North American ecorcgions showed occurs in 20 eastern and midwestern stales and has 
the majority (917, | listed im asue species as a crit- recently been observed in Washington. Colorado, 
ical threat to biodiversity at the site |TNC. a. 2000). Idaho, and Alaska (Nuzzo. 1995. (20()0|: [Tu «K 
Xm ""* l( "'-' I'l^'is. inva>,ve specie- were the site- Hire. 2001 . 20<)2| ; Rice ei Johnson, in press). Car- 
Implementing a conservation area plan often in- Ocularly spring ephenierals. and reduces survival 
volves direct ecological management of natural of tree seedlings in some areas (Anderson et al.. 
communities. In ]998. The Conservancy's Invasive 1096: McCarthv. 1907). Evidence suggests the 

^ ,,l,( ' ,< ; (i » *»'•"*>' of all TNC weed also reduces the fitness of al least two native 

responsi- butterflies because, although garlic mustard attracts 

ses were these insects, eggs laid on the plant have lower 

>hip than those laid on native host plants 

2 million acres. The (Bowden. 1971: Porter. 1991). Of additional worry 

usually labor-intensive, often difficult, and some- 
times impossible | Nuzzo. 2000J. 

Eeral pigs {Sus scrofa L.) provide a graphic ex- 
lace with invasive animals. The pigs cause damage 
by uprooting and killing native vegetation and 
thereby opening areas for colonization by invasive 
plants that further alter the system. Eeral pigs 

the U.S.. ranging from the (deal Smoky ' 

in the U.S. 

s [TNC. cj. A total of 110 

«-d Iron, -tall working at 916 

managers identified 216 

invasive plant species in- 

festing 810 (88%) of the conservation areas coy 

by the respondents. Tabh 

• 1 provides an illustr. 

list ol invasive plant- <•< 

insidered a serious threat 

al many ol these place-, oi 

• over a large geograpl 


area. The particular in 

problematic varied greatl 

V among TNC's const 

•rv a- 

lion areas. Of the 216 species, managers ideal 


175 as a serious lineal lo 

at least one conserv; 


area. Most respondents (over 80%) considere. 

1 in- 

vasive plants one of their 

top 10 management , 

National Park in Tennessee and North Carolina to proposed species introductions, and to detect and 
Santa Cruz Island off the coast of southern Caliloi- manage new incipient species invasions. It also will 
nia to Haleakala and Hawaii Volcanoes National require national and inlernational policies that flex- 
Parks in Hawaii (Aplet. IWO: Hratlon et al.. 1982: and rapidly integrate these advances into op- 
Hone & Stone, 1989; Hone. 1995; Singer. 1981). eralional practices |e.g., National Invasive Species 
In Hawaii, for example, subsequent recruitment of Council. 2001]. A new level of collaboration among 
non-native plants, such as the nitrogen-fixing Por- ecologists and other researchers, engineers, and 
tuguese fire tree {Myrua Java Ailon). into areas dis- policy experts thus will be essential to the success 
tuibed bv pigs idler* fundamental eco*y*|eni pro- of future prevention efforts. 

cesses and initiate* a cascade of other species Issues related to international trade demonstrate 

changes (Aplet ct al.. 1991; Vtorov, 1993). the need for lighter linkages between scientists and 

The compelling picture emerging from The Na- policymakers to improve prevention. Because of the 
ture Conservancy's efforts al all levels is that in- dual trends of globalization and trade expansion, 
vasive species are a high-priority threat to biolog- incorporating measures to minimize new invasions 
ical diversity that must be addressed now. Many in is both more urgent and mote difficult. Inspections, 
the organization believe that invasive species pose quarantines, and other forms of screening must be 
a fundamental risk to the biological diversity that rapid and compatible with free-trade agreements 
TNC and partner organizations once thought was already in force or pending. This will require im- 
already successfully conserved on millions of acres proved lechnologv. transparent rationales for ex- 
and in numerous aquatic and marine systems. In eluding certain species, and control actions that 
complement any exclusion actions within each na- 
tion or trading bloc. Policies under development 
(e.g.. the new L'.S. National Management Plan [Na- 
tional Invasive Species Council. 2001 J) include sci- 
ence and research agendas specifically designed to 
meet current and emerging information needs that 

Developing criteria and systems for accurately 
predicting which species are most likely to become 
harmful invaders if introduced to a given area is 

one of the most important technical challenge- 
ahead. Weed Risk Assessment systems to evaluate 

have alreadv been implemented in Australia (Phel- 
; affected by species invasions— such ^^ <( ^ ' , 9W) , n a( | () ; tion representatives of 
. forestry, tour,*..,, pub- ,,„. ' ( s imrsrn ill( | lls trv and other user groups 
lie health, horticulture, and others— combine the.r (| ^ n ^ n[h mm(M | , () ( | (>vek)p a vo l unta ry "code 
energies in finding common solutions to shared ^ ( . (m(|m . r , OI . ,,,| llrmg t | ie risks of new plant in- 
challenges |e.g.. National Invasive >pecies < .oiuieil. |m( | l|( . (il)I)s (Ham | a || ,., a |.. 2002). These and other 
2001]. Understanding has also grown over the past ^ ^ |)n)i)()s( . ( | |(N l ,. s ,.. iri . l „. I . s (e . g ., Reichard & 
decade of the potential policy and management , |;m)i||nn _ , w7 . , { , jnia nek, 1996) are being scru- 
strategies for reducing the threat |m|/) () ^ |(() ^ iM( . „„„| 1 | l , ,„,„„ an< l use in Hawaii 
as well as a set of key questions that can only be ^ ^^ ^^ ni ^ mM (Daehler & Carino, 
solved by scientific research. Tins section 2{)(){)) ^^^ a|ia , vs ; s „ f lll(>ir performance will 
several critical science challenges that be ahead. ^ .^^ [h ^ sy ^ ms and offer insights into 
how systems for other laxa and other areas could 
PREVENTION, EARLY DETECTION. AND RAPID ^ ^"..^j (K() , ar & , ()( , ge 2Q02 ). 

RESPONSt Kven w i ttl improved prevention systems, at least 

new invasions will continue. Many might be 
lined or even eradicated if detected early 

response, TNC is pres 
ti-vear organizational 
vasive species threat. 
the initiative's Executi 

entlv mounting a major, mul- 
initiative to address the iu- 
(For further information see 

iv or contact 

Setting a Fiti ke Cot use 

A picture has emerged over the past decade of 
the full impacts of invasive species on the world's 
biological diversity, economies, and human health. 
Paradoxically, the very breadth ami depth of these 
impacts may hold the key to a solution as the many 

> ranging from t 

continental to the *iuall nature preserve or suD- 
tershed— will be the most effective way to minimi/..- enough, and general agreement exists that better 
the numbers and impacts of future species inva- early detection systems are needed [e.g.. National 
sions. This will require scientific and technical ad- Invasive Species Council. 2001 |. Research discov- 
vances that enhance our ability to understand and erics that improve our ability to search for and de- 
predict invasion pathways, to assess the risks of ted new invasions will be crucial to making such 

plans work. For example, spatially explicit models species and am,,..- specific invaded portions of 

that predict where new invasions are most likely to preserve [TNC. c|. The I .S. National Park Servi, 

occur and how they would spread from sites of ini- (NPS) supports an online invasive plant rankii 

lial establishment could -really improve the effi- system that calculates a score used for assignii 

ciency of searches (e.g.. Iliggins el al.. 20(H); Wads- relative management priority to each invasive pla. 

worth et al., 2000). Fasily accessible and species at a site (Heiherl. 2001 ). tl 

networked databases that accept reports of new NPS system evaluates species' impacts relative 

species and locations and thai automatically notify the conserv; n goals f or ., s|M .,.jfj r siu . ilM( | ,., 

officials in the affected area would allow such in- plicilly balances projected impacts of invasive sp< 

formation to be acte<l upon rapidly. Accessible tax- cies against management feasihilitv and cost 

onomic keys that haw multiple entry points or per- These systems warrant further testing anil improve 

haps even on-line services that can identify species ment. and similar priority -setting is needed for otl 

from scanned images would allow many thousands er groups of invaders (e.g.. insect invaders, fisl 

of land managers and other interested citizens to aquatic invertebrates). 

tion systems (e.g., Brooker et al., 1999). setting approaches is that co 

i pno 

the focus of many site managers remains ] 

Kven with well-supported prevention programs in on how to eliminate or sharply reduce tl 

place, a need wdl exist to manage invaders. Many dance of particularly harmful invaders, 

harmful specie, are already established. Others systems and with some invaders effective 

presently under human cultivation may some day has proven t 

start to spread beyond cultivation, become firmly impossible. I 

established in the wild, and cause problems after Irol of eheatgrass (Bromus t 

long lag penods. And al least a lew new invaders expanses of the inlermountain West have had very 

are likely to elude even the most exemplary pre- limited success despite many and varied attempts 

vention efforts. Consequently, invasive species are dating back over i() years (Mack. 1986; Ipadhava 

not a problem that can be "solved" or "engineered et al.. 1986; Young & Kvans, 1985). 
away"; rather this is a problem lliat will require In contrast, mo believe conservation practitioners 

continuous management into the future. II1US , ;1 || ,ate- ,| lrir .^sources first toward sustaining 

routinely face chronic- shortfalls of money and staff storing, or mimicking critical ecosystem processes. 
time in the face of urgent needs. Since they cannot such as fire regime or periodic Hooding (e.g.. Ivand- 
do everything, they musl set priorities and tackle all el al.. 1967; Stromberg K Chew. 2002a). Seme- 
me most unporlanl problems first, often leaving limes this ma\ mean allowing certain invasive spe- 
lesser problems unaddressed altogether. In many cies to persist. In others, it may even mean 

,1.1 -|.,| 

managing a non-native species 

species or the area infested already far exceeds lo- conservation value of an area. For example, in some 

cal management resources. Many natural area man- chealgrass-infested areas, rather than seeking to 

agers therefore take a "triage-" approach and al- control the eheatgrass. limited management dollars 

lempl lo sel priorities bv identifying which species, might better he used for planting other native grass- 

" ecosystem processes thev ar. lering seed periodically nsim r.liabl. s'easou- 

. Ih.-weed, 

ulahililv ..[foods required by native roclei 

. Modeling, demonstration | 
jects. and monitoring and adaptive management 
>rks are avail- will be necessary for this approach to succeed. 
. These frame- Managers are already using this lactic in llawai'i 

^ ^"' ,; " al,/r '■'""•" l '•'"'"^'■•'l ^-w ledge Volcanoes National Park, where fire-promoting in- 
current and projected distribution range, and ease ally dry" woodlands to alien grass say anna') Tunis'on 
mple. The Nature- Censer- et al.. 2001). Mather than seeking ... directly control 

vancy's Weed Management Plan Template 
guidance on how to set |)rioriti<-s among i 

glasses and restore- the original 

Imitation" — tli( iil( I Km mi svstems and tin small spatial scale ol 

fire-tolerant native trees and shrubs that potentially standard methods for detecting, monitoring, and 

will persist and spread in the savanna^ now domi- managing invasive species has and should continue 

nated by alien species. to fuel interest in technologies and management 

Approaches like these that involve managing methods that can operate at very large spatial 

mixed assemblages of native and non-nalive spe- scales. Remote sensing data applied in CIS for- 

cies test the limits of existing ecological know ledge. mats, for example, promise to provide cheaper and 

Practitioners rarely have information about the like- more effective ways to detect and track the spread 

ly long-term impacts of the management choices of new weed infestations — and therefore effective 

they make today on the Inline structure and (uric- ways to eradicate or contain invasions early (e.g., 

tion of ecological communities and ecosystems. Kveritt et al., 1005; Andrews, 2001). Landscape 

This is one of the chief shortcomings of current analyses of invasion patterns may help identify 

approaches and of the information base that has emerging pathways of species spread and provide 

been supplied by prior research. The field of bio- a way for targeting resources to reduce species tran- 

logical control, a particularly rich discipline in in- sit and to monitor for new invasions (Thompson, 

vasion biology, provides an instruclive example. Al- 1009). Creating "nodes" of native species that serve 

though recent advances in selecting introductions as source populations and seed banks might en- 

of control agents promise to reduce ecological im- hance the recruitment and persistence of native 

pacts on non-target species (McKvov ei Coombs. species in certain systems dominated by invasive 

1900). methods do not vet exist lor assessing effects species (Tunison et al.. 2001). Finally, considerable 

such as the recent I v documented impacts in Hawaii scriptions. sudi as large-scale patterns of grazing, 
on food web dynamics (llermeman cv Memmoll. prescribed lire, or flooding and hydrology, that tar- 
2001). An emphasis on long-term research and the get abiotic or biotic ecosystem processes will pro- 
development of spatially explicit models and fore- vide cost-effective «,n> I., suppress populations of 
casting approaches to predict tin- long-term ecolog- invasive species and yield conditions that lavor na- 
ical impacts of todav s management choices and de- live species (e.g.. Koebel, 1005; Randall ct al., 
faults could help to remedy this gap. 1007; Thysell ,\ Can-v. 201)1: Stohlgren et al., 
1000; Stromberg & Chew, 2002b). 

An important paradigm shift is under way in the 

focusing on conserving areas Ours is a lime of dramatic and escalating change 

that are large enough to sustain dvuaniic ecological in the world's biota. Over the i 

processes and functioning ecosystems in addition species and habitat loss ami fragmentation uue 10 

to populations of rare species and examples of nal- land use change and changing resource exploitation 

ural community types. This shift, from preserves of patterns will continue to bo the major drivers of 

hundreds of acres to conservation areas thai are such change (Vitousek et al., 1096; Mooney & 

tens of thousands to millions of acres in extent. Hobbs. 2000. and papers therein). Cess well un- 

brings new challenges in addressing invasive spe- derstood. but at least equally important over the 

cies. For example, controlling local infestations of long-term, will be the biological changes wrought 

vegetation are by nitrogen deposition, enhanced C0 2 levels, and 

relatively straightforward lot small conservation ar- changing temperature and precipitatic 

eas like The Nature Cnnservanev's 70-acre Blowing (e.g.. Moonej & llohbs. 2000). The magnitudes and 

Rocks Preserve in Florida where mechanical re- qualitative details of most global change processes 

nioval. spot applications of herbicides, and direct have as yet proven difficult to predict, 

planting of native species are standard practices Nevertheless, all of these global change process- 

(Randall et al., 1007). Consider how infeasible es will interact with one another to yield cumulative 

such approaches are for large areas, such as ihe impacts on the distribution and abundance of spe- 

250,000-acre Demonstration Weed Management cies. the composition of ecological communities, 

Area along the Snake River thai encompasses parts and the functioning of ecosystems. Climate change, 

of Oregon, Washington, and Idaho! nitrogen deposition, and ambient C0 2 levels all will 

The profound mismatch that currently exists be- have profound effects on the pool of potential in- 

tween the large spatial scale required to conserve vading species, potential invasion pathways, and 

(Mooney K llobh- :'.iii... 

es, 2000). 

imate change is expected to result in the di 
nbly of many terrestrial and aquatic ecological 
as species ranges and abundances 
1 response to changing temperature and pre- 

abitats for colonization 

M.ttila. C. k.. C. C. Daehler. N. K. Rank * 1). R. Strou 

nd papers therein). The 
from the pre-industrial 

1998. Greater male fitness of a rare invader {Spartir, 

3pm today may already 


ain plant invaders like 

. R. A. King. C. Ferris. 1). R. Ay res & 1). 1 

Strong. 2(MK). Reciprocal hybrid formation of Sport in 

ppm within 100 years 

in San Francisco Bay. Molec. Kcol. 9: 765-771. 

Aplet. (;. II. |90<i Mi,.,,,,;,,,, ,,| , ;,,||, u ,, ml rummiini 

biomass by the alien Myrini Jaxn. Oecologia 82: 411 

frequency and other disturbances (e.g.. Chapin et 
al., 1997). This will greatly enhance opportunities 
for new invasions (e.g.. YVallhe.. 20(H)). The process 
also is likely to change the ecological impacts of 

even the possible and desired endpoints of eotisei- 
viilion niauageinenl activities. An important, and 
largely unanswered, question, in light of the com- 
bined impacts of various global change processes. 
is whether some of the invasive species that we 
choose to control t- .. I.i \ will become acceptable oi 
even desired species at some sites tomorrow. 

Invasive species have emerged as one of the key 
threats to the world's biological diversity. Today's 
richer scientific understanding of invasion process- 
es and of management and policy approaches for 
reducing impacts hums a strong foundation for tak- 

linkages between scientists and policy makers will 
be essential over the near term to improve the pre- 
vention of new invasion-, and In .nable more rapid 
detection and response to the new invasions that 

elude developing , -, ,, logical forecasting approaches 
to help set management goals for mixed species 
assemblages, building methods to effectively ad- 
dress invasive species and promote native species 
and communities al Luge spatial scales, and deter- 
mining bow the 1 1 it. t ,ii lion- among various global 

consider and respond to species invasions. 

rson, R. C. T. C. Kelley & S. S. Dhillion. 1996. 
pects of the ecology of an invasive plant, garlic mus- 
.1 ( Mliaiia prlinhil,/). in central Illinois. Restoration 
al. 4: 181-191. 

ews, M. M. 2001. Use of AVIRIS Hyperspectral Re- 
te Sensing to Map Mustard (Hmssica nigm) and Fen- 
[Fociwulum vulL'iirr). I iipuhlished M.S. Thesis, 
iversilv of California. Davis. 

lee. Kcol. 8: 1179-1186. 
{ovv.len. S. R. 1971. American white h.ltterllies (I'icn- 
dae) and Fnglish food plants. J. Lcpid. Sue. 25: 6 12. 

20(12. <:„■„- 

Callawav-.l. C. i\ \1. \. .1— ,-K n. I 002. Tlie introduction 
and spread of smooth cordgrass (Sparlinu alternifloni 
in soudi San Francisco I'.av. Fstuaries \7>: 2111-226. 

Chapin. K S.. III. II. Walker. R. J. Il„hbs. I). D. Hooper. 
J. H. Ixiwlon. (). K. Sala & I). Tilman. 1997. Riolic 


Cliiilim. \\. .1. 1999. F.\e< olive Order 1 51 I 2 of February 
5. I<W: Invasive Specs. I'd. Reg. hi: OlRo olJW, 

Craig. G. R. ]>><>:>,. The diaspora ol llie \sian hgci ,„„>- 
.|iiil". I'p 101 120/,, II \ Mcknight (editor) Bioloi 
ical Pollution: The Control and Impact of Invasive Kx- 
olic Species. Indiana Academy of Science. Indianapolis. 

Daehler. ( ( XI). \ Carino 2000 IVd g ,m ,. u 

plants: Prospects for a general screening system based 


! hi 

Kvcrill.J. II..",. I . \ndcrso„. I). F„ Escobar. VI. R. Davis. 
N. R. Spencer & R. J. Andrascikel. 1995. Use of re- 
mote sensing lor delecting and mapping leafy spurge 
{Euphorbia rsiila). Wed Technol. 9: 590 ,,0«l. 

Kwel. J. J.. I). J. (TDowd. J. Bergclson. C. C. Daehler. C. 
M. D'Antonio, b. D. Gomez, 1). R. Gordon. R. J. Hohhs. 
A. Holt, K. R. Hopper. C. E. Hughes, M. LaHart, R. 
R. B. Leakey, W. G. I^e, L. L. Ixwpe, D. H. Lorence. 
S. M. Kouda, A. K. bugo. P. B. McEvoy, D. M. Rich- 
ardson & P. M. Vitousck. 1999. Deliberate introduc- 
tions of species: Research needs. BioScience 49: 619- 

Gordon. I). It. 199!!. Kffccts of i,,v asiv c. non 

Florida. Ecol. Applic. »: 975-989. 
Cabas. G. P. & T. VI. I.avertv. 1999. I li« 
loosestrife (Lythrum salicai ' 

• ! , I | --«-«, i, J Id 2 12 

Crosholz. K. 2(K)2. Ecological and evohili 
queuees of coastal invasi 

J. Holilis (Editors). 2(MK). 
ling World. Island Pro. \\, 
1993. Distribution and s 

: _>7. 

iroves. C. R.. I). B. Jensen. I. 
M. I. >hafler. J. M.Scott. J 
-ins. VI. W. Heck & M. (J. 

cue- into practice. BioSoie. 
leiherl. R. 2001. Prioritizi. 

Ecol. Kvol. 17: 

II. |{rd|,„-d. 

r.'u^l.iiMv/r.-.o,., , 21100/ \I'K- Mil- HIM I 

rV. V* I).' M. Biehardson & B. M. Cowling. 2000. 
igement of alien plant invasions. Ecol. Applic. 10: 

|(Bieb.) Cavara « 
115-124 in B. N 
itimi: The Control an< 


^cies in the United States. U.S. Goi 
llicc. Washington. D.C. 
ilications of introduced garlic mil: 
lata) in the habitat of Pieris virg 
. 1,-pid. Soc. 48: 171-172. 
Williams & S. K. Halloy. 1999. 
nt model for use as a bioseeuritv lie 
Irodui lions. J. Envuonm. Manager 

>. \. I,,.,-, Service Pro,-. 5: 04-73. 

, R. R. Uewis, III & D. B. Jensen. 1997. Ecolo 

estoration. Pp. 20,5-219 in I). Simberloff. D. 

w'ildlile S.V. Bull. I 7:' 'l 19-125. 
koebel. J. W.. Jr. 1995 

Conservation Biol. II: 195-203. 
Kejmaiiek. M. |9«)f,. \ llieorv ol seed phml invasiveness 
The first sketch. Biol. Conservation 78: 171-181. 

. D. M. Richardson. \1. C. Barbour. \1. J. Cravvlev. 

P. B. Moyle. J. M. Randall. I). Simberloff 
Williamson. 2002. Biological invasions: Politics 
■ ,al terminology. Bulk 
83: 131-133. 
Rice. B. A. & P. Johnson. In Press. \otevvorthv collec- 
tions — Mliaria petiolata (Bieli.) C.i\;ii;i >\ Crande in 
Alaska. Madrono. 

M.. P. Pvsek. M. Rejmanek, M. G. Bar- 
mella £ C.J. West. 2000. Naturalization 
of alien plants: Concepts and definitions. 
6: 93-107. 
C. Thill. 2001. Biologv of common 
two Mediterranean win- 
North America. Weed 

I). E. MH.aiilcv. P. (TNeil. I. 

& S. C. Welle. 2001. The 


ual Rev. 

IcEvoy, P. B. & E. M. Coombs. 1900. Biologi. 
of plant invaders: Regional patten 
and structured population models. Ecol. Applie. 9: 
387- KM. 

■IcNeelv. J. A. (Editor). 2(K)I. A Global Strategy on In- 
vasive Alien Species. II CN Gland. Switzerland, and 
Cambridge . I k.. in collaboration with the (dobal In- 

, Pul li< 


H. A. & E. E. Cleland. 2001. Thee 
t of invasive species. Proc. Natl. 
.98: 5446-5451. 

C. J. Whelan. 1999. Effects of exotic 

13: 1502-1506. 

Schmitz, I). C., I). Simberloff. R. II. Ilofstelter. W. Halle, 
«\ I). Sutton. 1997. 'Hie ecological impact of nnn'mdig- 
cnous plants. Pp. 39-01 in I). Simberlnll. D. C. Srhmilz 
v \ I. (.. Iln.vvn (editors). Strangers m Paradise: Impact 
and Management ol Nonindigenoiis <pccie> in Klorida. 
Island Press. Washington. D.C. 

Simberloff. I), vK B. Von Holle. IVf). Positive interac- 
tions ,,f nonindigeiioiis species: Inv.isional meltdown? 
Biol. Invasions 1:21-32. 

Singer. E. J. 1981. Wild pig population in national parks. 
Environm. Managem. 5: 363-370. 

lens In l.uaKrs. lii, ,Si, me 19: 19-27. 
•Inhlgreu. T. J., L I). Schcll * 15. V. llt-tivH. I'W. H<»w 
grazing and soil <|iialil> alloc! native and e\olie plant 
diversilv in mckv mountain grasslands. Fcol. \pplie. <>: 

tromborg, J. C. & M. k. Chew. 2002a. Foreign visitors 
Univ. Arizona Press and the Ari/ona-Sonora IWn'viu- 

. 2(M)2I). Hood pulses and restoration 

11-19 in B. A. Middleton (editor). Flood Pulsing in' 

Wetlands: Restoring the Natural 

John Wilcv and Sons. New York. 
Thompson, H. C. 1999. Study finds adjacent laud uses 

are key to predicting the number and type of exotic 

species in forest gaps (Maryland). Keol. Restoration 17: 

Thysell, I). K. & A. B. Carey. 2001. Manipulation of 


Irmesia communities. Pp. 189-502 
. The Population Structure of Vege- 
Publishers. Dordrecht. The Nelh- 

ol ripan 


nternet Resources 

Person. A. J.. P. P. Fuller & C. C. Jacono. 2001 |. Sum- 
mary Report of Nonindigenous Aquatic Species in U.S. 
Fish and Wildlife Servi.e Region 4. U.S. Geological 
Survey, Florida Caribbean Science Center. <http://www. 
Icsc.usgs.^/B lfinalreport.pdf); viewed 30 September 

Brooke,-. M.. .1. K. Connors & A. V. She. I999|. FU- 
Cl.ll): KiicalvpN mI S„iiil,-| a-l \u-lralia. <lillp://www.; viewed Septem- 

ber lleibcrl 12001) in Literature Cited. 

land, K. \. K k. Cra.ldoek lim U iedih.i 
tificalion & Biology of Non-Native Plant 
Natural Areas. I FAS Publications, No. 
ity «f Florida, <http://aqiiall.ifas.ufl..- 


I ft . 

:>& R. k. Loh. 2001. 

. Turkingto.i & I). Mcllvride. 1986. 
ladian weeds 75. liromiis lerinnim I.. 
66: 689-709. 
& C. M. D. D'Antonio. 2000. Con- 



'I'lllo ■! pop ! li ::■ :ll •• . 

os 57: 7-13. 

.. Foopecv B. Weslerbrooks. 

,mer. Sci. 84: 4f38-478. 

I"93. Feial |M" removal: F 
>ods in a Hawaiian ran, f, 
. 57: 87.5-880. 
Ft. A.. Y. C. Collingham, 

v P. F. Ilulmc 2000. Sj„ 

Willi... I!. 

37(Suppl. I): 28-38. 
Aallhcr, C. |{. 2000. CI, mat,,' forcing on the dispersal o 

exotic species. Phytocoenologia 30: 109-130. 
Ailcove. I). S., I». Bollislem. J. Dubow. \. Phillip. t \ I 

lx)sos. 1998. Quantifying threats to imperiled specie 

ui the I ruled Stale. MmScenoe I,'!: 607-615. 
Williamson, M. & A. Fitter. IWi. The varying succcs: 

ol invaders. Fcology 77: 1661-1666. 

iouncil. 20011. Mc< 

30 September 200: 

|TNC. b]. How we w 


viewed 23 April 2002. 



/illiams. J. I). & C. \. Mcllc. |«XW,| \ ,.. , 

Species. Status and Trends of the Nation's Biological 

Na-vice. < III I p://bi..log\ . u^s.^m /s - l/S\T/ m de\.htm): 
viewed 20 September 2002. 


1634-1860 1 

mronliveearlv enlonisl- fell In Iran-plant European agriculture into (he Weslern Hemisphere. 

, 1 1„ natu.ali/ed flora grew as spe. ies f,.t I<mmI, forage, 
M-i-..nm«<- and ,,,.<) i< in.' woe imported cultivated, and . m aped 1 hi- bounds ,,l cultivated holds. Importation ..! what has 
hrenme ili," mo-l .-ominoii oal.-orv „f naturalized species, erslvvliile ornamentals. Iia<l a modest beginning by the mid 17th 
ceuturv The lirsl recorded invasion, llie spread aid prolileuilinu „i Unarm vulgaris in the Mid-Atlantic clonics, was 

n,-o»..i/.-(l hv die mid Kill, oonlnrv. and lichens vulgaris vva- rampant in s he... New England before li«M> Rotan.eal 

„, oids. in.,,, I Ho... I,.. ,m. ...... I. mo.< ...... « I. < ■' " • <« ntury and reveaU 

I „k,.,v, hv liWI ...d .... 1.1.1. d 
liromussecalinus. Crnoglossum officinale. Calinn, aparine. and >,,„•, i„ vulgaris. (Ml.., alien -pecs had. however, become 

invasive bv the UllOs suel. * Khinm migar, m\„ inia -| lhalwer, lo lorn, dovastalm, uiva-iorw m lie I n.t.,1 

Stales from I Hot) onward (e.g.. «««»»« teelorum. Euplunlna esula. Umuera ,,,, -i ■ > •nquenervia) had either 

not arrived hv I <">60, were un.leleeled. or were ,„,| reported i-luu. rs II. ion • .out), ol the naturalized 

flora and the'snhse«|nenl nii.nher of invasive taxa was oertainlv facilitated, and probably sparked, l.y the enormous growth 

" Kr!Tonl\: Hal I >• • .1 n "I 'I", I'.,, I, .in L, hum, vulgar,. Unaria vulgaris, naturalization. 

. Behold. I have given you every herb bearing seed. World, Emm the outset, each ".roup transferred 

eh is upon the la- -"> tree. ( . m|)s , (M( | ( j (m „. st j ra ted aniriiiils upon which they 

e fruit of a tree yielding see< : to you | )ast . ( | tn eir n . so |ve to establish European agrieul- 

,is"l: 29. Bible. King James version lure in the New World (Mack, 2001, and r, 

, /- '/ /-, • 

The first European human immigrants to North T(u , s<> ( ] e ,,. rm j„ed atlempts by European colo- 

America had special reason to believe fervently in ms(s an( | ,| iri| . American descendants to introduce 

ilies to a New World about which they knew little. { m|( . | , S(a|i .^ h;m . had pro f oun d and lasting influ- 

they needed to believe that a Higher Authority ( . I1( . ( . on u ,r Honstie eomposil ion of North America's 

would provide for their welfare. Most also firmly na , ul ..,| communities, and have largely determined 

believed that God provided lluougl. die deliberate vv | n( .|, S p ( . r ics have proved troublesome in agricul- 

planning and industry of the faithful. Each band of |mv (o| . ( . sts ranf , ( .|ands, and inland waterways and 

European colonists to North America, whether the |. ik( .^ , VA ( .„,| )mn U. 1998). The size of the current 

16th century Spanish or Portuguese c.louisls (Cros- liallira |j Z cd, i.e.. permanent non-indigenous, flora of 

by, 1972), 17th century English colonists to Man- ,, |r ( s js |m| kn()wn precisely but exceeds 2500 

land (Hall, 1910: 92), or the Pilgrims to New Eng- s|)( . ( . j( . s (Karlrsz & Mcacham," 1999). Of these, a 

land (Young, 1846: 42). did not trust then survival sm;i || inillon , v | iaV( . become invasive, i.e., prolific' 

to the happenstance acquisition of food in the New . |M( , ()( . ( . 11 p villf , significant new (i.e., non-native) 

■ National Agriei.llural Library, as well as R. 

Ann. Missouri But. Card. 90: 77-90. 2003. 

ranges. I'lioe species, along wild main oilier nat- 
uralized s|urir>. Iia\c caused enormous environ- 
mental and ee mi. damage dial ranges from 

tlirealening native biodiversity |„ changing llie op- 
eration of major ecosystems (Vlack el aL 2000; Pi 
mentel et al., 2000). 

Furopean settlements in North America arose 
from fledgling colonies of immigrants thai stemmed 

were established at different limes and al widely 
s f )aced locales (CUiinn, 1990). This mixture of cir- 
cumstances and events suggests that the current 
n;ilurali/.ed llnra has had a di\er>e origin and vai- 
ied cm umslaiices surroiiiiding ils eventual persi-- 
tence. This diversity spawns a series of questions. 
What has been the chronology ol plant iialurali/.a 
lions? More importantly, what can be deduced from 
that chronology about the circumstances that led 

Knrope.m>' history in North \uieiiea. while other 

b ■ 

species, if any. formed pre- 1 860 
what sparked the demographic transitions among 
those few naturalized species thai became invasive 
(Crooks A Soule. 1090; kowarik. 1995)? What role 
have species introduced long ago into the U.S. con- 
tinued to play in their new ranges? Kor instance, 
have they been supplanted by more recently arris - 

portal, on tlul so affected the I'.S. for the rest of (he 
19th century and beyond (Kuznets et al.. 1960; 
Yleinig. 1080). These questions about plant entrv. 

immigration histories form part of the infor- 
on upon which the prediction of future inva- 
s will continue to rely (National Peseareh 
ncil, 2002). 

Any assessment of species thai became natural- 
ized long ago In intrinsically retrospective, and the 
record is incomplete and often fragmentary. Four 
general sources that vary substantially in potential 

(I) Herbarium specimens provide unequivocal 
evidence for accurate species identifications. 
Consequently, they may form the best cate- 
gory of evidence, provided the collection 

ly, the number of existing herbarium speci- 
mens across the time period under consid- 
eration here is meager. Few pre- 1860 U.S. 
specimens still exist: even fewer have sur- 
vived from pre-1800. Their use remains an 
unfulfilled goal here. 

(2) Pollen records can prov ide some supplemen- 
tal information (Mrugam. 1978) but are lim- 
ited in their usefulness because many pollen 
taxa are impossible to identify with light mi- 
croscopy and for some families, such as IV 
aceae. only a few genera are possible to iden- 
tify in this way (Moore et al.. 1991). 

(3) Contemporary regional floras became much 
more common after 1800, e.g.. Pursh (1814) 
and Torrey (1819). and can be enormously 

(4) The most common tool examined here con- 
sists of contemporary records and correspon- 
dence of those who provided first-hand ac- 
counts of the species they saw, their ranges, 
and abundances (e.g., Josselyn, 1672; de 
Schweinitz, 1832; Dwight. |1821] 1969; 
Cray. 1842). 

Herbarium specimens, floras, and other records 
van substantially in accuracy and scope, and until 
the later part of the 18th century, almost all em- 
ployed only common names. I have cross-checked 
the common names (e.g.. Fernald. [ 19f>0| 1987; kar- 
tesz ci Vleacham. 1900) and avoided a species ci- 
tation unle^ | could be confident of I he identifica- 
tion. I have been similarlv conservative in a — igiiing 
an a real extent to any non-indigenous species. For- 

coiisislenllv lo i iiiauv Fiiropcan rude ra Is and medic- 
inal planls lor more than 400 years, e.g.. elecampane 
' <" I i. h< nbaiK i//i-/v miiiiiis niger I..). 
borehound (Mnrruhium rulgure I..), anil St. John's 
wort (Hypericum perforatum I..) (Harvey, 1974). In 
addition, a surprisingly high percentage (> 75%) of 
the species encounlcrcd in these early records were 
first described by Linnaeus, and many of those 
names have been retained in plant syslematics. Civ- 
en the limitations in the information that can be re- 
liably deciphered about pre- 1 860 naluiali/alioiis. 
the resulting record will, however, remain a mini- 
mum estimate of the lull scope of plant entry and kartesz and Meachams (\W)) Syn- 
thesis was employed as- 

First Plant Arrivals: Pre-1700 

Preparations for colonization of nev 
North America illustrate the care and f 

i lands in 

«• • • ! I ■.•lit 

the immigrants placed in the i m ■ I « ■ 1 1 ; i k i 1 1 u . l-o- 
I on a continent newly discovered by Europe- 
connected to Europe only by infrequent and 
ly unreliable ship traffic, each colony's plan- 
knew that they needed to become self-suffi- 
t almost from the outset. Acquiring sustained 
ces of food and medicine figured foremost in 
■ planning (Leighton, 1970). The most reliable 
i information we have of plans for colonizing 
h America stems from the establishment of 
1620. Although u.- 
vere in the original 
fest, several lists of plants ordered in Britain 
1628 and 1631 have « 

i Colony, beginning i 

ived. The Endic 
dude "Wheat, r 

barley, oats, a boghead of each in the ear. beans, 
pease, stones of all sorts of fruits, as peaches, 
plumes, filberts, cherries, pear, apple, quince, ker- 
nels, pomegranates, woad seed, sal Iron heads, li- 
quorice seed, (roots sent and madder roots.) pota- 
toes, hop roots, hemp seed, flax seeds against 

winter, coneys, currants plants . . ." [\ ig. I<". I<>: 


This intent to send woad seed (I sat is tinctona L). 
hemp seed (Cannabis sativa E.). and flax seed i /./'//- 
uin usilalissimum I..) forms the earliest record we 
have of the introduction of non-indigenous species 
that remain in the naturalized or at least the adven- 
tive flora of the U.S. Dyers woad. Isatis tinctoria, has 
not apparently become widely naturalized in New 
Kuglaml. despite its exceptionally early arri\al I Ma- 
gee & Abies. 1999: 552). but it has become invasive 
in Utah and other arid regions (Karah et al., 1988). 
Cannabis salira \- widely naturalized in the eastern 
half of the U.S. north of the 37° latitude (Hancy & 
Kazzaz. 1970). In contrast, flax appears commonly 
as an adventive in abandoned fields but may be nat- 
uralized only locally. 

In 1631. just three years after the Endicott Ex- 
pedition, John Winlhrop. Jr.. bought a detailed 
group of seeds from a London grocer with the intent 
of transporting these seeds to the small colony at 
Plymouth. \\ inlhrop's (Massachusetts Historical So- 
ciety. 19.13: 17-48) list also included species that 

(Rnim-x nispiis I., or Rumcx acelosella U.. or both), 
'Tansy" (Tanacetum mlgare I..). "Rockett" (Eruca 
vesicaria subsp. sativa (P. Mill.) Thellung [Eruca 
salira]). "Huglos" (e.g.. Anchusa arrcnsis (L.) 
Bieb.). "fennell" (Eoeniciilum ml pare P. Mill.), 
"dill" (Anetlium grareolens L.). "sweet maioram" 
{Origanum majorana !.. or Origanum vulgare L). 
and "nipp" (Nepeta cat aria E.). or advenlives. such 
as "summer sauorC" (Salureja hortensis L.), "Clary" 
(Salvia si larva L). "hysopp" (llyssopus officinalis 

E.). "marigold" (Calendula officinalis I,) and "hol- 
lihocks" [Althaea rosea E.) (Eernald. |1950J 1987; 
Magee & Abies, 1999). The identification of other 
species on his I i-t i- more equivocal, such as "mal- 
low." which may have referred to Malm moschala 
1... Malm neglecta Walk. Alalia rertivilluta L. or 
Malm parriflora I.. The listing of "|)0|>c>" (Mas- 
sachusetts Historical Society. 1913: 47) may refer 
to Papaver somniferum L., the opium poppy, but 
could also be referring to Papaver rhoeas L., which 
had reputed medic anal properties. \\ inthrop's list 
also includes "pursland" or purslane (Portulaca 
oleracea E.). It is often considered non-indigenou> 
to North America (Magee & Ahles. 1WI|. hut ap- 
pears in the fossil pollen and seed record of Ontario 
in pre-Columbian levels (Byrne & McAndrews, 
1973). Although these species appear to have ar- 
rived with the first v\a\es of colonists in New Eng- 
land, the earliest dale wheal thev l>eeame natural- 

nsive ] 
introduced by 17th century c 
land is derived from Josselyn's 1672 publication 
Sen-England's Rarities Discovered and its 167-1 se- 
quel \n Account of Two Voyages to \eu-England. 
Based on Josselyn's accounts of two yisits. I<>38- 
1639 and 1063-1671 (Josselyn. | 1674] 1988: xiii), 
lie appears to have been a keen observer of the 
condition of the colonies in New England. Eurther- 
Uiore. he deliberately categorized the species that 
he encountered in his travels between Massachu- 
setts arid Maine, noting those thai were apparently 
native to New England, those species also found in 
England (and apparently introduced), and even 
those introduced species that did not thrive in their 

many that are mil among the manifest- ,,l EndicoM 
or Winthrop. The naturalized species include cel- 
andine (Chelidonium majus I..), goose-grass or 
"clivers'" (Calium aparine E.). "Our English Clover- 
grass" (perhaps Trifolium repens E.), speedwell 

laria gramincu E.I. N. John's wort {Hypericum per- 
foratum), sweet brier or eglantine ( eglanteria 
L.), toadflax ({Anuria vulgaris P. Mill), wormwood 
( \rtvmisia absinthium E.). and yarrow (Achillea mil- 
lefolium I..), \mong species now adventive are 
bluellowered pini|)ernel \Anugallis arrcnsis subsp. 
foermna (P. Mill.) Schinz & Thellung], "egrimony" 
(Agrimonia eupatoria E.). herb Robert (Geranium 
rohertianum E.I. '"Oak of Hierusalcm" (Chcnnpodi- 
um botrys E.), speed-well (Veronica officinalis I..), 
spurge tune (Polygonum persicaria E.), "Rew" 
(Ruta graveolens 

nun [\ l Hayek), and wild-mint (Men 
llm aqnuticu I..) (cf. Magee K Abies, l<W). Fur 
thermore, Josselyn (1672: 85) took th 
extraordinary slep of noli rig which species had at 
rived ". . . since the Knglish Planted and kept Cat 
tie in New-England" that were already occiirrin 

•| I mi i I with h urn, i In this list, Jos 

selyn provided e\|ili< il mfoi 'irialion on some of lh, 
first naturalized species (see below). 

The diversity of species in Josselyn's list reveal 

lhal I 


|i il - M ih h eaili.a. the small lists ill species thai 
Kndicolt and Winthrop had ordered for the settlers 
had been expand* ! - i ilfo Mi>-.| pn.niinenl are 
spe. ics that had repiile<l medicinal value ( \rlemisia 
absinthium, Ihp.-rieinn perjomlum. Inula helen- 
iiiin). .is well as plants lor seasonings (Anethum 
graveolens, Focniciiliirn vulgare P. Mill.. Salvia 
sclarea). Perhaps most surprising is lhal the colo- 
nists were beginning to feel confident enough about 
their survival that they permitted themselves the 
luxury of importing a few ornamental species, Lin- 
ana vulvas and Rosa eglanleria (Keighton, 1970). 
both of these speeies would heroine naturalized, 
and L. vulgaris would heroine a scourge In the mid 
lolh century. The arrival of these species is also 
significant because this early entry of ornamental 
species signals a trend that would grow substan- 
tially over the following 350 years as ornamental 
species became the largest single functional cate- 
gory of imported species (Mack K Krnebcrg. 2002). 

<>iii ability to delect the earliest nalut. hzalioi,- 
is severely limited by the fragmentary character of 
the first accounts of European colonists in North 
America. Kor example, the late 10th cenliirv Span 
ish settlements in Klorida included gardens and 
cultivated limes, lemons, and oranges (I, von. I9<;<>: 
'■>'■>): conlciiipoianeous nilrodiici iocs of ouiiigcs -non 
appear in other Spanish subtropical and even tern- 
aware of any information on the fate of these spe- 
cies that were cultivated in 16th century Klorida. 
Maine also r. caved Kuropean colonists very earl>: 

nirig in 0)01. The record referred to as the Kdgei 
ton Manuscripts (Lockwood, 139) is intrigu- 

long-abaiidoued settlement in Maine "on the Pcr- 
marprid River Alderman Alsworth of Bristole set- 
tled a oo., of people in 1625 In 1675 I found 

the Roots and Carder. Ilcrbes. and some old walls 

there when I 

herbs many years after the site's abandonment 
strongly suggests that these species had become 

The first, admittedly tenuous, records of natural- 
ized speeies in what was to become the IS. appear 
in the decades after establishment of Knglish col- 
onies in New Kngland. Among the praiseworthy 
leal u res ol these Knglish immigrants was then com- 

about their nascent settlements, including the fate 
of their crops. Only I I years alter the first colonists 
landed in Plymouth May. William Wood (Wood. 
| 163 1 1 1977) had published his account of the 
"New England's IVospecl": a comprehensive report 
ol information lhal would have proven invaluable 
to anyone intending to immigrate to New Kngland. 
Wood's comments entitled "Of the llerbes. Kruites, 
Woo.K. Wale,- and Minerals" prove particularly in- 
formative about both the Kuropean crops being 
planted and also ihosc already observed to persist 
outside cultivation. 

Woods ([16341 1977: 36) listing „f t |„. (>mps , 
both native and introduced, reveals species that 
had already been brought under cultivation and 
their status: "The mound, adonis yen good kitchen 
(;ardens. for turneps. Parsnips. Carrots. Radishes, 
and Pumpkins. Muskmillions. ls,,uoulers(|ashes. 
(cucumbers. Omvoiis. and whatever grows well in 
Kngland grows wed there, many tilings being belter 
mikI larger; there is likewise all manner of llerbes 
for irieale. and medicine, and thai not oulv in plant- 
ed gardens, hul in the IbW.v icillmnl e\llin the ail 
or (he help of man, as sieeel Marjoram, I'nrselane. 
Sorrel. I'enenall. )arron. Ylirtle. Saxifarilla. limes. 
etc" (italics added). The italicized statement 
strongly suggests that several of lire speeies that the 
colonists had planted in the previous decade had 
already escaped cultivation .itul were growing freely 
in the surrounding forest. These species include 
sorrel (linmex aeel„sa I .. and Uumex erispns) arrd 
varrow I \elnllea millefolium). These are the earliest 
accounts I am aware ol for any naturalizations in 
V.iil vmeriea. hleiil ilical ion of "sweet Mai |m am" 
is equivocal. Wood may have been referring to 
Origanum inujoruna, which is still referred io as 
\ I | I I I i I | I cultivation 

and is not considered naturalized in New Kngland 
today (Magee <.V1 Miles. 1999: R9R). Alternatively, 
he may have seen Origanum vulgare, now termed 

les. 1999: 873). "Peneriall" irr Wood's lisl may refer 
Io the native species Hedeomn pule^ionles |l..| I'cis. 
iMlhci than the K.uropean species. Mentha pulegium 


, i •. j - 

Ins of some ncul- .1 « .1 In .p< in species in New 
Engla iii : iv -iih-i. nil, ill \ ronoboml'-d and e\pa u! 
ed hy Jossebn (1072). His lisl "Of such Plants as 
have sprung up since the English Hauled and kepi 
Cattle in New-England" is the best account from 
which to determine the European species thai had 
heroine nal in a I i/ed bv 107 I and probablv inn. h 
earlier ijossehn. I<>72). Manv ol the species thai 
lie reported in terms indicating their newfound per- 
sistence in the New England flora are still promi- 
nent today (Magee & Aides. 1000). such as Hyo- 
scyamus niger, Stellaria media (L.) VilL Tara.xai urn 
officinale 0. II. Wehher ex Wiggers, Seneew vul- 
garis E., and Hrlita dioica E. ,|ossel\n"s (1072: 80) 
repoil llial pi ' ■■ii'" E.) was re- 

ferred to hy Native Americans as "English-Man's 
foot" has been quoted repeatedly Hantain was al- 

1 I prea 11 1 Ion il d n id> w< II aln id 

of English sellleiiienls and served as a harbinger lo 

the I gme- of the coming of the new settlers 

(Darlington, 1859: 219). 

Aside from the immediate needs lor importing 
plants for food and medicine, the earlv colonists 

hail del' ml 11 1 1 11 11 1 h il llieii liv< sloe k 

had suitable forage. They soon found, he 

lied by one colonist in Massachusetts who stated 
that his livestock ". . . grew lousv with feeding upon 
it, and are much out of heart and likeing . . ." 
(Hutchinson. [1704| 1972: 483). Help was on the 

wav. I he-e luii- 11 1 ii 1 '. 11 ruei- and In id-- 

men were well acquainted with an arrav of ■■English 
grass" that would meet their livestock's needs, and 
ihev aclivelv son-lit seeds of these English pashm- 
s[iecies for introdiielioii into their larms. Species 


17th century accc 
imon reference to £ 
is "English grass," 
e species could hav 

or in the ballast or debris off-loaded from ships with 
livestock, or both (Bidwell & Falconer, 1925). Spe- 
cies that were introduced through the 17th century 
dare E.. Holru.s mollis E.. I'oa 
pratensis L., and possibly Agrostis capilliaris L. 
(Carrier, 1923: 211). and Holms lanatus. Not only 
were these species becoming naturalized in New 
lu-laiid. but ihev were also being actively spread 
in colonies to the south. In praising I In mouth ol 
cattle and other livestock on Eong Island (New 
York), Denton (1670: 5) 1 

land likewise |was| producing excellent English 
grass, the Iced of which was brought out ol Eng- 

limkl (1085: 10) commented on the practice of us- 
ing sheep to disperse these grasses "bill if we sprin- 
kle a little English grass Hay-Seed on the Land 
without Plowing, and then U-cd Sheep on it, in a 
little time it will so increase, that it will cover the 
band with English Crass, like unto our Pastures in 
England. . . ." 

Naturalizations arising from the introduction of 
forage species were not restricted to grasses. The 
advantage- of sowing pastures with (lovers, espe- 

1 I 'mm repens (while clover), were common 

knowledge in England bv the mid I 7th century, and 
this knowledge was widelv transferred lo North 
America. As early as 1635 a tract written for those 
who planned to immigrate to Man land advised that 
ihev bring "Cood -lore of clavei grasse seede. to 
make good meadow" (Hall. 1910: 98). This sound 
advice had also been followed in New England: Jo- 
selvn (| 0.71 J 1088,: 131) commented that "Our En- 
glish clover grass sovven thrives very well." It was 
both deliberatelv -own and spread as sheep were 
moved among fields, as Scot (1085: 187-188) re- 
corded for New Jersey: "As soon as any of the land 
here comes to be cultivated, il over-runs with small 
Claver-grass. bv the pasturage and dunging of the 
cattle, and so supplants the naturall grass and 
hearbs." The movement of livestock and seed 
among the colonies was extending the new ranges 
for these European pasture species. 

By the mid to late 17th century at least 20 spe- 
cies were observed to have abends become natu- 
ralized in New England: manv of these had likeh 
become established further south in Pennsylvania. 
The actual number of naturalized species was likely 
much larger. For example. Josselyn (I (.72) listed 
under his category "Of such Plants as are common 
with us in England" many species that were aheadv 
in New England and for which we have records of 
naturalization in the 18th eenturv. How long before 
1700 Ihev were naturalized we do not know. Spe- 
cies in this calegorv of I7lh eenturv iulioihi. u.»i- 

Arlemisia absinthium. Cannaln 
aparine. Hypericum per/t 
Sepeta calaria. and \anll 

Any chronological divisions, such as those be- 
tween centuries, that could be u-ed to delineate the 
growth of a naturalized flora are arbitrary. The lore- 

t brought new specii 

i North \nieric, 

ally partitioned ; 

t(l COIllolTM 1(1 

breaks. Nev- 
ertheless, there is some justification lot -distinguish- 
ing between the nalnrali/atioiis in the 17th anil the 
18th centuries. Leighton (1976: 1) has argued that 
the necessity of establishing self-sufficient colonies 
in the 1 Till ceiilurv ua- ,i j •< >v\ . tf u I incentive to 
ensure that almost all the species introduced before 
1700 were Co i M 1 1 1 1 !. ii c ii i purposes. As noted above. 
-nine pi. ml- introduced even bclore I ( i.")0 mav ha\e 
had no purpose olhei than ..rnameiitation, such as 
/»'. eglonlerm. lint c\cn Ii. eglanleria may have 

been deliheraielv introduced to I Iiedgc> (Jo-- 

selyn, 1672: 90). In contrast, the assurance that the 
colonies would not only survive but also mow and 
prospei -parked miieh incentive in the 18th ccnturv 
for the introduction of species for the widest range 
of uses, especially for ornamenlation. This market 
ui ornuiui spc( ii - ;;iv« -leadiK throughout the 
b'lth ceiilurv ami llicreliv created opportunity for 
more naturalizations, and even invasions (Lock- 
wood, 1931: 12; Leighton. 1976). 

Perhaps the best single view we have of the 
growth of this naturalized flora in the 18th century 
was a retrospective written in the early part of the 
19th ccnturv but clearly drawn I mm information 
gathered in the late 18th centun. Hafines<|iie. an 
itinerant Lrencli botanist, collected plants vvidelv 
in the new United States. In 1810 he assessed the 
naturalized flora of the Middle Atlantic states (Raf- 
inesque. l."llL Mis work appears to be based 
largely on Muhlenberg's (1793) Mora of the area 
surrounding Lancaster. IVniisv l\ auia. Index Florae 
hincastriensis. which Rafinesque augmented and 
annotated to include hi- own oli-ei v alion- lor ill.- 
North American region that stretched south from 
New York State to Maryland. The mosl specific lo- 
cation inloinialion h given for species occurring 
near Lancaster. I'etin.-v Kania. and those found in 
the vicinity of Rallinmre. Man land. l!aluicM|ue"- 
home in the U.S. 

In a useful pattern that was to be repealed by 
other authors in the 19th centun. Rafinesque 
(181 1) categorized the species on their mode of in- 
troduction: plant- introduced In agriculture, plants 
introduced hv gardening, and "lotallv useless 

Plants in a fourth category for him had varied 
modes of introductions and were not native to bu- 
rope. A great advantage of this Muhlciibcig cum 

planl svslcmalie-. In addition. Raiiiiesque gave his 

abundance of the species, mak- 

cuntliemum rulgare Lam. [which he termed Chr\- 
sanllicmum leucanlhemiim]. and Yerhascum 
hlatlariu L.) were Loth widespread and abundant. 
Among deliheraielv introduced species are Can- 
nabis salira. lion/emu nilgnrc I... I, mum usilatis- 
s'mumi. I'lanlago lanceolata L.. and Trifolittm pra- 

- all listed t 

I ' i. ! I 

i.e.. species grown in small plots and presumably 
cultivated much more assiduously than field crops 
(e.g.. Cleoclmma hederacea I... Sinapis alha L.|. The 
second eategorv reveals the extensive naturalization 
of medicinal plants and those used for -ea-oiimg- 
in the 18th ccnturv: \sparagus officinalis L.. Cy- 
noglossum qfjiciiiale I .. Digitalis puij>urea I... Inula 
Helen i um. Marrubiitm rulgare. \epeta calaria, and 
Tragopogon porrifoliiis L. Most of these species 

dwellings and gardens. \ ng tin- large list of spe- 
cies are also those such a- Cicliorium iuhbus L. 
(chicory), which was listed as "verv common — in 
fields, roads, cultivated grounds. . . ,'" and /.annum 
amp/ex icaule I... considered, "even where com- 
mon—in fields." 

Many more species bad been imported for strict- 
ly aesthetic reasons in the loth compared with the 
17th ccnturv. and among these were some that -oon 
became naturalized. Ornamental species that had 
already become naturalized ca. 1800 in Rafin- 
esque's opinion included Fuonymus eurojiaea I... 
Liguslruin rulgare I... Rhainnus calhartica L.. and 
S\rin»-a rulgari* I... illustrating that woody orna- 
mental species were also becoming persistent. The 
remainder ol the -pecie- noted |>\ Raiiiiesque were 
considered accidental!) introduced, such as Hro- 
inus secalinus I... Ecbmm rulgare I ... winch was to 
become much more conspicuous later. \gro\tcnuna 
git ha go I... Clienopodium allium \ ... Com olrulus ar- 
rensis. and Spergulo arrcnsi.s I . 

Rafinesque (1811) mav have applied the appel- 
lation naturalized rather liberally, as he used it 
to mean those species thai 

listed some species, s M ch as Fagopyrw 
esculenluin Moench (buckwheat), as naturalize 
and common, vet lodav it is probablv only advei 
live as an escape from cultivated fields. Alleriu 
tively. its status mav have indeed changed in ll 
last !()[) veai- (e.g.. lion diminished cultivation 
or he may haw been simply noticing volunte. 

was Rubin tinctnna I,. (madder), a species referred 
to among -mm- of the earliest plant import mani- 
fests in the 17th century (Young, 1846: 42) but 
which is considered a- oulv a ran- e-cape from cul- 
tivation today (Magee & Ahles, 1999). Neverthe- 
less, the hulk of the species that he noted as nat- 
uralized are indeed persistent today, and his 
■ corroborated by 

The major consequences of non-indigenous -pe- 

speeie- lliai beet .me prolific in the new range, i.e.. 
these species become invaders. Their abundance 
and aggressive growth brim: about eu\ iroimienlal 
damage to the native species and alteration of na- 
tive environments (Mack et al.. 2000). These spe- 
cies often invade arable fields and pastures and 
consequent^ result ill severe economic damage 
(Bridges. 1992). (dven the growing array of non- 
iridigenou- species that were being i Ml induced both 
ae, ideiilalh and deliberaleb from the earls 17lh 

eventually reported in terms that we would equate 
with plant invaders. 

-idled from the introduction of Linaria vulgaris P. 

Mill l\ellow loadllax). In the extensive correspon- 
dence dial John Bahrain, the dovti of li'.lh cenluiy 
American botanv. maintained with his colleagues 
Peter Collinson and Philip Miller in Britain (see 
Darlington, [1849| 1967). his remarks about L vul- 
garis are revealing. In a report that both colleagues 
apparently received accompanying a letter in 1758, 
Bartram stated, "It was first introduced as a fine 
garden Mower: but it was never more heartily cursed 
by those that suffer by its encroachment" (Darling- 
ton, [1849] 1967: 384). He added. "It is the most 

northern climate. Neither the spade, plough, nor 

ture" (Darlington. |I8I«>| I «io7: 383). Bartram ex- 
plained that so desperate were farmers to control 
yellow toadflax that they would even ignite log piles 
in a field in the hopes of destroying it in the soil 
but to no avail. Bartram reported that by 1758 L. 
vulgaris had ". . . spread over great part (sic) of the 
inhabited parts of Pennsylvania'" (Darlington. 
[1849] 1967: 384)— a clear indication that it was 
not simply a local problem. Linaria vulgaris had 

formed an invasion. It apparently spread Initial 
afield: Pursh (1814) noted that it had become ". . . 

several parts of Pennsylvania and \ irginia." Barton 
(1818) went even further in his estimation of its 

spread as he rated it (which he termed a- \ntirrlii- 
num linaria) as ". . . extensively naturalized, in the 

the borders of fields. \,r\ common and abundant." 
Yellow toadllax had reached North Carolina by 
1832 because dc Schwcinitz ( 1832) reported that a 
few years after /.. vulgaris (as Antirrhinum linaria) 
wa> introduced into a garden il hail ". . . contami- 
nated the whole vicinity for many miles." 

Other naturalized species were also causing se- 
rious problems, although the extent of their new 
ranges is more difficult to determi 
in Philadelphia but had numeroi 
along the East Coast, so some o 
of the damage 

i part I 


example. Bartram in the same 1758 corres| - 

deuce listed other non-ind igen. >u- species in what 
appear- to be a dec lining order of prominence. He 
described /hjH'ridirn pcr/aralnm a- a "very perni- 
cious weed."" which had spread over pastures in 
eastern Pennsylvania, interfering with the growth of 
pasture grasses and causing injury to horses and 
sheep (Darlington. |I849] 1967: 384). This plant 
was apparently introduced repeatedly in New Eng- 
land and the Middle Atlantic colonics (Josselyn. 
1672: 44; Haughton. 1978: 318). so it may well 
have formed an invasion bv the mid I7lh century. 
Hypericum perforatum was also proving to be a 
troublesome plant in fields in New Kuglaud. Eliot. 
writing in his fourth essay on agriculture, which 
was first published in 1753. complained about the 
difficnllv of eliminating St. John's wort in fields 
(Carman & Tugwell. 1934: 94). Even though Eliot 
lived in Connecticut, he traveled widely in New 
England, and his essays were meant to be advice 
gleaned from decades of observation in the region 
(Carman <X Tugwell. 1934). If//, perforatum had 

18th century, it was at least a widely distributed 

rail rali/c <i specie-. 

Linaria vulgaris and Hypericum perforatum be- 
deviled 18th century farmers at ros- a broad region 
from Pennsylvania to New England, but these spe- 
cies' prominence appear- to have -nice declined. 
Darlington I 1850: 225) reported that L. vulgaris was 

our pastures and upland meadows."" He described 

farms."" but did not state it in terms thai suggest an 

Cheslei Counts. I'ciiii-\ Kama, he noticed that the 
plant was not detected at all in 1842, and was rare 
in 1843. but had become "as common as ever" in 
subsequent years (Darlington. 1859: 55). This 
slaloiih-ni may allude to the prominence reported 
bv 1>. rlran i'i ihe I « .1 ! '.'nine, |( ',-ira i A ln ; . -.1:111 . 


Although v\idel\ nalni-ali/ed in die norlheastern 
U.S. (Uhoads K Klein. 1993; Magee & Ahles, 
1999), neither //. perforatum nor /.. vulgaris ap- 
pears invasive to<lay. However, these reduced roles 
are largely the result of active control measures. 
Hnaria vulgaris is controlled wild the herbieide 
glyphosate in agricultural fields (Saner et al.. 1995). 
while the abundance of //. perforatum has been ef- 

fectively < 

■ ■:! 

Kliol and Darlington were |>i iin.ii i l\ interested in 
specie-, dial were hazards to agriculture. Fven bar- 
tram's observations appear often influenced hv his 
concern about agriculture. Alien plants that were 
extensive exclusively in non-agricultural soilings 
may not have been commented on. 

As damaging as L. vulgaris and H. perforatum 
could be, colonial farmers already had a much 
greater scourge to cope with lierheris vulgaris L 
(common barberry), the alternate host for the stem 
rust (I'uecitiia graminis f. sp. triliei). a devastating 
parasite of cereal crops (Peterson. 2001). lierheris 
vulgaris had certainly reached the level of impact 
to qualify as an invader In the late 18th century 
.i:i:l I! . «! jiiobably reached that status over a century 

earlier. Our kiio !' i ii | bable entry into 

North \merica hv the mid I 7th e.-nl urv i> through 

as a medicinal plant (Cerard. [ I033| 1075: 1320). 
Josselyn (1072) referred to ••barberrv trees" in a list 
of introduced fruit trees and also described the rust 
(termed "wheal blast" until the 20th century) on 
wheat in New Kngland m the 1 000s. A more de- 
provided hv John \\ inlhrop in 1008: "generally 
through all the plantations, both ol ve Vlassacheu- 
setts colony. Plymouth. & this also [the colony of 
Connecticut! I insomuch that the eroppe of wheat 
hath failed divers veares in most plantations. The 
come Nourished we I it came to be eared, and 

the cares also would appeare lane, anil as il lull. 
but no conic in them. There have lieeiic thousands 
of acres in that nianer every veare. What the cause 
was. whether nalurall. or a blasting fro heaven we 
know not. Our old husbandmen of England, some 
of them thought it a meldovv . . ." (bidwell & Fal- 
coner, 1925: 13). (liven the obligate link between 
wheat, stem rust, and barberrv. it seems a safe in- 
ference that where rust was attacking wheal. bar- 
As further evidence for the spread and impact of 
lierheris vulgaris. Connecticut passed legislation in 
I 720 to control barberry, followed by Massachu- 
setts and Rhode Island in 1755 and 1772. respec- 
tively (Fulling. 1943). These measures failed, as 
llwighl (| 1821 ] 1909) provided direct observation 
in 1795 or I 7 ( >0 ol the extent of/A vulgaris across 

Hampshire. Within the approximately 3000 km' 

noted ". . . the barberrv bush is spread, not uni- 
versally, but in spots, and those often extensive. In 
some fields tbev occupy a sixth, fifth, and even a 
fourth of the surface" (Dwight. |1821| 1909: 270). 
Clearly, barberry was exerting a major influence on 
wheal production across eastern Massachusetts in 
the 18lh centun to the point that bread made from 
wheat had disappeared from farmers" diets in much 
of New Kngland (Midwell cN Falconer, 1925: 92). 
Common barberrv would laid spread much further 
across the I .S.. and l>\ the lime control efforts were 
fully implemented against il early in the 20th cen- 

Central region of the U.S. (Hutton, 1927). 

Other biolic invasions were growing in North 
America by the late 18th century, e.g.. the spread 
of Trifolium repeus. even it the proliferation of an 
alien clover does not produce the usual anthrnpo- 
eenlric connotations. As stated previously. '/.' repeus 
hail boon spread both deliberately and accidentally 
through New Kngland and colonies, such as Penn- 
sy hatha and New Jersey, in the 1 7th century. The 
resulting transformation ol pastures, for which it 

soil nitrogen poo! in these sites. Such change can 
precipitate a hos| of other environmental changes. 
including a facilitation in the establishment of other 
alien species. For example, introduction of the ni- 
Irogeii-lixiug Fin-tree. Morella faui (Ail.) Wilbur, 
so raised the amount of biologically available nitro- 
gen in Hawaiian soils that I octree has favored the 
persistence of other non-indigenous species (Yitou- 
sek et al.. 1996). 

1,'MMI I8«><1 

All the incentives and stimuli that had operated 
before 1800 for the introduction of non-iin I i-em .1 1~ 
plants expanded sub-lanl i.ilb with the growth of 
the new nation's commerce and Iransporlalion in 
the 60 years leading up to the Civil War in 1861 
(Meyer. 1917). New s|)ecies were actively sought 

and for the (irsl time these searches were not left 
simply to private enterprise. In an often-quoted 
proclamalion. President John Quiney Adams in 
1827 instructed U.S. consular offices to gather use- 
ful species and I .S. na\al ship captains lo provide 
for the transport of these living cargoes to the U.S. 
(Hodge cK Lrlanson, 1956). Annual reports of the 
Commis-ionor of Patent-, who was responsible for 
federal involvement in agriculture before I8<>3. 
routinely chronicled federal interest in introducing 
new species in the I .S. for potential use (e.g.. I .S. 
28th Congress 1st Session. 1844: 68: U.S. 36th 
Congress. 2nd Session. Kx. Doc. 48. 186J). 

Kederal actions were dwarfed, however, by the 
private sector's economic incentive to import spe- 
cies new to the U.S. for all manner of use. parlic- 

uurseries and seedsmen in major cities, particularly 
Philadelphia, even before the American Revolution 
(Uockwood. 12). was followed b> a huge 
growth in this cottage industry through the first half 
of the 19th century (Leighton, 1987: 67: Mack. 
1991, and references therein). 1'hese seedsmen 
were remarkably diligent, not only acquiring new 
species from overseas bill also in building (hen 
teles that were not restricted lo the immediate vi- 
cinity of their businesses and gardens. By 1804 
Bernard M'Mahon, a Philadelphia seed merchant, 
was advertising his nursery col- 

lection species from such far-flung locales as the 
"South-Sea Islands." Asia. Africa, and Kurope 
(M'Mahon. 1804). The product of such industry was 
the availabililv of several hundred species for sale 
that had not before entered the U.S. (I.eighlon. 
1987: Mack. 1991). Many of these species were to 
become naturalized by I960, such as Casitarina 
,: 1 .. ( \p>ni\ rs, iilrnhis I .. IIkIm liittm 
ixanliiciitinum Shcpard ex ker-Cawl.. linker a ja- 
ponicd Thiuib.. /AMiiKicliid iniinmularia I... 1/o/WA/ 
fa ya. Tamarix spp.. and / lex curopaeus L. (Mack, 


ally single 

may have spawned a naturalizati 
ins (L.) Link, Scotch broom, is i 
ing shrub that rapidly covers ne 

« niiibin.i i tn ni • i ili • o i'.' limn and local 
seed dispersal from explosive capsules. Once it oc- 
cupies a site, it can lorm an unpenetrable thicket 
that reduces the prevalence of oilier plants and in- 

& Prasad. 1998). Although there are apocryphal 
reports of earlier introductions il.o.kvvood. 1931: 
32). the earliest clear reference to the shrub in Vir- 
ginia arise- in a lellei ( \noiivtuoiis. 1921) detailing 
the misguided generosity of J. M. Call, a visitor to 
the farm of John Cocke in Mount Pleasant. Virginia. 
in 1803. Learning that Cocke intended to introduce 
.in unidentified species to hi- farm as sheep fodder. 
Call wrote wilh news about what he deemed a su- 
perior choice: "When I was at your house you men- 
lioned your Intention of Cultivating the Pride ol 
China for feeding sheep. This will answer for the 
winter months very well. It did not occur to me then 
to recommend to your notice the cultivation ol 
N olch Broom, which affords an ample food for be- 
tween two or three -uuiiiier months lor sheep and 
hogs. . . ." Dr. Calt gave more than advice; be also 
sent along seeds of C. scoparins: "I have sent you 
seed sufficient to plant all \our hill -ides ihal vim 
do not mean lo cultivate in grain." The seeds had 
been imported by Call from a farm in Warwick, 
Kngland. a lew vears earlier. His account of the 
plant"- behav ior on the Warwick I arm seems partic- 
ularly ominous in retrospect ". . . it was original!) 
planted as a hedge by an old Knglishman — from 
which it has spread over some hundreds of acres 
of land by the Birds." To ensure thai the spread of 
Scotch Broom would be complele. Dr. (.alt help- 
full) added the following tip: "in Kngland the\ have 
a method ol Kxpedilim: \e-elalion of Broom — 
Hawthorn and lloll — by mixing the seeds with the 
i^\ of their horn'd Cattle & keeping the Cattle up 
until the\ have passed lite seed llrev then sprin- 
kle this over their Land & plough il in, in the fall 
season, in the spring the seed will vegetate . . ." 
(Anonymous. 1921). Unknown is whether the land 
owner, John Cocke. aelualU sowed the seeds given 
by Gait. However. C. scoparins remains naturalized 
in Virginia in scattered locales (Harvill et al., 

Records of the escapes and naturalizations of 
these new iiiiinigi.ini species, as well ;h confirma- 
tion ol lire naliuah/ed stains of mam other species 
imported much earlier, are established through the 

proliferation of published lo< al and n gi I I n - 

. I. i g much of the fast Coast of the U.S. and even 
al newK eslabl si ed inland set: rnx nls (Sulli; ar . 
1840). Pursh (1814) and Torrey (1824) exemplified 
ambitious . ailv atlempls to record floras that were 
nol (.inlined to ui ban seaporls. Their records are 

able because these e 
v assigned Latin bit 
and often reported 

l.S. i 

unambiguous terms 
whether these species were naturalized. Tims. 
Pursh (1814) was able to describe the noxious alien 

grass I'Jeusinc indica (E.) Caertn. as occurring in 
sandy soil from New Jersey to Florida. Fesluca ela- 
tior | probably hdium pratense (Huds.) S. J. Dar- 
byshire| as occurring in wet meadows in Pennsyl- 
vania and New England, and / rl'n a diou a a- loimd 
from Canada to the Carolinas. K\eu iftlie ueograph- 

these species is nonetheless \aluahle. e.g.. barton's 
(1818) Compendium Jlnrae philadelphicae, which 
was restricted to sites within 10 miles of Philadel- 
phia. For example. Parton (IH18) described Ra- 
nunculus hulhosus I., as so abundant that 
whole held- an often rendered \e||ow h\ the pro- 
fusion of the plant" and Allium rineale E. as com- 
mon, pestiferous, and "". . . impossible lo eradicate." 

As valuable as published floras of the early 19th 
century are in tracing the earliest record and fate 
ol non-indigenous species, authors such as Pursh, 
Torrey, and Barton were not concerned specifically 
with these species and their effect in the U.S. Com- 
ing approximately .'50 years apart, the accounts of 
de Sehweinitz (1832) and Darlington (1859) pro- 
vide extra.. rdinarilv valuable benchmark accounts 
that deal explicitly with the scope and status of 
non-indigenous species along the East Coast of the 
I .S. Mlhough the detail of then investigations and 
observations differ in geographic range, these two 
accounts provide perhaps the l>es| guides we have 

<»'" ihe growth of the n; ali/e,| ilura across the first 

half of the 19th century. 

De Sehweinitz (18.32) explicitly dealt with spe- 

of accidentally introduced species included Allium 
inutile, (.cm-.!:, ||»i I iM > , , .■,,... 

fmlanum siihsp. rul^irr (Mailman) Creuler & liui- 
de| |. and l.nlium perenne I.. These species may have 
been introduced in the 18th century as there is no 
mention of them before 1700. Species that were 
naturalized locally, i.e.. without the extensive new 
range occupation he observed for others, included 
Anagallis arrensis I... liriza media I... Hmmtis bor- 
daceus L. Daclylis glomerala I... and Inula he/en- 
mm. I nlorliinalelv. de Shueuiil/ did not describe 
these species in terms of their abundance and im- 
pact, especially in terms that would allow evalua- 

Weeds and Useful 

Dnrliiiglons (1859) ; 
Plants provides probably the best * 
of the composition and impact of 
nous flora ca. 1860. Darlington wa 

lized i 


agriculture: then as now most agricultural weeds 
are- non-indigerioiis (bridges. I0O2). lie recorded 
about 400 non-indigenous laxa that were estab- 
lished in the e.i.-lern thud ..I the country, but it is 
clear that his collections and observations are 
drawn from the Middle Atlantic States, the collect- 
ing ana <■! Ilafincsque 50 \.ais earlier. Darlington 
repeatedly referred to species as "naturalized" 
(e.g.. Sisymbrium officinale (I..) Scop.), in contrast 
to other species thai were merely present in the 
U.S. by the 1850s. for other species, it certainly 
appears thai he also considered them as natural- 
ized, e.g.. in referring lo ihe Scotch thi-lle ( hiopor- 
dum acantbium L. as ". . . very common along 
road-sides and in waste places in New England" 
(Darlington. 185<> : |0<>i. Inte. .-hug in lliis regard 
of all Galium species, which he 

contaminants (•■Introduced fortuitously wilh agri- 
cultural seeds") (de Sehweinitz. 1852: 151). De 
Sehweinitz provided a separate list of plants that 

div idual stales or urban areas. I urlhernioie. he pro- 
species: ". . . which are regularly reproduced, and 
gradually extending themselves, without present 
cultivation . . ." (de Sehweinitz. 1832: 149). His list 

ol deliberately introduced species included many 
thai had been introduced at least a century earlier: 
Anthoxantkum odoralum E.. \epeta calana. I'lan- 
lago maim I, ■ ' i i , 

lhapsus E. P.ul this category also included apparent 
newcomers, such as Harbarea vulgaris Ait. f.. Poa 

disiui-sed as ". . . not sufficiently important even as 
weeds to require notice" (Darlington. 1859: 164). 
lb- reported that the flax dodder Cuscutu cpilinum 
Weihe had become quite rare because of the de- 
cline in the cultivation ol /., 

host. He ah 
as Ailanthus altissima (P. Mill.) Swingle ( 
heaven), had both beneficial and detriment 
ities: providing urban shade but also esca 
vacant lots and even emerging from pavem. 

Plant Invasions by the Mid 19th Ceimtu 

ilv dependent on the im- 
isporle.l lo many suitable 
M Moody ci Mack. 1088). 
' species are readily car- 
iimals < Pulley. 1030) an.l 

Volume 90, Number ' 

c;ni rapidh fill a new range with little or no human 
as-i-lancc. 1 > 1 1 1 ihr spiead ol main other- within a 
potential new range is greatlv facilitated by human 
agencies. Coiise<|iieiillv. a- the network of roads, 
canals, and railroad route- grew in the 19th century 
U.S.. the spread of non-indigenous specie- also ex- 
panded. Some of these route- or pathways were be- 
ing developed even before 1800. For example. 

postal roads that linked the country from north to 
south by 1804 (Paullin. 1932). Additional roads, 
supplemented by canals and established barge traf- 
fic along major rivers, such as the Hudson, the 
Ohio, and the Delaware, extended this network 
(Meyer, 1917). Even by the early 19th century, 
commerce, including seeds and s 
cargo, was moving routinely throughout the i 



swell as the volume of goods moved in the interior 
F the country, increased markedly with the growth 
r railroad lines. The fust U.S. railroad routes were 
uilt in the 1839s. New lines were added rapidh. 
I most importantly, the 

reported dial ill the northern stales he had seen it 
onlv as an occasional roadside plant. Darlington 
(1859: 212) later reported that he had seen it in 
"i on-iderable <|iiantilie-"' in Mar> land and in abun- 
dance in New York. Vipers bugloss may ha\e been 
introduced deliberately among these widely sepa- 
rated sites because it was \alueil as a medicinal 
plant ( Parsons & Cuthbertson. 1992: 332). 

It is intriguing that this plant, which was so 
prominent in the mid 19th century in Virginia and 
elsewhere, would be viewed today as locallv abun- 
dant but not invasive anywhere in the U.S. (Uoren/.i 
& Jeffrey. 1987: 215). The abundance of a plant 
invader can decline precipitously, e.g.. Agrostemma 
githago in Britain (Clement & Foster. 1994). 
through a change in agricultural practices. Given 
the need to control plant invasions, understanding 

ginia becomes an important topic for experimen- 
tation. The ability to idenlilv this species" pollen in 

19lli-eeiilur\ sediments could aid in this investi- 
gation (cf. McGlone & Basher. 1995). 

The list of invasions under way by 1860 likely 

included more species than (.W/v//s sapurius and 

, that goods could I 

OVed I 

1.2(H) - 

1 of eviden 

the eastern half of the nation (Meyer. 1917: 573). 
Growth over the following ten years would dwarf 
even this total (Meyer. 1917. plate 5). 

At least one plant invasion appears to have been 
added in the early l { >lh centun to those that had 
begun earlier. Asa Cray, who was to become the 
doyen of American botanv in the second half of the 
19th centurv. was bv his estimation the first bota- 
nist to explore the Shenandoah Valley (Grav. 1842). 
I pou reaching Winchester, \irgiuia. at the north- 
ern end of the valley in June 1841. he traveled 
south. Throughout the broad vallev for over a hun- 
dred miles Gray was amazed to see immense 
amounts ol Kchium rulgare I., (viper's bugloss). a 
Eurasian biennial, occupying main sites, including 
cultivated fields. Arriving in late June Gray saw 
vipci'- bugloss in lull Mower and described how it 
formed a "... broad expanse of brilliant blue'" 
((way. 1842: 13). Gray's account of the geographic 
spread and prominence of K. rulgare at this time 

sion. He was surprised that farmers had allowed 
the plant to overrun their fields. Their reluctance 
to remove it mav have stemmed in part from the 
dilliciillv of handling it. as it causes contact der- 
matitis (Magee & Aides. 1999). Gray (1842) further 

(1859) was concerned primarilv with the spread 
and damage of weedy species in agriculture and 
onlv iik •idetilallv with those species" occupation of 
other sites. Nevertheless . he did describe the range 
and impact of several dozen species in such terms 
that suggest these were invaders. In fact, he de- 
scribed Aegopodium podagraria U. (goutvveedi a- an 
invader that ". . . should be carelullv watched and 
its spread arrested" (Darlington. 1859: 151). The 
strength ol the descriptors he used for a few species 
is a guide to their unpad, for example. Darlington 
(1859: 197) referred to Cirsium arvense (L.) Scop, 
as ". . . perhaps, the most execrable weed that has 
yet invaded the farms of our country." Similar lan- 
guage was applied to the spread and impact of C\- 
perus rolundus U. in its role on cultivated ground 
in the South, especially in sandv fields and sand 
drifts and along die seacoa-l. I.eucunthcmum rul- 
gare Lam. (which he termed as Chrysanthemum 
leucanthemum) mav have also reached the -lain- of 
an invader because Darlington (1859: 189) de- 
scribed it not onlv as a "ureal nuisance in our coun- 

chisive possc-sion of their pasture fields." It is 
apparent in Darlington's description of the slalu- of 
Kchium nilgare and Berheris vulgaris that these two 
earlier recognized invader- had maintained their 
role until at least the mid 19th century. 

Collectively these invasions had already sparked 

imaging contaminants (Darlington, 1859: ported seed and oilier cargo. Knrope's colonizing 

en if farmers had universally accepted the flora had been largely transplanted to form much 

t would have been difficult to carry it into of North Americas naturalized flora. Although 

Threshing equipment in the early 19|h formed by happenstance, this link remains. 

(e.g.. Arena fat ua L. Broniits secalimts) that con- tally spread in commerce among the eastern con- 
taminated crop seeds, and a great wave of new in- nies beginning in the 17th century, a few species 
troductions were headed to the U.S. (Mack. 1991). became so widespread and naturalized that they 
As a result, many of the naturalized species thai formed invasions. Although the known list of ap- 
were to become ruderals in the U.S. were not only parently invasive species by IfiOO is modest, other 
in the country by 1800. they had been spread species were pmbahlv plavini: that role but were 
throughout much of the eastern half of the U.S. by not described in terms that we can decipher as in- 
1860. In contrast, many other species that have be- vasions. The frequency with which widespread and 
come invaders in the U.S. were just being detected perhaps invasive species were reported ilirough the 
by this date, such as liromus tectormn K. (Bartlett first half of the I9lh century suggests that these 
et al„ 2002) and Lmiccra japonica (Seluereiibeek species* opportunity to spread and consequently 
et al., 1994) or had yet to be detected (r.£..Sahola proliferate was lied lo the growth of all forms of 
kali L). In a sense, the damage caused by plant commerce and the forms of transportation that fos- 
invaders in the first 200 years or more of Kuropean tered the spread of commerce. 

colonization along the eastern coast of the United furthermore, the connection between which spe- 

States would be far outweighed by the damage eies received cultivation in their new range and 

brought about by species introduced or deliberately those that became naturalized appears high (Mack 

spread post- 1860. & Krneberg, 2002). This historic link between cul- 
tivation and naturalization (Mack. 2<MX)) provides 

Conclusions an important harbinger for the future. If the history 
of plant naturalization between 1034 and 1860 is 

Several timely observations can be gleaned from any guide, future naturalizations will be largely 

tracing the growth of the naturalized flora in the shaped by (1) the often idiosyncratic human moti- 

U.S. between the early 17th and mid 19th eentii- various for iniporling alien species. (2) the degree 

. From the beginnings of Kuropean colonization of cult i\ at ion provided lo these species upon their 

in Norll. \n 

entry, and (Ml human> in transporting those 

s shaped strongly by the species Im- species to many new locales and habitats i 

ts selected for their transplanted ag- ranges, thereby enhancing the opportunities 

Even if a naturalized species did not owe tablishment on sites where they can persist v 

arrived as a contaminant in the seed ..[ a deliber- 
ately selected species; Pursh (1814) maintained Literature Cited 

that Anllioxanthiim adaratiim and Festuea elatior Anonymous. 1921. Scotch Umom. Letter of Dr. J. M. 

[probably hdiiim pralense (Huds.) S. J. Darby slure| Call. Dec. 9. li'.o:',. Ty IcrV »,»narler|y Historical ami Ce- 

arrived in this manner. Although the link belween n.-al-.-i.-;,! Marine 2 |.'i) : 216-217. 

naturalization and the <leliberate introduction of l!:l111 '"- ' ■ ^ • s - 1 X "^ lk * ""• V ^« k - -™- C-netic 

species for food, forage, and medicinal use eontin- in „ u . ( ,, strm , „,„,,, S| „,., ,„„.,, , |5o( ,, (); (m _ 

ued, even in the I . lb < i iilui \ some species appear 612. 

to have been imported for aesthetic reasons. The fiaiton. \\. P. C. UH8. ( :..n.|..n. i nun Mora.- Philadelphi- 

list of species selected as ornamentals has expand- ,ar: Conlaiiiinj. a Description of t' 

. . • r • i r i i r I\:.l...- ; .li^,l l>l : ,„ls l-',„,,wl ,.!.!,;., .. ( 

ed ever since; in fact, it dwarfed the number of ne 

immigrant species in more utilitarian categories 1 

at least 1860, if not earlier (Mack cK Krneberg. Bidw.-ll. WW. K J. I. Falconer. 1 023. History of Agri- 

2002). culture in the Northern United States, " ~~ 

The link between agriculture in western Europe 

and the establishment of colonial agriculture was ""^ vMStete. Weed Sci! 

so strong thai by ea. I8:>0, the weed and Inderal Champaign. Illinois. 

and Philadelphia. Vols. 1 & 2. M. Carey & Sons. 

Institution of \ 

1692. Crop Losses Due to Weeds ii 

Briigam. U. M. 1078. I»c .1 l.-n indicators of laud-use change 
in southern Connecticut. Oual. Bes. 0: 349-362. 

Budd, T. 1685. G I Order KslaUished in IV.insilvania 

y\ New-Jersey in America. \\. Bradlord. I'ln lad. l| ,liia. 
(Microfilm. Ann Arbor. Michigan. I diversity Microfilms 


!<»,.». I',r < ,:\ 

Byrne, R. & J. H. McA 

purslane (Portulaca ale 

ture 253: 726-727. 
Carman, H. J. & R. G. I 

|tcrs. I 7 MM 762 l)\ Flint Jared. ' . I n .i; 

i".'. i. i:-^- 

Hid (l||,r, I'., 


Carrier, L. 1923. The B 

ica. McGraw-Hill, Ne 

Clement, E. J. & M. C. 

(.rooks. J. 

. K. Soul 


\m. new, Weeds and I selnl Plants. 2nd 
In G. Thiirber). A. (). Moore. New Wk. 

Bradley. London. (Microfilm. \ni 
University Microfilms [n.d.] Ameri 
Reel 7.68.) 
Dwight,T. [1821] 1969. Travels in Ne 

Cress. Cambridge. Massachusetts. 
Fatah. K. ().. A. F. Tanaka & N. E. 

Iiiu-I,'>ri,i) Weed Sei .',(»: 186 'lO',. 
. (195()| 1087. Cray's 

Culture Series. 

Joyce vK S. II. Curry (editors). The Bolanv and ( liem- 

istrv of Cannabis. J. & A. Churchill, London, 
larvev. J. 1971. Karlv Nurserymen. I'hdlimore. l.on.l..n. 
Ian ill. \ M.I B B.adlev. < I Nevens. I F W ie- I). M. K. Ware. I). \\. Ogle. <;. \\. Kamsev v\ <,. 

I'. !■ hailing. 1002. \Has ..I die \irginia Mora. 3rd ed. 

\irginia Botanical \ssociales. Kai nn ille. \iiginia. 


1978. Gre 

ca. Harcourt I'.rae.- Jnvanovieh. New 

H.tdge. W. 11. cN C. (). Krlanson. 1956. Federal plant 
introduction— \ review. Fcon. Hot. 10: 200-334. 

T. 1 1764| 1972. The History of the Colony 

Mutton. L 1). 1927. Barberry eradication re. 

JosseKn. J. 1672. New-Kngland^ Ivunlo-. I)i 
Buds. Beasts, fishes. Serpents, and Pla 

r l,c:,i„:„>. I',,. 103-125 in 0. T Sandlund. I'. J. Sclie, 
& A. Viken (editors). Invasive Species and Biodiversity 
Management, klnw.r. I >onl,echt. III. Netherlands. 

Crosby. A. W. 1972. The Columbiaii Ia< liangr: Biological 
and Cultural ( :<in>e.|ueii. e- of I 102. Creenwood. West- 
port, Connecticut. 

Darlington. \\. |I8I0| 1067. Memorials of John Bahrain 

I" . ,',.., I I in II .in , \ 

of Two Voyages to New -Kuglaiid. 

Ii.lien. VI. II. y\ VI. W. Griffiths. 1908. Biological Conl.ol 

Weeds. 111. ed. CABI Pub., Oxon, U.K. 
Karles/. .1. T. v\ C. A. Mcacham. 1000. Synthes,. ,,l the 

Noith Vm.aican lloia. ver. I.d. Ninth Carolina Botan- 

ical Garden. Chapel Hill. 
kowarik. I. 1995. Time lags in biological invasions with 

I . Mi in I 1 . I'vsck. K. I'rarh. VI. |{. |„ia.iek & M. Wade 

Problems. SBB. Vn'islcrdam. 
ku/nets. S. S.. A. B. Miller & B. A. Fasterliu. 1060. 

I'opulalion redistribution and economic growth: United 

Male.. 1870-1950. 2. Analyses of economic change. 

Mem. Amer. Philos. Soc. 51. 
Keiuhlon. \. 1070. Karlv V mci lean ( .aniens: "lor Meale 

or Medicine." Houghton Mifflin. Boston. 

1076. American Gardens in the Kighleeiith Cen- 

"I ,, I -. 


1087. American (hardens of the Nineteenth Cen- 

tekwood. V. (.. B. iKdilor). I'i:;|. (.aidens „| Colony 
and Stale: (.aniens and Gardeners of the American Col- 
onies and of the Bepultlie Before 1810. Vol. l.C.Scrib- 

(Kdilorl. 1031. Gardens ol Colony and Stale; 

obi..- Before 1840. Vol. 2. C. Scribner's Sons, 
New York, 
tren/i. H. J. & K. S. .lefferv. 1087. Weeds of the I nile.l 

Sei. & Arts, Ser. 1, 42: IM9. 

I.von. K. 1006. Settlement and survival. Pp. 10-61 in M. 

Hall. C. C. 1010. Narratives of Karlv Maryland. S.-rib- 

Cannon leditor). The New History ol Florida. Univ. 

ner's, New York. 

Florida Kress. Gainesville. 

Ilanev. A. vN K. V. Bazzaz. 1070. Some ecological itnpli 

Mack. B. N. 1080. Temperate grasslands vulnerable to 

cations of the distribution ol hemp {Cannabis salira U.l 

plant invasion: Characteristics and consequences. I'p. 

in the United Slates of America. Pp. 30-18 in C. R. B. 

155-170 /„ J. A. Drake. II. A. Moonev. K di Castri. B. 

[terser of weeds. Econ. Hot. 45: 257-273. 

. 2(KM). Cultivation fosters plaiil naluralizati 

reducing environmental slochaslicily. Biol. Inv; 

The U.S. naturalized llora: 

i and control. Eeol. 
Appl. 10: 689-710. 
Magee. I). W. & H. K. Aides. 1999. Flora of the North- 
east. A Manual of the Vascular Flora of New England 


Massachusetts Historical Soeieh. 194.3. Winlhrop Ca- 
pers. Vol. 3. 1631-1037. Massachusetts Historical So- 

McGlone, M. S. & U. R. Basher. 1995. The deforestation 

Zealand J. Eeol. 19: 53-66. 
Meinig. I). W. I9i!0. The Shaping of Amrrii-a: A (-co- 
graphical Perspective on 5(M) Years of History, Vol. 3. 

Pursh. F. IJUI, Flora \mcncac Scptentric 

■luilis. White. 

Cochrane, London. 

guinn. 1). 1990. Explorers and Colonies. 

1500 lr.25. 

Ilanililcdon. London. 

Ralinesque. C. S. 1H1I. An essay on the , 

now grow sponlaneousL in the Middle St 

ale of North 

2: 330 3 15. 

lihoads. A. F. X \Y. M. Klein.' 1993. Th«- \ 

,,-an Philosophical Socetv. Philadelphia. 

Ridlev, H. \. 1930. The Dispersal of Plants Throughout 

the World. L Reeve. Ashford. Kent. U.K. 

Saner. M. A.. 1). H. Clements. M. P. Hall. 

1). J. Doohan 

& C. W. Crompton. 1995. The biology 

..1 Canadian 

weeds. 105. Unaria vulgaris Mill. Canad. J. PI. Sci. 75: 


Schicrcibcck. K. A.. R. N. Mack & R. R. S 

1 II. i Is ..1 herhivory on growth and biom; 

,~ allocation 

75: 1661-1672. 

S< hweinity. 1 .. 1 >. lie. 1.3.32. Remarks on 

the plants of 

less degree, in the United Stales. Ann. 

Lyceum Nat. 

Hist. New York 3: 148-155. 
Scot. (;. 1085. The Model of the Covornment of the Prov- 
ince of East-Jersey in America: and Encouragements for 

burgh. U.K. (Microfilm. Ann Arbor. Michigan. Univer- 

Torrev. J. 1819. A Catalogue of P 

ide\ florae Lancaslnensis. Trans. 

il. 2002. Predicting Imasions 
saml Plant Pests. |l .S.) Nation- 
Washington, D.C. 
. Cuthberlson. 1992. Noxious 
vala Press. Melbourne, 
is of the Historical Ceography of 
nited States. J. K. Wright (editor). Carnegie Inst i- 

i of Washing X \mcricau ( ieographieal Socie(\ 

■u Wk. Washington. D.C. iK New York, 
n, D. J. & R. Prasad. 1998. The biology of Ea- 
rn weeds. 109. Cyllsus scoparim (L.) Link. Canad. 
. Sei. 78: 497-504. 

uliin. C (>. 1932. 

Yitousek. P. M.. C M. ITAntonio. L. L. Loope & 
Weslbrooks. 1 990. Biological invasions as glob; 
vironmenlal change. Amer. Sci. 84: 468-178. 

Weslbrooks, R. C 1998. Invasive Plants. Changi.i 
Landscape of America. Federal Interagency Conn 

vood. W. | 10.3 1 1 1977. New England's P. 
with an introduction by A. T. Vaughai 
Massachusetts Press. Amherst. 

oung, A. 1846. Chronicles of the First 
Colony of Massachusetts Bay. from 162 

raucous Printed Documents, and lllustr 


. D,.,um., (mm these examples a 

-en ants and plants ratine from 
facultative, loose associations to s|tecies-s|)ecilic. 
obligate mutualisms, and innumerable <|uciioii- 
lia\e been posed about the costs, benefits, and evo- 
lutionary implications of these relationship- ( II. ,il 
tie. 1985; Bronslein, 1994; Jolivet, 1996). As in- 
vasive ants spread, their interactions with plants 
arc inevitable and have potent iallv great implica- 

tent are ant-plant interactions altered by the intro- 
duction of these novel partners? 

Invasive ants possess a disbud combination ol 
traits relative to native ants (Holway et ah. 2002). 

In tins p ;l per. I hypothesize some ol the-e hail- 

and provide a basis for predicting the outcome of 
these novel interactions. I locus on the potential loi 
differences between native and invasive ants in the 
way they protect plant- from lierbivorv. tend llo- 
moplera. and interlere with plant reproduction. 
These examples encompass the more facultative. 

increasingly likclv to participate as they spread 
around the globe. 

\ltliough I explore the dil Tei ctiecs between na- 

ne ant can interact directly and 

e same plant in multiple ways. 

benefit the plant in one interac- 

! plant in another. The 

net outcome for the plant will depend on the rela- 
tive balance of a range of interactions, all of which 
in turn will be influenced bv the same set of vari- 
ables that Ivpicallv iiilluence ant-planl interactions. 
I Itimately. vve will want to know whether invasive 

) the bala, 

enllv ] 

( >l the approximately 10.000 species of ants in 
the world, about 1 . >0 have been introduced around 
the world with the help ol humans; these may be 
termed exotic, or tramp ants (McGlynn. I '>*>*>). In- 
vasive ants are the small subset of introduced ant- 
that are able to establish and penetrate areas out- 
side of human-modified habitats (Holway el al., 
2002). Well-known examples in the United States 
i in' hide the red imported fire ant, Solenopsis invicta, 

* Symposia 

I pn.y ii|e.| helpful 

o K y, Corson Hall. Ithaca. New York 14853. I i.S.A. Ijlt3@ 

\nn. Missouri Bot. Card. 90: 91-108. 2003. 

Missouri Botanical Garden 

ami Hie \ I genl inc anl. I.iiit ■pilheim/ luimile. The lit- 
tlc lire ant. Wasmaniiia aiiropiuielala. the big-head- 
ed ant, Pheidole megacephala, and long-legged anl. 
Anoplolepis gracilipes, also appear to be highly s 

I |] i 

See Table 1 for a suniiuaiv of dish ibiilmti-. 

One hallmark of invasive ants is their ability to 
become extremely abundant in then- new habitats. 
I'illall haps in areas invaded by Solenopsis inricta 
(Porter & Savignano. I WO). Lmepithema humile 
(Hoi way, 1998). and I'heidole megacephala (Hoff- 
man et al.. 1999) have all shown that the number 
of invasive anl workeis greallv exceeds thai of na- 
tive ants in nearby un-invaded areas. Several 
mechanisms are likelv eonl ribul ii ig lo invasive Mills" 
aehievemenl nf iiiunerieal Mipoiiorilv including es- 
cape from natural enemies and eompetilors and 
changes in eoloiiv -Inn hue Inward mil 1 1 i -< |ii< -• -ned. 
niiilli-nesl sii|)ercoloiiies that lack inlra-peeiiic ag- 
gression (Holvvay et al., 2002). 

The physiology and behavior of invasive ants also 
likelv plav a role in llieir lenileucv lo achieve high 
abundance (Holvvay et al.. 2(K)2). Kcologicallv dom- 
inant arils, including invasive mils such as l.inepi- 

i nil i i | i ii II i ill n 

to take in more liquid foods, including Moral and 
exlialloral ucclar and homopteran honevdew ll isnei. 
1957: Davidson. 1998). The ability to harvest these 

' oh le.di lie-rich re-oiuves mav be especiallv illl- 

porlanl in liieling ihe high lempo aclivilv ol a large 
workforce, ihereliv iiiainlairiitig a high dvnamir den 

ill create s 

tellite nests at the base of 

en lending 

phids (Kaakeh & Duleher, 

ligh level o 

aggression is another cliar- 

non to invas 

ve ants and likelv enhances 

ss as predat 

rs(Holvvav el al.. 2(H)2). Ac- 

lydrale-rieh resources and a 

force are hot 

. the cause and effeel of su- 

dilative and 

ilerfereuee compel il mm i abil- 

d. whirl, 

e anls and other eompetilors 

splaeed (Hoi 

vay et al.. 2002). 

ily documented consequence of an invasive ant 
oduction (Holway el al.. 2002). Linepithenm hu- 
'. for example, has displaced several species of 
;aeic ants in California (Erickson, 1971; Ward. 
7; Holway, 1995; Human <S Gordon, 1996). 
:ugal (Cammell et al., 1996; Way et al.. 1997). 
South Africa (Donnelly & (iiliomee. 1985). 
mannia auropunelala has expanded its territory 

t * § -r~ s i f 1 

; l*ji 

21 Mil 

II I 1 


I ^ 

rich food affect how they might protect plants from 

itats in Australia (Majer, 1985; May & Heterick. plant reproduction. 
2000; Hoffman et ah. 1999; Vanderwoude et ah, 

2000). Snlmopsis ^■minala (¥.) and other ants na- W| , |, R()TKrnoN n{()M H ERBIVORKS 
live to the southern U.S. have become much less 

common following iuva-ioii h\ .S. inrirlu (Porter & Protection of plants from herbivores is perhaps the 

Savignano, 1990; Cotelli & Arnetl. 2000). () | ( | est recognized effect of ants on plants. As long 

Invasive ants also affect other invertebrates and ago as x .„. ; M) i. Chinese citn.s growers facilitated 

even vertebrates, usually negalivelv. \nopIolcpis t |, r S p ie ad of the yellow citrus ant, which preyed on 

gracilipes is blamed for diminished populations of | 1(T | )ivores i„ their orange orchards (Huang & Yang, 

various invertebrates in die Seychelles (Haines K p )!!7) Sim . e ,, lelK . ||1|s < ahi|i|v ,„ prey on< or simp i y 

Haines. 1978) and is generating ecosystem-level [)a| .. lss wou ] ( l-be herbivores has been noted in many 

changes through its impact on the red land crab on svslems (Beattie. 1085: Holldobler & Wilson, 1990). 

Christmas Island (Creen et ah. 1900). Lim-pithnm, ,„ S()|nr rases _ p|an|s ,. nti( . e an(s | )V ()ffering food or 

humile is associated with decreases in some insects s|)(i||(i| . F()| . ( . x ., m[) | ( ._ ;, ; s [h(m „ h[ that one of {he 

in California and llawah (Cole et ah. 1992: Bolger ^-^ ^^ ,„,. ,,„. pi ,. s( ,, ( ,. „,- extraflora l nec . 

et ah, 2000). The great ,ed„c|,„u in .-nd.-m.e „,- |;mr> js , (( a(|ra( ., . mls jn ,. X( ., ia|lf , t . f()r protection 

sects ,n lowland Hawaii at the end ol the 1800s |mm , ir| ., m( , |VS (Brnt ley, 1977). 

has been attributed to mvas.on In I'hrululr nu^u- ^ ^^ ((f |rajls ass()< . ia(( , (1 „;,,, imasive ams 

cephala (Zimmenn. * .wv u,r,rta has ^ v ^ ,.„, ^ {q ^ ^^ { ^ 

in', i-i >-|» ■ |. "-., s nN. ■ rd 'iiai 'iia - in die southern 
United States (Holwav el ah. 2002). Wasmannia au- 
ropuiictdta has eliminated or reduced terrestrial in- 
vertebrate populalions in the Calapago- l-land- 
(Lubin, 1984) and New Caledonia (Jourdan. 1997). 
The mechanisms lor these effects are not always 
clear but likelv in\ol\e some combination of prc- 

-i/e. high levels of worker aeli\ ilv. and aggressive- 
ness (Davidson & McKey, 1995: Bronstein. 1998). 
\ large number of verv active workers ma\ result 
in increased plant visitation b\ ants leading to de- 
creased herbivory (Caume et ah. 1997: Bronstein. 
1008,: Caume et ah. 1998; Uinsenniair et ah. 2001) 
because herbivores are located more rapidly and 
have a shorter residence time on the plant (I)uarte 

J'"" "'" Kocha & Codov Hergallo. 1992). Benefits to the 

plant probably level off or even decline at some 

colony size, however. An extremely large colony 

\M) Ol k.omi-.s <>l niav imposr a ( . ()S , on the plant if the ants are steal- 

KACIIONS ing floral nectar, tending Homoptera, interfering 

established that the na- vvilh ,lu ' ,ialmal enemies ° f herbivores, or excavat- 

ng the plant's root system. Kesearch to determine 

vhere the cost-benefit trade-offs lie with respect to 

ture of ant-plant interactions is dependent on l 
variables including ant behavior, ant colonv 

and stage, host plant attributes, ami the abundance iinl ,ol,,n > si/t ' ,1;,s > el to 1m * (l(>l 
an.) behavior ol other organisms in the svslem \ggressive behav ior also aids 

(Buckley, 1982; Beattie. 1985; Keeler. 1080; Cash- ph»» '"»»' herbivores (Fiala et ah, 1989; Letour- 
man, 1991; Davidson & McKey. 1905: Bronstein. '"'an. 1985: Bronstein, 1998). Bentley (1977) re- 
100 1, 1998: Jolivel. 1096). As explained above, in- laled ant aggression on plants to three factors: pred- 
vasive ants tend to have larger colonies and exhibit "lory hehav ior. ownership behavior or the defense 
erritorial boundaries (Way. 1965), and swarming 
lavior or the ability to rapidly recruit workers, 
three behaviors may be enhanced at the colony 
■I bv having a large workforce. Aggression is 
i negatively correlated with proximity to an ants' 
I: disturbance near an ants" nest or food source 
I elicit a stronger response than a threat en- 
nlered farther away (Way, 1965). The vagile, po- 
oii ion- ne^ling bebav ioi ol invasive ants may al- 
them to nest In closer proximity than a native 

moie aggressive behavior than nativ 

e ants. Mo 

over, invasive ants appear to have a 

stronger p 

dilection than many native ants for 

.!. ,!, ,', 

resources, which are invariably obtained directly 

imlirectlv from plants. Since native 

ants are f 

<|uenllv displaced when invasive ants enter a n 

habilal. it will be liail- ol the invad 

ers that inl 

ence future outcomes for the plants. 

\ppl\ing . 

isting models ol ant plant inleiacho 

•ns, 1 (level 

hypotheses about how invasive ants' < 

■levaled abi 

:ice, the left side of the equatu 

greater henelit margin for the plant whei 
ants are present. 

;ft side of the e 

ant would, possibly facilitating a grcalci a^uressive 

Some ants provide protection from herbivores 
while coll. -cting extrafloral nectar. Keeler (1981: 
190) modi-led the conditions under which the mu- 
tualism is favored: 

P [A(1 - D)H]>I. 
where p = probability that ants will find the plant. 
A = effectiveness of ant-defense. I) = effectiveness 
of other defense*. II - intensity of herbivorv. and 

plants' perspective, the miitualisi 
when the left side of the equation, the benefits to 
the plant, exceeds the right side, the costs of at- 
tracting the ants (Keeler. 1981). A schematic of the 
relationship is shown in Figure 1. 

Given the numerical superiority, attraction to 
carbohydrate resources, and aggressive character of 

invasive ants, we might expect that p , MVi , > p, mlm . 

an( l A imilM „. > A„„ m .. Since I) is primarily a factor 

of the plarrt and the hcibivore(s). I ) and I) „. 

are not likely to differ. ll im( . K , may exceed H„„ 1W . if 
inlraguild puliation i- a factor, i.e.. il ants prev on 
or inlerlere with the natural enemies of herbivores 

(Fubanks. 2001). However. II , may be less than 

H„, lu , if invasive ants diminish herbivore popula- 
tions via direct or indirect interactions that affect 
herbivore eggs, larvae, pupae, or adults away from 

dueible defense (Koptur. 198,9: Ness. 2001) I, will 

increase with H. For example, if H is greater 

than H im ,. (W . then extradotal nectaries will be in- 

will be greater thai 

Ant tending of honeydew-producing 1 
•luding scale, mealybugs, apbids. 

(Way. 1963; Carroll & Jan/cn. I 073; l!„eklev. 
198,7: Holldobler & Wilson. 1990; Davidson & 
Vie key. l00. - {). Homoptera extract phloem from the 
host plants and excrete it as honeydew. Ants feed- 
ing on the honeydew ollen protect die-,,- llomoplera 
from parasites and predators (Wav. 1903: liuckley. 
1987). The ant-Hon.optera mutualism may be det- 
rimental lor the plant, robbing it of phloem and 
sometimes leading to mold accumulation (Bach, 
1991; Lewis et al.. 1976) or phytopathogen trans- 
mission (e.g.. pineapple mealybug will disease 
(Ueardslev et al.. 1982). I'hylophlhora pod rot 
(Fvans. 1973). swollen shoot virus (Taylor & Ade- 
dovin. 1078)) while conleriing no benefits in ex- 
change (Ih.ckley. 1987). Less commo.dy. the inter- 
action may benefit the plant if ants deter other 
herbivores while lending Homoptera (Carroll & 
Jan/en. 1973; Messina. 1981: Compton & Robert- 
son. 1988). 

As with plant protection from herbivores, the 
characters common to many invasive ants will af- 
fect their Homoptera-lending abilities and conse- 
quently the outcome for the plant. The need for 
carbohydrate resources and ability to harvest co- 
pious amount- o| liquid provides the basis for the 
allraetion ol iuva.-ive ant- |o llomoplera. \nl ag- 
gression, particularly ownership behavior, is an ef- 

pre<lators of Homoptera (Wav. 1963). I'olydomy and 
nest vagililv may allow invasive ants to reside in 
elo-er proximity to these carbohydrate resources 
than would native ants. 

Ant abundance, however, may be the primary 
trail affecting inva 
as it may affect 
mechanisms. Higher aut numbers may Iraiislali 
into a higher probability ol locating the tendees. 

Homoptera C h 

nils iii a byproduct mutualism with the plant (2a) and when it 
ws indicates the direction of the effect and the positive and 
f the arrow indicates the strength of the .•Keel. Dotted lines 
■ Homoptera, II. \. and I) arc as defined in Figure 1. In 2a, 

h> doler herl.iv.ire> mi the plant hears the cost of hosting the 

and a higher ant : homopteran ratio may make ants 
inure effective at deterring their partners" enemies 
(Sleyn. 1955). Tending ants are a limiting resource 
to honevdew-producing Homoptera in many sys- 
tems (Addicott. 197K: Sudd. 1987: Cushman & Ad- 
dieott, 1989: Breton & Addicott. 1992: Fischer et 
al.. 2001). and well-tended Homoptera may grow 
faster, reproduce more rapidly, and produce more 
young than unten.led ones (Way. 1963: Brislow. 
1984; Morales, 2000). Where the ant : Homoptera 
ratio becomes loo low. Homoptera may become ant 
prey or be more easily parasitized or preyed upon 
by natural enemies (Way. 1963. and references 
therein: Breton & Addicott. 1992: Sakata. 1991: 
Jolivel. I') ( )()|. The superabundance achieved by 
invasive ants may keep the tender: tendee ratio 
from falling below 1 1 1 1 — din -hold. In such a case, 
we would expect limits to the abundance of both 
mulualists to he imposed by the host plant. 

Impacts on the plant. however, also may depend 
on whether the ants tending Homoptera deter other 
herbivores. In this byproduct mutualism hn. 2<M)| . 
the plant benefits indirectly from the ants' pres- 
ence; the ants benefit from the Homoptera hosted 
by the plant and possibly the herbivores, if they are 
captured as prey. In this scenario. Homoptera arc 
analogous to extralloral nectaries and Keeler's mod- 
el (1981) can be modified as: 

p[A(l - D)H , J > C h 

it side of the equation (formerly the iu- 

eeplibilily to phylopalhogens. Other lerms remain 
the same: p = probability thai ants will find the 
plant, A = eflecli\i-nes> nl aiil-.l.Ten-e against 
luoii-homopleran) herbivores. I) = effectiveness of 
oilier defenses. II = intensity of (non-homopteran) 
herbivory. Considering ihe trail- ol niva-ive ants 
that favor high homopteran populations, it is prob- 
able thai in main cases C,,.,,,^,,,. will be great. -r ihan 
Ci,-„ a .,„- The probability that the ant finds ihe plant, 
[t. may vary not only with the abundance of the ant. 
lnil |M.s-il>l\ will: the ability of different Homoptera 
to recruit ants (Del-Claro & Oliveira. 1996). It is 
unclear, though, whether in\a-i\c ant- may ie-p.>ml 
to attraetanl en.- differently than native ants. Hy- 
pothesized differences in \. I ». and 1 1 for native 

with the lood-lor-prolection mutualism involving 

only be niainlained if ihe benefits, in terms of de- 
terred iioii lioinopl. ijti herbivores denoted on the 
left side of the equation, exceed the costs exacted 
on the plant by the Homoptera. the right side of 
the equation (Fig. 2a). If costs to the plant exceed 

toward a parasitic relationship with its host (Fig. 
2b). possibly resulting in reduced fitness of the 

plant. that l.iil to detect any change in plant 
fitness with Homoptera outbreaks associated with 

Seed predalion and Inn-resting. Impacts of ants 
Seed set. Willi few exceptions (e.g.. Yam.. 1994: on s, '«' (1 disposal and seed prcdation have received 
Puterbaugh, 1998). ants commonly have been re- »"»"" attention than impacts on seed set. In myr- 
garded as unwanted guests in flowers (Kerncr. mecochoroiis muliialisiiis. ants disperse seeds away 
IH78: Buckley. P>f!2). Attracted In floral nectar. Imm lh <- P 11 "' 1 " P"«»'l- ollen burying them in their 
ants mav damage floral structures, and depress pol- nvs[. '" «'xclinnge lor the lipid-rich eliaosome al- 
ien viability with their antibiotic secretions (Kerncr. belied to llie m«,I (Buckley. 1082; Beattie. 1985). 
18)78: Peakall el aL 1991: (ialen. |000|. \> with I be di>per»ed -eed- iherelore mav ex-ape compe- 
Homoptera and extrafloral neetaiy lending, super- titi() " W1, '> siblings and parents and are less sus- 
abundanee and an affinity I'm •< -arboh\<lrale resourc- <<'I>lible to picdalion and other threats, such as fire 
es combined with lii-h levels ..I' agression would (bu.kley. 1082: Beattie. 1985). Ants may also be 
lead to the expectation that invasive ants would be see(l harvesters, consuming the seed itself. Seed 
exceptional at recruiting colony members to flowers harvesters, however, do not eat all the seeds they 
and exploiting floral nectar. collect, and therefore mav act as accidental seed 
If numerous aggressive ants are present inflow- dispersers (Buckley, 1982: Holldobler & Wilson, 
ers. their presence may affect other floral visitors. 1990). 

Pollinators may be forced to reposition frequently Various species of ants may be attracted toeliao- 
to avoid attack by the ants, and this mav result in somes and opportuuislicalK lake part in myrme- 
increased transfer of pollen and consequently high- I'ocliorous mutualisms, whereas harvester ants tend 
er seed set. Alternatively, pollinators mav avoid the "> ll; '\<- special adaptations for harvesting, consum- 
flower altogether, or reposition too superficially for m ^- «»"• storing se, -,b 1 1 loll.lohler & Wilson, 1990; 
effective pollen transfer, ultimately resulting in de- Keeler. 1989). In both cases, the ants need to (a) 
creased seed set (Wyatt. 1980). \' plethora of floral <bseover the seed. (!>) recogni/e the seed as a re- 
antagonists, including pollen consumers, and other source, and (c) be abb- to carry the seed back to 
florivores, however, may also be deterred, perhaps tl ie "<*st. s <^ (l discovery will be dictated by ant 

The ants" ultimate impacts on seed set will be between ant activity and seed availability, and the 
dependent on the attraction and availability of floral attractiveness of the eliaosome. if present (Beattie, 
nectar in the context of the ants' seasonal diet pref- 1985; Keeler. 1989). Seeds are typically high in 
erences. activity patterns, and abundance, the iiuin- lipids and proteins: recognition of the seed as a 
ber of pollinators and floral antagonists and their desirable resource may depend on the relative 
susceptibility to ant deterrence, floral structure and abundance and composition of other food sources 
defenses against ants, and llie pollination require- m the environment and the dietary requirements of 
merits of the plant (Kerncr, 1878; Koptur. 1979; the colony when the seeds are available (Beattie. 
Vinson & (ireenberg. 1980: Huxley K Culler. 1991; 1985). In every my rinecochorous system studied, 
Klinkhamer & de Jong. 1003; I, an/a el aL 1905; seed-dispersing ants comprise only a small subset 
Koptur & Timing. 1008: Pulerbaugh. 1008,), as well <>l the .ml specie* present, and then behavior to- 
ns the relative difference between ants" (-fleets on ward seeds can vary from one day to the next (Beat- 
polhnators and Moral antagonists. Kor example. I tie. 1085). suggesting thai recognition of the seed 
would expect nectar-robbing by invasive ants to as a valuable food resource is not constant among 
have little effect on seed sel across a plant popu- anl species or oyer time for a single species. The 

tition for the resource, e.g.. if nectar is available in pend on llie ants" foraging behavior and size and 
excess of demand by pollinators. I would expect shape complementarity between seed and ant ( Kec- 
invasive ants to be more likely to decrease seed set lei. 1080). Harvester ants must also have the man- 
in a plant species that has few Moral anlagonisls dibular strength and agililv to ingest the seeds 
and is dependent on a few species of small polli- (Beattie. 1085). 

nators, than in a species that sutlers from many How do we expect invasive ants to compare to 

floral antagonists and has a diverse pollinator guild. native ants in these requirements? Traits of invasive 

Moreover. Mowers in which nectaries are in close ants that have formed the basis lor hypotheses 
proximity to the stigma and anthers relative to t 

paper. namely elevated a!>uii( lance, aggn-ssiou. and 
attraction to carbohvdrali •>. ma\ not have as much 
infliicm -e on seed predalion and dispersal. \iimer- 
it-al superiority of invasive ants may increase the 
probability llial they will discover seeds. However, 
common trails of invasive ants oiler little capacity 
lor predicting liou invasive anls will respond to the 
seeds tliev encounter. Other traits that may be 
linked to seed discovery and dispersal may not vary 

lenopsis imidn mcphala are at- 

tracted to oilv. lipid-rich lood sources (Vinson cK 
Greenberg. 1986; Sanders et al.. 1992). and seed 
harvesters occur in both genera (Holldobler & Wil- 
son. 1990). Tims, we might expect these invaders 
to be attracted to seeds, and even have some of the 
mandibular adaptations helpful for ingesting -rnls. 
but perhaps no more so than native ants. 

noted above, numerical superiority, aggn-s- 
■ behavi 

likely to account for all differences 

t may vary s 

among ants, hut an- unlikely to vary consistently 
among invasive and native ants such as tempera- 
ture tolerance, daily activity patterns, colony cy- 
cles, and seasonal preferences for food types will 
certainly affect ants" relationships with plants. Kven 
for a single ant species, associations with plants 
may change spatially or temporally. For example, 
the derive ol plant protection is largely linked to 
ant foraging patterns, which mav change depending 
on the nutritional requirements of the colony (e.g.. 
in relation to reproductive cycles), and the avail- 
ability ol ■ ah. hvi .... ofein resources that 
require less foraging effort than those on the plant 
of interest (AM & Reagan. 1985: Vinson & Green- 
berg, 1986; Stein et al., 1990; Rapp & Salum. 
1995; Cornelius & Crave, 1997). 

Attributes of the potential prey items. flomop- 
tera, and the host plant will also play roles in de- 
fining ant interactions with these organisms (Way. 
1963; Cushman, 1991; Huxley & Cutler. 1991; 
Bronstein. 1991). and invasive and native ants mav 
diverge in their responses to these variables as 
well. Many herbivores (Van Per Coot. 1916; kaak- 
eh & Dutcher, 1992; Pavis et al., 1992: Cunawar- 
dena & Bandumalhie. 1993; Cornelius & Rernays. 
1995; Montgomery c\ \\ heeler. 2000: Brinkman et 
al.. 2001) and flowers (Kerner. 1878; Willmer & 
Stone. 1997; Gha/oul. 2090) have effective chem- 

ical or physical defenses against .nils. \\ helhei dif- 

ance to phytotoxins, for example, will enabl 
certain groups of anls to evade these defenses i 

ml therefore might lack specific ( 


Do observations 
tions support pred 

I'l : <)\ • ■ 

i Inn. I.vnt 

asive ant-planl iulerac- 
that the distinct comhi- 
s will affect the outcome 
r interactions with plants? Specifically, be- 
of their combination of trails, are invasive 
lore likely than native anls to deter herbi- 
teud llomoplera to the detriment of host 
and interfere with plant reproduction? There 
>een few direct tests. Most research on ant 
:tions with herbivores has been done in 
n how pest popula- 
1 crop yields are affected by changes in ant 
ml has not explored any particular aril at- 
ributes influencing the interactions. Still less is 
i ants affect seed set, pre- 
ersal. Nonetheless, examples below 
vidence of the importance of the 

scoring the influence of ant biology and the 
other organisms in the system in determin 

outcome of these interactions lor the plant. 

ol examples of invasive ; 

rarely do they pinpoint any particular characteristic 
ol the ant responsible for effects on herbivores. 
Some insights can be gained from cases in which 
the invader fails to deter herbivores, lor example. 
Anoplolcpis gracilipes affords protection from the 
-licking bug. 1,., ■ '.'-' < Inn. i). in Sol- 

omon Islands coconut when it reaches high abun- 
dance in the trees. But when prey are abundant on 
the ground. A. gracilipes fails to forage in the trees 
and premature null. ill ensues due to A. cocophaga 
damage (Creenslade, 1971). In Keeler's (1981) 
terms, in this case a high A, ability to deter her- 
bivores, is incou ii I H I In i, i , I i ii il tin 
aging on the plant (p) is low. 

reducing the herbivore attack on the 
cnt(ili>ti< : > other natural 

ries are induced by 

Several studies have shown that different invasive inva 

ants are attracted to . \( i ;tlh»i ;il ne. lar and can de- and 

crease herbivory on extrafloral nectar-hearing are 

plants (see Holway et al.. 2002. for a review). The of extrafloral nectar is an ir 

high population density of Plwidole mrgarcphala system; extrafloral nectar r 

and its tending of extrafloral nectaries are hlamed response to herbivory, therel 

for the difference in populations of shrnhs between and parasitic wasps (Mess 

P. megacephala and native ant-inhabited sites in ments have shown that .S. i 

Australian rainforest. The native Vrena lobata L. 

and the introduced Senna nblnsi/alia (I,) H. S. Il- 

lative ant's interactions with the her 
specific ; i 1 1 1 i J till*-—, ol the invader lliat 

little lioni lolivory in area-- invaded by /' megace- 
phala. whereas in areas with native ants, the shrubs 
are small, isolated, and heaviK attacked bv herbi- 
vores (Hoffman et al.. I WO). Other studies dem- 
onstrate that Salenapsis inricla (Fleet & Young. 
2000). I.inepithema hnmite (Koptur. 1070). and 
Wasmaiuaa (iiimpmn -lala (Meier. 1 00 I) also are al- 

inlroduced habilal-. sometimes to the benefit of the 
plant. Kill these studies, as with most, do not com- 
pare the behavior of invasive ants to that of native 
ants, nor do they explore specific attributes of in- 
vasive- ants that may be influencing the interaction 

Research exploring the interact ions of Sol, ■/, apsis 
inricla with Catalpa hignanioides Waller provides 

anls and preys on the parasitic wasps c 
Ceratoma catalpae (Hoisdm ,\ 

nectar production coinciding will) the stage of col- 
ony cycle when workers prefer protein-rich resourc- 
es. So/rnnp^is inn, la i- apparently as effective a 
predator of C. catalpae as are the native ants be- 
cause it preys on pupae and pre-pupal iiislars as 
well as larvae, and because of its exceptional ag- 
gressiveness when prey are encountered. Thus, 
while S. inricla disrupts the mutualism by pre- 

visiting extrafloral i 

and protecting Catalpa 

creased herbivory (Ness. 2001). 

Pulling the example in the c 
(1981) model (Fig. 3), even with 

■ load 

, will be 1 

1 The difference 

duced herbivory and the investment in extradural 

present than when native ants are present. 

It is worth noting that p. the probahditv ol lo- 
cating an- 1 loraging on the plant, may he inlluenced 

sard offered. As such, an t 
Solenopsis inrirla may 

1982). Indeed, high abundance of S. ir, 
been offered as an explanation for the la 
difference in its foraging frequency between nec- 
taried and nectariless isolines of cotton (Agnew et 
ah. 1982). Analogously, it is possible that high den- 
sity of S. inrictu in the invaded Cutalpa bignonioi- 
des stand (Ness. 2001) facilitated the ants" foraging 
on the plant, notwithstanding the absence of its at- 
traction to extradotal nectar. In these cases, if abi- 
otic or other conditions ever result in diminished 
abundance of S. inrirla. the invader may become 
less effective at deterring herbivores than ants that 
may not be as abundant, but are lured to the plants 

\noplnlepis grarilipes. Lincpilbcma hum He. Phei- 
flulr mcaiK-cphtilit. and Wasmannia auropunrlala 
all have been noted for their ability to cause Ho- 
moplera outbreaks in various parts of the world 
(Holway et ah. 2002). Solenopsis inrirla may some- 
times obtain carbohydrates directly from plant tis- 
sue (Vander Meer et ah. 1905). but also has been 
associated with increased llomopleia populations 
(lofgren. 1980; Holway el ah. 2002). The data col- 
lected in these studies often fail to discern any par- 
responsible for Homoptera outbreaks. While some 
evidence of the importance of abundance, aggro- 
examples below, evaluation of my prediction that 
these common invasive aul hails play a major role 
in determining their interactions with plants will be 
enhanced by further detailed study. However, soun- 

der versus herbivore deterrent. These studies reveal 
that both types of outcomes outlined above occur, 
those in which the ant-Homoptera mutualism is 
para-ilic Inward the plant and those in which the 
ants lending llomopleia enter into a byproduct mu- 
tualism with the pi. ml l.v deterring iion-homopteran 

Linepilhema humile^ abundance, aggression, 
and tending of aphid* have all been related to its 
ability to control populations of the pine p reces- 
sionary moth (Tlianmrlopoea pilvorampa Den. & 
Schiff.) in Porluual pine plantations (Way et ah. 
1999). The aphids attract foraging L humile to pine 
tree crowns, baler in the season, pine pioce-.-donarv 
moth larvae on lives with /.. humile are fiercely 
attacked, whereas those on native ant-occupied 
lives are ignored. ( oiiseqiienllv. parts of plantations 
thai are inhabited by /,. humile escape the severe 
defoliation caus.-d by '/.' pilvnoimpa in native ant- 
iuhahiled areas, figures la and b contrast the in- 

/.. humile are present in terms of Keelers (1981) 

mo<hlied model derived above. 

moptera in cacao (Theobroma cacao L.) is a study 
in contrasts, and illustrates the pos>ibilil\ lor dif- 
ferent outcomes to occur between the same plant 
and ant. In West Africa. Wasmannia auropunrlala 
is actively spread among cacao plantations because 
it effectively deters pestiferous mi rid bugs despite 
its association with high levels of scale and psvllids 
(Knlwistle. 1972). In contrast, in its native Brazil. 
If. auropunrlala tends the mealybug Planororrus 
all i (bissol and mans other Homoptera in cacao 
and fails to control p.-| herbivores possibly be- 
cause it does not always achieve dominance in the 
ant mosaic (De Medeiros et ah. 1905: De Souza et 
ah. 1998). In the parlance of the modified keeler 
(1081) model, in West Africa. C h . the cost to the 
plant of hosting the llomopleia associated with 11. 
auropunrlala. is outweighed by the high A. the 
ability of W. auropunctata to deter key herbivores, 
and the ant can be smd to have entered a In product 
mutualism with cacao, but in Brazil. U. auropiinr- 
lala lending /'. rim imposes a high (.],, and a small 
\. resulting in a cost to the plant that exceeds the 

Simla! Iv. Pheidole megacephalu in its native 
West Africa tends the mealybugs. Planororrus ritri 
and I', njalensis (l.aing). which arc associated with 
swollen shoot virus in cacao (Taylor v\. Adedoyin. 
1978; Campbell, 1994). As part of its tending be- 
havior. P. megarephala transports soil from the 
ground to create tent shelters for the homoplerans 
and thereby acts as a vertical vector of Plnlopluho- 

|Hll,). Tlie.h.eetioiMil ll 

Vl '"> "l>l"'k liiils vy on the major herhi 

"i I "Umk II) illustrates ili« I .\ | >|-.mIii.-| inulnaliMii betueeii /.. humilv aral the pine 
to the lives l>\ aphids and preys on /.' pi/rotam/,,, to the Ixtutit of the plant. Nr 

anship (4a) and the Linc/iilliciuii Immile-T. 
indicates the direction of the effect and the positive 
row indicates the strength of the effect. Doited lines 
igures I and 2. In la, the native ant is nol attracted 

ra spores, llie eliologie agent of Mark pod rot 
(Kvans, 197:5). In this ease, C,, comprises not only 
the direct cost of hosting the mealybugs. I • 1 1 1 also 
Id.' associated increased likelihood of bene in fee I 
ed with two of the worst diseases of cacao. While 
no cacao studies have ever reported relationships 
between \\ mrgarrphala and m,,m homopleraii her- 
bivores, it is unlikely that strong deterrence of her- 
bivores (a high A) would have gone unnoticed by 
farmers intimately aware of any effect on their 
yields. Phcldolc megaceplmlus fre(|uenl association 
with high Homoptera outbreaks to the detriment of 
plunt Illness places it firmly in the pest category in 
West Africa (Taylor & Adedoyin, 1978). 

interaction between I lomoplera. cacao, and \noplo 
lepis gracilipcs in Papua New Cuiiieu. The ant re- 
its abundance (baker. 1972). It displaces the native 

black pod rot (Room cv Smith. 1975; McGregor ei 
Moxon, 1985; Way & Khoo. 1992). Moreover, un- 
like the ants it displaces. 1. L ;nu ilipr*. hnia-ses 
a<lult cocoa weevils and persistently disrupts egg 
laying and foraging of several inirid and eoreid 
bugs, all of which are major pests (Knlvvistle, 1972; 
McGregor & Moxon. 1985; Way cK Khoo. 1992). 
but the same ant fails to forage in trees in Malay- 
sian cacao and so is not effective against herbivores 
(Wax ,\ khoo. I , ),'! , )|. indicating that a low p, prob- 
ability of finding and lorat-.iirj mi the plant, can 
ibility to deter its herbivores. 
ie examples above suggest that ant-llomoptera 
lalisms that are parasitic on the host plant are 
likely to occur in the invaders' native range. 

mutualism benefits ihe plant 
the introduced range. If this 
uld expect (hat studies that 

effects of the ants 

a. a bl- i-l< -. 

Hut this is not t 

fitness is noted, il is usually reported as decreasing 
(Beallie, 1985; Holway et al.. 2002). suggesting that 
hi most r.ivrs ih, invasi\c nil I Imnoptera mutual- 
ism is parasitic for its host plant, beatlie (1085) 
further points out that the vast majority of data de- 
sciilini'j lii iii 1. 1| )l« -I ;in damage to plants is from 
agroeeosyslems. or other heavily manipulated cn- 

of Homoptera may be (jiiile depauperate. Discern- 
ing whether there are differences in how the inva- 
sive ant-Hoinoplera mutualism affects plants in the 

rii ultural versus less manipulated environments is 
dilioiial research. 

with exlrafloral nectaries and Homoptera. the 

set. seed dispersal, and seed predalion precludes 
attempts to lest the predictions outlined above at 
present. A brief review of the current slate of 
knowledge can. however, point to early trends and 
identify areas lor future research. 

Although several species of invasive ants have 
been observed to forage in various (lowers (Adams. 
I'>8(.: I.olgrei,. 1080: Bins. 1000; Hara K llata. 

]<m: \ieolson. 1001: I lata el al.. I<W5). little re- 
search has explored invasive ants" attraction to 
flowers and llieir interactions with pollinators. Li- 

.• . V" -- ' ■' '..'.•'.'". I- I 1 ' to , |.L Il 11.11 ll IICI 1,11 

I i l'>!!7). is associated 
with lower insect di\ersit> in I'ralca nilidn Mill. 
flowerheads (Visser et al., 1990), and is hypothe- 
sized to reduce seed set In deterring pollinators 
from some crops (Potgieter, 1937: Durr. 1952). But 
to date, no studies have been published that corn- 
pair- /.. huiiii/c. or am oth.-r invasive ant. to native 
ants with respect to their attraction to Mowers, in- 
teractions with floral visitors, and any subsequent 

The impact of invasive ants on seed predalion 
and dispersal has received a little more research 
attention. Linepithema humile in the South Uriean 
fynbos is slower to discover eliaosonie-bearing 
seeds than the native ants thai it displace- ilium! 
& Slingsby. 1984). Moreover, when the seeds are 
found, the invasive ants eat the eliaosomes hut fail 
to disperse and I tun the seeds, leaving them vul- 

essarv to determine tin imp I I'. 

S. inricla, and W. aiiropunclala on seeds relative lo 
native ants, and any implications for plant com- 
munity dynamics. 

The ability to predict the nature and outcomes 

of interactions between invasive ants and plants has 
the potential to yield many rewards. Agricultural- 
ists and conservationists, in particular, may benefit 

Itv knowing whelhei a | aitial invader is likely to 

have impacts that will thwart their management 
goals. This lorcsighl mav provide lurlher impetus 
lor pulling policies in place to prevent the invasion. 

I the specific ant traits that in- 

lh plants mav 

reveal options for 

mitigating un- 

I'siraltle effects. 

If a key to many 

effects of inva- 

ve ants is their 

extreme abundanc 

e and pen. haul 

,.d -I . 

duced seedling 

•X Slingsltv. 1984). A recent experiment reported 

that the disruption ol ihis mv rmeeoehoioii- mulii- 

tian, 2001), but this study did not take into account 

other potential effects of the anl on the plants (e.g.. 
pollination) that also might have been responsible 
for the observed differences between invaded and 
iniinvadcd sites. In Coisica. /,. lunatic is more like- 
ly to find and remove Ancliusa crispa V i v. seeds 
with eliaosomes than without, whereas the domi- 

seeds with and w ilhoiil eliaosomes in the seeds |he\ 
remove, ll is unclear what impact the differences 

mav have on A. crispa population dynamics (Ouil- 
ichini & Debussche, 2000). 

■i ,\\u\ mix la also 
appear to be attracted to seeds in their adopted 

I ,!.., mi /'. •/' ,. aaccphala 

■ In of 
lh .111- 

; (Majer, 1985). So- 

crops and other plants (Ready & Vinson, 1995; 
Morrison et al.. 1097) and is attracted to most eliao- 

some-beai ing seeds, but often destroys or scarifies 
them (/clller et al.. 2001 I. Wasmannia auropunc- 
i poor seed disperser 


A s,-h,mske 

ic its seed harvesting or 

:■ halu' il- I' ha- i'.;. 
lie. lurlher study is nee- 

area may not be necessary to prevent outcomes that 
arc counter to land management goals. Agricultur- 
al. sis have realized the link between invasive ants 
and llomoplera outbreaks, for example (I landers. 
1951; Prins et al., 1990; Reimer et al., 1990). Pre- 
venting the ants from foraging in trees via use of 
sticky barriers or other means decreases the llo- 
moplera population below pest levels and deprives 
ants of this source of carbohydrates (Samways. 
1990). Therefore, although the ants may still be 
present they nit longer contribute to yield loss via 
theii relationship with homopterans. Deprived of 
their carbohydrate source, they may not be abun- 

or native ant community (Addison & Samways, 

foraging conditions can also be manipulated 
naturally sit thai the ants arc less or 
even aid in achieving management goals. In Zan- 
zibar coconut, riicidolc mcgaccpliala is a [test be- 
cause il does not prey on the primary agent causing 
premature nulfall. the bug Pscitdolheraptus wayi 
(Brown), as much as the native weaver ant it dis- 
places. I'hcldole megacephala is attracted to the 
palm crown bv nectar, pollen, and various homop- 
terans. Betaining ground vegetation in palm plan- 
tations provides I', megacephala with enough for- 
aging opportunity that il does not ascend llie palm 
trees in search of food. The weaver ant is then free 
to inhabit the canopy and prey on I'. wa\i (Bapp 
& Salum, 1995). Similarly, in Louisiana sugarcane, 
a l<> . in- ioa.ll. a :• ■ ■<!■■ to gi •■ i lh. cat \ pall 

.mil i- capable 


dlcclnc Helen- 

— is seed-limited 

di-plaei-il -.'.-(I 
ing native ants. 

1 dispell 

ilif.1 - iS| 



js T T §e T 

1 1 a | 

5 -s-ol 1 1 1 

1 J ■! 

ill! -s Is 
IKi H 1 hi 

L ^ i ~ I i 1 1 "a 1 "J 

■ In a carbolmh 

.in" cirbolmlra 
Her pollinator. 
g pollen: 

= ■; ii-i'f i N = j j j 

MilirJ Hi! 





I i IJ ].! Ms Hi 1 1! 

J 1 U H Jill, nin It 

£ = = : 

1,1*1! UilJifJi 


-? ill! v I I! !?]< M ji 

I 1!j 

si .till e 

S a s S s j fc | 

J.fl.f »<fjj 


1 ni 

If IN 
. - -= j i 

1 :N M i!-l 

i ? 4i^ ill- 

7 e = z 

|| IJ 

\ r £ r " 1 

of the season facilitate- an increase of .S. m/i'rto 
population- hccaii-o il provides al.iiiiilan! prev (Ali 
et al.. 1984). As the weeds die bark with the clos- 
ing ill' the sugarcane <-.iru>|>v. tlie anl> transfer their 
foraging to I lie sugarcane where thev are effective 
piedalors against the -ugarcaiic I to re r. Dialled sar- 
charalis (F.) (Ali & Reagan, 1985). 

Traits assoeiated 
evated abundance, 
high-carbohydrate r< 

ctly and indirectly 

dance of invasi\e ,nil- letnl> lo be higher than na- 
tive ants, but as shown in the examples, it is where, 
how. and when the hulk ol lhe-e abundant worker- 
tend lo forage that iillimalelv iiilhiences interaction 
outcomes. As described above, while holh \n>>pla- 
lepis grariliprs and Phciilolc megarcphala can be 

coconut palms are quite different depending on 

ing ol the ants. However, the relative scarcity of 
Solrnopsis inrirla on Calalpa liignoiimidrs appear- 
not to be a factor in the protection of the plant from 
its mam herbivore because it is offset by the ant's 
foraging elliciencv. Invasive ant colony cycles and 

determining foraging behavior, and may or may not 
coincide with plants' or llomoptera carbohvdrate 
production and need for protection. In the case of 
N. inrirta on C. bignnnioidrs, had it been the time 
of year when >'. inrirta seeks carbohydrate-rich re- 
sources, or had the exlrafloral nectar contained a 
prolusion of amino acids, perhaps visitation fre- 
quency would have been higher and an even great- 
er decline in herhivorv wilne-.-ed (Ness. 2001 ). for 

n plants, it will be useful to knov 
or other cues trigger the ants' ag 

i:re— ive behavior. Win. for example, doe- \implo 
lepis grarilipcs di-placc die coconut bug \mhlypelt< 
rarapha^a in Solomon Islands coconut ((becnsla 
de, 1971), but fail to deter the coconut bug I'sru 
dotheraptUS inni in /.mi/iImi r..«„mil (Was. ITvll 
Likewise, carbohvdrate resources can vary greatl; 
in eompo-ilion and then-lore in their attraction o 

■ hints precludes any general- 

opporlum •• to el se among Pinions floral nectars, negative 

extradural nectars, or homopleran exudates. The ization al 

options and preferences may vary over the season on plants. Indeed, the same ant may affect different 

and result in different outcomes for the individual parts and processes ol the same plant in different 

plants and the plant community as a whole. For ways at different times (Lolgren, 1986). Nonethe- 

example, aphids attract Solrnopsis imicla to nee- less, the existing data offer a starting point for pre- 

the season nectar is pivlerred and the more likely than another for each type of interac- 
ants are seldom observed tending aphids (Agnew el tion. In Table 2 I offer predictions about under what 
al., 1982). The amount and types of amino acids conditions an individual plant may face a high or 
and sugars, and the balance ain..ti- them have all low risk of an adverse effect or may benefit from 
variously been suggested as all,-, ling attraction of association with an invasive ant for interactions in- 
ants to nectar and honiopieran exudates (Lanza, Yoking plant protection. Ilomoptera tending, seed 
1991; Lanza et al., 1993; Koptur. 1979; Vander set. seed dispersal, and seed predalion. A plant 
Meer et al., 1995; Koptur K Truong. 1998). Further may simultaneously have a high risk of adverse out- 
research may reveal whether invasive ants are at- comes in some categories and a low risk or polen- 
tracted to certain carbohydrate sources more than Hal to benefit in others. Net effects will depend on 
native ants, and why an invasive ant may be more whether impacts combine or compensate for each 
attracted to one carbohydrate source over another. other. For example, (he reproductive capacity of a 
as well as how any observed preferences change plant may be severeK curtailed if an invasive ant 
with other variables (e.g., availability of other re- both disrupts pollination and interferes with seed 
sources, ant colony needs). dispersal. Alternative! v. there mav be no net effect 
Flucidating the attractiveness and availability of on a plant's reproductive success in the presence 
different carbohydrate resources is likely the key of an invasive ant that increases seed set and in- 
to predicting when and where invasive ants will be terferes with seed dispersal (e.g.. Horvitz & Schem- 
'"" , - "'«■'•' '""'■ "'<' potential for a s ke. 1984). Plants mav have then ,m.i eompensa- 

ubsequenl effects. Of the three allraclanls. lolA mechanisms: angiosper.ns that rely on * 

^ra, extrafloral neclai. and floial ne< lai. it specialized pollinators mav be clonal or extremely 

Homoptera are the most important lures for long-lived, thereby decreasing dependence on seed 

ders because ol their near ubiquity and production (bond. 1994). 

. Extra- 
floral nectar, while considered generally attractive vas ,| v ( |j|| (Trnl potential net on 
to ants (Carroll & Janzen, 1973). is not as widely lhe Iieed for resear , h that takj 
available. Moreover, plant, that posses inducible imi|| ; n | e m( ,. ha nisms through y 
exlralloral nectaries may offer nectar 

may affect plants, and ill. consequences lor the 
appease the sweet tooth of invasive ants. ,,,„„ ;| , |H1| , ll | a|i|H1 ail( | ( ,, mmunitv levels . The 
P er hap* complexity and context-dependency of ant-j)lant in- 
teractions mak<- predicting the net effects of inva- 
sive ants on plants a lormidable challenge. None- 
theless, incorporating characteristics common to 
in\asi\e ants into pre-existing models ol .ml plant 
interactions provides a framework for gem-rating 
testable hypotheses about how invasive ants mav 

defense's can be thwarted (Gucrranl ov Fiedle 
1981; (d.azoul, 20(H)). Once attracted to the plan 
the invasive ant may deter herbivores, as part < 
the food lor-proleelion mutualism with exlrallor; 

imIik cl e|( ( -c| ol the ul-' -i — ncia on will, Hon , . 

... c.e. <™. ul ..„. w.u. .....■,..,■- lnU . ra( ., wi , h , ants aIH , lll( , ( , )nst ._ |,„ ,| |( . 

logons byproduct parasitism, al- |)|;m( y ^ „_. ,„ ^ _ ^ <<t >lll( . ( |,_ 

though not vet reported in th< 
plausible; invasive ants lured to plants by 
Ilomoptera oi exlralloral nectaries may deter pol- 
linators or a key herbivore enemy. Since evolution 
did not play a role in shaping the interactions be- 
tween invasive ants and plants in theii adopted 
habitats, we should no| limit ourselves to consid- 
ering only those outcomes that would be evolution- 
ary plausible. 

redly to conservation and agriculture, and provide 
insights applicable to both invasion ecology and our 
iindci-landing of ant-plant interactions. 

ddicott. J. F. 1978. Competition 
and ants. Canad. J. Zool. 56: 
ddison, P. & M. J. Samways. 2000. A survey of a: 

Ih,'- Western Cape Province, South Africa. African r 
tomol. 8: 251-260. 
gnew, C W., W. I,. Sterling & D. A. Dean. 1982. Inf 

predators. Km iron.... Kntomol. 11: 629-634. 
li, A. D. & T. E. Reagan. 1985. Vegetation man.pulat 
impact on predator and prey populations in Louisia 

J. Econ. Kntomol. 78: 

1 .;: ; ; hi . 

in ii . li ill ! i i< I i « i -i " ' -- ■'• l) - ' an 

; ! , I \\'.. Jr.. T. II. Su. K L MeKwen & I), Ceding. 

I<>82. field um-ligalinii- of the interrelationships of 
thi' hig-headed ant. I'lwidole mcgaccphala. the gra\ 
pineapple mealybug. Ihsnucoccus neobrcripes. and the 

i »ah. I SA. Pro< 

Hawaiian Enlomol. Sue. 21(1): 51-58. 
5eallie, A. J. 1985. The Evolutionary Ecology of \nl 
Plant Mutualisms. ( '.amhridge I ni\. Press, Cambridge. 

Buys. P. 1987. 

Competition for nectar 

between Argentine 

ii {Inilimni 

■vbees (Apis nwlli- 

ironbark (Eucalyptus .< 

dderoxylon). S. Af- 

i , an .1. /onl. 

22: 173-174. 

honevhees in 

Sonll. \liiea. Pp. 519 521 m IC K. Van- 

der Meer & / 

\. Dendeni, iedilois). \pplie<| Uyr.neeol- 

ogy: \ World 

Pei^ieelne. Westview 

! i, u,|o'. 

Caiimiell. VI. E.. 

, M.J. Way & M. IC Paiva. 1996. Diver- 

es associated with 

oak. pine, em 

calyptus and arable hi 

ibitats in Portugal. 

In-eetes SocK 

iux 43: 37-46. 

Campbell. C. A. 

M. L994. Homoplera 

associated with the 

and Tetramori 

'urn aculeatum (Hymen, 

iplera: Kormicidae) 

;hana. Bull. Kntomol. 1 

{es. 84: 313-318. 

I). J. Janzen. 1973. I 

ml Rev. Ecol. Syst. 4: 

of a biological in- 

\aime 113:635^639. 

Cole. K IC. A.< 

:. Videos. 1 .. 1 . Loupe K \\. \\. /uelilke. 

Reullev. 1 

!. L. 1977. 1 

Kxlrafloral nectaries 

and protection 

In pug, 

lucioiis bodyguards. Annual Rev. 

Ecol. Syst. 8: 


Bulger, 1) 

T., A. V. Su 

arez, K. R. Crooks, ! 

S. A. Morrison 

& T. J. 

C.-.m . .-(MKI 

Arthropods in urba 

„ l,ab,lal hag 

:alifornia: Area, age 

fects. Ecol. Applic. 

10: 1230-1248. 

Bond. W. 

J. 1 < )' ) 1 1),, 

\-se-slllg the 

and disperser disruption on plant 

on. Philos. T 

rans., Ser. B 344: 83 

& P. Slingsby. 

1984. Collapse of at 

—The Argen 

•uchorous Proleaceae. Ecology f>5 

P. ele.l. 1! 

. M. & J. K. / 


C.nanl, P. & C. IIimi.iiii.i 20(1 

h,h, (Mmik i I I ' ' ' ' ! " " 

land of'lluwai'i. bishop Mus. Ocas. Pap.6l:2l 22. 
Cornelius. M. L. & K. A 
plant chemistry on the 

nun,!,; lIKine,.,, 

: Kormicidae). J. Insect Bel 

•milkman. \1. \.. \\. \. Card.,,-, K < ,. I). Buntin. 2001. 
Effect ol red imported lire ant 1 1 l\ meuoplera: Kormici- 
dae) on Rlum>c\lliis nmicus (Coleoplera: Curculioni- 
dae). a biological control agent of musk thistle, liiol. 
Cuulrul 30: 612-616. 

kislow. C. M. 1984. Differential heneliN Iron, ant alien- 
dance to two species of Homoptera on New link iron 
wed. J. Anirn. Ecol. 53: 715-726. 

Srun,lein. I. L. 1991, Conditional outcomes in mutual- 
istio interactions. Trends Keol. Kvol. 9: 214-217. 

. 1998. The contribution ol an! plant pmlecliou 

studies to our understanding of mutualism. Biolropiea 
30: 150-161. 

I OchelAlim glabcr (Mayr) 

ng resource for aphids. Oeeologia 79; 315-321. 
dson, I). W. 1997. The role of resource imbalance 

nl. J. Linn. Soc. 61: 153-181. 

— . 1998. Resource discovery versus resource dom 

i,ade-oll. Ecol. Enlomol. 2.5: 484-490. 
— & D. McKey. 1993. The evolutionary ecology ci 
mbiotic ant-plant relationships. J. Hymenoptera Res- 

1,1 lll( " raitif'or«-sl tier Irommloxa a/rirana In the mi 
\< ): ff an1 Petalomyrmexphyl , Oecologia U2 209 21( 

''I'x'lr!'"!,, ,V , »nni 1 ) | l ' V, '' n '' ' <J><X> NotHvl.-w flickin- — S S. Terrin. IW. Ant-planl-homo 

Anim. Behav. 51: 1 07 1- 1 075. 

I.. Davis A S. Cover. 2000. Exotic anls in phvlo host. Pro.-. 'box. Soc. London! So, |{ B,ol ' 

265: 569-575. 
Ghazoul, J. 2(XK). Repellent flowers keep plants atti 
live. Hull. Brit. Kcol. Soc. 31: 4. 
'ii in Hi- .hmkorshoek \ alley. J. Fntomol. Golelli. N. J. & A. K. \r.iell. 21)00. |!in l! e,, i; ..iphie ell. 
1 l;l: 2 ' ' -'•'• of red fire ant invasion. Ecol. Lett. 3: 257-261. 

(.. I. A II. Go.U Bcrgallo. |W2. Bigger (keen, P. T.. I). J. 04W.I iK l». S. Lake. 1999. Mien 
reduce hcrhivon and herhiw.iv residence invasion and ecosvsletn collapse on Christmas Isla 

■- "I -i" anl-plant: l-/,v„ murlirri \s. Cor- Indian Ocean. Miens 0: 2-4. 

r«/.v on r;«T«/i/« pmlnstuclnu. Oecologia Greenslade. I'. J. VI. 1671. Interspecific competition ; 
frequetiev change- anion- anls in Solomon Islands 
l"'.»2. I he Vrgenline ant. Irulomvrnwx conut planlalions. J. \,,,.|. hoi !!• 323 .452. 

. farming South \frica 27: 381-384. Ouerrant. K. (). & P. L. Fiedler. 1081. Flower defen 

'■ V '"'"paralm- morphological sludv of against neclar-pilferagc bv anls. I S i, .( r, ,,,,<-. i (Nippl.l 

■nhis ol anls (IKinenoplera: Formicidac). 25-33. 

1072. Pests of Cocoa. Longman Group. 

I.ricius. (Hemiptera: Coreitk 

. Chem. Feol. 19: 8514561. 
' ' l ; ul " '. }!: -'>,-206. Hames. I. H. K J. B. Haines. 1078. Pest status of the 

uhaiik-. \l I). 2001. Fstunales ol the direct and inch- crazy ant. Anoploh-pis longipcs (Jerdon) (H\ nienoptera: 

reel cllccts of red imported lire ant- on biological con- Foi mioidael. in the Sexcl,4l,-. Hull. Lnlo„,4. lb- Ul\- 

trol m held crops. Biol. Control 21: 3:5-43. 627 638. 

:vans, H. C. 1073. Inverlebrale vectors ol I'lnlnphihnra llara. A. II. ,K T. V. Hata. 1992. Ant control on protea 
g hlaek pod disease of cocoa in Ghana. in Hawaii. Sri. Ilort. 51: 15.5-163. 

Haskins. C. P. & F. F Haskins. 1965. I'heidolc mega- 
cephala and Indomyrmex humilis in Bermuda: Equilib- 
ria. T. V. A. II. II.,,.,. I!, k. s. Ih,. I! f Kanek,'. X \ 
L. Tenbrink. 1995. Fxeluding pests from red ginger 

poKetlulene hags. J. Fcon. Fnlomol. 8P^303-5'»7. 

Impact ol' an introduced ant on native forest inverte- 
bialc: 1'hfnlnlr iufgacvphuhi in monsoonal Australia. 
(•colon,,, |2(»: 59.5-r>04. I!, a F. (). Wilson. 1900. The Ants. Belknap 
Press. Cambridge. Massachusetts. 

Holway. 1). A. 1995. The distribution of the Argentine 
ant [Linepitlu-mo humilr) in northern California. Con- 
servation Biol. 9: 1634-1637. 

. 1998. Factors controlling rate of invasion: A nat- 
ural experiment using Argentine ants. Oecologia 115: 

. L. Lack A. V. Suarez. N. I). Tsutsui & T. J. Case. 

2002. The causes and consequences of ant invasions. 
Annual Rev. Fool. Svst. 33: 181-233. 

Horvilz. C. C. X I). W. Schemske. 1981. Effects of ants 

erh Fcologv (>5 I «>9-l 578 

1986. Seed dispersal of a Neotropical 

c B,oi,o,„ca 18: 319-323.' 

I. T. & P. Yang. 1987. The ancient cultured cit 

. Bi<,Science 37: 66.54)71. 

. Gordon. 1996. Exploitation an. 
invasive Argen 

Oeeol„gi ;l 1(15: 105-112. 
Huxley. C. H. & I). F Cutler. 1991. Ant-Plant Inlerae- 
lions. ()xf„rd Lniv. Press, Oxford. I.E. 

Ann. Appl. Biol. 75(3): 331-3 15. 

ib.c. (, ,\ \\. L. Brown. Jr. 1978. The 

dnclion of the pesl ant Wasmannin ,uno, 

17: 1.1" -M2. 

-Howes, J. R 1999 Ixoti. ants in \sia Is 

il; , . o,l.,„l 

at risk? The case of Hong Kong. Aliens 9: .5-4). 

ala. B.. 1 . it/. T. V Pong c* \. ,|. Helbig. 1084 


Oecologia 79: 193-170. 

seller. M. K.. K. II. Hoffmann & W. Volkl. 

petition for nmliialists in an aul-hoinopicra 

nx'di; 1 bv hi. ran hie-, of ant attendant-. 


tinders. S. F. 1951. The role of the ant 

control of homopterous insects. Cariad. Entomol. 83: 


ajor an 

problem- of 

x. \and, 

■r Nicer & \. 


A World Per- 

Press. Oxford. 

es: Predation bv nec- 

relation lo van 

lloral form of 

rild :■ 

i. 1',,1,-nioimin, 

m. Oik.,s85: 

. Mckev & M. Anstett. 


Benefits con- 

timid" a 

nts: Active ant 


»■< | -.-lion 

s. Leiden. Netherlands. 

. Threats on Pacific islands: The spread 

brown ant Pheidole mega(cphala ( Fabricius|. mi the a 
fauna of the Perth metropolitan region. Western Austr 
ha. Pacific Conservation Biol. 6: 81-85. 
rtcGlvnn, T. P. 1999. The worldwide transfer of an 

.& i. F. Moxon. 1985. I'otentia 

pactions. Pp. 207-242 in 
^lant-Animal Interactions. 

I Their Unbidden Guests. 

e Jong. L993. Attractive- 
dilemma. Oikos 66: 180- 

61 in D. F. Williams (editor!. Fxolie 
Impact and Control n| Introdueed Spee 
Press. Boulder. Colorado. 
Messina. F. J. 1981. Plant protection as a 

. E. I.. Va 

Preferences ol' I 

. Hot. 66: 1016-1020. 

. Kntomol. 16: 203-210. 
o. S. Pulim & Y. Z. Chi 
e fire ants Solenopsis invi 

Formicidae) lor < 

enrod (Solidago sp). Kcol.^x 62: I llt-l 1 10. 
lontgnmerN. B. R. & C.S.Wheeler. 2(XK). " ' 

aoh\il\ of the wee\il <h w./-.v liliosu: \ biological control 
agent of Melaleuca ipuiupienerria. .1. Insect lieha\. 13: 

Morales. VI. A. 2000. Mechanisms and dc,,-il\ depen- 
dence of benefit in an anl-membiaeid mutualism. I. eul- 
ogy 81: 482-489. 

mirrisnn. J. K.. Jr.. I). F. Williams. I). II. ()i & K. N. 
Potter. 1007. Damage to dr\ crop seed b\ red imported 
[,,c am lIKmenoplcra: Fonnicida.-i. .1. L< on. I nlomol. 
90: 21 {5-222. 

Sess. J. H. 2001. The Calalpa hignonioitio h-.d Web: 
Implications of \ariable Interactions \mong four fro- 
pliic Levels. Ph.D. Thesis. I nnersilN of t.oorgia. \th- 

. Haines. J. B. Haines & P. L. 

Mathias. 1 070. The era/A ant. \noplolepis longipes 
!|, ,d.i lILin. nophaa. Formieidael in Seychelles, and 
its chemical control. Bull. Lntomol. Res. 06: 0,-1 I I. 
ansenmair. k. F... M. Heil. W. M. Kaiser. B. Fiala. T 
Koch X \\. Poland. 2001. Adaptations to biotic and 

(Coleoptera: Cureuli 
Peakall. R 

. Diaprepes abbreviate L 
) the fire-ant Solenopsis 
■rmicidae). J. Chem. Ecol. 

lei & A. J. Beattie. 1991. The 
portance of. ant pollination. Pp. 
421-120 inC. IL lluxles ex |i. f. < ulloi ■ (editor). \nt- 
Plant Interactions. Oxford Univ. Press, Oxford. 
Porter. S. D. & D. A. Savignano. 1990. Invasion of po- 

. IL<;. Bobertson ,n 

. Prins. 1990. Pest ants 

shal. .1. I col. II): 31-18. 

Spread of \rgentineants|/ J /He/«7/iem«/»/- 
,„., aal reterence to Western Australia. Pp. 

.ntrol of Introduced Specie 

. Colorado. 

terick. 2000. Effects of th< 

Applied Mvrmeeolo^: \ World Perspective. Westvi, 

Press, Oxford. 
Puterbaugh. M. N. 1998. The roles of ants as llower \ 

ilors: Fxperimental analvsis in three alpine plant -| 

eies. Oikos 83: 36-46. 
Quilichini. A. & M. Debussche. 2000. Seed dispersal ai 
' germination patterns in a rare Mediterranean island c 

..i|.|.. (. ». M S. Salu.n. IW,",. \rn I:,,,,,.,, ,,,-si .hmiii^t- Austral. Keol. 25: 253-259. 

and yield u, relation In iho dci.siu „l weeds in c-oc-onul Veeresh. (,. k. X Cubbaiah. IWl. A report on the Vrazx 

s.tes ... Zanzibar, lan/ama. .1. \ppl. K mol. 1 19: 15 anf ( \nn,,lol,;,is lon^ipr.s .hid menace in Kan.alaka. 

^ J. Soil. Hiol. Kcol. 4: 65-73. 

";•;;';;,' i r | M ^, l! '"";; '"" ■ ,' J ,i,m i ^ - s ' |! ^ (..,,,,!.,,, i<> ;: , i„, N . |mNs . 

|I "' laboratory kimmmn I n I " 1122 I 1 '. I i, - I! \, (l ,ni I , , ' ,'i"'l '" ",' „,',!' i n'pi, I and ( ,,,m~,l 

{eimer, N.. J. W. Uranlsley M (,. Jalin. I')<>0. Pesl anls of Social Insecls. Praeger, New York. 

Meer & A. Dendcno (editors). Applied Myrmecology: A .■fleet of the \rgeiuine ant.,;,iil,nn„ luuuilv i Max 1 1 

World IVispcclne. Weshiew Press, Oxford. (Hymenoptera: KormicidaoL on Ilnwer-x isilim. insect. o| 

loom, P. M. 1975. Diversity and organisation of the I'rotca ,u,i<la Mill. (IWaeeae). African Knlo.nol. I: 

:-'" - '<■ • ' '. ' >i l<> I i dand and tree 285-287. 

crops in Papua New Guinea. Austral. J. Zool. 23: 71 - Ward. P. S. I9J57. Distribution of the introdueed Argen- 

89 - tine ant {Iridomyrmrx humilis) in natural habitats ..i' the 

-~ ;^ J'-- S. <;. Sniilh. I<>,5. U.-h.l.v. abimdan, , and lower Nicamcln |{j w . r \ a || rv an( | j, s effe( . ts „„ ,„,. 

cocoa ecosystem in Papua New Guinea. J. Appl. Keol. Way. M. J. 1953. The relationship between certain ant 

1 ' '"" species with particular reference In biological control of 

•akata, II. I<*)|. I low an aul decides to pr. 
allend aphids. I.Vs. I'opulal. Keol. 0: 45-51 

producing llon.oplera. Annual lie\. Kiilo. 

Vlyriiiecology: \ World Perspeclive. Westview Press. opcltis : llivolmuiuu- damage and anls Willi special relor- 

, ''xlord eriee to Malayan cocoa sinallhohlings. J. PI. Protect. 

1 \- I- S- <:l \ K m , A N \ „,„„■;, Tropics 6: 1-12. 

I'><>2. I I acceptability and distribution in the colony & . 1992. Hole of ants in pest manage- 

of the bigheaded ant, Pheidole nu^atrphala (Kahr.) met. Annual Key. Knto.nol. 37: 179-501. 

(Hymenoptera: f ormicdae). Proc. Ilawanan Knto.nol. . M. K. Cammed. VI. IK Paiw, K C. A. Colling- 

■ S<K ' ' M u 6 m^' n ,i ■ WOOf1 ' IW7 ' l)lslnl " 11 "" 1 '"" l ^"amics of the Argen- 

.peneer, H. 1941. Wasmanmu ... .alms !>.„■ ml t.,n>-p,iliem„ {/rid»m\ni,c\) hum He (Ma\r) in r<- 

. B., H. (;. Thorvilson & J. W. Johnsor 

,x ,„| unpolled . M. H. Paha t K M. K. Cam. 

■ |..>|.|||; ; ||. 

1987. Ant aphid mutualism. Pp. 335-365 in 16,5-167. 

inks & P. Ilarrcwijn (editors), Aphids: Their Wyatt, |{. |y«0. The impact of nectar-robbing ants o 
Natural Knemics and Conliol. Vol. V. Klsevier, pollination system of \sdcpius ruiassarir,,. Hull. I 

Porter. 1991. Comparison of diets Voung. (,. !{..(,. \ |{,.||is. <;. |{. |} nm „ fl K . S 

1,1 "> ""I' 1 '■' ' " " '<■) s "l«l 2001 Ih. ,.a/\ ml I;. ./)/,./, /.,s »„„■///,« s|S 

llls - ■'■ Kolomol. So. 26: 150 165. menoptera: Formicidae) in Kasl V.nhem Kan 

). further investigations regarding lia. Austral. Kntomol. 28: 97-104. 

rlance of the Gramang-ant. Rev. V... I). \\. 2(M)K Parasite ol niiiln ,I,m,,~. I!„ 

73-395. Soe. 72: 529-516. 

eawright. /etller. J. A.. T. P. Spira & C. K. Allen. 2001. 

t (Hyme- mutualisms: Can red imported fire ants sou. 

°"""" li "' ) |»lt.i»iisliiiiiiLii,i „■*, sc K. ,,„ lion.liip? ISi ol. Conscnalion 101: 249-253. 

. Klorida Kntomol. 78: 144-154. Zimmerman, K. C. 1970. Adapt ve rad Hi n 

, C, K. A. h.bry I)e liruyn & P. N. House. will, special reference to insects, liiotropica : 


onsidered, behind habitat destruction, the second largest threat 

of I he invasion at the family, species, and/or population levels. 
i the I niled States. A synopsis of molecular systematic advances 
ipact on the control of Tumarix is discussed. \lso. a preliminarv 
using ehloroplasl l)\ \ sequence data. This analysis investigates 

- tin,, ugh molecular svs|, ■malic and |>o|>ulal inn-lcvcl studies will prove I 

The invasion of habitats by noti-naliv<- organisms and a large-scale biological control project con- 
is considered the second largest threat to biodiver- dueled by the United States Department of Agri- 
sity worldwide behind habitat destruction (Wilson, culture. Additional legislative control may be re- 
1997). In the United States exotic plants now rep- quired, as cultivars of Tamarix are still available 
resent 17.3% of the flora (Kartesz & Meacham. from numerous horticultural suppliers. The effee- 
1999). and approximately 400 of the 972 ph 
and animals listed by the Fndangered Species 
are at risk primarily due to competition with 
predalion by non-native species (Stein & Fl; 
1996). For these reasons, the control of invas 
is becoming an integral part of ecosystem stew; 

Methods of controlling invasive plants incl 
manual removal, lire. herbicides, biological control "^j ( . ^^ ^^ ^ bg [^ for ^ 
and legislation of nttpor. and sal,-. Filed, w control (>f h( , Kt . switching , )V , onlr()ntin g the 
of invasive plants often requires explicit character- ^ ^ ^.^ ^ dogely ^^ pl; 
ization of the invasion at the family, species, and/ 
or population levels. 

Several species of the genus Tamarix L (com- 

I'hylogenetic relationships of the invasive plant's 
family are important when biological control is pro- 

In the past, Tamaricaceae usually v 
in the plant order Violales of the Dilleniidae (e.g.. 
Cronquist. 1981). but recent molecular sequence 
data analyses have altered the traditional ordinal 
) are. as a group, considered one ol the «,„>t |;1( . ( . m(1|l( „, m;mv ,„,„ , ami | i( . s , an(1 Tamarica- 

1 - ! c ^ Jt. ceaearenowincludedintheC aryophyllales(APG, 

I9<)«'i). These changes will alter the plant taxa to be 


- . h t l» 1 ' I " i! ' I' ' ' ' lM«"s,onS.-v,M g, 

i I!. -.1 ...I J <.a-k.n x.i I ,.--.., . c foi Research and Kxplorat.on gi 

MMM <><) to J. (,askin. Il„ M. II i I < nnd - • n| 'or! ol Missouri Botanical Garden graduate students, and an t 
Science To \chi. I. ,.' |ST\I! o.diial I, H> hi| - .1 Caskin. 

Missouri Hot. I i.,imI.-ii IMi I'm, '■'■« ,i ' -. \ I'i.- I'l'<- DSDA-AI 

M'VliL I'.O. I5..x In.",. Sidi.ev. Montana .7^70. I .S. \. ig.,sku,("'sidi, 

Ann. Missouri Bot. Gard. 90: 109-118. 2003. 

T. afrieana Poir. 

T. aphylla (L.) H. Karsl 
T. canadensis Willd. 
'/: chinensis Lour. 

morphologically similar l<> /. rananmsis anil /. #«//»•« IM ;1( . sl j, 
rarely cultivated, not eM.-nsi\«-l> naturalized (Maum. 1 ( W)7) 
irn.r pliolo-i.-alK dissimilar to all oilier I .S. Tamari.x 
morphologicalb .inula, to /. gallica (Crins. 1989) 
' phologicalh similar lo /.' ramasissima (Crins. |««{<)| 


s> iii.ii> in of 7.' chinensis (Haum. 197H) 

'i rpliolou ilk li simil., | ,|| il , ,\ isive U.S. 7«hhi/m 

T. pentandra Pall. 
'/: letrandra Pall. 

synonym of /.' ramosissima (Haum. 1978) 

U.S. invasive specimens with llus name considered to he T. par 

7. letragyna hlirenb. 

nalurahzed m easier,, I S.. no, u-| imasiw (Crins. 1 W>) 

fed m ffie risk analysis of biological control llie issue of inl'ia-pc-ilie -.n^iapliical variation. 
!nt host-switches. Many species of Tamari.x are widespread in Eurasia 

(Baum. 1978). and it is unlikely much of the 
SPECIES LEVEL genelic diversity of any one species was imported 

Hie Tamari.x invasion consists of many snecies. lo ,,u ' Vni[t ' l{ S,aU ' s - Hist(,li <al <^ords «<» not re- 

; of many species. 

some of which arc ,,„„ pliolo^cal h mmv sunilar. ^M^cise ongms or genetic information 

The specific identifies of u.xa imohed in the in- "* llu " " 1 "'«>< l '"' li '^ (Horton, 1964). Th. 

vasion are controversial (Crins. IW). in part, he- a ^"' s '"''"P ,, ' s "' <l <«*+'" » a "^ (lar lea 

cause most '/immn.v species cannot he distin- Dwrhabda elongata) may not \v 

h-optimal control. For these reasons, it wot 

graphic- origin of the invasive species and its co- " Srl " 1 '" kmm ll,,u " ,;m > .^''"o Is pes are represented 
lived biological control agents. Tamari.x is one of '" l,l< ; l ^ l,UilsloM - ;,,ul '<» * liat degree we can 

Additionally. 'Tamari.x is still bei M u lioi t i< nil iii- 

challenging genera among 

uni. 1978). and intermediale 

nosological charae.eis used •'l'-l-'nbu.ed in .he I ni.ed Policy 
in species ideiililicalion. These characler stales can '"''' ,l "' (l <'lenn,.,e ,1 llic .cnolv pes currenll 
often vary on a single indmdual from season to l' I «»"" 1 '' 1 l""do!,u,,ale „, llie pop! 
season (Rusanov. PIP)), and hybridization mav l{as, ' ,l "" ,l "'" s|,,,,lar '"'"P 1 ^^- 'nvas 

play a role in the taxonomic confusion (Kusanov. J"""/ " ""; ' '-""gnishable from cultiw 

1949; Wilken. 1993). Improper species identific 
lion could lead lo searches for biological conti 
agenls perhaps outside llie nafive range of the i 
vasive plant. The failure of morphological data 
■ ■hieiilali- the idctililiev ol in\,i-i\c Tamari.x speer 

3) Population Level 

Population4evel invesligation of any of the in- 
»asi\e Tamaii\ ma\ he neee>sar\ if the biological 
•ontrol agenls are extremely host-specific, and if aggressive subset has overtaken more than 
he invasive plant species has widespread origins. I .()()().()()() riparian acres (Brot hereon & Field, 
nitial biological control tests show that imported 1987). This infestation is expanding by 40,000 
nsects have differential effectiveness on what pu- acres per year (DiTomaso. 1998). eroding the bio- 
atively appears to be a single species of Tamari.x diversity of many western U.S. natural areas, in- 
i Ledeb.) collected from different re- 
■ U.S. (Or Loach & Tracy. 1997), raising 

Tamari.x is an ( 

)ld \\. 

>rld genus of approxitr 

51 species (Baum 

. 1971 

3). Eight to twelve of 


(Table 1) were in. 

1 to the United States 

southern Europe i 

,r \si 

a in the 1800s to be 


Volume 90, Number 1 

during wel periods in in riparian areas. Once cs- 
lablished. they can tolerate drought hy utilizing 
deep groundwater sources. They also exude excess 
salt Iroui saliin/ed water sources from glands in 
their scale-like leaves (Neill. 1985). which are sea- 
sonally dropped, forming a thick saline dull on the 
soil surface that inhibits the germination of other 
plants. In the U.S.. Tanuirix species are avoided bv 

& Field. 1987). and only two mammal species (the 

ivri\ upon them, with minimal damage to the plants 
(I)iTomaso. 1998). Tamarix invasions lower biodi- 
versity levels by displacing typical Southwestern ri- 
parian vegetation such as coltonwood and willow 
(Hughes. 1993). as well as the insects, birds, and 
mammals that these native trees support (Neill. 
1985). Their profuse growth alters stream and river 
dynamics by narrowing channel width i h'ohiii-ou. 
!<)(,.", |. and invasions can extend over 1 km on each 
side of a river (e.g.. Oila River, southwestern Ari- 
zona, and Colorado Kiver south of Hlythe, Califor- 
nia. U.S.: J. Caskin. pers. obs.). Dense stands of 
Tamarix. with then high rates of transpiration, can 

caused perennial springs and creeks to dry up. in 
some cases threatening regionally rare or federally 
listed species such as the desert puplish and the 
desert slender salamander (Kerpez & Smith, 1987). 
Tamarix invasions have proven difficult to con- 
trol. These plants cannot be killed easily by lire, 
bv cutting at ground level, or by herbicide applied 
to the foliage alone. Kffective removal is both ex- 
pensive and potentially damaging to the habitat, re- 
quiring mechanical uprooting, or culling at ground 
level with application of a systemic herbicide to the 
stump. Repeated treatments are often necessary 
(Neill. 198,5). Control is possible on a small scale. 
but land managers are often forced to live with 
large invasions due to prohibitive control costs 
(Stein <\ black. 1996). for these reasons biological 
agents were proposed as an alternative means of 

Well-researched biological control projects often 
come under heavy public scrutiny due to the po- 
tentially dire effects of control agent host-switches 
(Thomas & Willis. 1 99R). Therefore, biological con- 
trol researchers must unambiguously know the 
identity of the invasive Tamarix and its relation- 
ships to native species. Improper taxonomic iden- 
tification ma\ lead to searches for control agents 
outside the native range of the invasive species and 
thus wasted efforts or less-effective biocontrol 
agents. Inipropei identification of the invasive 
could also lead to the collodion of biological con- 

Irol ngeuls that have historic lies to sympalric con- 
gener plant species or to genotypes with a different 
phenology or developmental liming, again yielding 
ineffective biological control. Considering that the 
<i\r\,\ : j,r biological control research program -pans 
many years at a cost of hundreds of thousands to 
millions of dollars (Cillol. 19<)5). it is logical and 
economical to predicate a biological control project 
with precise taxonomic knowledge of the invasive 

Knowing the genotype ol an invasive plant is es- 
pecially important when choosing a fungal, bacte- 
rial, or viral control agent involved in a gene-for- 
gene resistance/virulence interaction (Kerr. 198, |. 
Even insects are often species-specific, and in some 
cases, host-specificity can reach to the level of the 
plant genotype. \n example is the differential her- 
bivory of the Hessian lb \\la\riiola </cslni<tor\ on 
different genotypes of wheat (Trilimm aeslirnm U.) 
(Schoonhoven et at.. 1998). Also, differential her- 
bivorv on plant populations has been detected in 
willow trees (Salix) under natural conditions (Rank, 

The saltcedar leaf beetle (Diorhabda elongata) 
from western China is already being investigate. I as 
a potential Tamarix control agent in i|uarantined 
and field releases (C. J. DeUoach. pers. conim.). In 
no-choice lesls. newly hatched I), chmiiala laivae 
were placed in vials, each with leaf material from 
a different plant specimen. The plants were col- 
lected from different areas of the I nited Stales and 
grown in common garden plots. Using morphology. 
all -pecunetis were determined to be the same spe- 
cies I'/] ramosissima). Observations on the feeding 
and life span of the insects were recorded, and sur- 
vival of the insects to adulthood on different plant 
specimens varied from WW to ()<7r (DeUoach & 
Tracy. 1997). The reduced survival on several of 
the T. ramosissima plants may have in part been 
caused bv less than optimal physiological condition 
of some of the plants (DeUoach & Tracy. 1997). but 
geiiolvpic differences in llic plants also may have 
influenced the results. 

The search for Tnmari.\ biological control agents 
continues, as the \gri< ■ullural lb-search Service of 
the United Stales Department ol \griculture does 
not expect that the current control agents will 
achieve satisfactory control ol saltcedar in all areas, 
and perhaps as many as 8, to 12 additional insects 
as specific herbivores will be required (DeUoach <\ 
Tracy, 1997). This is based on biological control of 
other invasive plants, such as cacti, lantana. and 
leafy spurge, which have required up to 15 or more 
insect species introductions (DeUoach & Tracy. 
1997) for effective control. 

Tamarix ramosissima is commonly s.,1.1 today ;l s \ recent study used \)\ \ sequence data to de- 
li ornamental il . Cultivars til '/.' ramosissima lerniine how many invasive Tamari.x species were 

lclude Tink Cascade'. Kosea'. -|,'uhia". and -Sum- n; ali/ed In the United Stales and to see if the 

ler Glow'. The most common of these is '/.' ramo- molecular data were with the morpholog- 

ssima 'Pink Cascade', known for its dense, dark ical distinctions currently used to segregate taxa 

! to mostly compound. (Oaskin ci Schaal. in press). The taxonomy and 

not simple, iiifloresceu. e racemes) and finely lex- morphology of the I2|)iitativ 
tured bluish foliage. These cultivars and invasive marix species were investigated (Table 1). Three oi 
populations are almost identical in Moral and veg- tlie sj)ecies names had been designated as syno- 
etative structures, and may only differ in the inten- nyms, and two were not yet considered invasive. 
sity of flower color, dciisic «i nifm-scences, and leaving seven putative invasive taxa. A moleculai 
phylogenelic analysis of these and other selected 
species in the genus was performed 1 1 . mi samples 
collected in the western U.S.. Argentina, and wild 

lound, making the density of inflorescences also n<a (voucher information is listed in \ppcndi\ I I. 
highly variable. Invasive foliage color can vary Phylogenies from both nuclear ribosomal ITS 
within populations from dark green In die Llue-gia\ and chlorophisl ini^-ltnt, inlcrgenic spacer se- 
found in the T. ranumssimn Tink Cascade' cultivar quence data were constructed and compared. Por- 
(J. (Raskin, pers. obs.). tions of the phylo-cnies presented in Figure 1 
As a wee<ly species, the Tamarix cultivars are illustrate incongruence with earlier laxononuc un- 
easy to grow and tolerant of poor soils. They are dcrslanding of the genus. For example, note that T. 
available through main nurseries, catalogs, and <-lmirnsi\ Four, and T. ramosissima, thought to be- 
from internet sales (e.g., Gerlens Online Shop, long in different sections of the genus (seels. (9//- Tamarix ramosissima is not Ic- A'"'/''""' and Tamarix. respectively), have identical 
gaily available in Colorado, Nevada, Washington. placement on both phylogenies. Additionally, the 
and Wyoming, where it is listed as a noxious weed mosl "'cent sectional classification of the genus 
(USDA, 2002). (Paum. 1978) was not significantly similar to either 
the chlorophisl or nueleai lop.. logics found in Gas- 

TAMARIX ^ or ma "> s;im |)les there was incongruence be- 

tween the chloroplast and nuclear evolutionary his- 
I Wlin I.KVKI. lories. For exani[ile. in the nuclear phvlogenv of 

Figure I, T. ramosissima spe. iinen Schultc 1 was 
Biological control avails are assessed for their in a clade with all of the other T. i 

f host-switching In placing lliem on U.S. na- i„ the chloroplast phvlogenv it appeared in a clade 

live plants that arc closely related to Tamarix. Ta- with T. canariensis \\ did. (Caskin 3049) and T. gal- 

maricaceae, along will, the sister family Frankeni- /,;,.„ |.. (Caskm /.'UWI. Similarly, a /.' ranarwnsis 

aeeae, had historically been placed in the order specimen (Kirk 2) was in a chloroplast clade with 

Violates (Cronquist, 1981 ). Therefore. U.S.D.A. re- another T. canariensis (Cask,,, 3020), but in the nu- 

searchers tested if tin- Tamarix control agents elear phylogeny il was found far from specimen 

would feed and reproduce on plants Irmii other (am- Catkin :U>2(). as the sister to the T. ramosissima 

ilies in this order, such as Frankeniaceae and Fou- Hade. These incongruences of chloroplast and nu- 

quieriaceae (I)eFoach cK Tracy. 1997). liecent UNA r | r ar evolutionary histories, which were significant 

sequence data analyses strongly suggest that Ta- based on the Templeton lest (Templeton, 1983), 

maricaceae and Frankeniaceae actually belong to- supported a hypothesis of hybridization (W hitte- 

gelher in the order Caryophyllalcs (A PC, 1998), more & Schaal. 1991; Soltis & Kuzoff, 1995). 
closely aligned with families such as Droseraceae The study concluded that morphology within Ta- 

and Polygonaceae (Fledd el al., 1998). This phy- marix is often misleading as a means of identifying 

logenetic rearrangement requires a substantially specimens. Also, though not all putative invasive 

different set of test plants in the greenhouse, which species could be distinguished with molecular data, 

may provide significantly different assessments of there was enough phylogenelic resolution to rec- 

the risk of host-switching. ognize four invasive Tamanx entities in the U.S.: 

Caskin and Srliaal (in 

(11 '/: aphvlla (I,) H. Karst, (2) 
(3) T canariensislT. gallica, and 
ramosissima. Additional!}, then 
introgression bet 
sis, and T. gallica, which is a likely s 

III- ill 'l il <■ < iiai',M l(M-|/;iliul Ol so.'IH 

\asioa- (« ■ ; i > k i 1 1 tX Seha.T in press). 

To examine the Kurasian origins and relatic 
ships of T. chinensis and T. ramosissima invasi 
i i pn -nice of ci 

livalcd haplotypes in the invasion, the highly v; 
iable fOOl bp chloroplasl trnS-trnG intergei 
spatter is analyzed using the primers of Hamilt 
(1999). A gene tree, which infers genealogical i 
lationships of DNA sequence haplotypes (allele 
is constructed to represent the popnl v . 
their relationships (see Fig. 2). 

A total of 59 cultivaled. invasive, and native snmi oi T. rhinrnsi.- -p. < mens was colic. 
ed, with 33 samples from llie New World and 
I mi ic • • 1 1 J Y' odd I'l c ill ii ii 's nl mo si >pe. 
mens wore* determined using Baunis (1978) mi 
p'lolngical descriptions and keys. Voucher infi 
mation is listed in Appendix I. 

I . I •■ LI. o| -i -■.] ii.- i lined data set. ' 
(9.3%) of the sites are variable. There are 
(1.2%) single bp changes, three single base ins< 
in is /< < I lis. one 2-bp indel, and 

ent (a fifth base). \ 
most parsimonious gene lice (or minimum spanning 
network) of 22 steps was assembled by hand, rep- 
resenting the fewest imitations explain the re- 
lationships of the specimens (Fig. 2). 

The molecular analysis presents population-level 

represented l>\ a total of sc\cn haplolspos. marked 
A through G. on the gene tree (Fig. 2). The speci- 
men- and their origins are also presented in the 
boxes. The lines ■ ing the <. n. r. e boxes rep- 
small circles represent inferred intermediate hap- 
lotypes that may be extinct, may not have been 

existed if mulatto i m II il i - i I 

steps. Interesting results include the following: 

(1) Of the seven haplotypes found, four are rep- 
resented in the western U.S. Tamarix invasion. 
Haplotype A is very common, representing Id 
(7i!G) of the specimens sampled. The native hap- 
lotype A specimens were collected i i t In licoi hi c 
of Georgia. Iran. Turkmenistan. Kazakstan, Ghina, 
and South Korea. The naturalized U.S. specimens 
were collected from California, east to Texas, north 
to Kansas, and wc-l i<> Was liugloii I i ■ ;>, ;di - a. ad 

nature of this Ii iplel , II i «. ( ■ pi i| 

ii I l-i i- o- I I l a l i I I'M i '■ - J tin n 

revealed population structure or if the lui|>lol\ pe A iutroiis. I plan to continue >. ■< picixiii<i selected Ta- 
plants are genetically similar across Kurasia. II muri.x that exhibit resistance to biological control 
these plants are genetically similar across the na- agents. If they are determined to he genotypically 
tive range, collection of insects from am area of distinct from the susceptible Tamarix, their Eur- 
Eurasia will be equally likely to find control agents asian origins will be provided to the biological con- 
that have evolved with this haplotype. trol exploration project, knowing the number of 
(2) Haplotvpe C is rarer than \. representing haplotypes that comprise a plant invasion, their or- 
only two of the specimens (fig. 2). and was found igins. and the ability of cultivars to contribute to 
once in southern California and once in kazakstan. the invasion arc powerful tools to document and 
This haplotype is only one mutation different from control problematic exotic plant species. 
the common A haplotype. but that mutation is a 
prominent 9 bp indel event that was not found in CONCLUSION 

am other samples. This presents evidence that at . . 

1=.* * ,11 ....... ... .l,~ ;„„o l„„ mu „ U u ,^ ;,. ..... Mnlecidai an, ,ls-- ,mII | l; ,\e an increasing role 

The plant containing this haplotvpe «,aslan f,J) l,l, » l,, - l,;1 ,n,ll,()l Host-switching At the species 
morphologicallv resembled T. rana.sissnna. In a dif- l ^' L ™^"^ systemalu-s will help elucidate' ,,,- 
ferent studv. this haplotvpe was lound to be com- — ' ^^ " , """ 1 '- i "" 1 : "; \^ U f> 

cryptic hybridization events. At the population I 
el. molecular -vsleinatics will allow the unprece- 
dented characterization of invasive taxa as geno- 
types, allowing precise mateliing of biological 
control agents with llicir target-, and elucidating 
links between cultivars and invasions of [.hints. 
Tlie-e advances in understanding plant invasion* 
will enhance control efforts and contribute to the 

specie's. T. parriflora (C, 

Scha. il. in pro*). Tamarix /larri/lara is an invasive 
species with telramcioii- lloral slimline, morpho- 
logically very distinct from the peiitamerous floral 

incongruence between morphology and haplotvpe 
may be' due to hv bridizalion. as was found in the 
genus-wide studv iCaskin <£ Schaal. in press). 

, |) Haplotvpe K was found twice' in \rge-n.ina P™ 1 ^"" 1 "< n «™ l»«liv^ 
invasions by Tamarix. but never in the U.S. This 
haplotvpe is genetically epiite- distinct from the Literature Cited 

common A haplotvpe. differing hv II mutation-. \|'(, i Vngio-periii I'l.ylogeny (,n»ii|>). I 
including two notable 8 and 55 bp inde-l events. classification for llic lamilies ill' lUir 
This genolvpe' has not been found in Eurasia, in- Missouri B,,i Card. 85 531-553. 
(heating that further sampling ol nalive lam, an || H - I nited Stales ; .nd Canada. Hailev; 
populations is needed. . ]<)-<>,. The (; em is Tamarix. is 

(5) All cultivated I .S. specimens of Tamarix Sciences and Humanities. Jei n-.ih m 

contain haplotype C. This haplotype was not re-cov- l^n.therson. J. D. & "-J"'M- '^- J<» 
ered in Eurasian sampling. Haplotype C 
haplotvpe \ by at least 

3: 110-112. 
1989. The Tamarieaceae ol the soiilheas 
tes. J. Arnold Arbor. 70: 403-125. 
-■vents. The haplotype C. as representative Cronquist. A. 1981. An Integrated System of Classil 
■ introgression. was found once, as an in- tion of flowering Plants. Colombia Univ. Press, I 

• Salton Sea in California (Caskit 
■ of thi 

< al.ilit 

DeCoaeh. C. J. & J. I.. Tracv. 1007. The Kneel* ol Hrn- 

70). Even though the presence of this genotv p, ,- ,, ^ , „, ^.^ ,,„„„„„ 

ot common in the invasion, its ability to invade is K.tdangered Species. Biological Assessment Draft, 

ow confirmed. Am presence of ciiltivar haplotypes I SUA Agricultural Research Service. Temple. Texas. 

i invasions should serve as a strong forewarning • K. '• Cam.lhe,-. J. K. I.ovieh. T. L Dudley & S. 

..... ,. . , ■ , I). Smith. 2(KK). Pp. 819-4573 in Y >pe,iee, |<-.lii<.rl. 

i future policy deeasions n-garrhng .1 III- ^ ^ (§| |||( , ,,„,„, ,„„„.,, S , In | )0silim im ,,;,_ 

ual use ol invasive taxa. k)gi( , al ConlI . () | , (f NW(k Montana State University, 

The prece-ding chloiopla-l sequence marker dala P.o/enian. 

How us to begin to delve into the genetic structure Diloina-o. .1. M. I00J1. In.,u.<l. biology, and ecology of 

of the T. ramosissima/T. chinrnsis invasion. This 
prelimin ■ x ill f small samph 

Mates. Weed Techl.ol. 12(2): 320-330. 
..jskin. J. F. c\ H. \. Schaal. 2002. Mole.-ular phvloge- 
in-deplh population analysis is in prepara- nHl( . investigation of I .S. invasive Tamarix. Svsl. Hot. 

ing variable nuclear DNA sequence (in press). 

ui'N ii. \1,.|,-r. i;« ..I. ft: 321-523. 

■ . I - I ivm.1. Iu,i Collins, Colo- 

Illghes. L K. 19W. The Devil's Own Kin in- 
lands 15: 151-155. 

.arlrs/.. J. T. K C. A. Mecham. 1999. Svnlhes.s ol ihc 
North \meriean Mora, wis. 1.0. North Carolina Botan- 
ical Garden, Chapel Hill. 

:«T|)«-z. T. A. K Y S. Smith. 1987. Si.llr.-dar Control lor 
Wm ,i ll.h.lal Improvement in the Southwestern 
I niled Stales. I ,S. l-'isl. and Wildlife Service. Kesource 
!';.!. I , , ,1 I . , ,, I 


liev. I'lntopalhol. 

: ;:: i 

: Willow lealheelles in the eastern 
:. A. Hall, V. I)ovI.--JoI..h-s K H. 
natural Hislon ol Kasl,-rn Cahlor- 

iol.inson. T. W. 1965. Introduction. Spread, an. 
Kxl.-iit ol Sallcedar iTamarix) in the Western 

I >. C.-ologi, al Smvev inolrssional pap.-i I'M- 

Olhce. Washington. D.C. 
{usa.iov. K N. 1949. Sre.lniveaziatskie Tamariks 

k. ill. ! laiiiaiisk- ol C.-ntral Asia.] 
vhooi.hoven. L. T. Jenny <!< J. van Loon. 1998. 

I'l.iiil l!iolog\. Cha|iiiiaii and Hall, New York, 
iollis. I). K. & K. K. Knzoli. 1995. Discordance I 

hut necessarC Trends Keol. 1 \ ol. 13: 32.)-32 ( 
Whitlemore. A. T. & B. A. Sehaal. 1991. Intel 

gene flow in svmpatrie oaks. hoc. Natl. Acad 

88: 2540-2544. 
Wilken, I). H. 1993. Tamaricaceae. P. 1080 i 

Hickman (editor). The Jepson Manual. Univ. C. 

Press, Berkeley. 
Wilson. K. 0. 1997. Strangers in Paradise. Mi 

Washington. D.C. 

IV | Hi, Nature ( ..iw m.iiii \ I 2(M)2 <hltp://tn 

I SDA (United States Department of Agriculture 





•- ii 



I I ! 




1 „ 111111111111111133113111113 





I i lllilllllll 



SCC8E., i|eJJS3|| SSSSSi . .333CC, , 






Tropica] i, 

ippear to be especially vulne 

rable to invasive species as indicat. 

id by the often high 

e hypotheses offered 

- -I- ' i 1 ;;' 1 " 1 ;',,;; 

land- and suggest a simple svnthesi 

pn.pap.le supply. ; 

ies of exotic and island species. This 

review suggests that 

l> two interacting | 

'roceLts-hi'J'n,' 1 ' 

availability and poor ability of nativ 

species. In addition. 

In- ' -i i.mI!'. ! 

ugh rates of introd 

action have provided opporlm 

lily in the form of a diverse and abu 

ndant propagule rain 

rable to naturalizing exotics growing 

•back of the canopy 

ely managed and a 

lien propagule pressure reduced, ihev will be liiglib modilied b\ c 

qianding exotic plant 


o isolated ecosvslellls. \s m< 

unlanil natural areas become Iragm 

ented. degraded and 


islands, including limited habitat art 

ning species diversity, and disturlied habitat-. \ 

our attempts 

to protect both nu 

from the impacts of exotic species. 

Key words 

asive species, invisibility, island eo 

munities, tropieal islands. 

"Hi- whit admits tin- doctrine of the creation of each ||,,. x acquire maux nl the ecological attrililltes of 

■;;; / " / ;;';; ..L.nck including li,„ite<l habitat area, mining 

•,.. N .',„'„„,'.„' lunctioual groups, declining species diversity, and 

linmill) s/,h led tlu-m far more fully and perfecth than disturbed habitats (l.aurance cK Rierregaard. 1997). 

did nature" A better understanding of invasions on islands mav 

— C. Darwin ([1859] 1972: 347-348) ini p rove our attempts to protect both mainland and 

Island ecosystems appear to be especially vul- islan<1 ecosystems from the impacts of exotic spe- 

nerable to invasive species. Reviews cite the high Cles - 

numbers and percentages of exotic species on Here I reexamine hypotheses offered to account 
heavily visited oceanic and continental islands '<>'• the apparently, high invasibility of tropical is- 
(e.g.. Vitousek el al.. 1997). and invasive species lauds and suggesl a simple synthesis based on re- 
often are implicated in species extinctions in island source availability, propagule supply, and relative 
ecosystems (Simberloff. 1995: DWntonio & Dudley. competitive abilities of exotic and island species. 
1995). Tropical islands as well often are character- My focus is on invasions of exotic terrestrial plants 
ized by high alien species densities, unlike tropical into native island ecosystems. In the following dis- 
mainland ecosystems where the incidence of alien cussion. I describe some of the hypotheses pro- 
species is low (Rejmanek. 1996). The apparently posed to account for variation in community inva- 
high invasibility of islands might suggest that is- sibility: however, in developing the synthesis I have 
■ssons for the prevention relied on few assumptions about the presence of 

ecosystems. However, as mainland natural areas actions in equilibrium coiumiiiiilies. Rather. I as- 
become fragmented, degraded, and depauperate. sume that communities are open to the 

1 I am grateful to F. Hughes. S. Cordell. and S. DeWall for comments on an earlier draft of this manuscript and for 
23 K. Kuwili St.. Hilo. Hawaii 96720. U.S.A. 

Ann. Missouri Bot. Card. 90: 119-127. 2003. 

that they are c 

changing with variation in 
ill.' abundance of compet- 
es, ami predators (Hubbell. 

itors, mutualists, < I i ?- 
2001). The composit 

iiiiiiiilii-s seem largely allriliulaUc In differential re- 
sponses of individual species to resource a\;iila!»il 
itv. Iiiihitat conditions, and pest pressures, bevoml 
these indiv iduah-li. patterns. assembly rides for 
plant couimiuulic- are poorly understood and dif- 
ficult to demonstrate (Wilson. !<)<)<)). Thus, I as- 
siiinc that some degree of invnsibilily characlei i/e- 

open to the estahlisliment of new species than olh- 

munities particularly vidnerahle to the estal 
rnent of novel species. Communities with low r 
species diversity, missing functional groups 
harmonic community composition, poorly con 
tive speci<\s. and low pest pressures are se 
pro\ide few harriers to the estahlisliment of i 
land species adapted to ecosystems with h 
pressures from competitors, pred; 
(Kejmanek. 1996: SimherloK. 2000: Mack 

d pe: 

As D'Antonio and Dudley (1995) observed, gen- 
eralizations about the vulnerability of islands |o in 
vasive species often fail to dislinguish anion- ui- 
vasibilitv. opportunity, and impact, anv or all of 
which may contribute to observed patterns: (I) is- 

wilh similar opportunity exotic species may be 
>'■«»''" I k--l> i«> • - il'l -I 'irv. |... i II. - da !•.:' 

i land 

i. >, . 

ir new colonist.- may be greater on islands because 
dands may be exposed more frequently to inputs 
I seeds and olliei piopagules ill in inalnlai I hah 
als: and (.'>) i In unpad of exotic invasions on is- 
ecosystems may be more substan- 



eluding pasture grasses, limbei lives, food crops, 
ornamentals. a\u\ sources of fuel and liber (Mueller- 
Dombois & Fosberg. 1998). In addition, rates of 
iniiiilenlioiial nil rodiiclioiis ,ur olleu high be.-.iu-' 
Islands have lieen i mpoi pi , .u-h.iung slat i. .us 
for transoceanic shipping traffic exposing island 
habitats lo specie- from wide and heteiogeiteoii- Moreover, invasive species have strong im- 
pacts on islands, because the spatial extent of eco- 
systems and the population sizes ..I species are nec- 
essarilv small .nu\ ihus vulnerable iSimbei I. .11, 
1995. 2000: Sakai et al.. 2002). Here I focus on 
the innisibili/y of tropical island ecosystems and 
ask whether community proees-e- on islands make 
lh. in paili. -nlarlv vulnerable to the establishment 
and spread of alien species and. if so, whether 

our understanding ol invasion ecologv. 

1958). The proposition that islands a 

Both opportunity and impact are high on many 

tropical i-laiid- be. an-e islands often lack critical 
sources ol food, forage, and liber, colouisis bom 
early Polynesian \ovagers lo r leni government from island biogeograpliy theory (MacArthur 
& Wilson. 1907). which proposes that two inter- 
acting pioees-e- affect species richness on island-: 

likely to be lower than in mainland habitats of com- 
parable size (MacArthur & Wilson, 1967). 

Island-mainland comparisons of native plant 
species diversity are scant, however. Frequently 
. iled -uppoi'l loi lh< pall. -in of low plant diveisih 
on islands is still Mac \rtliui and \\ ilson's original 
treatise, although often the generalization is regard- 
ed as sufficiently widely recognized as to need no 
documental ion. However, mam laclors affect native 

easily separated (Carlquisl. 1971). Geological age. 

and pioximilv to source pools all affect rates of im- 

.species coexistence (Kicklels K Schluter. 199.?: 
Si.iiberloff. 2000). Farly human settlers on oceanic 
i-lands not only brought new species, but caused 

that native lowland Moras in paili. are difficult 
to reconstruct (Steadman. 1995; Kirch & Hunt. 
1997; Mueller-Dombois & Fosberg, 1998). These 
factors interact lo produce high variation in species 
richness among islands and mainland sites and to 

Weeds in Paradise 


ire effects of isolation and area on 

di ers tv 


ast one study suggests that islands n 

lay not be 

significamlv less rich than mainlanc - i ' s ' •-• .. 

1 1 999; found i <» ii l< i -in e in the species 

- 1! chlC-S 

of is] 

land and mainland native floras from 104 is- 


and mainland sites alter area was taken into 

int. Similarly, syntheses of earlier sti 

1( lies sllg 

g es1 

little difference between species-ar 

ea curves 

for i> 

dand and mainland sites (Rosenzwe 

it;. 199.1). 

and I 

ow regression coefficients in species- 

area plots 

of is 

land plants suggest that correlates < 

>f species 


sity are more varied than can be accc 

..mlcd loi 

by a. 

-ea and isolation alone (Gilbert, 1980). Island 


5 may not always be more species 

poor than 


■ of mainland habitats. 


species diversity is assessed at the 

patch or 


1 level where most ecological interac 

lions lake 


■. low diversity may be a salient aspect of is- 

iiu. i I i ions of species in plots or 

microcosms suggest that more diverse assemblages 
may exploit resources more efficiently and resist 
cut of new species (e.g., Levine & 
D' Antonio, 1999: Levine, 2000: Naeem et al.. 
2000: Tilman et al., 2001: Kennedy, 2002). These 

I i n iil i i ii I. • i i- I 'ini t \perimcntal 

i ' i , ies may b< n solved by con- 

sidering process and scale (l,c\inc Cv l)*-\nlonio. 
1999|. Both net resource availability and species 
richness may increase along a resource supply gra- 
dient if supply increases faster than uptake In the 
/ (Fig. 1). Thus the effect of diversity on 

, d: „el e 

-> stems neveilheless. Uland floras often 
cleri/cd bv high levels of endemism. Lven 
ehini .... ! (nigeneric species 
distinct and restricted to single islands, 
or valleys ((larlqiusl. 19, |; I lias.son. 
1995; Wagner et al., 1999). Although a few species 
rr widespread am: pob iMorpiHc. many I ,:i-.< U gli ■■ 
restricted. non-overla|)ping ranges. For example, 
specialion within II r ui in ' >. > , ■•■lid (55 species. 
(h m < riai « i ) a .■■ ii - ! p i nines and gulch- 
es, appears to have been driven In the isolation 
produced bv the dissected topography of highly 
eroded islands (Carlquist. 1974). In most mainland 
tropical and subtropical forests, in contrast, wilhin- 
sland diversity is high (e.g., Heaney & Proctor. 
p " I) . < . a iciei h ■ 1 dr.. rsity within fam- 
ilies and < o-oi ■ i < oi generic species 
(Croat. 1978: Hartshorn cK Hammel. 1991). Thus 
at a stand level, if not al a regional level, specie- 
diversity is likely to !>■• low r island ecosystems. 

The thesis that species-rich communities are less 
ii \asiiile tha'i spec es- huh e ( i!,iinimities was first 
i< . i 1 1 I 1 I 1 1 in ( I' . ) i i ii- become a cen- 

209(1) II . i hi ii I s.i ip< i oi tins hypoth- 

i. i .i i In .us \| the land- 

scape and regional scales, several recent analyses, 

ten also have the largest number of exotic species. 
\nai\s - if islands ( I ,oiisi i!< I >9«)). riparian eco- 
systems (Levine. 2000). ai d la igclands (Slohlgren 
el al.. 1999) all show strong positive correlations 
between native and exotic species diversity. The 
authors conclude that likely the same factors that 
promote a rich assemblage of native species — 

ej i .1 limiting 

cililaie llie esiahl 4 men! of e\o' 

the context of resource supply and deim 

s it a\ lilabil p< i- - m hm ss pri 

se likely community invasihililv (Shea 
& Chesson. 2002). Lor example, low diversity for- 
ests al Seinliki forest Reserve. I ganda, were found 
to be no more u • a - hit than h gh i iversily stands 
(Kejmanek. 1996). Similarly, observations of Davis 
el al. (1998) showed that competition intensity be- 
tween tree seedln ml icrl on gel nil- i I 
correlated with net resource supply rather than bio- 
mass ni gross resource supply. These studies sug- 
gest two correlates of community invasii h ill: 

new species are most likely to become established 
where limiting resources are available; (2) more re- 
sources are likely to be used or preempted when 
species richness is high than when it is low. I ndcr 
conditions of constant resource supply, high species 
richness may reduce divisibility (e.g.. Levine & 
irAntonio, 1999: Tilman el al.. 2001). On oceanic 
islands, low stand-level diversity likely contributes 

ability, and poor resistance to the establishment of 

new individuals (e.g.. I\i| |9<>,v k iavam, *X 
Itow, 1999). 

In addition to their effects on species diversity, 
long-distance oceanic dispersal and novel environ- 
mental conditions constitute a strong ecologic I! (sl- 
ier on island hiola. \s a result, insular floras are 
often depauperate in certain laxonoinie lineages, 
lunetional groups, life forms, dispersal charactei 

istics. or en\ iron i-nnl ,i/, i. i a s. a pailen, thai 
(",arl(|uist (1971) described as disharmonic. Lor ex- 
ample, rain forests on oceanic islands may lack tall, 
large-seeded, shade tolerant canopy trees that dom- 
inate many mainland rain loresls. In Hawaii, native 
palms arc confined to a single genus (/Vd. ■'•uiniia): 
in contrast, dwarf, understory. climbing, and clonal 
ui'ics ii . emmo, i i .ni la i ; loi. 'sis Similarly. 

Environmental Gradient 

ire 1. Hypothetical p;iii, -ni- 
ce>: K.„ = resour. 

- ■ Minium ,| . Hot- of ;lll co-u 

I groups. K ■■ 

mainland ecosystems, holh of these functional 
groups can have important effects on lores! slrue- 
I ii re and regeneration processes (Devvall et al.. 
2000; Schnitzer et al.. 2(KK): Karris-Lopez et al.. 

The availability of \acant niches arising Iiomi 
this disharmony in the structure of island I. iota is 
cited fre<|iienll\ as providing opportunity for inva- 
sive species (e.g.. Mack et al.. 2(H)2: Shea cK Ches- 
sou. 2002). Anions plants, however, the concept of 
functional -roups mav he a more useful model than 
niches, since unique liahilat and resource require- 


1995; Denslou. 1990). Kor example, \1\rica 
Alton (Myricaceae) is an alien. N-fixing tree 
nvades recent lava flows in Hawaii. Addition 

suggests thai this functional i 

of widespread Vlixing shruhs in the fossil record 
(James. 199.,) indicates that this role once may 
have heen Idled. Indeed the ability of alien invasive 
species to alter nuliiciil siipplv. di-liu Lain •■• re- 
gimes, light environments, productivity, and other 
ecosystem properties mav he taken as evidence of 
unexploitod opportunity m, I nursing or poorly rep- 
resented functional groups. Similarly. Kilavama 
(1990) and Kilayama and How (1990) suggested 
thai low stand productivity and low above-ground 
hiomass in spite of hLh n-oiucc av ailahilitv were 
linked to low species diversity and missing func- 
tional groups in Hawaii and the Lalapagos. How- 
ever, a deterrent effect of native species on the es- 
tablishment of alien species with similar habitat 

...av be difficult to demonstrate. In Hawaii, for ex- 
ample, much of the diversity among woodv plants 
resides in uiiderstory shrubs, yet this group of 
plants also is well represented in the naturalized 
exotic flora. 

High levels of apparent functional redundancy 
anion:: plants in ma i u lain I ram !ore-N -u^jc-l that 
there may be few barriers to 
many species playing similar I 

example, about 1700 species of v 
have been recorded in 1536 ha at 
Biological Station. Costa Rica, of w 
44 species of Piper (Piperaceae). 39 species offsv- 
rholria (Kubiaceae). and 25 species of VI icon ia 
(Melaslomataceae) (Wilbur. 1994), almost all of 
which are imderslorv shrubs and small trees. There 
are 107 species of lianas and 325 species of trees 
(Hartshorn & Hammel. 1994; Wilbur. 1994). Dif- 
ferences among these species, lor example, in light 
i ■ ■. |u i t.-mi-i ilr-. mouth form, or climbing mecha- 
nisms, are small, suggesting strong overlap among 
them in requirements. In addition, studies 
on ihe 50-ha permanent plot in Panama fail to re- 
veal significant hahilal differences among the ma- 
jority of (..existing trees and shrubs lliere (Hubbell 
et ah. 1999: Harms et ah. 2001). These patterns in 
relatively homogeneous forest- >uggest that the 
presence of many ecologically similar species is not 
a strong deterrent to the occurrence and persistence 
of planl species in mainland tropical loresl. In ihis 
context, il also seems unlikely that the presence of 
native species would be a significant harrier to the 
estahlishineiil of ecological l\ similar alien species 

Native species on islands often appear to be poor 
competitors (Darwin. [1859] 1972: Carlquist. 
197 1|. for example, wet and mesic forests through- 
out the Hawaiian archipelago are ilominaleil by a 
single highly polymorphic species. Mclrosidcros po- 
Ixmorpha Gaud. ( Myrlaeeae). These Mclrosidcros 
forests are characterized by relatively open cano- 
pies, widely spaced crowns, ami inefficient light ab- 
sorption (Cordell K Goldstein. 1999). As a result, 
considerable light reaches the loresl understorv 
(Cordell & Goldstein. 1999). where native and alien 
grasses, herbs. 

dispersal mechanisms and habitat requirements. 
the alien R. clliplicus exhibits hisler giowlh rales. 
canopy configuration, and greater 
seed production increasing the likelihood that il 
eventually will replace the native species in these 
forests (Denslovv, unpublished data). 

The reasons why island species should be poor 
competitors are various.. I .oss of resilience ill the 
gene pool may be a consequence of loiiudei .Heel-. 
inbreeding depression, small population sizes, 
strong post-establishment selection, drift, and low 
validity ol pollinators and dispersers (( ,arl(|uist. 
1971: Looped Mm llei-Doinbois. |<)!!<); kaneshiro. 
1995). The steep environmental gradients over 
short distances thai characterize high tropical is- 
lands may constrain development of specialized ad- 
aptations, for example, Kilayama (cited in Den- 
slovv. 2001) has suggested that the dominance of 
oceanic island forests by species with wide ecolog- 
ical ranges may he linked to then low product iv ilv 

ranges th, 

in did species 

in Hawaii. 

1 le suggested. 


that the borne 

an species may be better 

adapted I 

o their environments am 

1 thus present 

ad !ili 


and 1/. po/vmorpha shows little plasticity in re- 
sponse to increases in light or nutrient supply (Cor- 
dell et ah. 2001: Austin & Vitousek. 2000). \\ here 

pared, lower rates are often measured in native Ha- 
waiian than in comparable alien species il'mi-on 
et ah. 1998: Baruch & Goldstein. 1999: Durand & 
Goldstein. 2001). Similarly, the exotic Himalayan 
raspberry. Kubus clliplicus Sin., is replacing the na- 
tive "akala. Hulius litiiniiiensis \. Cray, in the tree- 
fall gaps both require loi establishment in montane 
rain forest. Although the two species have similar 

stronger barriers to alien species than the more 
broadly adapted species in Hawaii. The links 
among adaptive ranges, competitive ability, and re- 
dds regard. A large proportion ol island endemics 
are threatened or endangered dm- in part to their 
small ranges (Simberloff. 2000: Sakai el ah. 2002), 
and the positive correlation between range size and 
local abundance has been widely acknowledge! 
(e.g.. Brown £ Maurer. 198k Kelly. 1990). These 
studies -nggesl that on islands, depauperate and 
di-harinome Moras and poorly competitive species 
mav result in low productivity and high ecosystem 

Loss of dispersal efficiency in island species may 
conlribule to their poor competitive abilities. The 
generality of this pattern in many unrelated plant 
groups (Carlquist. 1974) suggests thai selective 
pressure to avoid loss of reproductive output and 
reduce investment in dispersal mechanisms is 
strong on islands. Compounding the evolutionary 
los- ol disper-abilitv has |„.,'ii the historic los> ol 
avian frugivores in Hawaii (James. 1995). loss of 
pollinators, and lack of a persistent seed hank in 
many native species (Drake. I')"!,",), all of which 
contribute to low seed availability. \s a result, pop- 
ulation growth of native species may be strongly 

habitats than aliens with better dispersal meeha- Kosberg, 1998; Hiehardson et al.. 2000). If islands 

nisms, greater seed output, and larger seedbanks. are depauperate in natural enemies or dominated 

ephemeral, establishment sites such as treefall gaps eies may be under little control by their natural 

and nurse logs. Dispersal limitation contributes to enemies. The net effect of low pest loads on native 

less-than-full exploitation of limiting resources in and alien plant species on islands is then difficult 

all ecosystems (llubbell. 2001). but may be partic- to predict. We lack critical information on the role 

ulaily uupoitant on islands. of natural enemies in limiting plant population 

growth, abundance, and distribution in both inam- 

EFFECTS OF LOW PEST LOADS land and island habitats. On balance, however, the 

more rapid growth rales and plastic physiological 
The enemy release hypothesis proposes that responses of alien species may result in their great- 
alien plant species in their introduced ranges ex- er population growth where pest pressures are low 
perience lower pest loads than co-occurring native (Keane K Craw lev. 2002). 
species and than they do in their native ranges 
(Keane & Crawley. 2002). As a result, populations Sn Mill sis 
of exotic species may be released from control by 

natural enemies, enhancing their competitive po- Net resource availability is an important eom- 

sitions relative to native species, which may remain ponent ol community mvasihililv. I he role of re- 

uiider pressure from specialist and gencralist cue- source pulses in the spread of alien species has 

mies. There have been few appropriate tests of this been noted widely (Vitousek et al.. 1997: D'Antonio 

hypothesis and in a few cases only were significant et al.. 1999; Mack et al.. 2002). The generality of 

impacts of either gencralist or specialist hcrbiyorcs this relationship can be extended usefully to iu- 

Crawley, 2002). Classical biological control— the ability, such as that provided by disturbances, and 
introduction of specialist pests and pathogens to to affect native as well as alien species. Seedlings 
control invasive species in their introduced rang- of both native and alien species are likely to be- 
es—is based on this premise. There are several ex- come established where resources arc chronically 
amples of intentional and accidental introductions or temporarily under-used, as following natural or 
of pests limiting I he populations of their hosts (I. on- anthropogenic disturbance; under conditions of nal- 
da et al., 1997: Strong & iV.nberton. 2000). and ..rally high levels of resource supply; where re- 
Louda (1982) and DeWalt et al. (unpublished ins.) sources are incompletely exploited by the existing 

itat distributions of weedy plants. In addition, data augmented. For example, it is widely recognized 

presented by DeWalt el al. (unpublished ins.) sag- that both natural and anthropogenic disturbances 

gest that survival of Clidemia hirta (I..) D. Don (Me- facilitate the establishment ol alien species (Kej- 

lastomataceae) is more strongly affected by insect matiek, 1989; I lory ilz el al.. 1998; D'Antonio et al., 

and pathogen impacts in its native than in its in- 1999). in part because disturbance reduces eom- 

Iroduced range. The role of pest and pathogen pres- petition and increases local resource a\ ailabilitv . In 

s ' gaps, and riparian habitats arc important sites for 

Furthermore, indigenous island species, which seedling establishment (I'latt, 1975; Denslow, 
also originated as waifs, are thought to be under 1087: Levi. ic. 2000). Similarly, where disturbances 
low pressure from natural enemies as well (Carl- are ecologically novel or unusually frequent or in- 
quisl. 1980). In the process of dispersal to new and tense, wo can expect that some native species as 
isolated habitats, insular lloras are thought to ha\e well as some exotic species may lind suitable en- 
left their specialized pests behind. While subse- vironmenlal conditions. Diversity and abundance of 
quent diversification has produced many special- both alien and native species also are often highest 
ized associations ol plants and their natural ene- in resource-rich habitats (e.g.. Stohlgren et al., 
mies (Swezey, 1954), the common lack of defensive 1999). am" 

compounds and slrucliiies among indigenous island bv resource limitation (in Hawaii, 

species suggests that plant pathogen and herbivore alpine ecosystems, 

impacts are not high. As a result, island plants may flows, and dryland forests) 

be particularly vulnerable to the introduction of ex- scarcity of exotic invasives. 

otic pests such as ungulates ( Miieller-Dombois t\ posed to increase coininunil 

Volume 90, Number 

as special examples of a more general pin 

In isolated tropical island habitats, resource 

availabililv likely is high because indigenous spe- 
cies do not effectively use or preempt them. Low 
local species richness, low diversity of functional 
groups, and low redundancy within functional 

groups suggest llial few species mas be a\ailable 
lo lake ad\anlagi i h i n opporluuil ie- ««r 

istie of all plant communities. nia\ be purliculuily 
strong on islands, reducing the occupancy rate of 
suitable habitats and leaving sites and resources 
available for others. Where propagule pressures are 
high, the vulnerability of island communities to es- 
tablishment ol exotic species will be particularly 
apparent (Rejmanek, 1989; Drake. 1998; Levine. 

of habitat and species. Limiting resources are avail- 
able relative to the physiological and morphological 
capacity of plants to exploit them. Low growth plas- 
ticity in response to resource availability in many 
native species contributes to the underulili/ation of 
resources and thus invasibility of some island com- 
munities, and exotic species with more plastic 
growth responses may compete more effectively for 
natural establishment sites such as tree fa 1 1 gaps 
and decaying logs and take belter advantage of nov- 
el disturbances such as pig foraging sites. 

\ssiuning that fundamental tradeoff- between 
rapid growth rate in high light environments and 
persistence in shaded environmc 
alien and native species (Reich el al.. 1997 
bell, 2001), both alien and native species 
be able to thrive across a full spectrum of e 

, of disturbance-adapted specie 
rapid growth in resource-rich environments a 
pious production of well-dispersed seeds): however. 
in part this pattern may reflect historic pathways of 
introduction associated with development ol agrr 
culture and rangelands (Mack, 1992). rather thai 
inherent characteristics of invasive species. A ris 

m ■ .1, il 1 , in, i I ! i ill rade is ii In. nig spe< a- 

that are shade-tolerant, spread vegelalively. and/oi 
have large seeds (Reichard & Hamilton, 1997) and 
thus are potentially invasive in intact forest eco- 

This review suggests that fundamentally two in- 
teracting processes — high net resource availability 
and poor ability of native species to preempt those 

ulnerable to 

the establishment and spread ol alien species. His- 
torically high rates ,,| introduction have provided 
opportunity in the form of a diverse and abundant 
propagule rain of exotic species. The combination 
produces a scenario that is not an optimistic ..tie 
for island ecosystems. It suggests that these native 
ecosystems on islands are particularly vulnerable 
to naturalizing exotics growing on their borders, and 
that while disturbance from a variety of causes, in- 
cluding pigs. fire, grazing, and natural dieback ..( 
the canopy dominants, increases the opportunities 
for exotic incursions, even intact forests are not im- 
mune, f nless these forests are aggressively man- 
aged and alien propagule pressure reduced, they 
will be highly modified by new exotic introductions. 
There are implications of this scenario for main- 
land ecosystems as well. With rising human pop- 
Ho wilderness are- 
change, mainland 
preserves will assume many island-like character- 
istics. Habitat fragmentation and large-scale habitat 
loss will increase habitat isolation, reduce the ex- 
tent, and increase the edges. I i.ignientation of for- 
ests is widely associated with increased distur- 
bance, alteration of edge and < 

(e.g.. Laurance K Rici n -guard. I<)97). Kffects of 
population reductions and species extinctions will 
simplify ecosystem structure, reduce population 
sizes, produce disharmonies in species assemblag- 
es, and reduce functional redundancy. I .oss of high- 
er trophic levels, in particular, may result in in- 
creased populations of herbivores and edge species. 
Increased intensities and rates of exploitation will 

free res ves (light, space, nutrients) and increase 

oppoitunilies. lor exploitation bv novel species. Ris- 
ing rales of human traffic through suburban expan- 

population movements will provide rising and re- 
pealed exposure to a diverse seed rain from exotic 
species. Tropical islands are an effective early 
warning system of the impacts successive waves of 

systems. As island managers develop > 
preventing and controlling these inv, 
will be watched by many on the mainl 

Literature Cited 

Austin. A. T. & P. M. Vitousek. 

nposability of Metrosideros 

Brown. J. II. & B. Maurer. 1984. 

between aJ>iiiiilaii<-t- and distribuli 

Naturalist 121: 255-279. 
BurK.ii. I'. J. 1982. The effect of t- 

on Mclrosidcro.s polymorph! seed germination. Pac 

Sci. 36: 229-240. 
& I). Vlueller-Dombois. 1984. Response of .1 

Irositlrros pohrnorplia seedlings to expei imenlal cam 

opening. Ecology 65: 779-791. 
Carlquist, S. 1971. Island Biology. Columbia Univ. Pre 

New York. 

ieal Botanical Cardei 

sitlrros pohniorpha I 

r& P. M. Vilousek. 2001. 

)ii in uutrieul-limilcd Urlmsidcms pohinorphn 
ees in Hawaii. Tree Phvsiol. 21: 13-50. 

1978. flora of Barro Colorado Island. Slan- 
. Press. Stanford. California. 
C. M. & T. L. Dudley. I'W.S. Biological in- 

lands. |>p. 103-121 in P. VI. Vilousek. I.. L. Pooped 

Kcosvstem Function. Springer. Stanford. 
. K M. Vlack. I 999. DiMurbanee and bi- 
ological invasions: Direct effects and feedbacks. Pp. 
113-152 in I.. B. Walker (editor). Ecosystems of Dis- 

Harms. K. K.. B. Condit. S. P. Ilubl.ell K B. B. foster. 
2(H) 1 . Habitat associations of trees and shrubs in a 50- 
ha Neotropical forest plot. J. Kcol. 89: 917-959. 

Hartsliorn. (;. S. X B. K. I lamm.-l. 1991. Vegetation tv pes 
and floristic patterns. Pp. 7.5-89 in P. A. McDade. K. 
S. Bawa. II. V. Ilespcnhcidc X <;. S. Ihulslmru (edi- 
tors). Pa Selva: Kcology and Natural History of a Neo- 
tropical Bain forest. I niv. Chicago Press. Chicago. 

Heaney. A. & J. Proctor. 1990. Preliminary studies on 
forest structure and florist ics on Volean Barva, Costa 
Bica. .1. Trop. Kcol. 0: 307-319. 

Iloml/. C. C. J. II. Pascarella. S. VlcVlann. V. Kreedman 
X B. II. NoMellrr. 199!!. fun.llonal roles of ,nv as,v e 
non-indigenous plants in hurricane-affected subtropical 
hardwood forests, f',,,,1. Vpplic. 8: 947-974. 

Hubbell. S. P. 2(M)I. The I nihed Neutral Then ol Bio- 
diversity and Biogeography. Princeton Univ. Press. 
Princeton. New Jersey. 

. D. B. Foster, "S. T. (KBrien. K. K. Harms. B. 

<;o..dil. B. Wechsler. S. J. Wright & S. Poo de Pao. 
1999. Light-gap disturbances, recruitment limitation. 
and tree diversity in a Neotropical forest. Science 283: 

ames, H. K L995. Prehi 

changes eanic isla 

sek. L. L. Loope & H. 
ological Diversity and 

■ mU. Pp. 87- 

is. M. \.. K. J. Wrage & P. B. Reich. 

letition between tree seedlings and herb 
ition: Support for a theory of resource su 
nmd. J. Kcol. 80: 652-601. 
islow. J. S. 1987. Tropical rainforest g, 
pecies diversity. Annual Bey. Kcol. Sysl. 

. 1990. Functional group diversity a 

. disturbance. Pp. 127-152 in G. H. Orh 
, J. II. Cushman (editors). Biodiversity ar 
n Tropical Forests. Springer. N( 

-. 2001. 
. Kvol. 

wland forest. J. Trop. Kcol. 

). B. 1998. Belalion>l,,|,- among the ~eed ,;,,,,. 
iank and vegetation of a Hawaiian forest. J. Veg. 
: 103-112. 
L. Z. X C. Goldstein. 2001. Photosynthesis. 

ivasive tree ferns in Hawaii. Oecolugia 126: 3 15 

i Function, i 

. C. S. 1958. The Kcology of Invasions of Vnin. 
. London. Kngland. 

Biological Diversity 

Kennedy, T. A.. S. Naeem 

the Pacific Islands. Yale Univ. Pre* 
Kilayama. K. 1996. Patterns of spe 

on species diversification. J. Veg. !• 

. H. 1995. Effects of div< 

. Pp. 177-186 in P. 

nutrient pools of a Scalesin pcdiiiiculala montane forest 
on Santa Cruz. Galapagos. Kcol. Bes. I I: 105-108. 
:uirance. W. I.. Jr. v \ It. ( ». Bicregaanl. 1997. Tropical 
forest Remnants. I niv. Chicago Press. Chicago. 

M. 2(KK). Species diversity and biological in- 

Science 288: 852-8,5 I. 
& C. \1. If Antonio. 1 999. Klton revisited: A re- 
view of evidence linking diversity and invasibililv. ()i- 
kos 87: 15-26. 

Lonsdale. W. VI. 1999. Clohal patterns ..I plant in\a-i..n- 
and the concept of invasibililv. Kcologv 80: 1522-1530. 
I "'. I . I . i\ I). \luell.-r-Dombois. 1989. Characlcris- 

Volume 90, Number 1 

Denslow 127 
Weeds in Paradise 

waii. Pp. 257-280 in J. A. Drake. K. DiCastri. R. H. 
Crows. F. .1. Kruger. H. A. Mooney. M. lirjmanrk X 
\l. II. \\ illiatn-on ledilors). Biological Invasions: \ 

Fonda. S. M. I««!2. Disl.'ibii i ecology: Variation in 

seed predalion. Feol. Monogr. 52: 25-11. 
. 1). Kendall. .1. Connor y\ 1). Simberloff. 1997. 

logical conlrol ..I weed-. Sic, e 277: 1088 

X M. Rejmanek. 2000. Plant invasions— The role of 
mutualisms. Riol. Re\. Cambridge Phiios. Soe. ,5: 0.5- 

Kieklefs. R. F. & D. Shinier. I993. Species Diversity 
in Ecological Coinmiiuilie-. I m\. Chicago Press. Chi- 
li .-■■■/«. i_ M I I 995 s |)r( i r> | > i v * - r - i I > in Space and 

Time. Cambridge 1 niv. Press. Cambridge. 
Sakai. \. K.. V\. I.. Wagner ci I.. \. Mehrhoff. 2i«»2. Pal- 
terns of eiidangermenl in the Hawaiian flora. >v-t. Pool. 

1 Park Resources 

i G. Orians. R. Din 

•rs). biodiversity and Fcos 

51: 276-302. 
■chnitzer. S. A.. J. \\. Hailing X \\. P. Carson. 2(KM). The 

cmopv gap-: Evidence lor an allernati\e pathway of 

•hea. K. & P. Chesson. 2(H)2. Community ecology lh.-,.n 

a- .i Ii work for biological invasions. Trend-. F.-.T 

Evol. 17: 170-176. 

!•><).). \\ by do iulroduced -pene- appeal 

I pideniinlogv. (.lubal ( a ,n-e(|uence- and Conlrol. Eco- 
logical Society of America No. 5: Issues in Ecology. 

. I) Sinbeiloff. \\. \l. Lonsdale. II. Evans X M. 

Clout. 2002 Biolic invasion-: ( jinx . pid. mi..l.. -y 
Ecol. \pplic. 10: 

to devastate 

islands in 

ore than mainland areas? Pacific 

Sci. 49: 87-97. 

. 2000 

. EMillcli. 

nl implications. The Raffle- Hull. 

Z....I. 1!',: 1 

Sleadma,,. D. 

\\. 1095. 

Prehistoric extinctions of Pacific 

Eland bird- 

.: Biodivei 

sitv meets zooarehaeologv. Sci- 

enoe 267: 1 


Nohlgreu. T. .1 

.. 1). Binklev.G. W.Chong. M. \. Kalkhan. 

Reich. P. IF. M. IF Wallers & D. 

From tropics to tundra: Global convergence in plant 
innolmnnm. Pro.'. Natl. Acad. Sci. I .S.\. 91: 13730- 


\. Drake, F. DiCastri. R. I 
F. J. Kruger. H. \. Mooney. M. Rejmanek X 

Perspective. Wiley. New York. 

. 1996. Species richness and resistance to 

sions. Pp. — 

. Otsi 

Ra-hkm X V Son. 1009. Exotic plant -pecies invade 
hoi -pol- of native plant diver-itv. E.col. Moiio-m. (■'': 

D. R. & R. W. Pemberton. 2(KX). Biological con- 
lrol of invading species — Risk and reform. Science 288: 
ezev, 0. 1954. Forest Entomology in Hawaii. Bishop 

Special Publ. 44: 1-266. 

D., P. B. Reich, J. Knops, D. Wedin, T. Mielke & 
C. Fehman. 2001. Divereitj and productivity in a long- 
term grassland experiment. Science 291: 813-815. 
lousek. P. \F. E. IF Walker.!.. I). W hlteaker. I). Mu. II- 
er-|)omboi- i\ P. V. Mat-on. I * >J { . . Biological inva.-ion 
bv \l\rini fino allei- oeo-v-tem development in lla- 

, C. M. D 1 Antor 

rl.Weslhn.oks. 1997 

. Ecol. 21: I 16. 

Foope, M. Rejmanek & 

■caused global change. New Xea- 
ner. 1090. Man- 

Hawaii Press. 

Vilbur. R. F. 1991. \ascular plants: \n inh-rim che.-k- 
1,-1. Pp. 350 378 //, E. \. McDade. K. S. Bavva. IF \. 
Hespenheide v\ G. S. Hartshorn leditor-l. Fa Siva: 
Ecology and Natural lli-torv of a Neotropical Rain Eor- 

\il-on..|. IF \<><><). \— mbb nil.- in plant communities. 
Pp. 130-165 ill E. Weihe, y\ P. Kedd\ (editors!. Eoo- 
logical \s-emhlv Rules: Perspectives. Advances. Re- 
in al-. ( ainbiidge I niv. Press. Cambridge. 


1753 00296 5579 


iTOR has recently jimoiinccu tin release of its Ec 

ore than 1 million pages of iinpc 
The Ecology & Botany Collec 

of the 

2003 * 

Volume 90 
Number 2 

Volume 90, Number 2 
Spring 2003 

The Annals, i.ul.l 

Bed, independent re 


ere. \ 

uthors should write 1 

he Ma 

naming K. 



tioii concerning am 

u-k (» 

7w\ { 

'or publishing in the 
ist issue of each voir 


.S. Instru 
d are alsc 

ctions to 




I ditoiial (oinmiltee 



In ftofaro 

■a( Ga\ 




icheuler McPhei 
png fiitoor, 



'//c AV//7r,/- 

cal Ga 


//7 BotanU 



UIU, i 


Press Assis 
iri Botani 

-til Ga 



ra Mack 



\. \1-She 
iri Botani 

■<d <;,t 


Missouri Botanical Garden 
Roy E. Gereau 
Missouri Botanical Garden 
Peter Goldblatt 
Missouri Botanical Garden 
Gordon McPhereon 
Missouri Botanical Garden 
P. Miek Richardson 
Missouri Botanical Garden 
Charlotte Taylor 
Missouri Botanical Garden 
Henk vanderWerff 
Missouri Botanical Garden 


Volume 90 
Number 2 

of the 



Michael J. Zanis, 2 Pamela S 
Yin Long Qiu, 4 Elizabeth Zir 
and Douglas E. Soli is" -' 

- |>m\idr<l pool resolution a 

Using a comp n nil i 11 | I 'n I .1 j li lo;>, n. 

n-niL' sei|iieuces Iron, >i\ genes |o\er 12.000 l>|i per laxont I mm all I 
l,",s'r|)\\. 2oS il)\\: mitochondrial- -m„l\i. <//,, \ I. Trees icsiillmg 
of the compartmentalized dala arc identical. We find strong support 
aiitihi-|.nni-: \ mlmicl lacae. \\ mphaeaeeae. and an \ustrobaileyale* 
\u-liuknlrwn i-ii.-i. Wli.iva- mn-t recent studies using multiple gen 
tionships among the remaining basal angiosperms (Ceratophy llaeeae 
perales. monocols. Vlaguoliales. I.auralesl. with comparlmentalization, we find liigh levels (> 007r ) n 
lor relalionships among these chides. Canellales and I'iperales form a strongly supported (100<7r) s 
in turn, sister to a well-supported I I (II )' , l cladc ol Liurales and Vlaguoliales. Canellales + I'iperah 
+ Uaurales form a well-supported magnolia! chide. Ceratophy llaeeae are strongly sup|)orted (KK)' 
monocols; the monocol/Ceralophs llaeeae chide is well supported {iW/t) as sislei to all remaining a 
anthaceae. the magnoliid chide, and eudicoisl. Tin addition ol entire 20S il>\ \ sequences clearly 
increased internal support. We examined the diversification of perianth phy llola\i-. n 
us. Ancestral perianth phvllolaxis and men 
strohailevales grade; however, an undiflere 

<imum likelihood analyst 

lifferentiation using 
vocal for each node 

ith is reconstructed 

Fulbright Distinguished IY< 

Kuzoff. and Shelley McMah 
ice with sequencing and pi 
iiingbae Suh and Sangtae h 

Iment of Kcologv and Kvolutionary biology. University of Michigan. Ann Arbor. Michigan 48109. U.S.A. 

itor\ of Molecular S\ sternal ic~. Smilhsoiiian luslilulion. Washington. l!C. 20.V>0. U.S.A. 

Iment of botany and' the (,enelics Inslitule. University of Honda. Uamesv ill,-. Florida 3201 1. U.S.A. 

Ann. Missouri Bot. Gard. 90: 129-150. 2003. 

l Amhorella. \vtii|iliaeaeear. and Austi 
V'.v words: Itasul angiosperins. floral e 

Understanding phylogenetic relationships among 
extant basal ;i i iiii* isjhthis is critical loi reconstruct- 
ing the evolution of numerous trails, including 
wood anatomy, chromosome numbers, pollen mor- 
phology, .in.l Mural morphology I Undress. 1980. 
1991a: Undress cN Hufford. 1989; Undress cN Ig- 
ersheim. 1997: Igersheim & Undress. 1997; Qiu et 
al.. 2(MK): P. Soltis el al.. 1999a; Thome, 1992: 
Walker & Walker. 1984). In recent classifications, 
most basal angiosperms have been recognized a- 
siibcla — Maunohidac |( aoriqui-l. I ' »."> I : I aklil.i |an. 
1969. 1997; Thome, 1992). However, these groups 
haw l\picall\ formed a grade (Donoghue iK Dov |e. 
1989: Doyle." 1091; Nandi et al.. 1998; Savolainen 
et al.. 2000; I). Soltis et al.. 2(M)0; P. Soltis et ah. 
1999a). and only rarely a clade (Chase el al.. 1993; 
Qiu et al.. 1993). in phylogenetic analyses of an- 

Previous analyses of relationships among basal 

angio-peniis ba-cd on morphology have geneiallv 
provided conflicting results regarding the early 
blanches ol angiospenn phvlogenv. I 'hv logeuel i. 
aualv -e- <>| morphological data (l)onogluic X Dos le. 
|«)j:<)| suggested dial Magnoliales were sister to all 
oilier augiospeim-. although in trees only one step 
longer Ny inphaeaeeae appeared in tin- position. In 
that same study, paleoherbs (Piperales. Nymphae- 
aceae. and monocols) lormed a .lade Chloraiilha- 
ceae and Austrobaileyaceae were nested within 

baileyaeeae. In contrast. I.ocoiile and Stevenson's 
I I"" » h |T\ logen.-lie aualv --is. ..I inurphological data 
placed ( 'alvcanlliaceae as sister to all remaining 
angiosperms: as in I glim- and l)ovle ( P>8')). 

naly/e phy- 
logenetic relationships among basal augiospci m- 
using both RNA and DNA sequence data. Phylo- 
genetic aualv -e- u-iug 18 s - i 1 1 \ \ sc(|iieiices (I lam- 
bs Ox /tiiunei. I ( )')2l suggested that N v inphaeaeeae. 
Piperales. and monocols v 
groups to all remaining angiosperms. In 
with I he broad aualv si- of '.">()( I ihr\ sequence- I roll I 
seed plants (Chase et al.. 1993). Qui el al.'s (1993) 
analyses ol rhi\. sequence- focusing on basal an- 

giosperins suggested that Ceralophy llaceae were 
sister to all other angiospenn-. with the remaining 
ba-al aiigio-peim- lorming a clade sister to the eu- 
dicots. ( v )iu et al. (1993) also recovered four major 
lineage- wilhin tin- large ba-al augio-pcim « lade: 
Magnoliales. I.aurale-. Piperales. and Nymphae- 
aceae. Combined analyses using morpho|.,gical and 
rRNA se<|iience data (Doyle et ah. 1991) found 
Nv inphaeaeeae. monocols. Piperales. and \n-lolo- 

est of the Magnoliidae formed a 
clade sister lo the eudieols. \nalvses of complete 
18S rDNA sequences with greater taxon sampling 
than employed by llamby and Zimmer (1992) 
placed Anihorellaceae. \uslrobailoy aceae. Illici- 
aeeae. Schisandraeeae. ,\\n\ Nvmphaeaceae as the 
earliest-branching angiosperms (|). Soltis el ah. 
1997). The analysis of a large t,lp\\ data set (Sa- 
volainen el ah. 2000) also placed Amborellaceae, 
Nvmphaeaceae. and \u-lrohniley ales a- the earli- 
esl-bi. inching angio-pei in-. In all ol these analyses, 
internal support ia- mea-iired by bootstrap value-) 
at deep levels within the phylogenies was low (less 
than 50%). In addition, some of the above studies 
may have experienced long-branch attractions of 
highly divergent ta\a or chides. 

By combining data from multiple genes and, in 
some cases, sampling numerous la\a. recent studies 
have greatly enhanced our understanding of rela- 
tionships aiming basal angiosperms. Recent studies 
not only agree on the same groups ol early-branch- 
ing angiosperms. but also provide strong internal 
support for niauv lelalionships (Craham <N Olm- 
slead, 2000: Mathews K Donoghue. 1999. 2000: 
ah. 1999; Oiu et ah, 1999, 2000; D. 
2000: P. Soltis el al.. 1999a). These 
- used sequences from a variety of 
genes representing the ehloroplasl. mitochondrial, 
and nuclear genomes. In addition, different ap- 
proaches were employed, including parsimony. 
maximum likelihood, and a duplicate gene rooting 
strategy (Mathews, \. Donoghue. 1009. 2000). These 
analyses reveal that \mborellaeeae. Nv inphaeaeeae 
(comprising Cabombaeeae and Ny inphaeaeeae 
s. sir., the two families constituting N v niphaeales 
following Ues el ah. 1999). and a clade of lllici- 
aeeae. Sch i-andi aceae. \n-l roba i lev aceae. and Tri- 
ineniacae (hereafter referred to as \ustiobaileyales) 

to the rest ol ll low ever, despite 

gencralk strong -i, ill l< bis braced n irdei 
an anaksis emplov ing I! \S \ 1 1 Aniis-Weiler <l al.. 
1996) indicated that Nymphaeaeeae + Amborel- 
laeeae mav form a chide sister to all other angio- 
I ii I Iciiki i ii ' i 'Hi 1 1 contrast, analy- 
ses of a large iniilligene data set found strong 
I Ml | l< > • - -I. I lo ill oilier 

angin-perius Using li.ilh maximum parsimonv and 
maximum likelihood methods, and a hypothesis lest 
using parsimony rejected the Amborella \ \yni- 


lira 'ii'. - e.| 

V. aiphai 


ree, relationships 

among the n i i i i - i i- em nn un- 
certain. As reviewed aho\e. lite nana initio hasal an- 
giospernis forma number of verv strongh supported 
clades: Piperales. \\ inleraceae/Canellaceae (Canel- 
lah's: referred to a- \\ interales bv -omc). Magno- 
liales. monocols. I.auralo. and Odlorantliaceae (or- 
dinal composition sensu A PC II. 2003). 
Uclationslnp> among thee clades were uncertain 
in main previous studies. Imt are hecoining clearer 
(e.g.. Qiu et al.. 1999. 2000; Zanis el al.. 2002): a 
magnoliid clade comprises Magnoliales + Laurales 
ami Canellales + Piperales. although the positions 

magunliid clade are not well supported. The place- 
ment of ( aaatopln llaeeae has varied among studies. 
Ceratophyllaceae were sister to all other angio- 
sperms in anakses based on rh< I , sequences |( idase 
et al., 1993; Qiu et al., 1993: Savolainen et al., 
2000): however, in the three-gene. 567-taxon anal- 
ysis (D. Soltis et al.. 2(XX); P. Soltis et al., 1999a), 
< .. -ralophv llaeeae appeared as sister to the eudicots 
twilh 53 r /r jackknife support). Other studies have 
placed Ceratophyllaceae closer lo the monocols 
(Qiu et al.. 1999, 2000; Savolainen et al.. 2000; 
Zanis et al., 2002) or as sister to Winteraceae (Par- 
kinson et al., 1999). 

A clear understanding of relationships among 
hasal angiospernis has obvious major implications 
for interpreting the morphologv of the early angio- 
sperins and siil)sc(|uent patterns ol Moral evolution, 
hark lipoidoses proposed lliat llie lirsl angio- 
sperms had large. /V/agno/i'a-like. flowers (Arber & 
Parkin, 1907: Bessey. 1897. 1915; Cronquist, 
1981, 1988; Takhtajan. 1969. 1997). However, 
Stebbins (1974) stressed that the concept of the 
earliest flower as large, slrohiloid. and )hiL'nuliu- 

• iheii thai oi « i;i> w ill i ' a • \1. 1', noliaceac. Nob- 
l.ins (1971) proposed that the earliest flowers were 
moderate in size. I sing a combination of informa- 
tion from doth extant .u\>\ fossil Ylagnoliidao. I'ai- 
ilress (1987c) suggested lliat the earliest angio- 
s[>erm was bisexual, but that the transition to 
iinisexualitv was relalivoh easv. the perianth was 
undifferentiated and could be easily lost, and the 
number of floral parts was labile. 

Early phylogenetie studies focused attention on 
paleoherbs (Nymphaeaceae. Pipcraceac. ami Cdlor- 
anlhaceae) as possible lirsl-brancliing cvlanl angio- 
sperms (Donoghue & Doyle. 1989; Doyle et ah. 
1994; Hamby & Zimmer, 1992). suggesting thai 
.■ail\ Mowers were small, wild a Irimeroiis perianld. 
au<\ wild lew stamens and carpels, beeent topolo- 
gies (Dovle & Kndress. 2000: Cradam & Olmstead, 
2000: Mathews K Donoghue. 1999.2000: Parkin- 
son el ah, 1999; Qiu et ah. 1999. 2000; D. Soltis 
et ah, 2000; P. Soltis et ah. 1999a: Zanis et ah. 
20021 thai place \mbon-lla as sister lo other angio- 
spernis suggest iiislead lliat the earliest Mowers 
were small lo moderate in size, with an undiffer- 
entiated perianld. stamens ho king a well-differen- 
tiated filament, and a gvnoeciuin composed ol one 
unilocular ovaries. The diverse array of 
earlv angiosperm los-ik ,s consistent with this hy- 
pothesis (Crane, 1985; Crane et ah, 1995; Friis et 
al., 1994, 1997. 2000). 

Patterns of evolution of -| .< . i li< floral characters 
in basal augiosp.-rm- have also been examined us- 
ing a phvlogeiielie framework (e.g.. Hullord. I00(, : 
Albert et ah. 1998). I sing the topologies of Dono- 
ghue and Doyle (1989) and Chase et al. (1993). 
Hufford (1990) found thai the laminar stamen mav 
have evolved independent k several times in the 
Magnoliidae. I sing tde rh< \ . topologv for angio- 
sperms (Chase et ah, 1993), Albert et al. (1998) 
reconstructed perianth architecture to elucidate the 
evolution of the bipartite w dm led perianld Ion ml in 
the eudicots. Albert et al. (1998) found that the 
single whorled perianth optimized as the ancestral 
character stale loi the angiosperins. based on a 
[ilacemenl of Ceratophv llaeeae as sister to all an- 
giospernis in the rbcL tree. 

sperms further, we added a sixth gene. 26S rDNA, 
to the data set of Qiu et al. (1999). A compartmen- 
talization approach (Donoghue. 1994; Mishler. 
1994; Rice et ah. 1997) for these six genes was 
employed: litis method la. ilital.-d the use of maxi- 
mum likelihood and more thorough parsimonv 
methods of phylogenetie inference. Using our six- 
gene phvlogeiielie tree for basal augiosperms as a 
■subsequent!) investigated the evo- 

The utility of entire 26S rDNA sequences lor re- 
con^l i net inii . t niiM >-| ><i m |ih\ logeuv has reeenlly,. demonstrated (Kan £ Xiang, 2001; Kishbein 
el al.. 2(K)1: Kuzofl et al.. 1908). Although partial 

2(>S rl{\ \ anil rDN \ sequences ha\ casionallv 

heen used to infer phylogeny (Buehheiiii i\ Chap- 
man. 1001; llamby & /i miner. 1002: Ho et al.. 
1007). the groat length of the gene (over 3200 bp) 
had. until recently, precluded the use of its entire 
sequence lor phylogeny reconstruction. The gene 
comprises I >i >l 1 1 conserved core regions and more 
rapidh evolving expansion segments, thus eulian. - 
ing the ill i I i I > of the -cue for phvlogenelic inference 

ficult to align visually. Cluslal X (Thompson el 
1007). with gap opening set to 10 and gap ex 
sion set to 0.2. was therefore used to ohtaii 
initial 26S rDNA alignment, which was furthei 
fined by eye. 

\lishler (1001) proposed 
as a method to reduce a large ( 
more manageable size, to de< 
■■spurious homoplasy." and to r 
ol iiiloi'inalioii used in phylogei 
low nig different data sets to b 

We generated and analyzed entire 26S rDNA se- 
quence-, for I I aiigiosperm- representing Magnoliales. 
I.aurales. ( Tloranlhales. riperaceae. \i islolochi- 
aceae. \v mphaeaceae. Cam-Hales. \uslrobaileyales. 
moiiocols. and eiidicots. The follow m;j gv niiioq .rnii> 
served ;is oiilgroups: Ginkgo, Gnrlum. Ephedra, and 
hiri.x. All species for which 26S rDNA sequences 
were included In ihis analysis are listed in Table I: 

voucher inform; n and CeiiHank numbers are also 

provided. We attempted to use the same species, and 
in many cases the same DN As. used by Oiu et al. 
(1993) and I). Soltis et al. (1997. 2000). familial and 
oidinal (aieiiniM upturns inoslb |,,llou those given in 

APG II (2003). 

Amplification and sequencing of 20S rDNA gen- 
erally followed the methods of Kuzoff et al. (1008). 
for i»CK amplification, we used either ITS3 or N- 
M.-26SI as the forward primer and 3331 rev for I lie 
we amplified 26S 

. The .7 

amplified iisiii.. ,-ithei ITS;; ,,i \ n,2oSI as ihe for- 
ward KCK primer and 1830rev as the reverse prim- 
er. The 3' half of 26S rDNA was amplified using 
N-ih-26S7 as the forward primer and 3331 rev as 
the reverse primer. For the sequencing ol 2(>S 
rDNA. tin- following primers were used: N-uc20Sl. 
N-nc20S3. N-nc26S5. N-nc26S7. N-nc26S9. N- 
ne26SI I. N-nc20S13. N-nc26SI 1. 268rev. Ollrev. 
950rev. I220rev. lfOOrev. I839rev, 21.5 I ivv. 
2782rev. 3()58rev, and 333 1 rev (Kuzoff el al.. 

In brief, an initial global analysis is perfumed on 
ihe large dala -el. \\ ell-^uppoi le.l clades (compart- 
ments) are identified from ihe global analysis using 
bootsirap. jackknifo. or dceav values (Hremer. 
1088: Karris et al.. 1000; Kelsenstein. 1985). 
Smaller, focused analvses are then performed to es- 
tablish relationships within each of the compart- 
ments. I pon completion ol the smaller analvses. 

strained and relationships among comparlmenls 
> an be inferred, or a hypothetical ancestor for each 
comparlineiit can be reeoiisli ucled using par-imom 
or maximum likelihood methods (Vang et al.. 1003) 
and used to infer relationship- among comparl- 
menls (Donoghue. 1001; Mishler. 1091: Kice el al.. 

Using the five-gene dala set ol Oiu et al. (1999). 
we conducted a global pai-iiiion\ analvsis with 100 

multiple islands ol mosl parsimonious ire.-s i\lad 
dison. 1001) and TIM branch swapping. Internal 
support was eslimaled using I h< bootstrap (Kelsen- 
stein. 1085) will. 100 replicates vvilh 100 random 
taxon addition replicates. MUIPARS. and TBR 
branch swapping. I adi ..I ihe lollowuig . oiuparl- 
menls appeared well supported in our global par- 
siinonv analvsis. receiv ing great. a ihan '»(»', |„„,| 
strap support (see also Oiu et al.. 1000; Parkinson 
et al.. 1999; l>. Soltis et al.. 1999a): Nymphaeaceae. 
males. I'iperales. 

Volume 90, Numbe 

r 2 Zanis et al. 



Perianth Evolution in Basal Angiosperms 

,„, , Ge„ Bank Acce»i - f, 

>rthe20S rl)NA sequences us 







Acoru.s gramineus Ailon 

Kuroff (1998) 



Amhon-lla trichopol, Uaill. 

piunki'it. <;. 


Asimina triloba (L) Ihinal. 

piu /.s 

AY 09545 1 

;:-;:!",, li-lir:,) inillisii Holt. 

Oase 207 (NCU) 

AY 095 173 


1 aeloridaerae 

I.acloris frmandrziuim Phil. 

StoeMyrta*. 11784 (OS) 


\spl 'laceae 

Iliilhiiif Mirculi'tila Complon 

UCIArb. 7174 

Vi 095 171 


\„strub„ilvytt srandms C. T. White 

Qiu 90030 

■tt 095 152 


Kcrberidopsis conillinu Hook. 1'. 


1 >l J Nil. -« ,M'l- 

Chase 203 (NCU) 


0iu 97027 


C.lKr.inlli.l. eae 

Calycanthus occidentals Hook. & Arn. 

Cult. WSU 

A\ 095151 


C»' W^/ 7 

V, 095 155 


Ccratophyllum demersum L 

<?tu 95003 

Vi 0951.50 


Chlnmnlhus miiltislachys IVi 

A. ritfrfacA 920070 

\> 005 157 


llc,l\nsnuun bonplandianum L. 

£«dre S s 97-702 

\Mi'-,| ( ,| 


Kuptcli'ii polyandra Sieh. & Zucc. 

(Wu 9007 (NCU) 

\ 1 ';■,::> )j|<i 

1 •'iiniariai-car 

/>/Vr„/r« rv,/Nm Torrey 

Kezmcefc 9756 (MICH) 



Ginkgo biloba L. 

Cult. WSU 

U 095 175 

Kuzoffet al. (1998) 



Epfeerfra disiacbya 

KuzofT et aL (1998) 



\ Hi: if, 195 


lb hi, nidi, i inmpliacifolia L 

Univ. Zurich Hot. Card. 

\\ 095102 


1 ,// / ; h« Mgrareana (K. Muell.) 

fl 77. Wfeston 929 

U 095 159 


Magnolia dcnudala Desr. 

S. AVw /0/0 (NPRI) 

\' ' "": 


Plata. ueeae 


Strybing Arh. 


Hoot 9212 (UWM) 


Qiu 91029 


Cult. WSU 

U 095 170 

(Wu 97048 

V> 095 107 

Qiu 1 '90005 (NCU) 


Sugiura et al. (1985) 

Ml 1585 

Chase 573 (K) 


Wagner 6518 (HAST) 


SuA 128 (US) 

\M)95l(,i! /52 (NCU) 

V. 095 172 

ANBG 701680 

\M km 170 

Qiu 90026 (NCU) 


Nickrent 3013 (SIU) 


(J/u 90052 


five genes analyzed l.y Oiu el al. (1999) and our lution. In addition to the 

26S rDNA data set using a maximum likelihood each of the compartments 

approach as implemented in l'h\ logenetie Analysis compartmentalized data se 

using Maximum Likelihood (I'AMI.) (Yang. 1997); phyilum and Amborella, cril 

we eni|)loyed a general lime-reversible model with were not members of any c 

jr model of molecular evo- Maximum likelihood represents 

Missouri Botanical Garden 

sequences (Yang el al.. 199.3). employing infonna- 

estimates. Comparing holh parsimony and maxi- 

quences, Yang t*t al. (1993) revealed that when se- 
(]lienees were from closely related species, holh 
maximum likeliliood and parsimony nieMiods were 

or less \ariahle -lies, hotli parsimony and iikimimiiiii 

likelil I methods often produced identic, il ie- 

snlts. and both had high levels of accuracy in re- 
constructing aueeslrnl sequences. In eontrasl. both 
methods in. is he nnreliahle at recon>li in tin- an- 
cestral sequences for highh \ariahle -it <■- (Yin- el 
al., 1995). 

All of the hypothetical ancestral sequences lor 
the -i\ genes analyzed (live from Oiii el al.. 1999, 
plus the 20.S rl)\ A dala set presented here) were 
iillimaleh combined for subsequent pai -imoiix and 
maximum likelihood analyses; we will hereafter re- 
fer to this as the compartmentalized data set. Iioth 

meiitaliz.ed dala sel were conducted using the 
braneh-and-bound search strategy with the initial 
upper hoiiml computed \ ia stepwise addition with 
furthest sequence addition. 

Maximum likelihood parameter values were es- 
timated hour the single most parsimonious tree ob- 
lained from the parsimony analysis desei'ihed. We 
used the general time-reyersil)le model of molecu- 
lar evolution, accounting for invariant sites and rah' 
heterogeneity (Swollord el al., 1996). The choice of 
model o| molecular evolution was done usiim YIO- 
DKLTKST 3.0 (Posada & Crandall. I'Wi). Fifty-six 
models were compared using the \kaike informa- 
tion criterion (AIC). The AIC model comparsion 
method does not require models to he nested and 
selects models lor good lit lull penalizes models for 
unnecessary parameters (Posada & Crandall. 
1998). The -In likelihood score for the single most 
parsimonious tree was 39713.36269. the proportion 
of invariant sites was estimated to he 0.316797, 
and lire gamma shape parameter was estimated to 
he 0.7 1 133,3. These estimates along wrlh estimates 

ikelihood and parsimony analyses was < 
rsing bootstrap analysis with 100 repliei 
00 random (axon additions per repliei 
'BR branch swapping. 

Perianth character slates were mapped assuin 
equally weighted parsimony using MaeClade '. 
(Maddison & Maddison, 1992). We employed I 


topology dealing with relationshif>s among the ha 
was generated here (see below); t 
mships are those from the jackkn 
ree of P. Sol I is et al. (1999a) and 
(2000). Phy logenetic relationships c 
pieted within Nymphaeaceae follow Pes et 
(1999). Relationships depicted within the monoc* 
are a portion of one of the shortest trees obtain 
from a combined morphological and rbcL data i 
analyzed hv Chase et al. (1993). 

Keeause d< 'In 1 1 1 ions for ihe perianth vary (lan- 
dless. 1961a; Crey son. 199 1: Weherling. 1989). we 
refer to the perianth as the sterile structures above 
the bract(s) if" present, and immediately below the 
spore-producing organs in a flower, lire organs ol 
a perianth (i.e.. sepals ami petals) share similarities 
with bracts, a scale-like leaf subtending the organs 
of a flower, as well as slaminode-. structures that 
represent sterilized stamens. For example, sepals 
have many features in common with bracts: both 
are green and lack clear anatomical differ <-i ilialn.ii 
into palisade and spongy parenchv ma. Petals and 


(E>mi. I"0/o. Roth are laminar in shape and usu- 
ally colorful idne to eliionioplasts ,,r pigments) 
(Fsau. P'f ,.->). Descriptions of the perianth charac- 
ters -aid -tale- lollowed pivvioii- workers' interpre- 
tations for each ol the groups studied and are pre- 
sented in the Appendix. 

Perianth phyllotaxis is the arrangement of peri- 
anth organs (i.e.. sepals and petals) on an axis (En- 
dress. 1 994a). The states we coded are: absent (i.e., 
there is no perianth), a single whorl, two whorls, 
multiple whorls, and spiral. We also explored al- 
ternative character-slate coding hv ledrn nig the 
different forms of whorled phyllotaxis (i.e.. single 
whorl, two whorls, multiple whorls) into a single 
character state, whorled. Phyllotaxis is difficult to 
divide into stales because il is labile in many 
groups (I)oust. 2000. 2001; Undress. 1987b). Fur- 
thermore, the distinction between spiral and 
whorled is not always clear. Thai is, spiral and 
whorled may not he fundamentally different forms 
of phyllotaxis. Spiral and whorled perianth phvllo- 

Volume 90, Number 2 

acter (Stevens. 1991). Perianths with whorled phyl- 

lotaxis appear In ha\e organs arranged in the same 
plane. These organs appear equally spaced and 
uImi appear to have I m -. r i initialed simultaneously 
(Kndress. 19871.). However, ontogenetic investiga- 
tions have indicated that in some cases floral organs 
typically identified as being whorled actually resull 
from the presence of long plasloehrons [llic lime 

organ primordia; (Kndress. 1991a; Tucker. 1900)| 

inserted belween , inber ol small plasloehrons 

[e.g., Huber (1980) for Solanaceae; Krbar & Leins 
(1985) for Apiaceae|. Thus, both spiral and whorled 
phvllotaxis ma\ ha\e the organs developing in a 
spiral sequence (Kndress. 1987b. 2001). 

Perianth merosity refers to the number of 
anth parts present within each whorl. We recog- 
nized five character stales: three organs per whorl 
two or lour organ- per whorl, live organ- per whorl 
or an indeterminate number of organs per whorl 
Indeterminate refers to a range with no fixed num- 
ber ol parts: typically, these are perianths w 
melon- spirally arranged parts. 

Differentiated perianths are those having ; 
er whorl (or series) that is clearly diflerentialed 
from the inner whorl(s) (series). \ difierenliuled 
perianth is commonly referred lo as being com- 
posed of sepals and petals. In contrast, an undil- 
ferenlialed perianth is one that lacks clear distinc- 
tion between the outer and inner whorls: lliese have 
been I rai lit ioiial I v reeogni/.ed as lepals ( ( j'om |iiist. 
1088: Takhlajan." 1997). Albert et al. (1998) sug- 
gested that there must be al least two whorls pre- 
sent for unambiguous interpretation ol sepals and 
petals, that is. an outer whorl that is clearly distin- 
guishable I'ioiii Ihe inner whorl. Often, when only a 
single-whorlcd perianth is present, it is referred to 
as a calyx as a matter of convention (Cronquist. 
1981. 1988). for example, those Aristolochiaceae 
charaderized bv a single-vv horled peri anth are re- 
ferred to as having sepals |( a'niiquisl. 1988; Takh- 
tajan, 1997: Tucker & Douglas. 1996); develop- 
mental data for \sanuii (I,, in- t \ Krbar, 1985) and 
\rislolo<-hi,i (Oonzahv <K Stevenson. 200(1) ,i.;»poil 

• iisidered the trealmeiil ol 
(Mabee & Humphries, 
1993; Maddison. 1993; Nixon & Davis, 1991; 
Wiens, 1999). There are three options: (1) code the 

polymorphic slate as "missing."" (2» cod.- polviuor- 

strucling the aneesti'al slate from a phvlogenv of the 
lax.i in (|uestiou. and (3) code the polymorphic taxa 
as polymorphic. We coded several taxa as poly- 
morphic, for example, Canellaceae are coded as 
trimernus, letramerous. oi penlaineroiis (Wilson. 
1906). We tested the robustness of our results by 
using alternative codings (Done-hue c\ Ackerlv. 
1996; Weller et al.. 1995). 

In addition to uncertainty in scoring laxa. due 
either to difficult) in interpreting structures or to 
polymorphism, then i- imeei lai ntv in the tree to- 
pology and the impact that alternative tree topolo- 
gies may have on ancestral character-state recon- 
structions (Donoghue & Ackerlv. 1996). We 
examined the sensitivity of our character-state re- 
constructions |,\ examining reconstructions on al- 

Perianth differentiation can often be diffici 
score in those taxa having a perianth of a s 
whorl. We used the following character states i 
analysis of perianth differentiation: undifferent 
(no clear distinction into outer and inner wh 

outer whorls), single whorl, and absent. 

« 1:1 I 

i • .-> '.; and \vmphaeaceae as a claile sister to 
the rest of the angiosperms. 

We omitted outgroups from our analysis of peri- 
anth evolution. Previous morphological analyses 
thai included both extant and fossil -eed plant- in- 
dicated that the angiosperms were part of an an- 
thopbvle clade containing the dictates. Hennelli- 
tales. and Pentoxylon. This anthophyte clade was 
sister to the seed fern Caytonia (Donoghue X 
Doyle. 2000: Doyle. 1000. 1908). Recent analyses 
(e.g.. Bowe et al.. 2000: Chaw et al.. 2000; P. Soltis 
et al.. 1999b) indicate that gymnosperms are mono- 
phyletie. with the Onelales nested within the co- 
nifers or as the sister clade to the conifers. Kur- 
thermore, it is difficult to assess the homology ol 
the angiosperm perianth with perianth-like struc- 
tures in the gymnosperms (e.g.. Crane, 1988; Huf- 
ford, 1996). Nevertheless, we explored the effect of 

e- mi). in ii f. i ii i upl i stales lor ihe an- 

giosperms on our inferences ol perianth evolution. 


The compartmentalized data set had 14 hypo- 
with aligned sequences of 12.215 bp of sequence 

1230 wen- parsimony-uninformative, and 1318 
were parsimony-informative. Parsimony analysis of 
llif compartmentalized dala resulted in a single 
most parsimonious tree ol 121!! slops (CI = 0.741 
and Kl = 0.691). The topology obtained Iron, the 
maximum likt-lil d analysis ol llic cotiipaitinen- 

ii ■M 

only 36% (Fig. 2). Phylogenelie .ni.ilv-.- of tin- 2<>S 

ll'\ \ ■Li, I alulll' ((fee IK .1 dl.lVMl) IC\cal ,1 |il|)o|l.g\ 

luglih -iimiI.ii in. although loss resolved than, that 
obtained lor basal angiosperms based on three (I). 
Sollis et al.. 2000; P. Soltis et ah. 1 000a). live (Oiu 
t-l ah. 1000. 2000). or more genes (Zanis el al.. 

'.7 I."..: 


■..»lil| i i in 'i nia 

>d d, a 

lihood analyses of tl 

sistenl with the shortest tree of Qiu et al. (1090). 
as well as the tree we obtained in our first global 
analysis. The same early-diverging angiospetm- are 
present in the same order: \mhorclhi. Nymphae- 
aeeae. and an \uslrobaile\ales chide of Austro- 
eeae. Bootstrap support for these relationships is 

In reeenl studies <>l basal aii-io>perm ph\lo-en\ 
(e.g.. Oiu et ah, 1000, 2000; Zanis et ah." 2002). 
bootstrap or jackknife supptirl > 70% has been 

ing elades of basal angiosperms (i.e.. Piperales. Ca- 
nellales. Magnoliales, monoeots. haurales. and 
Chloianlhaeeae). However, some relal ionships. 
such as the sister group ol the monoeols or the 
position of the ( Ihloranthaeeae. remain uncertain. 
In the compartmentalized anabsis. bootstrap sup- 
port for relationships among the eoinparlnienls 
(elades) is much higher (> 00%) for most nodes 
than the internal support found in bolh urn jli-bal 

anabsis and in the reeenl eonesp ling parsiu \ 

analvses (On, el al.. 1000. 2000; Zanis el ah. 
2002). Canellales and Piperales form a strongly 
■mpporled i loo'.; i Hsler group thai is. in turn, sister 
to a well-supported (100%) FauralcsAhmnoliales 
chide. This magnoliid chide is also slronyh up 
ported (100%). Ceralopliyllaceae are stronglv sup- 
ported (100%) as sister tt, the monoeots. There is 
also strong support (H6%) for the placement of the 
monoool/Ceratophvlhiceae clade as sister to a large 
elatle comprising Chloianlhaeeae. Canellales. Pi- 
per.iles. Magnoliales. haurales. and eudicols. The 

|>lacement of Chloianlhaeeae within this elade. Al- 
though Chloianlhaeeae appear ,h the sister group 
to the large elade of eudicols and luagnoluds (Pi- 
perales, Canellales, haurales. ami Magimliah-sj. 
this placement receives bootstrap support • 30%. 
The complete 26S rDNA sequences have made 

internal support. For example, in a boolslrap anal 
ysis of the compartmentalized data set using par- 
simom with 2(>S rDNA se(|iienees removed, sup- 

single whorl. 

(Fig. 4). Fach node of the Amhorellaceae/Nwii- 
phaeaeeae/ \iislrobaile\ales grade is equivocal with 
whorled or spiral perianth being equalh parsimo- 
nious stales for the base ol the angiosperms. How- 
ever, the ancestral stale lot 1 1 1 . • Luge elade contain- 
ing the magnoliid clade. Chloianlhaeeae. 

whorled perianth I Fig. I). The mapping of number 
ol perianlh pails onto our topology, indieales I it 

aiilh wilh pails in threes or indeterminate (Fig. 3). 

perianth (Fig. 3). 

Mapping of perianth differentiation results in 
much of the tree being equivocal with multiple 

perianth differentiation (Fig. 6). Clearly a differ- 
entiated perianth e\ol\ed multiple times in the an- 
giosperms. Our analyses indicate that there may 
have been as lew as two to as many as six changes 
i" •' dill. i. niiai'-.l perianlh hoin an undifferentiated 
perianth (Table 2). Futhermore. some differentiated 
perianths ma\ have evolved from spiralis arranged 
perianths, whereas other differentiated perianths. 
such as in Suruniu ( \ristoloehiac.-ae). mav have 
evolved from ancestors thai had a perianlh com- 
posed of a single whorl. 

gies. Our anabsis of perianlh phvllolaxis and num- 
ber of perianth parts is robust (little effect on 

ancestral reconstructions) I,, alternative topologies 
(data not shown). However, the anabsis of perianth 






iative topologies. Laurales is reconstructed as having a differentiated 

ire sister to the perianth (data not shown). In addition to examining 

itaining the monocots and CemtophyHum, the sensitivity of our character analyses to alter- 

ancestral state for the clade containing the native topologies, we examined the effect of coiling 

Canellales, Piperales. Magnoliales, and terminal taxa as monomorphic versus polymorphic. 

fee led b\ changes from | .( »l\ UK >l| »l)ic lo a fixed c 

acter state. Lastly, assuming different plesiomoi 
ic stales for the angiosperms docs affect 
inference of ancestral character-stale 
tion. Tims, if we assume llial the tritnerous perianth 
is the plesiomorphic -late lor the angiosperms. then 
the Irimerous perianth maps as the ancestral slate 
for the deepesl nodes of the- tree. However, assum- 
ing thai ihe plesiomorphic slate was a perianth with 

from what is shown in figure I, If we assume that 
lacking a perianth is the plesiomorphic state for the 
angiosperms, we then find that there have l>een 
multiple origins of hoth a differentiated and undif- 
ferentiated perianth. Assuming other plesiomorphic 
character staler lor periaulh •lifferenlialion had III 
tie impact on our inference's of perianth evolution. 

i/.ation approach allowed us 
mher of taxa from 105 in the (^iu 
et al. (1999) data set to 14 hypothetical (ancestral) 
taxonomie units and three operational taxonomie 
units so that we could emplo\ hoth a more thoiough 
parsimoin anabsis and the maximum likelihood 
method. The topologv thai resulted from the com- 
partmentalized ;tli.l l\ -l- I- elllll'eb consistent with 
respect to the phylogenelic placement of basal an- 
gio>perni> uilh the tree> obtained in hoth our initial 
ulohal .iii.ihv,. .1,1,1 ||„. f i \ r ■ - - , • 1 1 , anaKsis of Oiu el 
al. (1999) (Fig. 1). Identical topologies were ob- 
tained using Ihe entire data sel of Oiu el al. (1999) 
or the hypothetical ancestral se((ueiices using both 

, I ,1 I I I I II 

i,.Ium-.I »ilh the whorled phyll. 

I. . I i- ■ i 

n ol pet Kintli pliyllotaxis using MaeClade 3.03. I'i\e 

ided into three separate character states: two whorls. 

maceae/Zosteraceae: CAN = Canellales; CHL = Chlor- 

methods on the fin, 

(Fig. 2). Nevertheless, further work is i 

understand the limits and bia 


Phylogenelic analyses of the 26S rDNA se- 
quence data ih.n i '• I i |. »1< < - >\ ilh the same 
caiK-bram Inn- angio>pemis. as well as the same 
major chides of basal angiosperrns (i.e.. I'iperales. 
Chloraulhales. Magnoliales. Laurales, Canellales, 

in >eots) obser\ed In analyses ol three-gene (D. 

Soltis et al., 2000; P. Soltis et al.. 1999a), five-gene 
(Qiu et al.. 1999). and Lugo, (Xanis et al., 2002) 
data sets. The addition of 26S rDNA sequences to 
the five-gene data set played an important role, in- 
creasing bootstrap support for relationships. For ex- 
ample, the addition of 26S rDNA sequences pro- 
sided high bootstrap support lor the placement of 
the tnoiioi o|s among the basal angiosperm lineages 
and the placement of Ccratophyllum as the sister 
group to the monocots (Fig. 2). However, compart- 
mentalization of the lliree-gene data set (D. Soltis 
et al., 2000; P. Soltis et al.. 1999a) did not lead to 
greater resolution or support among basal angio- 
sperrns (P. Soltis et al.. 2000). Thus, it is not the 
analytical approach 

; that accounts for the 

i- ■• -ened m other re- 
cent studies (Mathews cK Donoghue. 1999; Parkin- 
son el al.. 1999; Qiu et al.. 1999. 2000; D. Soltis 
et al., 2000; P. Soltis et al., 1999a; Zanis et al.. 
2002): Amborellaeeae. followed by Nymphaeaceae. 
and the Austrobnileyales .lade I \ustrobailcyaceae/ 

We found high levels of bootstrap support (> 
90%) in the compartmentalized analysis for all 
nodes, except the node supporting Chlorantliaecae 
as sister group to the clade of Magnoliales. baiua- 
les. Canellales. I'iperales. \ristolo< hiales. and eu- 
dicots. Our analyses show strong support for the 
sister relationship of Canellales and Piperales 
(100%); Laurales are sister to Magnoliales (100% ): 
l.aiuales/Magnoliales and < ianellales/Piperales abo 
form a well-supported magnoliid clade (100'ri; 
Ceratophyllaceae are sister to the monocots 

The increase in I Islrap support between the 

initial global analysis (as well as the broad analysis 
of Qiu et al., 1999) ,,,id the compartmentalized 

analysis is most likel\ due to both the minim /ai 

of homoplasv as a result of the eoustnietion of the 
hypothetical" ancestral sequences (Mishler. 1994) 
and. as noted, the addition of 26S rl)\ A sequences. 
However, seme areas of the phylogeny remain poor- 
ly resolved and -upporled de>pile the use of data 
for six genes and approximately 12.000 base pairs 
of se(|uence data per laxon. For example, the node 
>upporlinu the placement of die Chloranthaeeae is 
weakly supported in hoth the Oiu et al. (1999) anal- 
ysis and our compartmentalized analvscs ol the six- 
gene data set. The low bootstrap support for this 
node is likely the result of short internal l.ranch 
lengths. For example, the branch leading to the 
(.hloraulhaceae and all remaining angiosperms is 
only 17 steps with ACCTRAN optimization. The 
uncertainty surrounding the placement of the 
Chloranthaceae supports Mathews and Donoghue's 
(1999) Migg ( '>|iou that alter the ap|)earanee of the 
Amborellaceac. \\niphaeaceae. and \ustrobailey- 
ales lineages, there was a rapid radiation of the 
remaining ba^al :i ■ iv. i- — | » -i mi lineage*. The hypoth- 
esis of a rapid radiation i> plausible and consistent 
with long-standing hypotheses on the origin and di- 
vergence of the angio>perm> e-peeiallv given that 
fossils ol Chloranthaceae. I.aurnles. Ma-uohales. 

The origin of the perianth is uncertain, and \ar- 
ious hypotheses have been proposed. Weberling 
(1989) outlined four views regarding the origin of 
the perianth: ( I I the whole perianth is derived from 
foliage leaves (following (duck. 1919; Prantl, 1887; 
Velenovsky. 19 10): (2) the perianth is derived from 
sporophyll structures (Celakovsky. 1900; Nemejc. 
|9fif,); (.'{) the calyx is derived from bracts and the 
corolla is derived from modified stamens (l)rude. 
1887: (ioethe. 1790; Naegeli, 1884); and (4) the 
calyx is derived from bracts and the corolla has a 
different origin, neither calyx nor staminal (Weber- 
ling, 1989). Features of the perianth that have at- 
tracted attention are perianth phvllotaxis (Fndre-s. 
1987c). the number of parts (Kubitzki. 1987). dif- 
ferentiation of the perianth (Albert el al., 1998), 
and. recently, patterns of symmetry in asterids 
(Donoghue et al., 1998; Hee & Donoghue, 1998). 
We have examined here the evolution of ph\ II. .tax- 
is, number of parts (merosity). and differentiation, 
characters that are n< •eessariU independent. 

i phyllotaxis: Five states 

For example, Kubit/ki (1987) discussed the origin mg the inagnohid clade. Chloranthaceae. Cerato- 

of the trimerous perianth from a spiral (indetermi- phyllaceae, monocols, and eudicots is a perianth 

nate) ancestor. Thus, merosity and phyllotaxis arc based on a whorlcd ground plan (Fig. 4). However, 

sometimes intertwined. Moreover, perianth differ- we have treated perianth plnllolaxis as a discrete 

entiation is dependent on haung more than a sin- character rather than a continuous character (Ste- 

gle-whorled perianth (Albert et al.. 1998). vens, 1991), and interpretation of our results should 

Our analysis indicates that when perianth phyl- reflect this treatment. For example, the transition 

lotaxis is considered a live-slate character (spiral. from a spiral to a whorlcd arrangement brings floral 

two whorls, single whorl, multiple whorls, absent). parts close, together through the gradual shortening 

ancestral reconstruction is ambiguous for many of of the Moral axis (Takhlajan, 1980). 

the internal branches (Fig. H). However, the anal- Perianth merosity has been of particular interest 

ysis of perianth phyllotaxis as a three-stale char- with regard to the origin of the trimerous flower 

acter (absent, spiral, whorlcd) indicates that while (Kubit/ki. |<W57>. Out analysis of perianth merosily 

the ancestral state for the angiosperms is ambigu- indicates that the ancestral condition for the angio- 

i in Basal Angiosperms 

ratophyllaceae. bract in origin. Thus, the differentiated perianth 

monocots, and eudienl* is a trimerou* perianth ll'ig. found in \\ mpliaeaeeae. Kanimeulales. and Cary- 

5). There may have heen as many as four origins opln Males is considered to lie staminal in origin, 

of the trimerous perianth from a perianth with an whereas the petals of Magnoliales and Austrobail- 

indeterminale number of parts, based on our to- evales are considered to be bract in origin (Takh- 

pology (Fig. S). Also, the indeterminate perianth tajan, L 980). These are just two of several scenarios 

may have originated as lew as two or as many as for organ differentiation, 
five times (Fig. 5). (1987) suggested that 

the amount of morphological change required in the Conclusion 

transition from a spiral (indeterminate) perianth In Compartmentalization of a six-gene data set 

a trimerous perianth could be small: moreover, he vit ,| ()( . ( | „, vs eongruent w j t h many recent phylo- 

suggested that the trimerous condition is a mm- „,.,„.,;,. atul | vs ,. s ((; ia han. K Olmstead, 2000; Ma- 

phological constraint and that the return Iron, a ||)( . ws cK | )(M10 ^ llt .. |000. 2000; Qiu et al.. 1999. 

trimerous perianth to a spiral (indeterminate) pen- 2()()(); () SJ|is (l| .,, 2 ()()() : |» Sollis et al., 1999a; 

anth oi pentamcroiis penanth i- not possible. Sg- /am ^ <i , a , 2()()2) vw ., U() n |, serve d high levels of 

nificantly. however, our analysis of perianth iner- , MM , ts|rap Sll p, )()rt f or ni()st of the nodes in the com- 

osity indicates that there have been from one to five parlrnen t a |ized tree: the values we obtained were 

transitions from trimerv !o an indeterminate f)en- higher than those in anv previous study. However, 

anth (Fig. 5 and Table 2). Some of the changes from ^^ va|m , s shm| , d , (( . mtenH ,.,ed cautiously as the 

a trimerous perianth to an indeterminate perianth poU , ntia| | )iast . s n f ( . () .npartmentalization analyses 

occur in Laurales. For example, the perianth of Ca- are un | 1 . nown 

lycanthaeeae, which has an indelerminale number T , H> an( . eslra | slalrs [ or perianth phyllotaxis and 

of parts, may have evolved from an ancestor with a mer()siu aiv unt .lear. However, the ancestral state 

trimerous perianth. However. I.aurales are unique f((r ,, ie ). n .^ ( .| a(|( , ( . nlaining the magnoliid clade, 

among basal angiosperms due to the presence of a chloranlhaceae. Ceratophyllaeeae, monocots, and 

hypanthium, which may affect the development and eudk . ()ts is a w horled perianth, with reversals to 

arrangement of the perianth. spjral phvllotaxis atl( | UI , indeterminate number of 

Perianth differentiation has long been studied |)ar(s JM s()me | mra ^,. s . These results support the 

(Fames. 1931; Hiepko. I9(>5: kosiige. 1994). Re- hypothesis *if floral lability ("open organization") for 

cent investigations of the evolution of perianth dif- hasa] aiI{ ,j os|MTIlls: imricate synorganization oc- 

fercntialion have come from both phylogenetic and curred wUh the ((ripn ()f eudi( . ots ( e . g ., Albert et 

developmental genetic perspectives (Albert et al.. a , ^ 1998; Klldlvss . p n > 7( .. 2 ()()l). Our analysis of 

1998; Kramer & Irish, 1999, 2000). Differentiated [)e| . iaMt|l PV(1 | lltioll armss | )a . a | an gmsperms indi- 

perianths have been proposed to have evolved mul- ( .. ([es t | ia| ,| ]<1Ri atv mu hip],» m()S ( parsimonious re- 

tiple times from different struclures in dilleivnt r()lMlU( . ,„„,., „| ,„., i.inlli phv llolnxis and perianth 

groups (Hiepko, 1965; kosuge. |99| : Takhlajan. ( | in( . r( . M | jali , m . | |, mr , ( . r . the trimerous perianth oc- 

1991). Our analysis of perianth differentiation in- ( . u|Te( , ( , aHv JM . lll} ,[ ospei . m evolulion. In fact, the 

nitii! the mag- 

as many as six changes to a differentiated perianth no]iid (>lade Chloranlhaceae. Ceratophyl 

from an undifferentiated perianth (Fig. 6) (Table 2). monocots , an ,l eudicols is Irimerv. Much of our 

Thus, our results agree will, earlier interpretations mapping (> r p ( . naril l, ,|,ffcicnlialion is equivocal for 

that a differentiated perianth (Hiepko. 1905: ko- ([)r deepesl nodes of llle tree . although a differen- 

suge, 1994; Takhlajan. 1991) evolved multiple (ia(tM , p t>naill |, | ias elearly originated multiple 

times. Some differentiated perianths mav be ihoso |im(vs S()|M( . ( |j||,. nM1 |jated perianths may have 

in which the calyx is derived from braels and the evo | vod f roin sp i ra |l y arranged perianths, whereas 

corolla derived from modified -I. miens. ,, corolla (l ,her differentiated perianths, such as that of Sar- 

(Kozo-Poljanski, 1922; uma (Aristolochiaceae), may have evolved from an- 

differentiated perianths cestorg ^ had a [>tM - jaill h composed of a single 
may have a corolla mat is derived Iron, modified w h or l. 
bracts, referred to as "bracteopelals" (Kozo-Poljan- 
ski. 1922: Takhlajan. I' MM)). The ontogeny of petals Literature Cited 
and sepals (Kosuge, 1994) and the number of vas- 
cular traces (Carlquist. 1969: Melville. 1969; Puri, 
1951: Smith. 1928) have been used as criteria for 

lal genetics, and llic angiosperm petal. Pp. 3 10 ''7 1 in 

I). f. SolllS, I'. S. Soltis ,\ J. J. Doyle (ediloisi. \|„|, v 

ular Sysleniatics of Plants II. Kluwer, Boston. 

>I'C II. l^i Hi;;. \n update of t li. \iigiosperin l*h\ lo^en\ 
Croup classification ol the orders and families of flow- 
ering plants. Bot. J. |J n „. Sue. (in press). 

rber, K. A. N. t K .1. Parkin. I<)()7. On the origin of an- 
ginsperms. Hoi. J. I .inn. So.-. 38: 29-80. 

rm.| \.\ S.C. Tucker. 1980. floral development 
in 1/.//.W/.7/ (Mvristieaceae). \lller. J. l!o|. 73: 

& T. K. Wilson. 1978. floral morphology of 7/or.v- 

fu-ltlia (Mvrislicaceae). Amer. J. Hot. 65: 441-449. 

a. lev. I \\. X H. (,. I.. Svva.n>. I<)|,", \mhu,r/l„ In,!,,, 

":'"'■• I'.iill \ new morphological h| I \ ,-sselless di- 

eot.vle.lon. J. Arnold Arl.or. 29: 245-253. 

arkman. T. J.. C.CI.ene.N.J. |{. McNeal..). I Aons-Weiler 
<X <7 \Y. de I'amplnlis. 2000. Independent and eoni- 
liined analyses of se.|uenees from all three genomic 
compartments converge on the root of dowering plant 
phvlngenv. Pro,-. Natl, \ea.l. Si.. I .S. \. 97: j.3100 


esse*. C. I-'.. 1807. Plixlngem and taxonnmv of (lie an- 

»i.. N .ei.ns. Hot. Gaz. 21: 1 15-178. 
. 1915. The phvlogenetic taxonomv of dowering 

plants. Ann. Missouri Hot. Card. 2: 109-164. 

iwe. I.. M., (i. Coat * C. W. dePamphilis. 2000. Phv- 

'•n.Lll.-' ■ !■>--! ivl.ilnrs :,,,- .oiufe,.. Pro.-. \.i|l. 

Acad. S.-i.. I .S. \. 97: 1092-1097. 

emei. k ML",;;. I'hc Iii, ills ..I 1 1 r 1 1 1 1. , .i,-i,| M-.|iieu< .- data 

in ,iH.-m N „,,„ pf.xl.,,,-,,,-!,, i.m, t -.|,,i, Hon. Evolution 

42: 795-803. 

i/ieky. G. k. IT,'). \arial>ilil\ in the floral parts of O'o- 

m,„lr ;i <, K.oinorl.-gaeeae). Will.l.-nowia 2: 200-207. 

'< lih.-im. VI. \. i\ I!. |„ Chapman. I 99 | . Phvlngenv .if 

;. iiiii>. 

Y.-L Qiu, K. 

A. Kron, J. H. Rettig, E. 

'aimer, J. H. Manhart, K. J. Sytsma, H. J. 

J. Kress. K. 

G. Kami, W. 1). Clark, M. 

ren. H. S 

. Cant. H. K 

Smith, (i. H. 

Furnier, S. H. Strauss, Q. 

\l. Plunkelt. 1 

>. S. Soltis, S. M. Swensen, 


i, P. A.Gadck. 

C. J. Quinn, L. E. Eguiarte, 

0. H. Learn. 

Jr.. S. W. Graham, S. C. H. 

iva.ian.lan ,\ 

V. A. Allien. 1993. Phvlo- 

the pla-lld ge 

ne rbcL. Ann. Missouri Bot. 

. I). V\. Stevenson. P Wilkin cK P. J. Hudall. 1995. 

Monocotyledon svstemalies: A combined analysis. Pp. 

<"':■> : 30,/, P J. 1,'udall. |l. I Guile, ,\ C.J. Humphries 

(editors), Monoeolvle.lons: Syslematies and Evolution. 

Hoyal Holanie Gardens. Kew. 
:haw. S.-M.. C. I.. Parkinson. V Cheng. T. \]. \ ineenl ,K 

all three plant genomes: Monophvlv of exla. 
-perms and origin of Cnetales from conifers. |> 
Acad. Sei., U.S.A. 97: 1080-4091. 
Crane. P. |{. 1085. Pin logenelie analysis ..f M , 
and the origin of angiosperins. Ann. Missouri I 
72: 716-793. 

lefgvinnospemw. Pp .21!'. 272 in C. H. Beck (ed- 
Origin and Evolut, 
. Press. New Wk. 

. FriisiK K. R. Pedersen. 1995. The o 

1981. \n Integrated System of Classify 
g Plants. Columbia Univ. Press, New A 
. The Evolution and Classification of I 

. P. K Clifford K P. P. Vo. 1985. 
Monocotyledons: Structure, Evolut 
ger, Berlin. 

Dahlgren. H. 
families , 

Dengler, N. G. 1972. 

apex of Culwanthus occidcnlulis. Cana.l. J. Hot. 50: 
Donoghue. M. J. 1901. Progress and prospects in recon- 
structing planl phylogenv. Ann. Missouri Hot. Card. 81: 

>X I). I). AckerK. 1990. Phvlogenetic uneerlaiii- 

nalyses. Philos. Trans.. Ser. B35I: Press. Oxford. 
& . 2000. Demise of the 

polhesis? Curr. Biol. 10: BI06-HI09. 
. R. H. Ree & I). A. Haum. 1998. 

Trends PI. Sei. 3: 311-317. 

I'niirl.i.. \. Wei S. C. fucker. 1990a. Compa.ahv e lloi 
ontogenies among Persoonioideae including Hrllcndvi, 
(Proloaceae). Amer. J. Hot. 83: I 528-1 555. 

N-i. 157: 

3 72. 

Doust. \. N. 2000. Comparative floral onlng.uv in W i, 
teraceae. Ann. Missouri Bot. Card. 87: 366-379. 

in Drimv.s uinlrri (W interaceae). Int. J. Pi' Sei 16; 

Doyle. J. A. 1991, The ongin of ihe angiosperm flower: 

phvlogenetic perspective. PI. Sysl. Evol. 1.52: 7-30. 
. 1990. Seed pi. nil phvlngenv and I,, icl nshi 

of the Cnetales. Int. J. PI. Sei. 157 (Suppl.): S3-S39. 
. 1998. Molecules, morphology, fossil-., anil the r. 

lationships of angiosperins and Cnetales. Molec. Phv 

logenet. fvol. 9: 448-^162. 

netic analysis of basal angiosperins: Comparison am 
combination with molecular data. Int. J. PI. Sei. 10 
(N.pi'l.i: SI2I-SI53. 

X A. Ue Thomas. 1990. Phvlogenetic analysis am 
Hisl. Nal.. B. Adansouia 18: 279-331. 

oghurK K. A. /inline |00|. 


A. N.. I>. R. Crane A S. 15. I 

I ' <.:ml. : 

>•■<} 'at!.', ,- 

lion of non- 

M.-.-rlollld (]:c ; ,i, !: il:»n- << ll<li( ols). PI. Svsl. Evol. 8 

(SuppL): 93-122. 

hude. O. IJMJ7. Die sv stemalische mid geographische 
\unrdnungder Plmnorogamen. I',,. IT.". \'H, m S-I.enk 

■imrs. A. J. 10.51. The vascular anatomy of the flower 

U ; lfi I- 1 1 J t ■ i ill the theoiv of ( -ill-pel p< »i M : f c.rpl I i -I i ■ . 

Amer. J. Bot. 47: 147-188. 

aidless. P. K. 1077. unci \erwandtschall der 
Eupomatiaceae mid Hin 

iral structure and relationships of// 
eae). PI. Syst. Evol. 133: 199-221. 
ic early floral development of Ausl 

Kan. C. & 0. Xiang. 2(K)1. Phvlogenetic relationships 
uitlun Corniis |( !... inunirl l.ased on 20S rl>\\ se- 
quences. Amer. J. Bot. 88: 1131-1138. 

Karris. J. S.. V. A. Albert. M. kallersjo. I). I.ipseoml. & 
A. <;. Kluge. 1000. Parsimony jackkniling outperforms 
neighbor-joining. Cladislles 12: 99-124. 

Kelseustein. .1. 1985. Confidence limits on phv Ingenies: 
approach using the bootstrap. Kvolution 39: 783- 


II ,i 


ishbein. \1. C. C. Hil.sch-Jetter, I). K. 

lord. 2001. Phv logons of Saxilragalcs (angiosperms. 
dicots): Analysis of a rapid, ancient radiation. S 
Biol. 50: 817-847. 
T.ster, A. S. 1963. The 
Circacasler. J. Arnold Arbor. <- 

Viis. K. \1.. P. R. (lane & K. H. Pedersen. 1907. fossil 
history of magnnliid angiosperms. Pp. I 2 I - I .TO /// K. 
[« .tsuki & P H V Ml I d irsl. KvoIuIiom and Diver- 
sification of Land Plants. Springer- Verlag, New York. 

, K. R. Pedersen K P. K. Cane. 1001. Angiosperm 

floral structures from the early Crelaceous ol Portugal. 
PI. Syst. Evol. 8 (Suppl.): 31-49. 

. 1987a. The Chloranthaeeae: Reproductive st 

lures and pin logenetic position. Bot. Jahrb. Syst. 

Floral phyllotaxis and ev. 

Jahrb. Syst. 108: 417-438. 

Coethe, J. \V. v. 1700. Versuch del Metamorphose der 

. 1987c. The earlv evolution ol the angiosperm 

flower. Trends Keol. Kvol. 2: 300-301. 

Gonzalez. K & 1). \\. Slevensnn. 2000. Periaull, devel- 

. 1901a. Diversity and K.volul ionary Biology of 

opment and sv sternal ic> of \ns!nlorhi,i. flora PC; .''.,() 

hi|i 1 1 .. , i ( .1 ndge I niv. Press, Cambridge. 


. 1904b. Evolutionary aspects of the floral struc- 

Graham. S. \Y. & R. G. Olmstead. 2(MK). Utility of 17 

tures in Crmlnphyllum. PI. Syst. Kvol. 8 (Suppl.): 1 75 

chloroplast genes for inferring the phvlogenv of the bas- 

al angiosperms. \nier. J. Rot. 87: 1712-1730. 

. 2001. The llowers iii e\iani basal angiosperms 

Crevson. R. 1001. The Development of flowers. Oxford 

„l nil. ii in in ii il If ■ mi 1 I'l s ; i 102: 

Univ. Press, New York. 

1 lambs. 1! K. X K. A. Zin.mer. 1002. Ribosomal RN\ as 

& L. Hufford. 1989. The diversilv of stamen 

a phvlogenelic lool in plant -v sleinalics. Pp. 50-91 in 

-Iru. line- and drlu-reiice pallems among Magnoliidae. 

Bot. J. Linn. Soc. 100: 4.5-85. 

ilia, Systematic- of Plants. Chapman and Hall. New 

& A. Igersheuu. I'XIT. Cvnoecium diversilv and 


.,ys|emalic, ..fll.e Ian, ales. Rot. J. Linn. Soc. 125:03- 

Hayes, V., K. L. Schneider K S. Carlq.iisl. 2 Floral 

& . 2CXX). Cvnoecium structure and evo- 

J. PI. Sci. 161 (Suppl.): S1R.VSI0I. 

lution in basal angiosperms. Int. J. PI. Sci. K)l (Suppl.): 

Hiepko. P. 1065. Yergleiehend-morphologiselie ,,iid rn- 


twickluugs-geschiehlliche I nlersii. Iiungen iiber das 

& K. B. Sampson. 1083. Floral structure and re- 

Perianth bei den Polvcarpicae. Rot. Jahrb. Syst. 84: 


Arbor. 64:' 447-473. 

Hoot, S. B., J. W. Kadereit, F R. Rlattner, K. R. Jork, A. 

, A. Igerslieim. K. 15. Sampson ci C. E. Schal/. 

E. Sehvvarzbach v^ P. R. Crane. 1007. Data , onm u.-u. c 

il M i 1 "i 1 Hs -\s|, main 

and phytogeny ol the I'apaveiaceae s.l. based on lour 

position in Wnilrraceae. \nu. Missouri Rot. Card. 87: 

data sets: <///>R and rh<\< sequences, IrnK restrielion 


sites, and morphological chaiaclers. Svst. Rot. 22: 575- 

abar. C. & P. Feins. 1981. /,„ Spirale in Magnolin- 


bluten. Beitr. Biol. Pflanzen 56: 225-211. 

Huber, H. 1993. Aristolochiaceae. Pp. 120-137 in K. Ku- 

bitzki, J. (T Rohwer X \. Ritlrieh (editors), The Kami- 

bei einigen Magnoliiden. I5..I. Jahrb. Syst. 10.5: 433- 

lies i„,l Gen ra ol Cscala, Plains \„| 2 tlonenng 

. 1985. Studien zur Organsequenz in 

i. Bot. Jahrb. Syst. 105: 379-400. 

i & P. Leins. 1999. Development and 

alomv. John \\ ilov X Sons. New 

vophvllid Families. 

springer. R. 


Huber, K. A. 1980. 

^ 1 > • i , « i l ' " 1 « - *-■ I ~ 

an 81 

slanden von Solan 

Lindl. (Nnlanaeoae): Dissertati* 

ones 1 

Cramer, Vadus. 

Hufford, L. 1996. Thi 

- morphologi 

3 . R. Crane & S. Blackmore (editors), Evolution, Sys- 
ematies, and Fossil History of 'the Uamamelidae. Vol. 
I. Clarendon Press, Oxford; U.K. 

rsheim, A. & P. K. Endless. 1667. Gvnoecium diversity 
u.d systematic* of the Magnoliales and Winteroids. Bot. 

rol, K. G., Y. Suh, G. E. Seliatz & E. A. Zimmer. 2(KM). 

Vlolecular evidence for the phylogenetie position of 

'hkhtajanid in the Winteraeeae: Inference from nuclear 

ibosomal and chloroplasl gene spacer sequences. Ann. 

Missouri Bot. Card. 87: 414-432. 

wler, P. J. A. L993. Menispermaceae. Pp. 402-418 in 

k. Kubitzki, J. G. Bohwer & V. Bittrich (editors). The 

■ainilies and Genera of Vascular Plants, Vol. 2. Flow- 

■ring Plants: Dicoh ledons: Vlagnoliid. 

,.ii\c.|ili\lli(l families. Springer. Berlin. 

-uge. k. PIM1. |V,;,I eyolnlion i„ Banune 

nst. Evol. 152: 185-191. 

'.o-Poljanski. B. VI \'*22 \ l,:i u i i, -■■ 

'a.ner. E. M. & 1 
I 11-118. 

flowers. Taxon 36: 21-28. 
, J. G. Holme, K \. Bitlrieh. I<)<>3. The Families 

and Genera of Vascular Plants. Springer. Berlin. 
Kuzoff, B. K.. J. V. Syyeere. I). F. Soltis. P. S. Soltis & F. 
\. Zimmer. I99B. The phy log.-uetic | -utial of euliie 

2dS ,|l\\ M-quriic,-. in plants. Mnlec. Biol. I AnL 15: 

Feins. P. X (.. I«WJ5. En. Beitrag /... Blulenen- 
twicklung der / 

/u den Mo.M.kntyl.-n. Bot. Jahrh. Sysl. 107: 
Fes. I). II. |<)88. The origin and affinities of I 

and IBS rDYV data. Sysl. Bol. 2 I: 28-16. Cladi-lio 7: 267-296. 
yons-Weiler. J.. G. A. Iloelzer K B. .1. Tail- li. !«»«»(.. 

Bel.Hiye Vpparcnl Sy napomoi pin VnaBsis (I! \S U I 

The >lali-lical measure ol phylogenetie -igiul. Molee. 

Biol. Evol. 13: 749-757. 
lahee. P. M.& J. Humphries. I 663. < ioding pol y moiplue 

dala: Examples from allozymes and ontogeny. Syst. Biol. 

llliple islands of mosl |M. in..:, , ,„-. - -■ I / ,, 

: 315-328. 

ers in phvloge.lelic analysis. Syst. Biol. 12: 576-581. 

& I). R. Vladdison. 1992. MacCla.le 3.1 

y>is of Phylogcny .iixl Character Fvolulion. 
Sunderland. Massaehusells. 

Magalldn, S., P. R. Crane & P. S. Herendeen. I< 

I lel.ieeon. Saulon 
Int. J. PI. Sri. 158: 373-361. 
Winchester. S. B. 1686. Vegelatiy 
phology of an exlincl plane tie 

>. S. K VI. J. Donoghue. 1999. The root of anj 
i phytogeny inferred from duplicate phvtoclir< 
. Science 286: 917-950. 
& .2000. Basal angiospei m plnlo^em 

g petal developmenl Yilme 3<><i ^ ^^ ^ 

-, R. 1966. Studies in f 
Kew Bull. 23: 133-180. 
. B. I). 199-1. Cladislie, 



floseley, M. F. 1961. Morphological studies of the Nyi 

phaeaceae. II. The flower of !\ynif>hae<t. Bot. Gaz. 12 

. 1972. Morphological studies of Nymphaeace* 

VI. Development of the flower of Nuphar. Phytomi 

phologj 21: 103-252. 
, I. J. Mehta, P. S. Williamson & H. Kosakai. 198 

85: 137-212. 

Vllieje. I. | <);,(,. II,, ||„. problem of ill igm .11, <| phy 

logenetic development of the Angiosperms. Acta. Mus. 
Nal. Prague 1 2B: 64-144. 

Nixon. K. C. & J. I. Davis. 1991. Polymorphic taxa. miss- 
ing values and cladislie analysis. Cladistics 7: 233- 

Padgett. I). J., I). H. Fes & G. E. Crow. 1999. Phyloge- 
nclic relalionships in \uphar ( Nx mpfi.i. ,i< e,,e|: I \ i 
deuce from morphology, chloroplast DNA, and nuclear 
nbosomal D\\. Vine. J. Bol. IK,: 1310-1324. 

Parkinson. C. I.., K. F. Adams & J. I). Palmer. 1999. 
Multigene analyses identify the three earliest lineages 

Pigg. K. B. >\ I!. \. s, (M . kt . N . |.«,|. I' phinl- 
from the Palaeocene of Vlherta. Canada. Bey. Palaeo- 
bot. Palynol. 70: 125-146. 

Posada. I). K K. VCrandall. I 998. Uodeltest: Tcsling the 
model of DNA subslilulion. Bioiulornialics 11: 817- 

Volume 90, Number 2 

morphological pro! 

Y.-L., M. W. Chas< 

olecular phvlogenel 

of the Magnoliidae: Cladistic 
nucleotide se(|iiences of the plastic! gene 
lissouri Bot. Gard. 80: 587-606. 
. E. BemaMoni-Ouadroni, I). K. Sollis. P. 
Zanis. K. A. / Z. Chen. \. Savolai- 

I ., I i . 

1 (Suppl.): S3-S27. 

VI. J. Donoghue. 100,'!. St. p mall ices an. 

ilmn of I opl.isv. Svsl. Biol. 17: 582 

<tebhiiis. (,. I. 1071. flowering Plant-: K\olution Above 
die Species Level. Harvard Univ. Press. Cambridge. 

itevens. P. I. 1001 Ch.ii.u In -lale-. morphological vari- 
ation, and phslngeiielic analvsis: A review. Svst. Hot. 
16: 553-583. 

nvamv. B. (;. C. 1949. Further contributions to the mur- 
ph« log: ,:1 Degeneriaoeae. J. \mold \.l«.r. 30: 10 :;!', 

nw.fiord. I). I.. 100!!. |'\| p! 1.0: l>l,v logenelic Analysis 

Using Parsimony. Sinauer. Sunderland. Massachusetts. 
, G. J. Olsen, P. J. Waddell & I). M. 11,11,-. |00 ( >. 

Phvlogenetie inference. I',, 107-127, in I). VI. Ilillis. 

( .. Mnril/ X B. K. Mable (editors), Molecular Svstem- 

alics. 2nd ed. Sinauer. Sunderland. "" 


persal. <>|u,., and Boyd. Edinburgh. 
. 1980. Outline of the classification of How 

plants iVIagnoliophvta). Bot. Key. 46: 225-359. 
. 1991. Evolutionary Trends in flowering PI 

Columbia Univ. Press, New York. 
. 1997. Diversity and Cla-ila . u »l I low 

Bo, K.. C. S. Kernel K B. \. MePheroii. I'>07 \1 .!ee,ihi 

I i log. [i i, I <>!' ih I if < a. I 1 i I i 1 > ol tin 

nuclear 26S ribosomal DNA in inferring mi. I , i "I 
relationships. Molec. Phvlogenet. Evol. 8: 117-127. 

Bobertson. K. E. & S. C. Tucker. 1979. floral onlogem 

. , I., II, Ml « I lo' ... 005-617. 

Savolainen. V.. M. W. Chase. C. M. Morton. S. 8. Hoot. 
I). E. Soltis. C. Haver. M. E. lav. A. De Hri.ijn. S. Sul- 
livan iK Y.-l.. (Jin. 2000. Phvlogenetics of (lowering 
plants based upon a combined analvsis ol | I 
and rbcL gene sequences. Syst. Biol. 49: 306-362. 

Smith. G. IE 1028. Vascular anatomy of Banahan flowers 

. 1,2 177. 
Soltis. I). E.. P. S. Solt.s. VI. \\. Chase. M. E. Mort, D. C. 
Albach, M. Zanis. V. Savolainen. W. IE Hah... S. H. 

Ilool. VI. E. Fa*. VI. \Mell.S. VESwensen. E. VI. Prince. 
\\. J. Kress. K. C. Nixon K J. S. Earris. 2000. Angio- 
-perm phvlo^env inferred from I8S rl)\\. rhr\ .. and 
«//,B sequences. Hot. J. Einn. Soc. 133: 381-161. 

, , I). E. Nickrent, E. A. Johnson. \\. II. 

Hahn. S. H. Hoot. J. A. Sweere. B. K. Ku/oll. K. A. 
Kron. \E \\. Chase. S. M. Swensen. E. A. Zin.mer. S.- 
VI. Chaw. I .. .). Gillespie. \\. J. Kress ,K K. J. Sytsma. 
1007. \niri.>s|..T.n phMogrm inferred from 1 8S ribo- 
somal IA\ sequence-. \u.i. Missouri Bot. Card. P, I : 

cots, and exhibit floral features of 
dicots. \ J. Hot. 90: 461-473. 
Soltis. P. S.. I). E. Soltis & M. W. Chase. 1999a. Angio- 

-p I"' ph\l u\ iuf -i red li:>ni iiiiiitiple "« e a- a < 'I 

forcomparative'biologv. Nature 102: 102-101. 
, . P. G. Wolf. D. E. Nickrent. S.-M. Chaw 

I li | in in |000|, II h, j li-.l .• .i -I in, 1 , in 

n ii 'I ii .m I'M |ii I. I 1 'i i. li I 

ol rDNA signal? Molec. Bi< 

. , M. J. Zanis 

eages of angiospi 

for floral evolution. Int. J. PI. Sci. 161 (Suppl.): 


'>'"^'"- ' 

. Basal lin- 

. Nnel. \e,.ls It. s. 

■| Mien . ..i d >el ' :-i | lie '.I.-. Pp. I I I- I .5 

in I). \\. BnlorcN. E. J. I liekey icdilors). Flowering Plant 

Origin, Evolution and Phylogeny. Chapman and Hall. 

New York. 
Van Heiisekom. C. I. 107 E Bevision of Meliosma (Sabi- 

aceae). -eclion h>ri>nz<mra excepted. living and lo— il. 

ucngrapln and ph\|o-en\. Bliiinea 10: :i55-520. 
\eleno\skv. J. 1010. \e.gleichende Morphologic ilei 

Pllan.-en I \L> pin ■ gie . I. I Itiiilc ■ !■ I I'll, . c ii^m e 

4: 733-1211. 
Vink, W. 1970. The Winteraceae of the Old World. I. 
■'•in and fhinns — Morphology and taxonomy. 

Blumea 18: 225-354. 
. 1977. The Winteraceae of the Old World. II. /».«- 

ogynum— Morphology and taxonoim. Blumea 23: 21" 

. En.lre 

. 2002. 

aceae. Int. J. PI. Sci 163: B 17-876. 

Walker, J. W. & A. G. Walker. 1984. 

Power Cretaceous angiosperm pollen a 

Gard. 71: 464-521. 

Weberling. E 1089. Morphology of Flowers and Inflores- 
cences. Cambridge Univ. Press, New York. 

Weller. S. C. M. J. Donoghue & D. Charlesworth. 1995. 

[1 ev. I i on o 'H angiopserms: A 

pliMogrnrlic approach. In P. Hod, K \. Ncphen-o,i 
(editors). Experimental and Molecular Appro 
Plant Hiosystematics. Monogr. Svst 
Gard. 53: 

Wiens, J. J. 

Wilson. T. K. 1966. The c 

Canellaceae. IV. Floral morphology , 

Amer. J. Bot. 53: 336-343. 
Vang. Z. 1097. PAME: A program package for Phyloge- 

1005. \ I1CW method ()f Cel'alophvlla, vac. < r, ul opln 1 1 II m has a vvh. 

.tide and amino acid se- dial subtend the reproductive organs, and t 
Soltis, S. Mathews & , 

. I'roc. Nail. Acad. s,,.. I >. \. o«) : (,;;iu <,;;:>.;. 

Chloranthaceae. Within the Chloranthaccae onlv the I'e- 

Outgrnup. With the exception of Curium, none of the male flowers of llcthosmum lia\e a structure regarded as 

extant gymnosperms has a perianth. The extinct Hen- a perianth (Kndress. I 087a). Thus, we scored Chlomnthus 

neltitales have a perianth. However. Crane ( 1085) sug- as lacking a perianth and Ihuhosmum as having a single. 

homologous to lh i perm | ill. M reovei phy- Kaclondaceae. I .acloridaceae have a single-whorled. 

dictum \* sister to the angiosperms (e.g.. Doyle. I')%. I.aurales. I.aurales consist of seven families: \lheros- 
1008). Thus, we have scored the perianth in llie oulgroup permataceae, ( .alveaiithaceac ( .oniorlcgaceae. Ilernandi- 
as absent. aceae. Liuraceae. Vlonimiaeeae. and Sipanmaeeae. The 
\mhorellaceac. bailev and Swamv I I '> K!) described the perianlh is complex within this clade. Also, manv mem- 
male and female Mowers in detail, noting thai llie perianth hers of lliis group have a hvpaiilhiiuii. \lherospermala- 
is not differentiated into caKx and corolla, hut mslead ceae have an indeterminate number of undifferentiated 
(imposed of lepals. The perianlh is arranged in a cyclic perianlh parts that are spiral or whorled. Calvcanthaceae 
or spiral anangemeut. and the iiumh, i ol lepals ,anges have an undid. ,. nli ,1. d ,.. m u.lli thai is .p, M ||\ a.ianged: 
from six to eight in the female flower. This work was fur- the number of parts is indeterminate (l)engler. 1072). (mo- 
ther corroborated by recent work done by Kndress and mortegaceae have a spiral, undifferentiated perianth with 
Igersheun (2000) which shows a spirally arranged undif- an indeterminate number of parts (Hri/icky, 1959). Her- 
lerentialed perianth; the nuniher of lepals ranges from naiidiaeeae have a perianlh w ilh one or two whorls of three 
lo I I in male Mowers and is , or 8 in female Mowers. to live undifferentiated parts (kuhit/.ki el al.. 1003). I.aur- 

trimerous perianth that consists of a single whorl, with lh.' each whorl cnns|..iing ol ihree paiN. Moiumiaceae have a 

exee|>lion of Siirunw. which has two whorls that are dif- complex perianth, with some members bavin- a well-de- 

ferenliated (Huber. 100.5; l.eins X Krbar. 1085). In sumc veloped perianth and others ha\ uig an enlarged receptacle 

species of Asaritm. pelals apparently begin to develop, but and diminutive perianth. The perianlh in all Unnimiaeeae 

the onlv traces are small, ihreadbke siruclures (beins X is undifferentiated and vanes ,,, uumber ol pails: some 

Krbar, 1985). (ion/ale/ and Slevenson (20001 demonstral- laxa have a spiral perianth (e.g.. ilorinnia. Kndress. 1080: 

'■<! dial the uuiseriale perianlh in \,,Mnl„rlii,i is denved however, see Dovle e, Kndress. _>00<» and olhers have a 

from the outer whorl of a biseriate perianth. whorled perianth. Sipanmaeeae have a perianth of a single 

\uslrobaile\ales. \us|robaile\ales are composed of whorl and a variable number of parls | Do\ le K Kndress 

\llsl,oba,leva, eae. I , mien ,.„ eae. S, II ,s.„„ I, aeeae. an, I II- 2000). 

lieiaceae (\l'(, II. 200.1). These families have all been Vlagnolialcs. Vlagnoliales ai v eomposeil of s, v. fami I les: 

scored as having an undifferentiated, spiralis arranged \nnonaceae. I )egeneriaceae. Kupomaliaeeae. Himanlan- 

perianlh with an indeterminate number of parts (Kndress. draceae. Vlagnoliaeeae. and \1\ nslicaceae. Annonaceae 

1983; Kndress fl Sampson. MW3: Robertson K Tucker. have a trimerous. whorled. differentiated penaiith Hollow - 

|()7<> >- ing Dovle X Kndress. 2000: Dovle fl Ke Thiimas. IW(i). 

Huxaceae. Huxaceae have unisexual Mowers. The sla- Degeneriaeeae have a differentiated. Irimerous. spirally 

minale Mowers have a perianth consisting ol luo or three arranged perianth iSwamv. I'M')). Kupomaliaeeae have a 

whorls of dimerous tepals (Drinnan el al.. I « >* > 1 1 . The pis- . alvptra: however, il is uncertain if this structure is sepal 

filiate Mowers typically have a spindly arranged perianth oi bra, I in origin (Kndress. 1077); llius. we coded it as 

with lo 20 lepals (Diinnan el al.. \<><>\\. However. Drin- ->". I limantandiaeeae have two ., pal-like siruclures thai 

nan et al. (1994) noted that the tepals in the pislillale surround the bud: however, il is uncertain if these slruc- flowers lures are perianlh or bracts (Kndress. |<>77l. Thus, we 

identical to coded llmiantaudraccac as ">". Some Magnoliaceae have 

described as dimerous and cyclic (e.g.. Drinnan et al.. e.g.. Ma.nn, »/,„). but in I niu,lcn<h„u and certain species' 

100 1: von Haltha/ar K Kmlress. 2002). We coded llie peri- of Magnolia, llie perianth is in three Irimerous whorls (Kr- 

anth of Ruxaceae as being undifferentiated and dimerous, bar & Keins. I 08 I , I 0[«; Tucker. I0(>()>. My nslicaceae 

whorls). & Tucker, 1986; Vrmstrong & Wilson, 1078). 

of the monocots as placeholders for our analyses. 

Nelumbonaceae. Nelurnhonaeeae have been considei 
to have a spirally arranged perianth with two (to five) 
pals and l<>-30 petals (Hayes el al.. 2000). 

\\in|.haeaceae. \ \ lll|)haeaceae are com|i. .-.•< 1 ol two 

ers are found in sonic genera of Papax erueeae. and some 

baeeae and \ wnphaeaeeae s. -Ir. ( \l'<- II. 2002: Pes el 

e,ous p,-riantlis are piesenl (llool el al.. PX»7l. Moieovei. 

some taxa in Papaxeraeeae laek petals (Hoot el al.. 1997). 

Kupleleaceae laek a perianlh. although the lower llowers 

in ih, nillore-i enee max iiaxe Ixxo small prophxlls ( Kll- 

,.| a!.. I'«!ll or uudilleienliated (Doxle xK Kndiess. 20001 

dress. lOiSO). Cireaeasleraeeae have llowers, uxxluel, there 

is no clear differenlialion ol Tllie perianlh. and the perianlh 

Willnn Vmphaeaeeae. peiianlli phxllnlaxi- ha- hern re- 

garded allematixolx as spiral (Albert et al.. 1998: K. 

Schneider, pers. comrn.) or whorled (Dovlc K Kndress. 

aid. trimcrous perianlh (I)rinnan et al.. 1991). Mcnisper- 

2000: K, ss. 2001: Vlosolev. 1901. 1972). \unhai and 

\\mphtira and \irtoria the diH'erenlialion i- gradual, i li« 

number of petals and sepals varies throughout die laiiulx 

'Ihncrous lUeadiani 1 !0|. I.'anmi. ul... ,.^ , Jul 

l<>00l. | hose \ „,,/„„ -peele- Willi 

be based on a trim 

to 14 petals. Ona 

petals, and \vmphaea. \ictoria. and Euryah 

sepals, with the number of petals ranging froir 

\\mph<>r,i and from 50 

are scored as hav 

are scored as polymorphic for 

Nymphaeaceae, each taxon was scored as polymorph! 

phvllotaxis and for number of perianth pa ' 


Piperaceae. Piperaceae lack a perianth. 



. 1991). (aon<|uisl (1981) emus. 

ended as hax 

developed sepal- thai alternate Willi -mall petals. 1 
and Crane (1989) coded I'latanus as having an undiffer- 
entiated perianth. Dnxle and Kndiess (2000) coded Pla- as haxing a one or Ixxo w hoi led. undifferentiated 
perianth, with the numbei of part- two. four, or five. Stud- 
ies of fossilized Platanaccae indicate that the p, nanlli ua- 
divcrse. with some members haxing a xxell-dex eloped im- 

eomposed of a single perianth whorl composed ol four 
tepals (Magalldn-Pucbla et al.. 1997: Ylagallon el al.. 
1097. 1009: Manchester. 1980; Pigg & Sloekex. 1991). 

anlh. polymorphic for number of whorls (one or two 
whorlsl. and polymorphic lor the number of parts (trim- 
crous. pentamerous. dimerous/tetramerous). 

Proteaeeae. The periantli in Proteaceae consists ol loin 
pelaloid tepals (l)ou-las K Tucker. I 090a. I99(,l,|. some- 
times rehired to as petal- (Takhtajan. 1997). We coded 
Proteaceae as polymorphic for 
-ingle whorl of tepals. or as ha 

I -pee,. 

e. The family comprise five genera 
. The llowers arc perfect and with.) 


genera. Tmcluxleiulron and Tclrucentron. The peuaiilh oi 
Tt'lnirrnlron ha- Ixxo whorls, each composed of two lepal-. 
Trochodeiuiron is usually considered to laek a perianlh: 
however. Kndress (|98<>l noted several small scales that 
may be interpreted a- a perianth, between prophxlls and 

, ./, a- la, king a perianth. 
\\ inlerai < ae lb p!,«.hi ol . i < indii il« - 

in •, lie -IK , e—l\e 

sisters to the rest of the family (Kami et al.. 20001. Pal- 
terns of variation in perianth structure and organization in 

ales. We used -e' 

lanll, Hilli >|.n;.l |.lnll. l l.iMMl>.HiM. J(MM). 2001: \ ink. la, I. ,,|ix I l„|;m> in Dunns mav wnx Ii.< -ainr planl 

"0. 1<>77). Takhtajuniu has a pnianlh [..-gills as (l)ousl. 2(K)I|. \\r thr Iannis a- li.mi.n :l ,1,11,,,-,, 

icioiis. ihcn changes In trtramerous. and evrntuall\ lit'- lialnj \»-i lanlh. 11, ran-.' all ,,l tin- .-.i 1 1 \ - 1 1 1 ; 1 1 1< ■ h i 1 1- nirin- 

„- |,.,.-r.,ns ,h„lM-,-i ;,l . 2000). iasmannia has l.rrs ol ||„- family have a uhurlol pin llotaxis. vvr <o«lnl 

" ■ ' / ' ' ' i -iii.i«Ii.iiI.-.I Wnil. i.h..i. a- uhorlnl lot plixlloluxis ami polymorphic 

MOLECULAR SYSTEMATICS, Marcus Kochr Ihsan A. Al-Shehbaz,' and 

EVOLUTION, AND Kla ™ Mummenhoff 


tin- 'context of ongoing and accumulating studies. The review i-. useful in \ie\\ ol tin- immenseb increasing work on 
■abidopsis thaliana, the model species ol plant molecular hiol.igy. and on important crop plants such as species of 
rassira. Traditional and mnlcculard.ased pin logenies are critically discussed, new generic alignments are pinpus,-,!. 
id groups in need ol m.T Mil ii ul d« lifi. I li I 'I I I. ...i inol u« n« H« - mil 

nelopment of 1. tlwlmwi is onlv \en douK creeping into the sv denial ic- of lira-ieaeeac. future directions ol 
search should move hevoml assessing generic iclal lo.iships or limits, and should also address character development 
homoplastic characters, the nature ol the genome, and many other new 

.. _ .,. . 

etailed molecular studies o 

I'.i ae< ic. (.11 

I), I 

,;il)i). Second, main spe- 
ema R. Hi., ilyssum L. 
, Draba L.. Erysimum I... 

forage and fodder for animals, or as vegetables Hesperus L.. /beds L., L>bularia Desv., Lunaria L., 
(Crisp, 1976; Simmonds. I9B0). The mosl impot- and Matthiola R. Br. are enllivated as ornamentals 
tant members belong to the genus Brassica L., in- (Al-Shehbaz, 1984). Third, Arabidopsis thaliana 
eluding varieties of B. oleracea I., (broccoli, bins- (L) Heynh. (thale cress) is considered to be the 
sels sprouts, eabbage. cauliflower, kale, kohlrabi. most important flowering plant for conducting ex- 
savoy). B. juncea (L) C/.ern. (Indian mustard). B. perimenlal work in various fields of biology, in- 
nigra (L.) W. D. J. Koch (black mustard). B. napus eluding plant genetics, physiology, development, 
L. var. napobrassica (I..) lu lib. (rutabaga). B. napus pathology, genetic engineering, and related fields, 
van napus (rape), and B. rapa I,, (summer turnip In recent years. 1. thaliana has become a model 
rape, Chinese mustard, Chinese cabbage). Other lo- system for plant molecular biology, which has cul- 
cally important crops are Lepidium sativum L. minated in the recent publication of its complete 
(cress), L. meyenii Walp. (maca). \rmoracia rusti- genomic sequence (The \rabidopsis Genome Ini- 
cana P. Gaertn., B. Mey. & Scherb. (horseradish). tialive, 2000). also available via the World Wide 
Raphanus salivas L (radish). Sinapis alba L and Web (< 
B. juncea (both used in the manufacture of table a_thaliana.html)). Arabidopsis thaliana has a small 
mustard), Nasturtium officinale R. Br. (watercress). genome, is fast growing, easy to cultivate with min- 
Corhlearin officinalis I., and Caulanane amara L space and care demands, and self-pollinating 

1 We are grateful to C. Donovan Bailee Neil \. I lainman. Tlioma- G. Lammers. Karol Marhold. and two anonymous 
reviewers for then critical comments on 'the manuscript. We thank Victoria Hollow-ell for her editorial advice and Mark 

\. r.eilslein. Waller lileek.-r. \mlieas I lan/.ke. and Herbert Ilnrka for fruitful discussions. 

- Institute of Holam. Iniversitv for Agricultural Sciences \ ieniia. Cregor- Vlen.lel Mi. T".. \-l 180 Vienna, Austria. 

Vlissm til Hot. Gahd. 90: 151-171. 200.'i. 

/or, frenetic modifications by transf'nr- to the septum), such as the members of tribe Lop 
itinely accomplished to generate mu- idieae sensu Schulz (1930). while in the majorit 
s particular genes. Stock ecu- ,,|' the family the fruits arc cither latiseplate (com 
proved parallel to llic -epliiml or llol ll.itleued lie 


mutants, populations. DYA libraries. ,-,.,,. or angled in cross section). On the basi 

YACs (yeast artificial chromosomes), and BACs length/width ratio, the fruits have been traditionally 

(bacterial artificial chromosomes) for molecular divided into silkies (length less than three times 

analysis in genetics and < A central wi( | th ) or s j|j qut . s (| (M1 , t |, „„„,. tllan ,| lrefl timrs 

, I , "' S,I,,M '"' ''"^ " - •"' h I"" ' - '■"« •» a PI^ width), but such division, though useful for iden- 

atory lines/strains of A. lhaliana) to wild plant lin |; u , Al-Sh.-hba/. 1981: \ppel c\ '\|_Sl7<hV 
•ecies. \ wealth ol i nloi iii.n ion on 1. lhaliana i 

•Ironically ({http://www.aspb. m( . |m , e embryologi( 

. is arbitrary and ha- limited pli\ l«»-«-i 
1984; A 

i ! >psis)L and it should be 

•onsulted by biologist- Iioiii , 

tar-gland morpl 

deisland the evolutionary process- .,,,'., i II.. i r 

-II- • lUIIllliei. I -..-,. ■■;! , :i il naloiin .lll.l -ml, lee. 

dth their e.r 
. work with 


from the early 19th to the i 
most notable of which are those of de Candolle 
(I 82 1), Prant I (1891). Hayek ( 191 1 ), Schulz (1 936). 
and Janchen (19 12). According to these systems, 
the Brassicaceac can be divided into am where 
from 4 to 19 tribes and 20 to 30 Becenl 

><>ini\is. annual versus perennial life cycles, and 
orpholo-i, al characters -ueh as leal architecture. 

,il diversity, development ol v b „„,. seed T ^^ {V *< ^ H i,L 2(MMa > S, W* 

oduction. and trichome morphologv. Therefore. ' Wt »^ tax„nom.«- subdivisions mostly do not re- 
i increasing number of molccuhir and evolution- |>h\ logenetic relationships. Molecular data 

\ biologist are Ionising on the study of the Jil'd sllon ^ lv ^pport a sister relationship between Cleo- 
latives of A. lhaliana (eg IVniii-i 200<h '"..idea.- << -ipparaceae) and Mrassieaceae (Rodman 
et al.. 1994). On the basis of a little generic sam- 
pling. Judd et al. (1994) proposed merging Bras- 
34o" genera and some 335(1 "specie's sira( ' ear an,J ^apparaeeae in one family, but recent 
ributed worldwide , sp, , ialb in t, i.ipera.e re- l,, '^<" la ' < )ala <■'■ < : - "all el al.. 2002) clearly sup- 
s of the Northern Hemisphere (Al-Shehbaz. l M,rl ,m " mailll, ' ll;,nn ' °' K'assicaceae, Cappara- 
i; Appel & Al-Shehbaz. 2002). The family. r, ' at *- aml ( ; l( '<»">acoae as three distinct but closely 
•h has long been recognized as a natural group, re, ated families. 

ell defined morphologically and has a fairly A Letter understanding of evolution within Bras- 

orm flower structure. The flowers are radially *"accae can be achieved only by a comprehensive 
metrical and consist of four almost always free review of sll,,hVs 'anging from the population level 
lis in two whorls, lour free petals (though some- '" analyses from the iiilraspecifie level to the entire 
s lacking), often six. free, telradvnamous sta- family. This should lake into account data from 
s (outer 2 shorter than inner 4: though mam analorm. morphology, .horology, cytology, popula- 
tion biology, ecology, and molecular systematics. in- 
cluding results from studies of A. lhaliana. The 

relevant literature from the past decade on inolec- 
slandingof phvlog. nv. evolutional v trends, and pal- 

species of th« 

• large genus h'pid, 

turn I, have four or 

only two stan 

lens), and a biearpellale ovaiv almost 

always with a 

false septum (a pa 

rlilioii divi 

.11 ng .1 ■ 

ovary into tw( 

» locules). On the. 


exhibit enormous 


in size. 

shape, and s 

tructure. They are 

the sour- 

e of the 

most important diagnostic charai 

sters for the delim- 

itation and identification of tax; 

s ranks. 

The fruits an 

J often dehiscent. 

2-valved , 


divided longitudinally by a septum, though 

In mam 

genera the fin 

and/or tin 

is totally lack 

ing. Some groups a 

re characterized by 

The review specifically identifies new challenges in 
the systematics and phylogenv of Brassicaceac. 
proposes a few generic alignments and groupings, 
ami determines group- mo-l in need of studies. 

Review of Brassicaceae Researc 

The Dii.mii dC polyploids ami llic mechanisms be- 
hind the establishment of newly evolve! popi la 
lions and ta\a air anions the main challenging 
questions in plant sciences iliamsev t!\ Scheinske. 
IW«; Thompson cK l.umaret. 1992; Petit et al.. 
1999; Soltis & Soltis. 1993. 1999). Recent studies 
of the overall genome -tincture of hybrid and poly- 
ploid taxa provide new insights about tin- dynamic 

these studies were generated on the basis of arti- 
ficial hybrids or by comparali\e mapping (Biese- 
berget'al.. 1999: Bieseberg «!x binder. 1999: Pag- 
ercrantz. 1998: kowalski et al.. 1994: Axelsson et 
al., 2000). Comparative genome analysis of some 
Brassicaceae (Acarkan et al.. 2000) has shown that 
structural rearrangements occurred with a signifi- 
cantly higher frequency m polyploid Hrassira than 
in diploid Arabidopsis tluiliana or Capsella rubella 
Rent. By contrast. Axelsson el al. (2000) have 
shown thai the genome ol I',, jiuirra remained al- 
most unchanged since the species originated as an 
li, ii I I nil" i ■ and /.'. m/Hi. 

Unfortunately, such studies air based only on a few 
cultivated species ol lUassun and on 1. llialiana. 
and hardly anything is known about other poly- 
ploids in the family, especially those with very high 
chromosome numbers such as Cardamiuc diphylla 
(Michx.) A. Wood (2/i = 256), which has the high- 
est known numb, i in brassicaceae (see Al-Mieh- 
baz. 1984). Polyploidy is widespread in the Bras- 
sicaceae. occurring in at least 37% of the species 
(Appel & Al-Shehbaz. 2002): some genera (e.g.. 
Crumbe P.. Moric audio PC. \clla P.) appear to be 
exclusively polyploid (Al-Shehbaz. 1984). If we 
consider the diploid species of lirassica and other 
genera (e.g., Physaria (Mutt, ex loir. & A. Cray) A. 
Cray, including l.rscjiicrclla S. Watson) as "palaeo- 
polv ploiif-" (see below), the percentage ol p.«lv ploid 
taxa in the family will be at least 50% (authors' 
compilation), \llopolv ploidy is probably the most 
common mode of hvbrid spoliation in the Brassi- 
caceae and accounts for the majority of polyploid 
species (authors" compilation). However, phyloge- 

(ii polyploids occurred al rales higher than in dip- 

Molecular methods have helped to elucidate the 
complex evolutionary history ol inanv illopob plo> 
groups in the family. Arabidopsis sitcrica (Pr.) Norrl. 
(« = 13). an allopolyploid species derived from A. 
ihaliaua (u - 5) and 1. arcnosa (P.) Paw bee (n = 
8) (Mumnienhoff \\ llurka. 1991. 1995: Kamm et 

al.. 1995: (Tkane el al.. 1997: Comai el al.. 2000). 
is one of the best studied species of hybrid origin. 
Other studies include lirassica (Palmer et al., 1983; 
Prickson el al.. 1983: cK Pvdiate. 1990). 
Arabis holboellii llorne.u. (Sharbel & Mitchell- 
Olds. 2001; Koch et ab. 2003a). Capsella Me.lik. 
(Mumnienhoff & llurka. 1990; llurka & Neuffer. 
1997). Cardamiuc Xiusucta Urbanska and C. 
scbuhii I rbanska (I rbanska et al., 1997; Franzke 
o, Mumnienhoff. 1999). Cardamine Xeuriquei Mar- 
hold et al. (Marhold et al.. 2002b). Diphlaxis DC. 
( Mum-,,,,, holf et al.. 1993). Ihaba (Brochmann et 
al., 1992a, 1992b. 1992c; Wiclmer & Baltisberger. 
1999a), Nasturtium R. Br. (Blocker et al.. 1999), 
and Rorippa Scop. (Bleeker & llurka. 2001). More 
rccenllv. research focusing on hybridization and po- 
.. i ainonj , ruciferous plants has been 
attributed additional importance because of the 
successful crosses between the model organism \r- 

diana and its w ild relative A. Ivruta 
(P.) O'Kane & \I-Shehbaz subsp. pvtrava (P.) 
0'Kane & Al-Shehbaz (u = 8) (see Nasrallah et al., 

A knowledge of the patterns of speeialion and 
species relationships within a group of plants is 
critical for the understanding of its morphological 
evolution. I.cpidium. vx liuh exhibit- , Ilmal di- 
versity than anv othei genus ol brassicaceae. shows 
that more than hall ol ils species lack the lateral 
stamens and most ol these also have reduced pet- 
als. The species with reduced flowers are distrib- 
uted primarily in the Americas and Australia/New 
Zealand. Prey ion- phvlogenetic studies with non- 
coding regions of cpDNA and rDNA ITS showed 
incongruences in mosl New World species relation- 
ships (Bowman el al.. 1 9O0; Mumnienhoff el al.. 
2001a). This, combined with the presence of many 
polyploid species, implied a reticulate history of the 
genus, but did not provide enough information to 

To address this question more thoroughly, sequenc- 
es of the first intron of a single copy nuclear gene. 
PISTILLATA (PI), were determined' from 43 Iri- 
dium species. The phvlogenetic analysis of the PI 
intron siiggc-is thai many species in the New World 
might have originated Iron, al lopolv ploidi/al ion cor- 
related with floral reduction. Interspecific hybrids 
were generated to I. -si this hypothesis, and the phe- 
notypes of P, flowers indicate allelic dominance ol 
the absence of lateral stamens. This suggests that 
propagation of dominant alleles through interspe- 
cific hybridization could account fc 
of the allopolyploid species without lateral s 
(Pee et ab. 2002). 

Hybridization and reticulation have been 

Missouri Botanical Garden 

concerted evolution of llic internal 
spacer regions 1 and 2 (Koch et al.. 2(K 
case it is remarkable that a|n 
played an important role in specialion 

„,! dii 

Polyploidy also played an important role in the 
evolution of species complexes. particnlarl\ .luring 
time periods greatly influenced by glaciation and 
deglacialimi. flu- holds true lor \nd>idopsis siteeica 
in Scandinavia (Mummenhoff & Hurka. 1991, 
1995; O'Kane et al.. 1997), the Cardamine prote- 
sts complex (Franzke «!x Hurka. 2000). Cardamine 
amara (Marhold et al.. 2002a). and Cochlear!*, L 
in Kurope (Koch et al.. 1900. 1 998a; Koch, 2002). 
Furthermore, diversification of most species of Dra- 
in, (Koch & Al-SI.ehbaz. 2002; Brochmann et al.. 
1992a. 1992b) is best interpreted by reticulation 
and polvploidi/.alion during the last lew million 
years. In the absence of significant glacial influ- 
ence, however, hybridization, reticulation, and 
pobploiib apparently have played an important 
role in the evolution ol old species complexes in 
the genera Yinshanio Ma & Y. X. Zhao in China 
(Koch & Al-Shehbaz. 2000) and Ylicrothlaspi K K. 
Mey. in the Balkans (Koch et al.. 1998c; Koch Ox 
Hurka, 1999). 

origin of (alloploid) hybrid taxa. But techniques 
such as genome mapping and OTI. analysis will 
allow us to sludv main detail- of hv bridi/alioii and 
speciation (Bradshaw et al.. 1995: Rieseberg et al.. 
1995). It is also important to measure in future 
studies the relative filne— nl hvbiids and parental 
taxa in a range of habitat- to further elucidate the 
role of natural hybridization (Arnold. 1997). Fur- 
thermore, to lest the evolutionary importance of hv- 


and genetics. Numerous phenotvpic traits have 
been analyzed in Brassicaceac. including local a«l 
aptalion aero-- climatic gradients in \rahis feeunda 
Rollins (VlcKav et al.. 2001) and Capsella (Hurka. 
1900; Hurka K Neuffer. 1907: Neuffer & Hurka. 
1900; Neuffer^ Hoflrogge. 1999). survivorship in 

■ibis laevigata (Muhl. ex Willd.) Poir. (Bloom et 
. 2001). glucosinolate accumulation during plant/ 
l in Arabidopsis tlioliono (Kliebeu- 
stein et al.. 2001). and herbivore resistance and 
pollination in Hrassiea rapo (Strauss el al., 1999). 
Capsella bursa-pa<H»ris (I..) Medik. was studied for 
maternal effects upon germination behav ior (Neuf- 
fer & Koch. 1996). mechanical stress (Neuffer & 
Mevei-Wall. 1000). flowering time (Neuffer & Hur- 
ka. 1980; Neuffer & Barlelheim. I9H9; Neuffer. 
1990). and leaf morphology (Neuffer. 1989). De- 
tailed analysis of host-pathogen interaction and 
their coevohition have been aiulv/ed in Brassica- 
ceac (Conslanti.iescu cK Fatehi. 2002). particularly 
in \rabis (Boy. 2001). and studio have focused on 
plH'iiotvpic plasticity, reaction norm, and ils evo- 
lution in A. thaliana (Pigliucci <S. Bvrd. 1998: Pig- 
liucci & Schmilt. 1900; Pigliucci el al.. 1999; Pig- 
liucci & Marlovv. 2(HH: Pigliucci. 2002; Pollard et 
al.. 2001). Mam of these im estimations have also 
used molecular markers, isozymes, or B \PDs to 
study intra- and inlerpopiilalioiial dilleiviiliatioii 
(see Appendix 1, parts II and III). 

The majority of these investigations have, how- 
ever, neglected the genetic diversity stored in the 
soil seed bank. Depending on the Ivpe of seed bank 
and the reproductive biology of the selected spe- 
cies, the seed bank may play an essential role in 
the recruitment and establishment of new genera- 
tions. The spalial genetic structure of the subpop- 
ulalion- (surface and ahoveground populations) 
represented by seeds may account for significant 
changes in the gem-tic constitution of plant popu- 
lations during their history. The soil seed bank not 
only reduces the rate of genetic erosion (e.g.. via 
genetic drift), but also may compensate environ- 
mental changes in space and time. Of the seven 
known studies focusing on seed-bank genetics 
(Tonsoret al.. 1903: Alvarcz-Buv lla & Carav. 1994: 
McCue cV lloltslord. 1008; Main el al.. 1999; Cab- 
in. 1996; Kvans et al.. 2000: Koch el al.. 2003a). 
the last three were conducted on members of the 

3. Cow i ta. i ■ yi Fvou no\ or \1< >i;i-iii )i cx.n \i 


Studiks Contradict Thaimtionai. Svstkms ok 
Tribal Classification 

Traditionally, the Brassicaceae have been divic 

ed into tribes and sublribes on the basis ol reh 
tively few characters (morphology, embryology 
Major disagreements are readily observed by con 
parison of the systems of Havek (1911). Schul 
(I9.-U)). and .lau.l,,,, ,PH2l. \l,,lcc,la. .lata den 

Volume 90, Number 2 

In 1 1n- family. 

tionally in these three classification svslcms docs 

in. I prov idc |.li logon in i I i i i live characters 
to distinguish several groups (Koch el a I.. 2(H) I a j. 
The major drawhaek of all ihe classificalioii sys- 
tems proposed thus lar is their almost complete fail- 
ure to accommodate convergence ol niorpl <>h>gi, ,i 
characters. In fact, molecular studies reveal ho- 
moplasv in almost even conceivable morphological 
character in the lhassicaceac (Price et al.. 1994; 
Zunk et al., 1996; Mummenhoff et al, 1997a; Koch 
& Mummenhoff. 2001; Mummenhoff et al., 2001a. 
21)01 li). Lor example. gamo-epalv. which is rare in 
the family, has recently been shown to have evolved 
independently in at least 12 genera (Al-Shehbaz. 
2001). As for fruit morphology and cotyledon po- 

sively used in every facet of taxono 
convergence is so widespread th, 
should be taken in classification i 
of taxa solely based on these eharact* 

Artihidopsis Heynh 
ized by linear an 
(•homes, and eoty l.-dous that are primarily aeeinn- 
benl. Recent molecular studies (Koch et ah, 1999a. 
2000, 2001a; Al-Shehbaz & O'Kane, 2002a; 
O'Kane & Al-Shehbaz. in prep.) have pro\ ided am- 
ple evidence that both genera are poly phy lelie. and 
an overemphasis on these homoplastic characters 
is the main reason for their artificial delimitation. 
(.oiiverseK. |a\a that appear to be quite diflereiil 
morphologically may in fact be very closely related. 

m al diversity in habit. Irichome type, flower color. 
in i - i, i i i, J "il , I, in u I olngy. but mo- 

lecular data clearij support itsmonophyb (KoehvK 
Al-Shehbaz. 2002). For Thlaspi I... molecular data 
(Mummenhoff K Koch. 1991: Zunk et ah. 1996; 
Mummenhoff et al.. 1997a. 1997b; Koch & Mum- 

ol i> I, _'U(»I I piovide strong support for the rec- 

ognilion of several segregates as proposed by Meyer 
(I97.'5. 1979) based on seed-coat anatomy. The mo- 
lecular phylogeny of Thlaspi is not congruent with 
the traditional classification of Schulz (19.50) based 
on fruit form. Several lineages in the molecular tree 
include Thlaspi species with diverse fruit types, 
whereas species with the same Iruil tvpe belong to 
diflercnt chides, thus prov idmg strong evidence for 
convergence in fruit traits. As for h-piiliiim. the 
cpl)V\ sc(|uence phylogeny one.- again demon- 
strates the widely artificial nature of the traditional 
classification schem 

this genus (Biiiggciiai n. 2001): Miiinine nholl el al.. 
200 1 a I. Numerous older examples of honmphisv in 

e an be demoiist rated in Cixhltnna L. (Koch et al.. 
1 000b; Koch & Al-Shehbaz. 2000). Arahis and Ar- 
ahidnpsis (Koch el al.. 1999a). Ihalm (Koch & Al- 
Shehba/. 2002). Cardamine L. (including Dcnlaria 
L: Franzke et ah. 1998; Sweeney & Price. 2000). 
Mirmlhlaspi (Koch & Mummenhoff. 2001), and the 
halimolobine clade (Bailey et ah, 2002). 

Some traditionally recognized tribes have been 
shown to be polyphyletie, including Arabideae e 
Lepidieae (Koch et al., 1999a, 2000, 2001a) i 
Sisymbrieae (Koch et al., 2001a; Bailey et 
2002). Moreover, boundaries between the tril 

Alvsseae are highly artificial (Zunk et ah, 1993; 
Price et al.. 1991). The tribes Brassiceae. Thely- 
podieae. and lepidieae have long been thought lo 
be natural groups (Al-Shehbaz. 1973. 1981: Zunk 
el ah, 1999). but the results of Koch et al. (2001a; 
strongly suggest that the Lepidieae are polvphy lel- 
ie. This tribe is based solely on the presence of 
aimiisliseptate fruits (Schulz, 1936), a feature that 
in fact evolved independently within lepidieae. 
brassiceae (e.g.. I'svchinc Desf., Schoiiivia DC), 
and Thelypodieae (Ctudanthus californicus (S. Wat- 
son) Pav-on). Moreover. Capsclla. Ijepidium, and 
Thlaspi s.h. which have been placed in Lepidieae 
in every system of classificalioii of the family, are 
eleailv uiuelaled. and Capsr/la shows more allini- 
lies lo \rabidopsis than to either Iridium oi Thlas- 
pi (Koch et ah, 2(K)la; O'Kane & Al-Shehbaz, in 

Moiiophvlv ol Ihe Prassiceae is shown in a series 
of studies summarized in Warwick and Black 
(1997a. 1997b). Members of this tribe are charac- 
terized by haying segmented (heteroarthrocarpous) 
fruits and/or conduplicate cotyledons (Appel >!\ Al- 
Shehbaz. 2002). and there is strong evidence that 
reversals in one or both characters have taken 
place. Segmented fruits and conduplicate cotyle- 
dons are among a few characters in the family 
where convergent evolution apparently has not oc- 

The Heliophileae. a tribe comprising six South 
African genera that mostly have diplecolobal coty- 
ledons (Appel .\ Al-Shehbaz, 1907), are also likely 
monopliy lelie. Preliminary ITS sequence data 
(Mummenhoff el al.. unpublished) strongly support 
the monophyly ol Heliophila L. and related genera. 
However, the phy logenetic relationships within He- 
liophileae are nines,, Led. hii, I ihe sinallei genera 
llniflncarpara DC. C\clopl\chis L. Mey. ex Solid.. 
St ■hlrrhleria Bolus. Silicularia Compton. and Thins- 

nltila and might not be laxononucallv distinct 
m it. The last genus exhibits far more diversity 
habit, leal morphology, (lowers, fruits, and seeds 
m any other genus in Brassicaceae. 
Outside of the Brassieeae and Heliophileae, the 
yledonary position is unreliable in tribal dela- 
tion. Within several genera (e.g.. Erysimum) or 
'I' speries (e.g.. h-pidium rirgiiiicum L) one 
ds both aeeunibent and incumbent cotyledons. 
■ most common types in the family, 
ippears thai except for these two tribes, the re- 
inder of the family is not readily divided into 
ge monophyletie groups based solely on single 

tribe Brassieeae ha\e identilied six reasonably 
well-di-lnied groups corresponding to sublribes. 

Ihmevei. laxc n\ (e.g.. Coniez-Campo. 

1W9; Comez-Campo & Brakash. 1999) did nol 
lake into consideration the overwhelming in,,le. ulai 
data that ha\e accumulated over the past decade. 
.1111 1 Hias.sHa. Diplota.xis, Enicaslriim, and Sinapis 
continue to be artificially ilelinnled following 
Schulz (1936). The ' 
era (e.g., seeds 

vs. ebracteate) exhibit enormous homo|)lasv 
throughout the family, yet they are taken in these 

ulardata (Warwick K black. I 9971,). species of all 
four genera fall into two inonophy lelic groups, the 
nigra and rapa chides, thai also include many 
smaller or inonotypic genera. In our opinion, laxa 
within the rapa anil nigra chides need extensive 
studies lo establish nionophv lelic groups and to ic- 

ualion of morphology. Warwick and Black (1997a) 
also clearly demonstrated that Cakilc Mill.. hides 
mils Desv.. and Enicaria Caertn. form an unre- 
solved chide, (.a/ale is distinguished primarily by 
having corkv (instead el nou-coikvi hints and ob- 
solete (vs. distinct) styles. Corky fruits evolved in- 
dependently in some species of Crambe and AV 
p/ianns. and this I, -aline i-, unreliable lor generic 

three genera were maintained by Appel and 
Shehbaz (2002), further studies may prove that th 
are congeneric. 

Molecular data (Warwick & Black. IW4; Cres 
et al.. 2000) provided ample evidence that Bolei 
Desv. and Euzomodendron (loss, are nested will 
U'Ua, which prompted Warwick and Al-Shehli 
(1998) and Appel and Al-Shehbaz (2(K)2) to re 
ognize one genus, Vella. Of the entire Brassicet 

ol .1 I ,. However. Ooniez-Campo | I 999 : |«)) 

tincl sets of adaptive characters . . . lor seed dis- 

distiuet vs. vestigial seed wing, seed number, fruit 
dehiscence) are also homoplastic elsewhere in the 
lanulv and. therefore, should not be overempha- 
sized at the expense of the extensive molecular data 
now available. In (act. Iruil iiidehiscence evolved 
independently iiuinei oils limes 1 1. .| ,,ul\ within 
Biassieeae but in the entire family (Appel <X Al- 
Shehba/. 2002: Mim.n.enhoff el al.'. 2(M)la). Of the 
.'*..' 1 7 genera of Brassicaceae recognized bv \ppel 
and Al-Shehbaz (2002). 125 are monotypic and 
about an additional UK) contain two lo lour species 
lohgohpii |. \lol( cnlai -Indies -how that several of 
the monohpic genera are indistinguishable from 

olhei larger getieia. I'ol example, both \tfallis IMiil. 

ami Tivisselmannia Al-Shehba/ are nested within 
and should be united with Tropidocarpiim I look.. 
wheieas llugueninia Rchb. should be united with 
Desrurainia Webb K Berlhel. (I'rice c\ \I-Shehbaz. 
20(«). Other examples include the union oUaco- 
carptts O. K. Schulz with Sisymbrium and \eobeclda 
Oreene with Rorippa (Price, pers. coiiiin.). Ilulruin 
and Euzoinadeadrnn with \ella (Warwick & Black. 
1991: (Irespo el al.. 2000). hi Oarn.-Jones K \\ \. 


Cardamine (Mitchell 


2002). Paehyphragma (DC.) Rchb. 
Krai with Thlaspi s. sir. 
2001b). and Drahopsis k. koch with Drain, (/ 
Shehbaz i\ koch. in prep.). We believe that up 
a combination of molecular studies and critic 
evaluation of morphology the vast majority of mor 
typic or oligolypic genera will eventually be nierg 
with other genera. 

Molecular data indicate that the largest gem 
of the family, including Draba (350 spp.). U>pidii 
(ca. 220 spp.). Cardamine (200 spp.). Erysimi 
(180 spp.), and Physaria and h'squercUa combin 
(100 sp|).). are monophyletic-. Studies are need 
monophy ly of \l\ssum (170-1' 
•lulling Meiiiaeiis Desv. and Plilalriefiiim 

A. Mey.. ami Rnrippa Scop. (73 spp.). /'.'/ \suuum i- 

I, ' KnnicalK one of ' ie most dilli< nil -nicra of I Lit - 

lamilv. ami il is much in need of extensive phylo- 
gem-tic -1uilii->. Phylogetielic -Indies on Sisym- 
brium (Warwick el aL 2002) show thai the genus 
is polvphvlelic. that il should consist of 43 Old 
World and I New World species, and that the re- 
maining 10 species in the Mew World belong to 
other genera allied to the I 'hel\ podieae sensu \l- 
Shehbaz (1973). 

As for Arabis, molecular studies (Koch et al., 
1009a. 2000: O'Kane & Al-Shehhaz. in prep.; 
Price, pers. comm.) clearly demonstrate that it is 
pol\ph\lelie and thai most of the species recog- 
nized by Rollins (1993) should he assigned to 
Hoechera A. Love & I). Love. Turrit is I... Pennellia 
Nieuwl.. and \rubidopsis. whereas some of the Ku- 
ropean species (sensu Jones & Akeroyd. 1003) 
should he assigned to Fourraea Center cK Hun let 
and \rabidnpsis. \rabis lias heen liadilioualb de- 
! 1 1 1 1 1 !• < I -o|i-|\ h\ having hranched 1 1 •ichonies. Hal 
and linear fruits, and acciiiuhent enly ledons. a com- 

ol IJra>sjcaceac The entire \rabis complex is much 
in need of further morphological and molecular 
studies, but all of these segregates do not appear- 
to be closely related to the core of ' Arabis that in- 
cludes the l\pe specie-. I. tdpina I .. 

Kxtensive studies on Lepidium (Hriiggemann. 
2000; Vluinnienhoff. 1003; Mummenhoff & Hurka. 
1991; Mummenhoff el al.. 1992. 1003. 2001a; 

genus should include C.urdariu Dc-\.. ( jirnimpus 
/inn. and Stroganouia bar. v\ Kir. and that the last 
• two genera are polyphylelic. Hased on such data 
and on a critical re-e\alualinn ol morphologv. \l 
Shehbaz et al. (2002) united all three with />•/>/- 
(Hum. Although Irpidtum shows more diversity in 
the reductions of petals and stamens than am other 
genus in the family (bowman & Smyth. 1908: How- 
man et al.. 1000). it is well defined by having an- 
gusliseptate (rarely terete or 4-angled) fruits and 
two Mihapical ovules one in each loculc. Other gen- 
era that show this combination of characters (e.g.. 
\eanlliocardamum Thell.. Delpinophylum Speg.. 
StubendorJ'/ia Schrenk ex Kisch.. C. A. Mey. c\ Ave- 
ball.. Wink/era Hegel) should also be studied, and 
it is quite possible that they. too. might be conge- 

The polycolpate clade (Al-Shehbaz & O'Kane, 
2002b: O'Kane vK \I-Shchbaz. 2002) is a mono- 
phyletic New World group readily distinguished 
from the remaining 90 r /r of Hrassicaeeae by having 
4-10-colpate instead of strictly 3-colpate pollen. 
The clade comprises Plnsaria (including l,es<pier- 

! yrea Man . Lyr- 

oearpa Hook. ik llarv.. Nerisyrenia Greene. Pay- 
sonia O'Kane & Al-Shehbaz. and Synthlipsis \. 
Cray. Se<jirence data of this group (O'Kane K Al- 
Miehba/. unpublished) ill GenRank have recently 
been taken by Kropf (2002) in his comprehensive 
ITS tree. The results support the monophy b ol ilie 
Physariu clade and place it between \ 
and Olimarabidopsis Al-Shehbn/ el al. 

Although no comprehensive study has been con- 
ducted on the cardaminine clade. a compilalion of 
various works (Bleeker et al.. 2002a. 2002b: I r..../- 
ke et al., 1998: bes. 1994; Mitchell & Heenan, 
2000: Sweeney & Price, 2000; Price & Sweeney, 
pers. comm.) shows that it consists of Cardamine 
(including Dentaria and ///'). \aslurlium. Rnrippa 
(including Neobeckia), Barbarea H. Hi.. Armorucia. 
lodanthus (Torr. & A. Gray) Steud.. L-arrmcorlliiu 
Torr.. Plumules Greene, and Selenia Nutt. The vast 
majority of taxa in this clade occupy wet or mesic 
habitat.-, are glabrous or rarelv have simple Iri- 
chomes. and have dissected or compound leaves. 
do i" i- d fruits, and mostly accumbcni coty ledons. 
A few exceptions occur, but this is a well-defined 
group that perhaps uiclude- the genera Lignariella 
Haehni. Xeomarlinella Pilger. Ornilhoearpa Hose. 
Pegaeopliylon Hayek M Hand.-Ma/.z.. Raphanoryn- 
cha Rollins. Subularia L, Taphrnspermitm C A. 
Mey.. and Yinshania. The analysis of Kropf (2002) 

nested within (Uirdumiuc and ///' (New /calaudi 
forming <i -i-ler group with the unrelated Bivanaea 
DC. (Mediterranean) instead of being nested within 
Cardamine (Mitchell & Heenan, 2000). 

The halimolobine clade is an exclusively New 
World group that consists of llalimolobos Tausch, 
Manena Wedd.. Pennellia NiewL Sphucmearda- 
mum Schaur. and. vet lo be described, a new seg- 
regate genus (Bailey el al.. 2002). Almost all mem- 
bers of this clade have dendritic Irichomes. plump 

leaves, and small white (lowers with suhtdradv mi- 
nimis stamens. The clade is related to other North 
American genera with branched Irichomes. includ- 
ing Hoechera. Cus'ukiella Rollins, and Perinea 
Price el al.. bill some South American species of 
Sisymbrium that have dendritic Irichomes. all ex- 
cluded from the genus (Warwick et al.. 2002). 
should also be tested. 

A morphologically well-defined group of primar- 
ily central Asian genera (e.g.. Chart, >lnma Rung,-. 
(daslaria Roiss.. Cnldbarhia DC. Isulis L, Litwi- 
mncin Worounw. /',/, <h \ph-r\isium Riinge. Pugion- 
ium Gaertn.. Samcruria Des\.. Sehimpera llochsl. 

Fiscli. ex DC.) has not been studied phvlogeneti- 
(•ally, and it probably forms a nionopliy li-li.- clade. 
All members arc characterized hv ha\ ing indehis- 
cenl. I- or 2-seeded. primarily angusiiseplale 
fruits, yellow or rarely v\liitisli Mowers, sessile often 
auriculate cauline leaves, and simple or no tri- 
chomes (Appel & Al-Shehbaz, 2(K)2). Only ITS 
data for /.satis are available, and the genus appear 
to be related to llm-e uilli august iseplate fruits and 
simple or no trieliomes (Kropf, 2002). 

Based on ITS data (Warwick et al.. 2002). most 
of the New World \meriean genera will) simple or 
no Iriehomes. as well as I'ringlca (si\ i>l;iii<l> in i li< - 
South Indian Ocean), form a large unresolved . -lade 
that includes members ol the Thelypodieae sensu 
Al-Shehbaz (1973). This clade encompasses taxa 

ular and morphological studies are needed to re- 
solve its generic limits. Among the genera that fall 
in this clade are most ol the South American Sis- 
Miihmun. Ihvoprlaluni \. Orav. Ibspcridanthus (B. 
L Rob.) Hydb.. Mostacilla.slrum 0. K. Schulz. Si- 
bam Creene, and Werdrrmannia 0. K. Schulz. Oth- 
er genera that should be studied are Cbilurarda- 
mum 0. K. Schulz. Dirt yophra limits 0. E. Schulz. 
Frvmodraba (). K. Schulz. \<>uoiitohotrys O. K. 
Schulz. Phlcbwpbnigmus (). K. Schulz. I'hh-bolob- 
ium (). K. Schulz. l>olypsi>cadium 0. K. Schulz. and 
Sanodraba Cilg & Vluschl. I'epper and Norwood 
(2001) have shown Caulanthus S. Watson to be 
polyphyletic and. logethei with Ciiillcnia Creene. 
is nested within Slrcplantliiis \ntt. In our opinion. 
these three genera, plus Strcptantliclla Kvdb. and S. Boyd & T. S. Ross, are peifectly al 
home in the more inclusive Slrcplantliiis. 

Trichome type is a feature used extensively in 
the delimitation ol genera of Brassieaccao. liirhoiiir-. apparently evolved indepen- 
dently at least two or three times in the family (Cal- 
loway el al.. 199B; Koch el al.. 2001a; Kropf. 
21)01') However, little is known about the genetic 
back-round behind ill.' -lull in trichome morphol- 
ogy. Knowledge of the molecular basis of trichome 
development in Arabidopsis tbaliana (see the ex- 
cellent review of Szymanski el al.. 2000) is rapidly 
expanding. Once the genes coding lor the devel- 

quenced. we should be able to use ihesc markers 
as potentially powerful tools in phvloge.ietic studies 
in Ihe family. Such knowledge will also help in un- 
derstanding lh,' origin ol glabrous from ihe pubes- 

acters (e.g.. segmented 
corky vs. non-corkv ones, dehiscent vs. indehis- 
ccut. seeded vs. seedless segments, winged vs. 
wingless seeds, flattened vs. terete valves) are un- 
derstood, the taxouomv ol the lamib could be based 
on more solid foundations. The differences in fruit 
characters often are overemphasized al the expense 
of more useful features an- largely ignored. Two 
examples demonstrate that. First. Ticissclmannia 
was shown by Price and \I-Shehbaz (200:'.) lo be 
itli-nlical lo 'thipidocarpiuii in their . liloi,.pLi~l gene 
ndhV and unclear ITS. has short ( 1- 
> mm long), oblriangular. 1- lo 8-seeded fruits su- 
pcrhciallv resembling those ol ' C.apst>lla bursa-pas- 
loris. whereas Trapidixarpnm has longer (10-70 
mm) narrowly linear or oblong. If)- 70- seeded 
fruits. These genera arc indistinguishable in every 
aspect ol habil. leaf, laceinc flower, and seed mor- 
phology. Second, based on nuclear and chloroplasl 
DN \ se<|uenees. Miunmenhoff et al. 12001b) have 
clearly demonstrated that Al/iaria (with linear, 
wingless, subterete to 1-aiiglcd fruits) is nested 
within Thlaspi s. str. (obcordate. winged. s| n ,iiglv 
angustiseptale fruits). Traditional laxoiiomv (e.g.. 
Schulz. 1036) places Mliaria in the tribe Sisvm- 
brieae and Tlilaspi in the l.cpidieae. Both genera 
have concentrically striate seeds, a character not 
known elsewhere in the Brassicaceae. and molec- 
ular data strongly indicate thai thev -hoiild be unit- 
ed, a position with which we agree. In this context, 
it is important to note that fruit dehiscence (also a 
diagnostic character for generic delimitation: see 
above) and the relative length of fruit are controlled 
in 1. (Indiana by only few M \DS-box genes, i.e., 
^IIYIIIKI'liOOF ami F/U ITFIIJ. (I.iljegren el 
al.. 2000; Kerrandiz et al.. 2000). We conclude that 
fruit characters alone may well lead to erroneous 

should be critically evaluated in light ol molecular 
and other morphological dala. 

\rabidnpsis represents a classic example where 
molecular studies have contributed lo a belter un- 
derstanding of lis laxoiiomv and generic delimila- 

assigned lo the genus are now placed in 14 genera 
(Al-Shehbaz el al.. 1000) and. as presently circum- 
scribed (O'Kane e> \I-Shehbaz. 1997: Al-Shehbaz 
& O'Kane. 2002a), Arabidopsis includes only nine 
species. Support for the new segregate genera pro- 
posed by Al-Shehba/ el al. I 1000) comes from ITS 
sequences (Koch el al.. I 000;,: O'Kane K Al-Sheh- 
baz. 2002. and in prep.), chalcnne sv nthase ( Koch 
et al.. 2001a). matK (Koch et at.. 2001a). ami al- 
cohol dehydrogenase (Miyashila et al.. 1908). 

Review of Brassicaceae Research 

nncmti li. lit. in relation lo the rest of Bias- i. aceae 
(Zunk et al., 1993. 1996. 1999: Galloway et ah. 
1998: Koch et al.. 2001a). Although only the highly 
variable 1 l , • • < llohen. (n = 14) 

has been studied, the kisal position ol the genus 
poses several challenging questions, hirst. Arthi- 

I i I! i ill i i i 1 i ii i i eiill-. ol ")0 

00 species (primarily Turkey and the Middle Kasl I. 
is much in need of detailed systematic and phvlo 
genetic studies (Appel & Al-Shehbaz, 2002). Sec- 
ond, fruit morphology in the genus langiisiiseplale. 
2- to 8-seeded) has traditionally Keen considered 
as specialized (Schulz. 1930). Third. Aethionema 

bers (n = 7. 8. 1 I. 12. Ik 10. 18. 21. 22. 21. 30: 
see Appel & Al-Shehbaz. 2002). Without critical 
plivlogenetic studies of \rlhioncma and several 
other southwestern Asian genera, it is premature lo 
speculate as to what an ancestral mustard might 
look like. It is interesting to note that members of 
the Thelypodieae (New World) were considered at 
one point as primitive (Al-Shehbaz, 1973) but now 
are treated as relati\el\ specialized (Callowav el 
al., 1998). 

Although substantial molecular data have accu- 
mulated during the past ten years, only data on ITS 
sequence- are somewhat more e\|en-i\e than those 
of other markers. Such data are being svulhoi/ed 
b> Bailey el al. (pers. coinm.), and it is premature 

to make any generalizations. The coinpil; by 

Kropf (2002; of largely unpublished sequences 
from GenBank is far from adequate. It would be 
rewarding to assemble similar large data sels on 
chloroplasl genes and compare them with those ol 
ITS -equelli es. \| (1 -| -.Indie- haye coneetll rated on 
North American and Kuropean ta\a. and except for 
the work of Mitchell and Heenan (20(H)) and Hurka 
et al. (unpublished) on the mustards of New Zea- 
land and Australia, respectively, hardly any studies 
exist on the South American. African. Himalayan. 
and central Asian genera of the family. Moleculai 
-.Indies on disjunctions, especially in cosmopolitan 
genera such as Curduminc. Lrpidiiim. and /,', >ripp<i. 
which occur on all continents except 
are needed, though the work 

function of many ol its gene- ha- been identified, 
it is highly desirable thai phylogenetic studies be 
based r.n sequence- of genes uilh definite func- 
tions, especially those controlling ihe development 
ol ti ■ichonies. leave-, and fruits, structures thai offer 
the taxonomic characters most frequently used. 
Equally important is to determine on the basis of 
molecular data the evolution of individual charac- 
ters and the distribution of their homoplasy on a 
family -wide basis. I oi example, we have no idea if 
the presence of multicellular mulliseriale glands 
represents a synapomorphv within the lanulv. and 
the evolulion of mauv other characters (e.g.. schiz- 
ocarpic or samaroid hints, palmatelv compound 
leaves, decurrent stigmas with united lobes, stellate 
Irichomes with webbing and numerous rays, seg- 
mented fruits, conduplicate or dipleeolohal cotyle- 
dons. ])roduclion of garlic smell, to name a few) 
needs critical studies. A good example illustrating 
this aspect is the determination of evolutionary 

(2001a;. lee , 

..Im.'I el ,1 


2002) and floral structure in h'pidiiim (Bowman et 
al., 1999; bee et al.. 2002) and fruit convergence 
in Thlaspi ( et al.. 1997a), but studies 

The absence of a given feature in \mhidopsi.s thal- 
itina should also stimulate research to trace if this 
species still has at least part of the genome re- 
sponsible for that character. The studies of self- 
incompatibility in Brassiitt and Us IioiuoIol in b- 
abidopsis (Conner et al., 1998; Kusba et al., 2001) 
are excellent examples to follow. 

3. 4, 

ined bv cloning and characterizing 

l Lepidium are ideal studies to follow. 
The vast majority ol molecular studies in Bras- 
sicaceae involved sequences of single markers, but 
it is far more important lo study multiple marker- 
separately and combined, as was done by Bailey et 
al. (2002) and Koch et al. (2001a). Now that the 
entire genome of \rabidopsis is well known and the 

region-. II there are high levels of genome ('..-lin- 
earity, as well as of the gene level (microsv ntenv ). 

comparative ge e -mapping experiments can 

serve as an eflieiei ol f..r transferring information 

and resources from well-studied genomes, such as 
that of \ralndopsis. to related plants. Sequencing 

ahnlops,-, Cei.ome Initiative. 2000: Blanc et ah. 
20(H); Vision et al.. 2000; Simil liciti el al.. 2002) 
have shown I lull al least 707, of llu ■ genome is du- 
plicated and that the original haploid number of its 
ancestors was probably lour. < l< hi I >l< -• I to eiidil. 1 1 1 • I 
rcdiiccil to li\e as a result ol'e\tensi\e chromosomal 

Brassica has been thoroughly studied using eom- 
|iarali\e mapping experiinenls (I .agercraulz vK I .\ - 
diate. 1006; Kovvalski et al.. 1004; Fagercrantz. 
1008). Other studies (Kovvalski et al.. 1004: Fag- 
ercra.itz, 1008; Orant et al.. 1008; Fagercrantz et 
al.. 1000; Cavil et al.. 1008; Jackson et al.. 2000: 
O'Neill & Bancroft. 2000; Axelsson et al.. 2001; 
Ouiros et al.. 2001) have focused on the Arahidop- 
sis Biassira i •oinpaii-uii. These results have sub- 
stantiated the view that "diploid" specie-, of Urns 
sua. />'. mm, i \n 8). />'. o/eraeea (// -■ 0). />'. rapa 
(n = 10), have largely triplicated genomes and most 
likely have descended from a polvploid ancestor. 
Comparative mapping experiments have demon- 
strated that approximately 00 rearrangements oc- 
curred since the div etgence of \rahidi>psi< and II. 
nigra, estimated to have occurred 14-20 million 
years ago (Koch et al.. 2000. 2001a: Vision et al.. 
2000). Higher degrees of genome co-liuearit.v have 
been observed in comparisons between 1. llmlitinn 
and 1. lyrala siibsp. pctrnca (diveiged ."> \1V\: 
De Haen et al.. 1000) and I. lhaliana and Capsclla 
rubella (diverged -0.2 to 0.8 MYA; Schmi<lt et al.. 
2001; Acarkan et al.. 20(H)). Co-linearity on a small 
scale (niicrosvntenv ) lias ben delected in several 
regions in Arahidopsis and Brassica (Sadovvski et 
al.. 1000; Sadovvski K Ouiros. 1008; Conner et al.. 
1008; Osborn el al.. 1007). although no general 
conclusions can be drawn from these studies. How- 
ever, similar homologous segmenls with drastically 
increased fragment sizes were found. Nevertheless. 

binalorial approach of analyzing microsvntcnv 
(based on DNA sequence information) and function 
(gene expression analvsis) has been perlormed on 
a large set of cruciferous plants using |)romoter re- 
gions of APFTAFA3. and (41S (Koch et al.. 
2001b). This analvsis among distantly related spe- 
cies mav help to predict gene and promoter fiinc- 

ons and specifiei 

nd may also help achieve a 
of the evolutionarv. sigmli- 
:s of mutations in coding re- 
gions (Slotz el al.. 2000; Bishop el al.. 200(n. ( ^un- 

deletions, and rearrangements, as well as for 
termining the evolution of duplicated genoi 
through time and the unification of genomes wil 
polyploids (A. E. Hall et al.. 2002). In our opin 
more studies of this kind, though both labor ml 

and the role polyploidy played duri 

Molecular markers have been utilized as a tool 

whose divcisihcatioii has not subslanliallv been in- 

DNA markers (e.g., the plaslome ol most angio- 
sperms: sec Harris »X Ingrain. |00| : IJebou.l »X 
Zeyl. 1994) can be used to trace the maternal lin- 

"BAI'Ds" (random amplified polymorphic- l)N As) or 
"AFFPs" (amplilied fragment length polymor- 
phisms), as well as l)N A se ( |ueiices of nuclear re- 
gions such as the frequent ly used internal Iran- 
scribed spacers of nuclear ribosomal DNA (ITS-1 
and ITS-2) (Franzke et al.. 1998: Francisco-Ortega 
el ah. 1900; Bleeker et ah. 2002b). These nuclear 
markers serve as tools to deled patterns of genetic 
diversity thai are inherited nialernallv and pater- 
nally. By using molecular markers, 
(source areas of genetic diversity, vicarianl pat- 
terns, migration routes, hybridization zones, sec- 
ondary contact zones, etc.) and related evolutionary 
(|ue>iioiis (specialion |irnir-Ms. poklopic origin, 
etc.) can be addressee!, and the data obtained s.tvc 
to develop more appropriate models that explain 
present-day distribution and diversity. A recent ex- 
ample for such analyses is the coordinated effort to 
develop a deepei n iider-laui I i ng ol i|is| i ibu I ion pat- 
terns, group differentiation, and the evolution ol 
plants in the antic and alpine regions (Stehlik et 
ah. 2001). Both BisniWlla F. (Dannemann. 2000: 
Treinelsberger el ah. 2002) and Covhlvaria (Koch. 
2002) include some species of alpine plants that 
have been subjected to such studies. Other studies 
on European Aralns alpina F. (I'lanlholt. 1005). 
Canlaminr (Franzke vK Hiuka. 2000). and Furasian 
Mirmlhlaspi (Koch el ah. 1908c; Koch & llurka. 
1000) helped to elucidate colonization routes from 
lelugial areas into formerly glaciated areas of north- 
ern and central Fnmpe. The analyses of \iahidnpsis 
"•■ :.".; khi J.i ..ii-ii 1 1 . -. | d -is bulion patterns of 

found in other plant and animal species, all of 
which demonstrate that the Iberian Peninsula. Italy. 
and the Balkans represented three major refusal 
areas during glaciation in central Kurope (Sharhel 
et al.. 2000). A comple\ specialion and migration 
scenario in lhaba was elaborated for North. Cen- 
tral, and South America (Koch & Al-Shehhaz. 
2002). Dniho also shows strong affinities to high 
alpine regions (e.g.. in the Alps. Scandinavia, the 
Himalayas. Rocky Mis.. Andes) and. therefore, its 
evolution appears to have been influenced by gla- 
eiation and deglaciation periods throughout the 
Pleistocene. Studies on the Chinese Yinslmnia elu- 
cidated migration over long distances and extensive 
ivliculalion during migration (Koch M \ l-Shehba/. 
2000). It is remarkabie that, in all cases analyzed 
to dale, polyploidization played an important role 
in migration and phy Ingeography. and this seems 
also to be the case in l.cpidiuiti. Comprehensive 
molecular studies of the biogeographie history of 
l.rpidium and Korippa on a worldwide scale were 
conducted by Muimnenhoff et al. (2(K)la) and 
Bleeker et ah' (2002a). respectively. The easily dis- 
persible mucilaginous seeds of species of h'pidiiim 
and the widespread occurrence of autogamous 
breeding s\-.|eiiis suggest a rapid radiation ol the 
genus by long-distance dispersal during the Plio- 
cene/Pleistocene, an interpretation supported l>\ 
estimates of divergence lime- based on cpDN A se- 
quence divergence. Climatic changes resulted in 
the establishment of arid/semiarid areas, thus pro- 
viding lavorable conditions for the radiation of //'/>- 
idium that led to its current worldwide distribution. 

frpidium is represented in Australia and New 
Zealand by 10 and 7 native species, respectively. 
ITS and cpDN A phylogenies gave strongly conili.l- 
ing signals and prov ided ev idence for biconlinental 
hvbrid genomic con-l ilul ion in l.rpidium in these 
areas (Miunmeiihoff et al.. submitted). Seventeen 
\u-lralian/New Zealand s|.ecies share a Californi- 
an cpDNA type. As for ITS. nine s|>ecies appear to 
harbor a Californian type and eight species share 
a South African type. This pattern is most likely 
explained bv two trans-oceanic dispersals of U-pi- 
d in in from California and \frica into \uslrulia/Ncw 
Zealand and subsequent hv hndi/alion followed bv 
homogein/alioti ol the ribosomal DNA either to the 
Californian or South African ITS type in the two 
different lineages. 

Some future directions of research should focus 
on the origin and early diversification of Brassica- 
ceae. There are two contradicting hypotheses: a 

western North American and Irano-Tiirani; rigm 

(Hauser & Crovollo. 1982). Molecular and morpho- 

persal of Brassicaceae. Other research directions 
should focus on the range extension of species, as 
recent global climatic change- and human activities 
will lead to the spread of species beyond their nat- 
ural ranges or the reduction of their ranges (Wallhei 
et al.. 2002). Studies on \rubidnpsis and Trrsda/m 
H. Br. demonstrate the influence of climate on the 
global ranges of species (Hoffmann. 2000. 2002). 
Such studies may provide information about the ec- 
o'.-eograpluc of species. Species of some 
genera (e.g.. \asturtium. Curdaminr) have been cit- 
ed as examples where hybridization may serve as 
a stimulus for the evolution of 
Hand & Schiereubeck. 2000). 

evolutionary divergence has been the subject of 
much debate (Sanderson. 1098: Sollis t\ Soltis. 
2001). Questions largely involve: (I) the possibility 
of varying substitution rates within a lineage. (2) 
differing substitution rales between lineages, and 
(."*) the lack of accurate evidence to calibrate the 
molecular dock (e.g.. missing loss i | ev id.-ucc) (Sin 
derson. 1998. 2002: Britton et ah. 2002: Syvanen. 
2002). Kstimates from Chs (chalcone synthase), 
\</li (alcohol dehydrogenase), and nuilk (malura-c 
K) indicate that the Brassicaceae appeared approx- 
imately .">(> million vears ago (MY A) (Koch et ah. 
2001a). Synonymous substitution rates of 1.5 X 


I Chs . 

1.7 ■ 10 " 

estimated (Koch et ah. 2000. 2001a). Siil.slilul ion 
rates obtained from these studies were used to cal- 
culate that \ndiidopsis ihalianu diverged from its 
Ciirdaminopsis-Wki- common ancestor approximate- 
ly 5 MYA. while the 
of polymorphic A. thaliana 
roughly 1.5 MYA (the average divergence lime be- 
tween Brassini and \ml>id<>p-,is has been calculated 
to be 14-20 MYA: Koch et ah. 20(H)). These results 
correspond to findings obtained from divergence 
time estimates of the mitochondrial gene for NADU 
subunit t (nad\) in Brassica species bv Yang el al. 
( 1009b). These data also help to calibrate other mo- 
lecular clocks of more widely used markers in 
Brassicaceae, such as the ITS or 1rn\, intron and 
IrnL-F spacer region (Mui.imenhoff et ah. 2001a: 
Bleeker et ah. 2002b). Despite the inherent uncer- 

lincd by Sanderson I I OOP,), ihis procedure has been 

lion Mini hiogeographv (Bohle el al.. 199(>; kirn et 
al.. 1998; Vargas el al.. 1998). Meanwhile, an ap- 
proach lo estimate divergence limes in the absence 
of rate constancy has heen developed (Samlerson. 
1997). As for Krassicaeeae. it would he desirahle 
lo calibrate siihsliliition rales of widely used nucle- 
ar. epDNA. and mtDNA marker systems by refer- 
ence to alternali\e/t r -uhstantiated fossil re- 
cords, geological hi->loi_\/e\cnl>.. .llld | icl in 1.1 1 n ■ 
data (Baldwin ei Sanderson. 1998: Richardson el 

& U. I.agercranl/. 2001. Mul 
flowering time OTCs within several Brassica spi 
could he the result of duplicated copies of one arret 
gene. Genome 44: 856-864. 
Bailey. C. I). & J. J. Doyle. 1999 Potential phyloge 
utility of the low-co|>\ nuclear gene pislillala in d 
vledonous plants: Comparison to nrDNA ITS and 

\l..l.. I'IinIm.;. net. Kv..l. 13: 20-30. 

. H. A. Price & J. J. Doyle. 2002. Systematic 

the I l.ilimolol .iti.- Pia — icaceae: K\ ideiice limn I 
loci and morphology. Syst. Pot. 27: 318-332. 

Pishop, J.. A. M. Dean K T. \1 it. h< 1 1-< ) 1 . 1 - 2000. liapi.l 

il -.mi. »-.|\ iii the Pras>icn- 

hs„trria. with a critical evalu- 

lection in plant-pathogen coevolution. Proc. Natl. Acad. 

Anri. Missouri Pol. Card. 87: 

Sei. U.S.A. 97: 5322-5327. 

Plane. (... \. Parakal. P. I.uvol. P. Cooke X VI. Delsnev. 

!i. 20o2a la\onom\ and phv- 

2(KK). Kxtensivp duplication and reshuffling in the Ar- 

las.icaceae) (22 \ugusl 2002). 

abittopsis : genome. PI. Cell 12: 1003-1 101. 

'.. \1. Vleve.owit/ (chin,-.). The 

Hleeker. \\. <S II. Iliuka. 2(H)I. IntiugresMxe livl.ridi/aliori 

ican Society of Phuil Pioloui-i-. 

in Rorippa (Prassicaceael: (.cue (low and its conse- 

oi/MUPMJ/iah. (MM)], (http:// 


Kcol. 10: 2013-2022. 

. VI. Hullimaiin K II. Iliuka. 1000. Evolution of 

hybrid taxa in Nasturtium P. Pi. (Prassicaceae). folia 

'rice. I0O0. Ceneric placemenl 

Ceohot.31: 121-133. 

in \iabiiiapsis (Prassicaceae). 

. C. Weher-Sparenberg & II. Hurka. 2002a. Ohio 

roplasl DNA variation and hiogeographv in the genus 

■nliofl" e\ (>. \ppel. 2002. Car 
ro t 

ez-Puvlla. E. P. & \. \. Garay. 1901. Population 
letic structure of Ccropcgia obtusifolia. a tropical pi- 
■er tree species. Kvoli.tion 18: 437-453. 
rson. J. K. & S. I. Warwick. 1999. Chromosome niiin- 

uee Iron, ,so/\,ne uuinhrr. PI. Sysl. Kvol. 215: 255 

. A lip. Pol. 
lamburg Zr. ab-SKi. 

& . 2002. Cruciferae. Pp. 75-174 in k. 

lants. Vol. 5. Springer-Verlag. Berlin. Heidelherg. 
hidopsis Genome Initiative. The. 2000. Analysis of the 
iMiome se(|iienee of the dowering plant Arabidopsis 
uiliana. Nature 408: 796-4115. 

old, M. L. 1997. Natural Hybridization and Evolution. 
Moid f niv. Press. Oxford. 

. Pouck. 2(M)I. Natural hybridization and litness. Tax- 

. 50: 93-104. 

Isson. I'.. C. M. Powma... A. C. Sharpe. I). J. I.vdiale 

liniii>;t,i Nop. (Prassf 

. A. Kran/ke. K. Pollmann. A. H. I). P.own & II. 

Hurka. 20021). Phv logenv and hiogeographv o( Southern 
Hemisphere high-mountain Cardaminr species (Pras- 
sicaceae). Austral. Svst. Pot. 15: 57.5-581. 

Ploom. T. (... .1. VI. hVkm v\ C. C. Paskin. 2001. Eco- 
logical life history of the facultative wiHHlland biennial 
\rabis laevigata variety laevigata (Prassicaceae): Sur- 
vivorship. .1. To.rev Pot. Soc. 128: 93-108. 

Bohle. I. -P.. II. II. Hilger^ \\. E Martin. 1990. Island 
colonization and evolution of the insular woody habit in 
Echinm I,. (Poraginaeeae). Proc. Natl. Acad. Sci. LCS.A. 
93: I 17 10-117 15. 

Pohuon. K.J. P.. D.J. Keith. I. A. P. Parkin. A. G. Sharpe 
& I). J. Cydiale. 1996. Alignment of the conserved ge- 
nomes of Brassica oleracea and Brassica napus. Theor. 
Appl. Cenet. 93: 833-839. 

Evolution ol 2> albumin seed storage protein genes in 
the Pra^H-aceae. Pioclicu. SM. v\ Ecol. 27: 223 -23 I. 
Powman. J. I.. X D. Smvlh. |00J>>. Patterns of petal and 
stamen reduction in Australian species of h'pidium 
(Prassicaceae). Int. J. PI. Sci. 159:6.5-74. 

■ & K. Miimmcnholl. 
changes in floral structure within 

929. eomliined analv-is oi H> urDN \ sequences and innr- 

Bradshaw. II. I).. S. M. ttilherl. k. G. Oil.. X I). W. phological dala. \nn. But. (Oxford) 86: 53-62. 

Schemske. 19%. Genetic mapping of floral trails as- Crisp. P. 1976. Trend- in the hreediiig and cultivation of 

sociated with reproductive isolation in monkey flown- ci uciferous crop-. I'p- 00-118 in J. 0. \aughau. \. J. 

(Vmii/kvl. Nature 376: 762-765. MacLeod & B. VI. (.. Jones (editorsl. The Biology and 

Bricker. J. S.. G. K. Brown & F. T. I.. Patts. 2(KK). Status Chemistry of the Cruciferae. \ca.lemic Press, I/mdon. 

of Dexenrauna larulnsa ( Brassicaceae). \\. Y Amer. Dannemann. A. 2(KM). Dcr Kinfluss von f ragmciiticruiig 

Naturalist r.0: 126-T32. und Population-gnU!, u die g, -netische Variation und 

Brill. Hi. T.. P.. Oxelmau. \. \ iniierslen K K. Bremer. 20(12. I Hue- urn scltenen Pllaii/enarlen lici-piel von His- 

Phylogenetic dating with confidence intervals using rntella laevigata ( Brassicaceae). Diss. Bot. 330. J. Cra- 

incau path lengths. Molee. Phvlogencl. Fvol. 21: 5B iner. Berlin. Slullgarl. 

65. De Haen. A.. F Mauri, ■.... I). ( kn lesvvorth, J. Crepau, C. 

Brochmann. C. 1992. Polyploid evolution in arctic-alpine handle. VI. Koch. T. Mitchell-Olds & 0. Savolainen. 

Ihalia I Brassicaceae). Somiiierleltia Suppl. 1: 1-37. 1999. Comparative mapping l.elween \nil>iil<>i>si~, ilia! 
. P. S. Soltis & I). K. Soil,-. |99|. Multiple origin- iana and \rahis pelraea. Proceedings of the VUlh Plant 

of the octoploid Scandinavian endemic Draba cariiinin- & Animal Genome ( oiileience. Nan Diego. Poster 131. 

urn: Kleclrophoretic and morphological evidence. Nor- Del-env. M.. J. M. Met, rath. P. This. A. M. Chevre & C. 

die J. Bot. 12: 257-272. F. Ouiros. 1990. Ribosomal BNA genes in diploid and 
. I). K. Soltis & P. S. Sollis. 1992a. Flectrophorclic amphiploid [imssica and relale.l species: Organization. 

relationships and phvlogenv of Nordic polyploids in polv morphism. and evolution. Genome 33: 733- , I 1. 

Draba (Brassicaceae). PI Sysl. K...I. IB2: .''.5-70. Fllslrand. N. C. K k. A. S.hierenheck. 2(MM). Hyl.ridiza- 

formation and polvphvlv of Nonli. polv ploids in lhaha plants Pro. Nail \cad S.i. I .S. \ 07 7015-7050 

(Brassicaceae). Amer. J. Bot. 79: 673-688. Frirkson. I.. R.. N. A. Straus & W. D. Beversdorf. 1983. 

. B. Stedje Ox I.. Borgen. 1992c. G one How across Restriction patterns reveal origin of chloroplasl ge- 

nomes in Brassica amphiploids. Theor. Appl. ( ieiiel. 05: 
'.vans. M. K. K.. B. W. Dolan. K. S. Menges & 1). B. 
(.ordon. 2000. Genetic diversity and reproductive l.i- 
ol.»uv in Waira ,-ailrn | Biassicaceae). a narrow l\ en- 
.I.i.iic Honda s.T,il.,,n,ii.;il. \mcr.J. Bol.87:372 381. 
crrandiz. C. S. J. I.iljegren <!v M. F. Yanofskv. 2000. 
Negative regulation of Slf YITI.IU'Rt >< )Ff genes l,v 
I-' HI ITFI LL during \rabidapsis fruit development. Sci- 
ence 289: 136-438. 

miiei-eo-()rlega. J.. .1. Kuertes- Aguilar. C. (i.iiiuv-Ciiiii- 
po. A. Santos-(;ucra & R. k. Jansen. 1990. Ini.rnal 

tl'jli-erilie.l -pa. el -c.|llencc phvlogenv ol Cnilllhr I . 
lliia-ieaccae): Vloleeiilar data reveal hw. Old \\,,rl,l 
disjunction-. Molee Phvlogenel. Kyol. II: 301-380. 

II: 62 69. Franzke, A. & II. Hurka. 2000. Molecular systematic and 

worth. I).. F. I.. Fill & F. /hang. 1998. The evo- hiogc.grapliv ol 'the' anlamitie pialensis complex (Bras- 

lution of the alcohol dehydrogenase gene familv hv loss -i.accae). PI. Svst. Fvol. 224: 213-234. 

of introns in plants of the genus h-arenuorthia (Bra-- & k. Mummenhoff. 199<). Recent livliri.l -pe.ia- 

552-559. lion in Cardamine (Brassicaceae) — Conversion of un- 
ci- M. H. Schierup. clear ril>o-on,aMT>-e<|i,ciice- in -tatu na-cendi. Then,. 
1 iimltiallelic -ell-,,c \ppl. Genet. 98: 831-831. 

I of Cardamine and allied 

Comai, L., A. P. Tyagi. k. Winter. B. Holmes-Davi 
Reynolds. \. Steven- ^ B. Bvcrs. 2000. Phenol} 

stability and rapid gene silencing in newlv formed lr- ( ,onie/-( .ampo. C. 1099. Taxonomy. Pp. 3-32 in C. Go- 

abidnpsis allotetraploids. PI. Cell 12: 1551-1507. mcz-Campo (editor). Biology of Brassica Coenospecies. 

region and its homeolog in \iabiilop\ts: Impl icalions for Pp: 3:5-58 /// C. ( -/-Campo (editor). Biology ol Bras- 
Cell 10: 801-812. " Grant. M. R.. J. M. McDowell. \. (,. Sharpe, M. de Torres 

Constantinescu. 0. & J. Fatehi. 2(K)2. Peronnspora-likc /ahala. I). J. I.vdiale iK J. F. Dangl. 1998. Independent 

fungi (Chromista. Perono-poiale-l para-itic on Brassi- deletions of a palliogen-resi-tance gene in Hrassiea and 

caceae and related h,,-l-. Nova llcluiuu 71: 291-338. \raludapsis. Proc. Natl, \ca.l. Sci. U.S.A. 95: 15843- 

Crespo, M. B.. M. D. Fled... M. F Fay & M. «. Chase. 15848. 

•nome: Opportunities In. 
Ksiol. 129: 1439-1447. 
. Sytsma & H. H. litis. 

. K li 

. Bot. 89: 1826-1842. 

. \. K.-ll....o.j.T. c.sta III. W. 
el. 1001. Ilaplohpt- .•li\fi-.u<Tn- l - 

ens: Bole lor both balancing 

Genetics 1 38: 811-828. 

1991. Chloroplast DNA ami 

The effects of intraspecific diversity and 

<sion. Taxun 40: 393-112. 

v J. S. Heslop-llarrison. 1005. Centro- 

meric icpelilive |)\ \ sequelli es in the nemis /.','. h-^v 

Thenr. Appl. Geuel. 00: 157-160. 

lauser. L A. i\. T. .1. Croxello. F»82. \ -i i. al jiuLi- 

ol trncric relationship- In Th.-b podieae. Sxst. Hoi. 7: 


Ia\ek. \. 101 I. Knlwurl cine- ( auciferens\slems aul 
phylogenetischer Grundlage. Beih. Bot. Centralbl. 27: 

Iccnan. P. 8.. A. I). Mitchell X M. Ko.h. 2002. Molec- 
ulai -\-lematics ol the New Zealand I'lirlnrladon 
ibm— icacoac) complex, Genetic circumscription and 

. Bot. 10: 543-562. 

n. M. H. 2000. Biogeographv and climatic differ- 
on ol two annual species of Tersdalia R. Br. 
. Biogeogr. 27: 989-999. 

I. Biogeogr. 29: 125-134. 
.2001. \llo/M,,e diversity and 
Japanese and Korean popula- 

iceae). Genes X (,encl. Systems 

. Evol. 206: 295- 

in. H.. F. Tujma, l{. Terauchi & T. Miyashita. 19%. 
itragenic recombination in the \dh locus of the wild 
laul \ndudup*is lh„li,,,ui. Genetic- I 13: 1761-1770. 
tson, S. A., Z. Cheng. M. I,. Wang. II. M. Goodman K 
Jiang. 2000. Comparative fluorescence in situ hy- 
ridi/.ation mapping ol a 131 -kh \rabidopsis lhaliana 

Kanim. A.. I. Gala— o. T. Schmidt A J. S. I le-lop I In i i-on. 
1995. Analysis of a repetitive DNA family from Arabi- 
dopsis arenosa and relationships between \rabidopsis 
species. II. \1olec. Biol. 27: 853-862. 

Kavvabe. A.. II. Innaii. B. Terauchi v\ N. T. Mivashila. 

1997. Nucleotide polymorphism in the acidic chitinase 
locus (C/n'A) region of the wild plant \rabidopsis ihal- 
iuna. Molec. Biol. Kvol. 14: 1303-1315. 

Kim. H.-G.. S. C. Keelev. I'. S. \ room K K. K. Jansen. 

1998. Molecular evidence for an African origin of Ha- 
waiian endemic Hesperomannia ( Asteraceae). I 'roc. 
Natl. Acad. Sci. U.S.A. 95: 15440-15445. 

Kliebenslein. I). J.. J. Krovmanu. I'. Broun. A. Figuth. I). 
Pederscn. J. Cershen/nn K T. Milclicll-Mlds. 2001. Ge- 

inolalcs. PI. Phvsiol. 126: 811-825. 

Koch. M. 2002. Genetic differentiation and specialion in 
prealpine Corhlearia (Brassicaceae): Allohexaploid 
Corhlearia bavarira (Brassicaceae) compared to its dip- 
loid ancestor Corhlearia pyrenaira in Germany and 
Austria. PI. Ssst. Kvol. 232: 33 16 

& 1. A. 2000. Molecular systematies 

ol the I Inn.-. Yin.shania (Brassicaceae): Kvidence from 
plastid and nuclear ITS DNA sequence data. Ann. Mis- 
souri Bot. Card. 87: 246-272. 

& . 2002. Molecular data indicate com- 
plex intra- and intercontinental differential ion of Amer- 
ican Draba (Brassicaceae). \nu. Missouri Bot. Card. 
89: 88-109. 

<X II. Murk. i. |00<). Iso/wne aual\-i- in the poly- 
ploid complex Mirrolhlaspi perfolialum 1 1..) F. K. Meyer: 
Morphology. biogeouraph\ and evolutionary history. Flo- 
ra 194: 33^18. 

& K. Mummenholf. 2(M)1. Tldaspi s. sir. (Brassi- 
caceae) versus Tldaspi s.l.: Morphological and anatom- 
ical characters in the light of ITS nrDIMA sequence 
data. PI. S\st. Kvol. 227: 209-225. 

. & K. Zunk. 1003. Isoelcktrische Fok- 

ussierung der I nlereinheiten der Bubisco in Tldaspi 
(Brassicaceae): Weitere Hinweise aul cine Formengat- 
lung. Feddes Keperl. KM: 371-381. 

. II. Murk., v\ K. Muminc, I. |0<),, ( .|,|o,opla-l 

DNA restriel.on site variation and BAPD-aualvses in 
Corldrano (Brassicaceae): Bio-v -Icinal us and sp,., aa 
lion. Nordic J. Bot. 16: 585-603. 

. M. Minimum., \ II. Ilurka. I »> , I / „ 

s|)eciation and evolution in the polyploid Corhlearia I.. 
(Brassicaceae). Bot. Acta 111: 111-125. 

. K. Mummenholf \\ II. Ilurka. 10081, ^-Icnahc- 

and evolutionary history of heavy metal tolerant Tldaspi 
i Western Kurope: Kvidence from genetic 
on isozyme analysis. Biochem. Sysl. & 

I',, .ft J. B. Akerovd. I <)'•:',. |,„/„,s. In; \\ (.. 
. Binges, A. (). C'haler. J. I!. K.linondson. \. 

I. D. M. M c. I). II. \alcntine. S. M. Wal 

\. Webb (editors). Fl. Furopaea. ed. 2. I: 
..imbriiKe I ni\. Press. Cambridge. New 

B. W. Sanders & M. J. Donoghue. 100 1. 

I in I --' I w.l 210: 207 230 

.15. Ilauhold X T. \l,lrl„ll-()Ms. 2000. Compar- 
ative evolutionary analysis ol chaleone synthase and al- 
cohol dehydrogenase loci in Arahidopsis. Arabis, and re- 
lated genera (Brassieaccae). Mole. Biol. Lvol. 17: 


JUKIh Molecular syslematics 
ence Ironi coding plaslidic 

, B. Weisshaar. .1. krovmann. It. Haubold & T. 

Mil.-h.-ll HhU. 20011,. Comparative phylogenctic anal- 
\sis of the CHS and upetiiluW promoter reveals new 
< hie- ioi promoter function determination. Mole.. Idol. 
Lvol. 18: 1882-1891. 

. M. Huthmann & K. (, Bernhardt. 200 ; , I 'op 

ulation dynamics of Cardaminc amara I.. (Brassieaccae) 
in dynamic hahilats: Genetic comparisons ol seed hank 
and surface populations. Basic Appl. Keol. (in press). 

. C. K T. Milchell-Ohk 2008b. Multiple 

hybrid formation in natural populations: Concerted e\o- 

i I'"- ' I l-l m N h \mi. .. , ., ' 



& I). I.vdiate. 1000. Comparative gen, 

ping in Brassira. Genetics 111: 100., 1010. 

, J. Putterill, (;. Coupland & I). Lydia 

Comparative mapping in \rabidopsis and lira. 

S. Ncgi. |00|.Stiiieli„ 
itions of the rl)\A inlergenic 
-. I'l. Mole,-. i:,ol.2l:01., <)27. 

hi I I ' < *, ultima, Genome 

lies. 10: 776-788. 
Lanner. C. 1998. Relationships of wild Brasstra -pe.-ie- 

i-on ,.l the D\ \ >e.|ueiieo . ,| ill,- i hloioplasl inlergenic 
region between Irnl. |l \A) and IrnV |G\ A). Canad. J. 
Bot. 76: 228-237. 

Lee. J.-Y.. K. MummrnliolT K .1. I - Bowman. 2002. \llo- 
ploidization and evolution of species with reduced floral 
-Iruelures in Irpidium I, ( Bia-ieaceac). Proc. Natl. 
Acad. Sci. U.S.A. (in press). 

Lenoir. A.. B. Courunve,. S. Warwick. C. heard X J.-M. 
Deragon. 1007. Involution ol" SINK. SI rctmpo,ons in 
Cmeilerae plant species. Molec. Biol. Lvol. I I: 931- 


Kropf. M. 2(K)2. Vergleichende Biogeographie europaisch- 

er Cebug-pllan/en: Molekul.uv mid morphometrische 
I nlersuchungen an der monlan-suhalpineii \i,ii,\lits 
molilalia L. (Fahaceae) und der hoehalpinen Prilzela^o 

alpiim (I..I (>. Kuill/e i I !i a»iea< r;„ - 1. Doctorate I )i»ei 

tation. Mainz University. Cuvillier Verlag Cotlingen. 
Kuiltinen. II. cK M. vguade. 200* I \ -Ciii C ; ,aiio- i 

Ceoelie- IV,: J !(>3 872. 

Kusal.a. M.. K. Dvvver. J. Ilciidershot. J. Vrebalov. J. B. 

\asrallah & M. L. Nasrallah. 2(M)1 . Self-inconi ihl I il> 

in I In- --tin- [rabidopsis: Characterization of the S lo,n- 
iii tin outcrossing 1. hrala and its autogamou- relative 
1. Ihaliana. PI. Cell 13: 627-643. 

I.ugercranl/. I . 1008. Comparative mapping between 1/ 
. >, > .in-l /. »><,<,<■! Iii Mi 

Hmssira genomes have evolved through extensive ge- 
nome replication accompanied hv chromosome ln-ioiis 

, I). H. 1001. Mole, ula, -v-leiriatres and taxonomy ol 
lake cross I \rabrrlaa aanalica: Bras^icaeeae). an im- 
,, nil,-,! c.maii. i, Hi-lard. Aquatic Bot. 10: 110-105. 

.rhova. J.. K. Marhold & B. Neuffer. 2000. Taxonomy of 
Cardamim- amara |( a ucilerae) in the Iberian Peninsula. 
Taxon 49: 747-763. 

aljegren. S. J.. C. S. Ditta. V Lshed. B. Sav idge. J. I.. 
Bowman K M. V. Nanofskv. 2(M)0. Shatterproof M \ D- 

Liude. VL. S. Diel ,\ B. Vullei. 2001 . flowering eeotv pes 
ol ' Capsrlla bursa-pasloris (L.) Mc.lik. (Brassicaceae) an- 
alvsed bv a coscgregation ol phenolvpie eharaeteis 
,(,111 i and molecular marker,. \rm. Bot. OKIordi ,'!,: 


Lowman. A. C. & M. I). Purugganan. 1000. Duplication 
<>il\irlirassin,obni«>«AI>ETAIAI\\ " 

and the evolution of domesticated eaulillow, 

90: 514-520. 
Mahy. C.. X. Yekemans & A. L. Jaequemarl. 

lalus/vnska. .1. X .1. S. I leslop-t lamson. |0<)2, |-|,v-„al 
mappiui; ol rD\ \ loi-i in llrassini species. (,cnome30: 


larhold. K.. M. Iliilhmamr cK II. Ilurka. 20<>2.e I volu- 
tiouary history of the polyploid complex of Cardam 

I., 28. 

I. Syst. Lvol. 

y. R. Grupe & B. \eufler. 
on in Cardaminc (Brassica- 
udeuee Ironi morphological 
. Linn. Soc. 139:275-291. 
/-Vlamo. 2000. Genetic re- 
,1 the genu- />//»/"'" v'.--(Bra- 

r quantitative trait loci sculpting the curd of Brassica 
Icra.ra. Ccnetio 155: 1027-1054. 

& . 2001. Comparative mapping of QTLs 

etermine the plant size of Brassica oleracea. Theoret. 
ppl. (ienet. 103: 383-397. 

, T. A. Del Monte, K. P. Reischmann, J. Hyman, 

. P. Kowalski. J. McLerson. S. Kresovich & A. H. Pat- 

;ene flow in serpentine endemics. 
Amer. J. Bot. 81: 1288-1299. 
IcCue, K. A. & T. P. Ilollslord. 1008. See. I bank iiillu- 

(Onagtaceae) \m, i I Bot P. 5 10-30 
1ckav. J. K.. J. Bishop. .). Z. Lin. J. II. Richards. \. Sala 

,K I. Milehell-Ol.U. 200I. Local adaptation across a 

Sri.. Ser. B 208: 1715-1721. 
vleekins. J. E. H. K. Ballard X B. C. McCarthy. 2001. 

ala. Bras-icaeeae). Int. .1. I'l. Sri. I<>2: l()l K.9. 
/lever. K K. 1973. Oon-pectiis <ler "Tldaspi"- \rten Ku- 
ropas. Urika- mi. I \oid.-r,i-ieii-. Ke.ldes Ueperl. HI: 


. 1979. krilische |{. 'vision der "Thhtspi"- \ 1 1- m 

Europas, Vfrikas nnd \. m I. i.i-i. n-. I. Cc-duchi.-. Mor- 

| >!■< il i •- i' ' nnd I ll logic, lo.lde- Heperl. 90; 129-151. 

.leverowilz. K. \1. I')!!'). \,a!>idopsis. a useful weed. Cell 

-, H. Hurka & H. J. Ban.lelt. 1992. Svstematies 

\ustralian Irpidium I Hra->iea. cad and imp 
cations for their origins: Evidence from IKK analysis 
Huhiseo. I'l. Svst. Kvol. 183: 99-112. 

. C. Esehmann-Crupe & K. Zunk. IW,i. Sul.ut 

polypeptide composition of Huhiseo indicates Diploid: 

nam,/,,. Pliv tnchr„u-l n 31: 129-131. 
. E. Kuhnt, M. Koch & K. Zunk. 1995. Systemal 

lis s 

. Xsl. 

t l)NA v 

. I'l. Svst. Kvol. 190: 75-J18. 

L Kranzke & M. Koch. 1997a. Moleci 

..mergence in fruit characters used in I 

ii ol Tldaspi -.1. llSinssicaceael. Hot. 

, /,,,/, 

d:\clad, >,, (lti.i--ii-.n-.Mrl compley li.l-rd on fluorescent 

\ I I - 1 * data. .). l!ov. Sir. New Zealand .52: 127-113. 

livashita. \. T. 2(H)! . |)\\ variation ,., the 5'upstrea.n 
region of the Adh locus of the wild plant Anilndopsis 
thalianu and Anifns t>emmifcra. Molec. Biol. Kvol. 18: 

. II. Innan X H. Terauolu. P>0(,. I,,,,,,. and ml er- 

-pecihc variation ol the alcohol dehydrogenase lorn- 
region in wild plants \rohi.s gemmifera and Anilndopsis 
lluiUana. Molrr. Itiol. Kvol. 13: I33-13<>. 

. \. Kawahe. II. Innan X K. Terauchi. 1998. Intra- 

and i-|ir. ill. D\ \ variation and codon l.ia- ol the 

alcohol dehvdrogenase (\dh) locus in \mhis and Ara- 
bidopsis -pocies. Mnlec. Mini. Kvol. 15: I 120-1 129. 

plasl |)\ A phylogenv and hiogeographv of the gen 
Irpidium (Brassicaceae). Amer. J. Hot. 88: 205l-20( 
. II. Coja & H. Hriiggeman... 2(H)lh. IWhyphra 

, <„o|„,|. 

Iridium I.. Iltia-ira.-e.u-): l-orlekl , i-rhr loklls-i- 
cruug-mu-ter de. I ritereinheilen von 1 .5-Bis- 
pho-,,hal (.arl.ovvla-.-/l)vvg.Mia-e (III HISCO) al- -v- 
temalischer nnd ph v h. genet i-rhri Marker. I'll. I). 
Thesis. University of Osnahriiek. Osnahriiek. 

. 1995. Should Cardana draho (I..) Itesv. I.e clas- 
sified within the genus h-pidiuni I .. (Brassicaceae)? Kv- 
i.lrn.r Iroin -ul. unit polvpeplide composition of RUB- 
IX (». I.ddr- UrprM l(Hi: 2.5-2H. 

& H. Hurka. P*X). Kvolution of the letra, I 

Capsclla bursa-pasloris ( Brassicaceae): Isoelectric fo- 
cusing aualv-,- ol Hul„-ro. I'l. ->v-!. I vol. 172: 205- 

& . |9<)|. Isoelectric locu-iug analv-i- of 

HUBISCO in Irpidium ( Hrassica<-eae). sections h V ,a. 
h-piocardamou. and Curdumon. Hiochem. Svsl. X Eeol. 
19: 17-52. 

& . 1991. Sulmnit polypeptide coinpo-i- 

lion of Huhiseo and the origin of allopolyploid \rabi- 
dopsis suecicu (Brassicaceae). Bioohe.n. Svsl. X Kcol. 

close relationship 

Yi-iallah. M. K... K. Wec-vvaran. S. Snvdei v \ J. It. Y 
rallah. 2000. Anilndopsis species hyhrids in the slu 

planls. I'l. Physiol. 121: 1005-1014. 

Neuffer. B. 1989. Leaf morphology in Capsclla (Crucif. 
ael: llrprudcurv on em iioument- and lnologioal | 
rameters. Beit.. Biol. Pflan/en. 01: 39-51. 

. 1990. Kcotvpc differentiation in Capsclla. \og 

tatio89: 16.5-172. 

. 1990. K API) analyses in colonial and ancesli 

population- of Capsclla Imrsa-paslorLs (I,) Med. (Bra 
sicaceae). Hiochem. Svst. & Kcol. 24: 393-103. 

-X S. 

Heeologu It: 522-527. 
& H. Hoffrogge. 1909. K.-olvpic ami allo/v in.- var- 
iation of ( ,, III hint i i I n arid ( rubella (Bras- 

■ ■. Bot. M.kIm.I '.7 29') 

1 1 •« - 1 ■ : 1 1 ■ I'. nni-iil, i. \u:ilc- J 

4 H. Hurka. 19H0. Variation of development lime 
owering in natural populations of Capsclla Imrsu- 
s (Cruciferae). PI. Svst. Kvol. 152: 277-290. 
Colonizing i 

12: 807-812. 

duetion dynamic- of Capsclla bursa-pasloris (Brassic; 

& . 1995. Allopolyploid origin of \ralu- 

ceae) in North America: Isozymes and quanlilaliv 

lopsis suecicu (fries) \orrliii: Evidence from chloroplast 

traits. Molee. Eeol. 8: 1667-16JU. 

ind nuclear genome markers. Bot. Acta 108: 449-456. 

& P. Jahncke. |9<>7. H API t analv si- „l hvl.ndiz^ 

& M. Jentzsch. 1991. Isoelectric focusing analysis 

if Huhiseo in the genus Erysimum E. Euearpia Cruci- 

Phvtotax. 32: 57-67. 

erae Nevvsl. 1 6: 16-17. 

& VI. Koch. I90(,. Vlalernal effects ..,. populatio. 

X \1. Koch. \<»>\. Chloropla-t D\ \ ..-In. !,<,!, 

of Capsclla hursa-pas/oris ( 1 .. 1 Med. from N-andinavi 

ile variation and phv h-genetic relation-hip- in the gr- 

and the Alps: Independence of seed weight and th 

in- ridasp, >r„-ll l.ilo lltra-lr.i, rnr|. S V> |. |-,,|. |l). 


& K. Zunk. 1991. Should Tldaspi (Brassicaceae) 

& M. Mey.-,-\\.,ll. \<><u, l.oivp,,' variation in n 

Volume 90, Number 2 

on to man made habitats in Capsella: Fie 
npling area. Flora 191: 49-57. 

-. S. Hirschle & S. Jager. 1999. The colomz 

ol Capsdla in Patagonia (South I 
r and adaptive significai 

, I .1 ,i i 

:, S. L.. Jr. & I. A. Al-Shehbaz. 1997. A synopsis 
„/,„/, ,;,, .IS,.,-...!.--... i N..x«»ii 7:323-327. 

- & . 2002. Paysonia, a new genus segre- 

1 from Irsauerella (Brassicaceae). Novon 12: 379- 

, B. A. Sehaal & I. A. 

■ III .11,-. 'n ! 

I b\ nuclear rD\ \ se(|ii< lice-. 


Syst. Bot. 21: 559-566. 

'. Comparative physical 

.1 i| | in i i ! 

,,,,. alboglabra thi I u i homo i logi us to sequenced re- 
gions ol chromosomes 1 and 5 ol Arabidopsis thaliana. 
I 'I I 2'. 233-243. 

Oshorn. T. C. C. Kole. I. A. P. Parkin. A. O. Sharp.-. M. 
Kuiper K I). .). I.vdiate. 1997. Comparison of flowering 
time genes in Brassica rapa. B. nupus and i,,-/,./..j: - 
Iholiana. Genetics 116: 1123-1129. 

(Kama. K.. VI. Ito. T. \ahara & M. Ono. 1993. Icu genetic 
'differentiation anion- populations ol \rai>is scrraia 

(Brassicaceae) al an all din I L'Vadient. J. PI. Pes. 

106: 143-148. 

Palmer. J. I).. C. B. Sl.i.-I.k I). P.. Cohen & T. J. Orion. 
19,",3. Chlnmplasl DY\ evolution and lh<- origin ol am- 
ol,,, olcid Brassica species. Theor. Appl. Genet. 05: 

liii i: 

(ili«- i 

molecular phvloge- 

.,n .«>;. 

. Oeeol- 


:»' ■■! I- -II - I 

\i ■;.';.. , In.: hii- \on Arabis alpina I 

Kuropa. \orderasien utid \lnka. Ph.D. I'hc- 

sitv of Osnabriick. Osnabriick. 
Pollard, H., M. Cruzan X VI. Pigliiicei. 2001 . C 

studies of reaction norms in Arabidopsis. I. V 

response to daylength. Evol. Ecol. Res. 3: 129-155. 
Pradhan. A. K.. S. Prakash. \. Uiikhopadhv ay & I). Pen- 

tal. 1992. Phylogeny of Brassica and allied genera 

based on variation in chloroplast and mil- ■ ■ d 
'\ \ MM in I, ! > in • I i" "" 

are incongruous. Theor. Appl. Genet. 85: 331-340. 
IVatill. K. 1891. Cmcilerae. Pp. I 1"> - ,,( "» V bugler iK 

K. Praull (edit, .i-i. Die naliii licheii Pllanzenlamilien. 

Vol. Ill, 2. Verlag von Wilhelm Engelmann. Leipzig. 
Price. R. A. & I. A. Al-Shehbaz. 2003. Tuisselmannia 

l Brassicaceae). a remarkable case of fruit modification 

— ' \. It'l'ilin. . o, I \ \l H,. 1,1 ,. lO-H > N s|em- 
alic relationships ol [Aralndopsis: A molecular and nior- 
pliological perspective. Pp. 7-19 in K. View ■m«,l/ X 
('.. Somerville (editors). \ rabidop-is. Cold Spring Har- 
bor Laboratory Press. New York. 
Purugganan. M. I). cX J. I. Suddilh. 1998. Molecular pop- 
nhilion genetics of the \rt,l>idopsh CM UFUWERreg- 
nlalon gene: \onnculral evolulion and nalurallv o.-cur- 

'•!",d -,', V- ■ i ! 

. Gen, 

. 1 ,1 

. E. & L. E. Norwood. 2001. Evolution of Cau- 
hus amplcxicaiilis var. barbarae (Brassicaceae), a 
: serpentine endem 
c perspective. Amer. J. Bot. 88: 

2001. Gen, llou 1- ,! I oi..,i . in two endemic 

plant species. Biol. Conservation 100: 21-34. 

'iuliueci. \l. 2002. Kcologv and evolutionary biology ol 
s i22 \ugust. 2002) In < B. Somerville & 
K. M. Vleverovvil/ (editors). The \rabldopsis Book. 
\inerican Soeielv of Plant Biologists. Kockville, Mary- 
land. ,lo,/IO.II99/lab. 0003. <hltp://u u 

& N. Byrd. 1998. Genetics and evolution of phe- 

,,',,',",.,' '." ' i," ,", ' i!„ - 

& K. T Marlow. 2001. Differentiation lor llouenng 

, A. L. Bovles & J. 1. Suddilh. 2000. \anal„„ 

selection at the CAULIFLOWER floral homeolic 

olemcca. Gem-tics 155:855-862. 
Quiros, C. F., F. Grellet, J. Sadowski, T. Suzuki, G. 

T. Wroblevvski. 2001. Arabidopsis and Brassica cor 
alive genomics: Sequence, structure and gene co 
in the \BII-Rps2-Ckl chromosomal segment an 
lated regions. Genetics 157: 1321-1330. 
Ramsey. J. & D. W. Schemske. 1998. Pathways. m< 
nisms. and rates of polyploid f 

Melee. Biol. 12: 93-113. 
Reboud. X. & C. Zevl. 1991. Organelle inheritance in 

planls. Heredity 72: 132-140. 
Il.dunbun. .1. I.. I'. M. Wellz. \1. V. f'av. O. B.C. Clone. 

11. P. binder. G. Beeves X M. \\. (dulse. 2001. Ilapid 

of South \lriea. Yiliire 112: 181-183. 

Rieseberg. L. H. yV C. R. lander. 1090. IM„,d ,-h.~ili- 
ealion: Insights from genetic map-based studies of ex- 
perimental hybrids. Kcologv 80: 361-370. 

, B. Sinervo. C. B. Under. VI. C. I ngere, X D. VI. 

. J. Whitton & K. Gardner. 1990. lb bud /ones 

l". . ",, - ,.. ,«s..„„, t,cs 152:713-727. 

Bodiiian. J. I-!.. R. V. Price. K. Karol. K. Conti, K. J. 

V.m. Missouri li„l ( .".< ) (,!!<, (><)', 

. K. <;. Kami. I{. \. Pnce. K. Conli K K. J. Svlsma. 

1991. Nucleotide sequences of rbr\. eonlirm the cap- 
paralean affinity of ihe Australian endemic Cvrnslc- 
mnnaceae. Austral. Svsi. Rot. 7: 57-69. 
. . & K. .1. Sylsma. 1996. Mole- 
cules, morphology, and Dahlgren's expanded order Cap- 
parales. Sysl. Not. 21: 289-307. 
ollins. B. C. 1993. The Crueiteraeol Continental \,,,ll, 
. Stanford Univ. Press. Stanford. 

Vcarkan, R. I lei 

eluding the RPS2 gene ii 

Theor. \ppl. Cenet. 96: 468-174. 

. P. Oaul.ier. M. Delsenv vK C. I. Chun-. !<■'«,. 

Oeiietic and phvsieal mapping in Hrnwini diploid -pe 
eies of a gene cluster defined in \r<ihi<li,psis ihulxiiKi. 
Vlolee. Ccn. Oenet. 2.1 1 : 298-306. 

anderson. M.J. 1097. \ non parametric approach to es- 
timating divergence limes in the ahsenee of rate con- 
stancy. Molec. Kvol. II: 1218-1232. 

P. S. Soltis. K. K. Soltis & J. J. Doyle (editors). Molec- 
ular Nvslemalics ol Plants II. kluwer, Boston. 

. 2IMI2. Insinuating alisolule rate- of molecular evn- 

V penalized likelihood ap 

80: 59-71. 

hierup. M. II. 1998. The ..fleet of enzyme lieh-.../M l „>iix 

eompalif)le [rahis petraru ( lir.i — i.-aecae) I leiodil.i- 
128: 21-31. 

prospects. Cur,-. Opin. I'l. Mini. 3: 97-102. 

. A. Acarkan cK K. Hon in. 2001. Comparative 

structural genomics in the Brassicaceae family. I'l. I'hv- 

gemeinsames Meeting zur Phv logcograplue von ar 
lisehen und alpinen I'flanzen in /iirieh. 1-3 Juni 200 
Naiihinia 15: 69-90. 
Slolz. II.. J. Bishop. C. \\. Bergmaun. \1. Koch. I'. All), 
sheim. A. C. Darvill vK J. M. Kabavitch. 2(MM». Kvidcr 

lul.ilors. I'hvsiol. Molec. I'l. I'athol. 50: 117-130. 
Strauss. S. V. I). H. Siemens. V1. B. lie, her 4 \ T. Mil, I,.- 

Olds. I'W. Kcological costs of plant resistance to h. 

Iiivore- iii ihe cm it'll. \ ol pollination. Kvoluli.m 5 

I 105-1 I 13. 
Svveenev. I'. \\. ,X. B. \. Price. 2000. Polvphvlv of tl 

genus Drnlaria (Brassicaceae): Kv idencc from lru\. i 

lion and mlhV se,|ucn.e data. Svsl. Bol. 25: 100 17! 
Svvanen. M. 2002. Balo ol ribovmial B\ \ ew.lulion a 

uniquely accelerated in Kukaryoles. J. Molec. Kvol. 5 


. A. M. Kloyd vH M. I). Mark.- 

molecular genetic anahsis ol I 
orphogeiie-is in \nihi(l»i>sis. In 

Sci. 5: 21 

Thompson. J. I). v\_ K. I.umaret. 1992. The evolul 
dynamics of polyploid plants: Origins, eslablis 
and persistence. Trends Kcol. Kvol. 7: 302-307. 

Toledo. J.. P. Dehal. K Jarrin. J. II M II , in,, ,X ". 
h V nlinm mru; 

markers. \,in, Pol. (Oxford) i!2: 523-530. 

Tonsor. S. .).. S. Kalis/.. J. Fischer & T. P. Itollsfonl. IOT,. 
A life-historv has I nil- nl »»pul n <;t'iielic struc- 
ture: Seed hank i< > lull in /'.' >i lunccoltila. Invo- 
lution 47: 833-843. 

Tremetsberger, K.. ( ;. komg. I!. Samuel. \\. Pinskrr K I. 
K Stuessv. 2002. Infraspecihc genetic variation in /,'/* 

PI. S\si. Kvol. 233: 163-181. 
Isiikava. II.. J. Vok.nama. II. Ikeda. II. kuroiwa. 4. ku- 
roivva & K. Iwalsuki. 109,. Morphological, pin -iotogi- 
eal and molecular genetic eliaraeteri/ali. ! \n<h«ln r , 

110: 15-23. 
Frbanska. K. M.. II. Ilurka. K. I.andoll, 15. Noulfer & K. 
\|„„ mcnlicn !')«>;. Ilvhridi/alion and evolution In 

- u , .i. '■!. I. ii I !<■« uku evidence. PI. 

Svst. Kvol. 204: 233-256. 
I VI. k. 2000. Kvolulionarv dynamics of self- 
Van der /wan. C S. A. Brodie cK J. J. Campanella. 2000. 

The inlia-pc. ill. | I I. i - I > , , 

in worldwide Mf.ul.iu ■ - i l.< i 

- I I I I: 

i L Constance. 1008. 

'cula (Apiaceae). Proc. Nat 

•ad. Sci. U.S.A. 95: 235- 

•• in « ' ' dii| II 1 ' ik. in *' ' ' ' •/» -< i ii 

290: 2114-2117. 

Wallhcr. O.-H.. K. Posl. P. Convex. A. Men/el. C. Par 
mesan. T. J. C. Becbee. J.- VI. Fiomenlin. (). Itoegh- 
Guldberg & F. Bairlein. 2002. Fcological responses t, 
recent climate change. Nature 416: 380-305. 

Warwick. S. I. cK I. A. Al-Shehba/. 1008. Cenerie evalu 

rea. I. \<»\on ii: 521 325. 

& L. I). Black. 1991. Vlolecula. sv.lemalic-o 

Brassica and alii. ■' • . • i iNihl. m !m ■> I i 

siceae) -Chloroplast genome and evtodeme congru- 
ence. Thcor. Appl. Genet. 82: 81-92. 

& . I00I5. Molecular relationships in sub- 
til !i -- « 111 « 1 !« 1 . 11 1 hi I 1.1 1 ... 1 « 

J. Hot. 71: 906-918. 

& . 1901. Kvaluution of the subtribes Mor 

icandiinae. Savignyinae. Vellinae and Zillion, lll.^i 
.i.i ml. I5i 1 . • - 11 i' . iil< 11 1 I ll' 1 
lion she variation. J. Pol. 72: 1092-1701. 

& . 1997a. Phvlogenetio implications ol 

,-lilr„,,p1a^i |)\\ rcstriclioi. >\\r variation in subtribes 
I! i| h nun 1 in. I I 1 I" 1 llha 1 1 I .. I '.. i - -11 

rea. i. Canad. .1. Pol. 75: 960-973. 

& . 1997b. Molecular phytogenies from 

ih. 1 r 1 in hi 11 in in 1 I If 1 I P.i 1 

siceae. P, as.ieacoae). Opera l!oi. 132: 159-168. 

, — K I. \guu,agaldc 1002. Vlolcculai ms- 

1 11 ■ i ■"■;<• m ' ni> I I • 11. 1 1 n ' 1 « Pi 1 

eae)— Chloroplast l)N\ variation in ll 
s. Theor. Appl. Genet. 83: 839-850. 
ihehbaz, R. A. Price & C. Sander. 200! 
'ssmluium I Brassieaceae) based on II 
uclear ribosomal DNA. Canad. J. Br 

- allopolyploid speciation 

e 39: 15 
iberger. 1999a. Molecul 

■ r . thus 

; origin of the 
(Brassicaceac). Amer. J. 
Hot. 86: 1282-1289. 

& . 1099b. Intensive inlraspcche chto- 

, , , 1 \ (, f ,|.\\i. 1iMi1.11 in llu alpine Drabaai 
zoides L. (Brassii a. . ic) Ih. ..!.,». 1 ln.n up ml 
population structure. Molcc. Fcol. 8: 1405-1415. 
fang, Y.-W., W. H. J. Kuo & 1. H. Wong. 199,", („ „eii, 

1 1 • s 11 - 1 1 1 . iii - 

RAPD markers. Bol. Bull. Vcad. Sin. 30 : 

s W. -H.Li. 1999a. 

m the internal tran- 
rDNA. Molec. Phy- 

ships in tribe kepidi 
roplast DNA restrict] 

lethods have been 1 

in m. 1 P. i . 1 ! << il 

according to the maikei sv-h-m- used, lis 
calls b\ la\a and then In authors. The lis] 
studies at the popiilalioual level or above. 
I. Isoelectric focusing of ribulc 
carboxylase (IEF-Rubisco): 

1. In 1 1 1 1 • ., • ' 

(M..i..m«>nlioir& Hurka, 1994). 

in ,11 mil. il '. I In 
|\1 immenholl el al., 1993). 

... |\l I I \ hlll/M-h. |99|). 

CM.. 1.I...II I'M'* l<i«)5 M..1..1 

Hurka, 1991; Mumm. iiIk.II et ah. 1902). 
,.!..■ ,. .,!.. I'»«: Vlu.nmenhoff& 

i .-i i.i.. 100.;,. 

s (Roy, L995; Roy & 

Wraut sul,sp. petraca, , 

. fecunda (McKay el al.. : 
. serrata Franch. & Sav. I 
'isculclla (Tremetsberger e 
W/,m, A. & I). Love, as 

Rieseberg, 1989). 
rassica (Chevre et al., 19* 

1995a, 1995b). 

Brassiceae (Anderson K Warwick. IW Simonsen K 

Hcncen. |005 a . 1 9051, |. 
r;«/«v/Ai (llurka. IOOU: llmka X l»iirin». 1991; llurka 

& Neuller, 1997: NeuflerK Hurka, 1999). 
6'. hum,-,mstoris (Neuf'fer, 1996: NeiifTer & Hoffmgge. 

1999; Neufferei Hurka. 1000; Neufler el al.. I'W). 
C. mhc/la | Vuf'ler K llollmgge. IW)|. 
.W,/„„„ril rbanskad al.. 1997). 
C. annua (Ko, I, et al.. 200.1a. Mai hold el al., 2002a). 
(A /wii/mm/a I., agg. (r'ranzkc K llurka. 2000). 
Cacldcana (Koch et al.. I'OlIa: Koch. 2002). 
Drain, ( Brochmann. 1002: Bruchmann el al.. 1991. 

,fcndlcni\. („aMS. Walson (Cabin. 1996). 


Voel, v \ I 

II, , 

■ Mill. 

► total nuclear DFNA: 


IV. Random ai 

Capsclla hursa-pasloris (Neufler. 1996; Neuller el 

1990: Van- el al.. 1998). 
Cardaminc (Neufler K Jahncke, 1997; Urbanska 

Cochlcaria (Koch el al.. 1996). 

/Mi/wi (Scheen et al., 2002). 

bepidium mexenii (Toledo el al.. 1998). 

V. V.uplified fragment length ,,„| 


Aruhidopsis llmliana (Sliarbel el al.. 2000). 
Can/,„nnn unuii,, I Ma. held el al.. 2002b). 

(Jars, .nam,, II. I >cliul/. I'm In. Union II,. 

100.-,: \an der /wan et al.. 2(MM)). 
1. /vf«/« subsp. /rtnira (van Treuren el al.. 1997). 
Brassica (Harrison X Heslnp-Harrison. 1995: Saal «■ 

Hiplohni* iMarlin X S.ii.rhe/- Vlamn. 2000). 
VII. I)NA Sequencing of 

I. Ue^l.i. -lion fragment U 
) ,l,lo,o,,lasi (ep) DNA: 

■ahidopsis. including Cardaminopsis I Mummeiiholf \N 

llrassna iSi.iui «-t al.. I«><>:>; War 

I00.{. IWI. |907a. I007I,: V 

Brassicmae (Warwick K Black. 

Cakilmae (Warwick & Black. 1007a). 

Cardaminc (Urbanska el al.. 1997). 

Diplolaxis (Warwick et al.. 1002). 

Drofea (Brochmannet al.. 1992c). 

I.epidieae (/unk et al.. 100.5; /, m k et al.. 1999). 

Irpidium (Miiintiiei.hol'l et al.. 1995). 

MUrollilaspi (Koch et al., 1998c). 

\l.,| (Warwick ,\ Black. lOOll. 

Kaphaninae (Warwick X Black. 1007a). (Warwick K Black. 1 00 I). 

Sisymbrieae (Price el al., 1994). 

Strcptanlhus [M^c-v & Soltis. IWI). 

Thel\podicae(/unk et al.. 1006). 

Thlaspidineae (/unk et al.. 1996). 

Thlaspi (Mummenhofi ,\ Koch. 1994: Mummenhoff et 

al.. 1007a. 1007b: /unk el al.. 1006). 
Vellinae (Warwick & Black, 1994). 
/illinae (Warwick & Black. 1094). 
(b) nuclear riboMimal DNA: 
Brass,,,, (llclscn el al.. 1 00( ) : Malus/wiska X lleslop- 

lirassicaceaclkoch d al.. 2001a). 

Cardaminc. including; Dcnlaria ( 
Price. 2000). 

; Tsukaya et al., 1997). 

Arabis (Hoy. 2001). 

I. holboellii (Sliarbel A Mitchell <H,k 20(i| 
/»'«.■. Inn, (Hon. 2(K)1). 

Cardaminc. including Dcnlaria (I ran/ke el . 
SweeiieN K Price. 2(KK): Meeker el al.. 2( 

C pralcnn agg. (kran/.kc K llurka. _> ,. 


Cochlcaria ( ko< 
Or«/>« (Koch & Al-Shehba/.. 2(K)2). 
Halimolobos Tauseh (Bailey el al.. 2002). 
h'pidium (Mummenhoff et al., 2(K)la). 
Mancoa Wedd. (Bailev et al., 2002). 
Pennellia Nieuwl. (Bailey et al., 2002). 
/,',</,,./>„ (Bleekei X llmka. 2IMII: 1)1, , k, i , 
■V" ".'".■„„ I, in, inn Shane, iBailcs A |l,,\l, 

ley et al., 2002). 
Strcplanlhus (Pepp«-r & Norwood, 2001). 
Yinshania (Koch & Al-Shehba/. 2000). 
(c) nuclear roiling alcohol «l. Ii\ . 1 1 . i-.n; 
Wr,W.y..w\ iMisa-lula el al.. I 000. lOOJi). 

; Waters ,K Schaal, 1996). 

h'arcnuorthui To it. (Charlesworlh et ul.. I'Mi). 

(d) nuclear coding S-allele>: 

ii, i ((.liaik ^worlh et al.. 2(MM)). 

■ .1 I ' Ml , 

(e) nuclear coding clialconc synthase (CHS): 
B.assicaeeae (Koch .'I al., 2000, 2001a). 

<('} nuclear codiiiii argininc dccurliow luxi- ( VIM.): 
Bm.-sicaccae (Gallowaj et al., 1998). 
(g) nuclear coding 2S albumin: 
Arabidopsis, Brassica (Boutilier el al.. 1999). 

ili I I i i J J'lHOl 

, il ninli-ai- coiling .-iciilic i-hiliii;c.e: 
\i.,l„ Ilii4,..|. e| al.. 20001. 

, HI II I < II ! II III V "'I 

ade, 2000). 

(k) nuclear coding il -al genes ! Vl'l I V 


I , nun ' n-_a.Mii. F>90; |'„- 

I'rnnvlii,, (Railcx el al.. 2002). 

Pringlea W Anderson ex Hook. f. (War 

Prltzelago (Kropf. 2002). 
Romaiischulzia (.). K. Schulz ( Warwick < 
Rorippa (Yang et al., 1999a; Bleeker et 

,,<-. Gn 


(Bailex & Dovle. I <>00 : Pa, lex . 
irwick et al.. 2002). 

Stanleya Nutt. (Wan 

SiivpKiiiilii'llii l!\(lh. (Warwick el al., 2( 

Strcphinlliiis (Pepper Si Norwood, 2001: 

rin-hpn,Ii,,psi> (Warwick et al., 2 
Thehpodium Kiidl. (Warwick et al.. 2(M 
r/j/as/*/ (Koch & Mummenhoff, 2001: \ 

1 DNA (ITS): 

Arabidopsis ((VKane et al.. 1997; Yang et al., 1999a). 

\mbis (K.m-Ii el al.. 1999a: Kov. 2001). 

/W/,em (Ron. 2001: Koch el al., 2<XK.b). 

Branca (Yang et al., 1999a, 1999b; Warwick et al., 

Brassicaccae llleenan el al.. 2002: Mileliell X Mr. mm. 

Cardamine. including Denturiu (Fran/.ke »S. Mummen- 
hoff. 1999: Franzke el al.. 1998; Bleeker et al.. 

C. pratensis agg. (f'rau/ke Ox Mnrka. 2(K)0). 

Caulanthus. including Guillenia (Pepper & Norwood, 
2001: Warwick et al.. 2002). 

Cochlearia (Koch et al., 1999b). 

,;,-„„//,c<r'ranci>cn-Orlegaet al.. 1999). 

Descuruinia Webb & Berthel. (Bricker et al.. 2000). 

Koch & \1-Sheh- 
baz. 2002: Widmer & Ballisberger. 1999a, 1999b). 

Dryopctalum A. Gray (Warwick et al., 2002). 

Eruraslnun C. IYes|'< Warwick el al.. 20(12). 

Halimolnl,,* Ml x el al.. 2002). 

Ilonumnia I kropf. 2002). 

Hymenolobus ( _!im>_... 

h-pitlium (Bou man et al.. 1999). 

Mamoa (Bailev et al.. 2002). 

I.'. ' ■ I - I 'I al. 20021 

\eolorulami Hedge <K J. Leonard (Warwick el al.. 2002). 
i ' ih i Loll et al.. 200lh). 

Brassicaccae (Koch el 
Itilttmnlobns I Railcx c 

Manr-on (Bailey et al. 
Pennellia (Bailey et < 

■ I . 20< .i 

. 2002). 
L, 2002). 

■i (Bailey & Doyle, 1999; Bail 

Irpidium (bee et al.. 2002). 

(q) coding mitochondrial n«d4: 


19 . 


i \ . ,ki. . I il - 2tMMI \n„„ , I ,1 . 200HI 

...'>! i:' hn. .,' I . : I""- < ■ I I I ""' < i 
ner et ab. 1998; Lagercrantz. 1998; bagererantz el al.. 
1996; ban & Paterson, 2000. 2001; ban et ab. 2000; 
Rossberg et ab, 2001; Ryder et ab, 2001: Schmidt, 
2000: Schmidt el al.. 2001: Sillito el ab, 2000). 
Capsrlla bursu-paslonsilAmU-el al.. 2001). 

ab. 2000: I 

Review (Schmidt. 2000: Schmidt et ab. 2001). 
IX. Short interspersed elements (SIM): 

Brassiceae (Lenoir etal., 1997). 


ccics arc a major glohal lineal lo l..,||, Inodiversilv and agriculture and llu 
science. Governments icco-mi/.- this ami aie devoting increasing resources lo\ 

as lend lo lie verv favorable lo in\asi\es. e-peeiallv the ueedv -|ieeies. We i 

umbers and with sued speed as today. Musi are 
enign. Itul a dangerous few cause major environ- 
lenlal piohlems. Invasive species inav thrive in 
leir new environment and dramatically change the 
vnanues and composition of die eco~v -tern. Ih 

The simplest answer to this question is "ever 
where"! Consider the flora ol Urilaui: It mclud< 
Polygonum amplcxicaulc I). Don from the Him, 
lavas. Dianthus caryophyllus L. from southern Ki I'a/miri sommlcnim I,, from western Asi 
,aus, of our la. k of v.plau,,, w, now suffer la,,,- '"'"'""'"" '"rr'l "l' ^ '«"" !""f' *""''"' 
. )nnn , n . ■■ , | i i • ■ >'■'»''/./ imnir-zrlamliur I. knk from \e« /ealan 
Iruusport (,.. B „, ,,a ,,,us,,.| W,™„ ,,,,/, r ,, ,„> ^ ^ ^ ^ ^ ^ ^ ^^ . ^ ^^ ( 

Asian clam (Corhiculu flii>„„,<<,,. ,, 

, uu .. . . , , , . include species from around the urn Id. 

2(MM)). and witness the degradation or even re- s[)( ,, u . s _ |)OU| . natjve flom r 

" (e.g., purple loose- a l flora. Species can he introduced 

to anywhere, it would seem. That 
; are more likely t, 

: species than others. 

strife {Lythrum saliraria L). Thompson et al., 1987; J* "anywhere^ 
Pimentel et al.. 2(HM)). Growth of the glohal trade , [m| sn||I<1 S|) " (J 

> ihe prohlem is inherently glohal. we musl () ur arguI11Ptlt is lnal 

! ' IL prime habitats I,,, invasive specu-s. l{ecogniz 

h focuses on iden- ,| le absence of an ideal database, we selected a 

that enahle them () f species from the Worlds Worst Invasive Al 

ade. The idea is to identify who will invade Species list compiled hv Lowe el al. (2001) to : 

sties of s 

I, i I ill.' V I, 

vital approach to solving the top three plants, and the i 

L. [51 
mta R. M. King & II. h 
■ocephala (Lam.) de Wit 
ders will he. The counts of c 

per takes an alternative approach. We attempt lo Chmmoluvim odorata \{. M. King & II. Koh. |88|. 
identify where they will invade rather than who the and h-uruetm Irucorephala (Lam.) de Wit |,T7|. 

"' l> " ol I c..l„g\ :l „d KwlulionaiA |;i„|„p. | mv.-ilv ,1 lenne.-ee. .,(,<> l);il.n,\ Hall. knovu 
.5. Wo. I ..S.A. Present .,,1.1, e.-: I.', \alural lo- ,ii„ , ■-.. I )e,,ai tl.lent ol 'Fisheries and \\ ildlile. Michigan Si 
I'm Landing. Michigan IS5H2 1-1222. I .S. \. |enk,ns(«, 

■■Nicholas School ,,1 Ihe rimmnmenl and I' .nil, Sciences. Duke Univcrsilv. Uos <><).',2o. ham. ' 

27708. U.S.A. Stnartlm»m(«aol.eom. 

Ann. Missouri Hoi. Gard. 90: 172-178. 2003. 

i>l;iiul nations. Nonetheless, all three species have 
in\ .)• I.mI large areas. Chromohicna odorata. native 
to tropical America and the Caribbean (Holm et al., 
1977: 212-216; Holm et al.. 1979: 85). invades 
pastures and croplands in much of tropical Africa 
and Asia, the Southeast Asian island nations, and 

Global assessments of landeover change provide 

series of somewhat overlapping estimates of dis- 

SuXXm. 'rTra (Holln 71 1077: 299-302: 'l Mm turbed habitats. A rough estimate of the size of the 

et al., 1979: 207). is a widespread 
turelands in South 

croplands, with (2) degraded grazing lands, and (3) 

Weed of National Significance), parts of southern the- areas of cleared forests t 

\ln, a uid \1 u _ Ik U. liih n uu an, and ex- 
treme southern parts of the United States, h-iuanui 
h-ticoccphala i- native to tropical \merica (Holm et 
al.. 1979: 214; Lowe et al.. 2001). but in its non- 
native range can form dense, aim 
stands that render large areas uni 

(1) Of the ice-free land surface of about 129 r 

All three species occur in heavily disturbed ar- 
as, as do many other widely distributed invasive 
peeies. such as Mimosa pi«m I... various species 
AOnunlia. and liirinus omniums I., \nother way 

the planet. Ml but I million km- were converted 
from naturally wooded or forested ecosystems 
(Pimm, 2001). 

(2) The world's drylands cover roughly 61 million 

shriiblanils. and savannas. The area varies from au- 
thor to author depending on what one means by 
"dry" (Pimm, 2001; Vitousek et al.. 1986: Olson el 
al.. 1983). 

Most of the world"- dn lands suffer from desert- 
to view this problem is to survey (hs.urbed areas. ifi( . atioii _ t| , a| . g( , |(()| . |i() | in „, moslK human-caused 
Throughout the tropics, it ,s our expenence that |)m|)|rms |h;|| (|( , |)n . ss |)|aM , ,, n>(luc tivity. Some 23 
disturbed habitats will have at least o...- and oil, n ^.^ , ^ ., o{ . ^ (|| . vhm(ls ,,.„,,. damaged vegeta- 
several of these listed Often they will be ^ (|)l . ( , gM(1 _ , 986) T|l ; s „<„.„ means the spre ad 
common; sometimes they will be the dominant spe- ^ j 11( . n . as ,. () f unpalatable plant species following 
cies. Searching for "worst weeds" on the World over ^ a/i ,^ , )V ratlU ._ gnals . am | ot her livestock 
Wide Web produces a list ol r. peal oiled, ■ - that ^^ & , (m ,| ;m< |_ | W() . . X ,,.| ie r, 199 4; Bahre & 
cause economic harm to croplands and pastures. She ] ton ^ jc,^. \ an Auken. 2000). Those will some- 
Most are exotics, but not all. t j mes i )e nal j Vr species, for instance, mesquite (na- 
In short, these and oilier examples suggest that |jv( . /> roSl)pis s|)|)-) | uls dramatically increased in 
invasive species occur predominantly in disturbed ( | ( . Msilv nv ,. r miu .j, „f [|ie southwestern United 
ecosystems. By "disturbed." we mean major eco- Stales' (Bahre c\ Shelton. 1993; Archer, 1994; Van 
system changes, such as conversion to croplands. \uken. 2000). More often, the weeds will be exotic- 
grazing lands, urban areas, or anthropogenic ones. i nvas j v «-s such as Opuntia in Australia and Africa, 
such as grasslands where there was once forest. ( 3 ) Another 40 million knr of the land surface 
We will not further belabor this connection be- has fort , s , s or U()()( |[ an( ] s of one kind or another. 

tats, for it is well-established and n bviously ( | er j II( .l ll( | ( . s wetlands and urban areas (Olson et 

controversial. Rather, we accepl the connection and £., 1983). 

move to the problem of estimating how large a frac- The conversion of forests to other habitats is 

tion of the Earth's land surface humans have mod- more complicated, since most of the world's crop- 

ified. Since invasive species in these disturbed lands were once forests — and so have already been 

habitats are often deemed to be weeds, we can re- counted above as croplands. Most of the converted 

phrase our question to: How big is the global weed forests are in temperate regions. About 2 million 

patch? km 2 of these forests have also been converted to 

We proceed by estimating the size of the weed grazing lands (Pimm. 2001). 
patch at first global, then regional levels. Globally. Modern human actions have shrunk the world's 

the- area of disturbed habitats is large. Regionally. tropical humid forests from an original area of from 

we find that our global estimates are too small, for 14 to 18 million km 2 to about 7 million km 2 at 

there is much disturbance thai thev miss. It is not present (Myers el al.. 2000; Pimm, 2001). Again. 

irface. The potential 

weed patch is huge. 

approach is inevitj 

ibly approximate a 

iere these disturbed 1. 

inds are. To both refi 

•stimates and provide 

• a check on their i 

we will mm exploit 

; detailed estimates 

ter. We .1., so first at 

a global scale, then 

.1 scales. Our analyses 

i relate primarily to t 

the exact numbers depend somewhat on what one to identify landcover in ca. I9<>2. As the primary 
means by "humid." Yel only about 2 million km- concern of onr study is disturbance (i.e.. areas vul- 
of croplands are in what was formerly, humid for- nerable to invasion), we locus on tin- disturbed 
ests. Some 5-9 million km- of humid forests have classes of the GI.CC (croplands, mosaics of crop- 
been converted to nominally grazing land, though lands and natural vegetation, and urban areas). The 

much (.1 il ha- \e,\ |„w slocking rales (I'nnni. lesoluli I' ihi- map is approximately one km-' al 

2001). the equator. 

Summing these three pieces suggests that l.-> We also identify areas that have changed land- 
in. 1 1 ion km-' o| present-day croplands. 2.i million cover, but not necessarily into croplands or urban 
km- of drylands. 2 million km- of temperate forest areas. For example, the conversion of forests into 
converted to grazing lands, and from 5 to 9 million grasslands lor grazing will „ot appear as disturbed. 
km- nl additional forested laud not producing crops but it obviouslv is. [Grasslands are a natural type 
have sustained sufficient changes to their vegeta- of vegetation, but not where the original vegetation 
lion In make them target areas lot invasive species. was a humid tropical forest.) We do this to assess 
The combined total is just under half the ice-free potential error causes and means of improving on 
our main analysis. 

plete and detailed analysis ol landcovei < hauges of 
all ecosystem types. The BIOMF and GI.CC maps 
use different classification schemes for vegetation 
that make matching corresponding classes between 
them somewhat arbitrary. For example, the CFCC 
map has an open shrubland" class that corre- 
once-loresled half of the planet since the remote sponds to souk- grassland in the BIOME map. How- 
sensing imagery on which we rely does not so read- ever, open shrubland also includes areas that are 
ily detect the damage to (Inlands. obviouslv not grassland, such as central Australia. 

which the BIOMF. map classifies as hot desert. It 
A Sl'VTiM.l \ Fm-I.KIT Cl.niui. \ssi:ss\1K\t (il- is uncertain thai these changes represent land deg- 

CoWKKTKI) FoUKSts radalion: more hkelv. thev represent differences in 

classification schemes. 
For the global analysis, we use a Geographic In- 
formation System (I BDAS Imagine v 8.5) to com- K ,, s| ,,,, VN|) | )|S(;|SS1()N 
bine a map ol presumed original vegetation with an 

estimate of current landcover. The result is a global Not surprisingly, the global analysis confirms 

map where each pixel has information about its that humanitv has disturbed a large fraction of the 
original vegetation and if it has changed, or not world (Fig. I). Of the ca. 129 million km-' of ice- 
changed, into a different type of landcover. free land. ca. 27 million km- appear to have been 
The original vegetation map is from the bite- converted to croplands, mosaics of croplands and 
grated Model to Assess the Global Environment natural areas, and urban areas. Table 
(IMAGK) project ( & van den Born. 1994; 
Aleamo et al.. 1998: IMAGF t 

17% of the laud. Disturbance concentrates in leni- 
ties, ror a detailed description of this model, see perate climates (temperate forests, warm mixed for- 
Prentice et al. (19)2) II md van d est. grassland/steppe) and the drier subtropical and 
Born (1994). The resolution of this map is one-half tropical vegetation (scrubland, savanna). Again, not 
degree of latitude and longitude. Color maps are surprisingly, most of this disturbance coincides with 
available from Aleamo et al. (1998) and IMAGF the world human populal . d croplands, n.nsl- 
team (2001). | y in the Northern Hemisphere. 

The current landcover map is from the Global Grasslands, scrublands, and savannas have lost 

band Cover Characterization (GI.CC) (Foveland et from a fourth to a third of their original area (Table 

al., 2000). A digital version is available al (l.ttp:// I), but these numbers may be misleading. Dislur-,nl). This project used bance from livestock grazing on such ecosystems is 

a one-year sequence of AVHKK satellite imagery difficult to detect by satellite. Most of the very cold 

,. Mack arras include croplands, mosaics ..i croplands and natural u-uelaliou. 
rent landcoxer is from the OLCC map. This map uses a geographic projection 

areas (tundra and boreal forest) anil the very dry However, the U.CC data do not include separate 

(deserts) escape major disturbance. categories for dr> habitats (such as grasslands) that 

Of the combined disturbed area.. 1 1,2 million are purported to originally have been forest (Love- 

km 2 are from oneedon-led area- (fable I. sum of land et al., 2000). What happens if we assume that 

disturbed boreal forest, cool conifer, temp, mixed these are also converted landscapes? We approxi- 

and deeid. forest, warm mixed forest, and tropical male the area of lores! to dryland conversion by 

woodland and forest). This seems In be a lov\ esti- identifying tropical woodlands and forests that are 

mate, for we should compare it with the 1 1 million now ,ma»lan<l-. -ax annas, or woody savannas. The 

km 2 of forests that are now croplands plus a further area is about 2.4 million km 2 . Adding in this piece 

minimum estimate of 5 million km 2 of cleared Hop- suggests that 16.6 million km 2 of forest have been 

ical forests not converted to croplands (above and converted. This i< dose to the lower estimate of 16 

|» imm . 2001). million km' based on combining independent es- 

Table I. Original area of each <,.■,, talmn 
vegetation, and as urban areas. Drylands are 
map), but are now grasslands, savannas, or vv 

( ro|. lands, mosaics of croplands and natural 
■opical woodlands or tropical forests ( UIOVI K 
. All areas are expressed as l<M)0> ol M|ua,c 

Temp, mixed forest 

Warm mixed forest 


27 11 

le. The area 

edge disturbed 

7525 km-, which is slill less than two-thirds o 
TRFIC estimate. Kven after correcting the C, 
ma|) lor unnatural drylands, it still misses a 
of the disturbed area. 

For the Atlantic Forest example, we map f 
cover in 1999 using Landsat 7 FTM + satellite 
agery. I sing standard .supervised classilic; 
techni(|ues, we classify seven Landsat images 
lores! and non-forest classes. We do not (listing- 

specific plantations (e.g.. eucalyptus) are not a I 

We selected two tropical forest areas in 
with di Hen lit disturbance histories to try to idenlifv 
what the global analysis is missing. 

The Amazon is relatively intact but has had high 
rales of recent deforestation (i.e.. within the last 30 
years. Skole cK Tucker. 1993). Northern Mato Cros- 
so stale, in the southeastern Amazon liasin. is our 
example ol such recent anthropogenic disturbance. 
In contrast, most ol the Atlantic Forest was defor- 
ested more than 30 years ago (Fundaeao SOS Mala 
Atlantica, 1998). The state of Kio de Janeiro and 
the surrounding area is our example of such his- 

For the Amazon example, we calculate defores- 
tation using forest cover maps from the Tropical 
Rain Forest Information ( ienlei (TKFK i. 2002). We 
use maps for 1992 to match the year of the GI.CC 
map. We also simulate the undisturbed condition 
by replacing the deforested class in the 1902 map 

il analysis significantly underestimates 
in this region. According to the CA.CC, 
example shows jusl 3913 km 2 of dis- 
. I hill is only about a third of the re- 
late in the previous paragraph. Another 

obs.). The World W ildlife Fund eeoregion map pro- 
vides an estimate of original forest cover (Olson & 
Dinerstein. 1998). 

In the Atlantic Forest example, the analyses of 
Landsat imagery show thai 91.993 km- of forest, of 
an original 127.850 km', has been lost to defores- 
tation. The area ol disturbed lands in the (d(.C 
map is just 51.851 km-. 56<7r of the Landsat-de- 
rive<l .•sinuate. However, adding in Ibiesi- eonverl- 
ed to drylands increases the area to I I 1,000 km'. 

In both examples, the best lit of estimates from 
CA.CC data to detailed regional estimates comes 
only when we recognize the conversion ol forests |o 
obviously disturbed habitats (such as croplands) 
and less obvious categories (such as grasslands and 
sav.uuiasl thai could be natural ei os\ stems, but are 

Two factors contribute to these deficiencies ol the 
CAXIC map. One. the AVHRR imagery used in 
making the GLCC has limited abilitv to discrimi- 
nate vegetation types. The AYHHK sensors spee- 
tral bands are loo wide and poorly positioned for 
mapping vegetation, yielding inevitable errors. This 
appears to contribute to the ove [estimation of de- 
forestation in the Atlantic Forest example. The re- 
maining lores) is simply misclassilied. The second 
laclor is thai the resolution ol the regional analyses 
(0.0000 km') is much liner than the GLCC dala (I 
km') in the global analysis. This finer resolution 
enables better detection ol small areas of defores- 
tation. This likely contributes to the underestimate 

small patches of deforestation. 

We do not know if these error rales are general 
for ihe global analysis. These are only Iwo examples 
from relatively small areas. W hal they do indicate 
is a need for belter global mapping of landcover. 
At the lime of writing, the GLCC was ihe best glob- 
al dala set available, but g I prospects exist for 

refining our results. FITorts are under way using 

html) and Landsal I i n - 'i 

ellite imagery lo map ' ■ 

30-meter resolutions. respeelbeb. \egetalion map- global 
ping is a primary consideration in those sensor"-, 
designs, so the resulting maps should hotter (lis- Conci. 
eliminate vegetation l\ pes and their level of dis- 
turbance. The improved spatial resolution should , . 

I i ' mi I i i I islurbanee. 

The Missing Edges 

We have mapped disturbed areas as best possi 

wood pateh is still missing Irom our estimates 
Laurance (1997) found forest e.lges to be vulnera- 
ble to imasbes because of "edge effects"" disturb 
ing the forest commiii i \ 1 i i i < cannot as- * ' 

i iir'. . -i . . .1 iii ma P '»' disturbed 

. . , , i ■ ■ some <>f the worst 

weed pateh. the one km- resolution is loo eo, 

we can estimate it for the Amazon and Atl; 
Forest examples. 

The regional forest maps have some errors 
we must first correct. They have small gaps of nuu- 
foresl. some of which ma\ be natural, but main of "'' * ,M '" ' " 
...i ■ l .. ..I....:*:...!.: .1 " :_ The advar 

isi\e species are a growing problem lor the 
both ecologically and economically. In re- 
■ to the problem, governments are devoting 
sing amounts of resources toward the preven- 
ontrol, and eradication of invasives in many 
)f the world. To enable efficient use of these 

icre invasives arc likely to be a problem or 

ie a problem in the future. 
Imbed ecosystems arc often favorable for in- 
ion confirmed by comparing our 

invaders. These disturbed areas. 

the global weed patch, occupy at least 29. 1 million 

km-' (23<7r) of the ice-free land surface. Other than 

the overgrazed dr\ lands, which our analyses arc 

detect, the numbers broadly agree with 

ii«, ««vantage of our approach is that it sin 

""'\'' """ Jl """ VI " ""'""■'■ • where those dislui I icd area- are and thus where 

. 6 . . " , , invaders arc likeb in be Combining tin— \Mlh 

- i. " < - ' lo —I * ' I" » i ■ - dge. . " . 1-1 I . 

.„ , . . , ■ i T . lormalion on who llie invaders are likeb to 

even it that edge is a single ,10- m pixel, lo account 
for this, we replace patches of iion-forest smaller 
than 2 hectares with forest. The use of 2 hectares 

sification errors, bill ma\ also include some truly 
delorested areas and natural gaps. The result is a 

should help in efficiently allocating resources I 

R. Leemans & K. Kreileman. 1998. Global 
(Jiang,- Scenarios ,,l tlir 21-1 < < ill ii i v : lirsulls Inim llie 

probable underestimate of edge lores,. IMAGE 2.1 Model. IVrgamon '& Klsevier Sc, 

In the Amazon example. 17.801 km- of forest is 
within 300 meters of an edge (fable 2). the distance 
that Laurance et al. (1998. 2000) detected com- 
munity changes in \inazoniaii forest fragments. 
However, in the original state this region already 
had 1986 km- of natural edge forest due to rivers 
and savannas, \dding the edge forest, minus nat- 
ural edge, to the earlier estimate of disturbed area 
yields 24.760 knr of disturbed area (Table 2), ap- 

In the Atlantic Forest example. 29.373 km 2 of 
the remaining forest is within 300 meters of an 

edge, \dding this to the delorested area yields 
121.366 knr of disturbed area, leaving just 6484 

(disturbed forest. The map of original for- 
F ecoregion) is too general to reliablv cal- 
' "natural"* amoiinl of edge as done for the 

cock & R. D. Pieper ( 

The Netherlands, 
knr of undisturbed loresl. The map ol original for- lumlirao SOS Mala -Vllantica/INPE/Instituto Socioam- 
est (WWF ecoregion) is too general to reliably cal- hiental. 1998. Atlas da Evolucao d( 

Ecossistemas \ss,,ciados 
Perfodo de 1990-1W5. Sao Paulo. Brasil. 

I,. I'lueknell. J. \. Pancho K J. P. II,-, 


When we incorporate these regional estimates of | )(M . gt .„ p >77 Th( , World's Worst 

edges into the calculations of disturbed area, then and Biology. Univ. Press of Hawaii. Honolulu. 

I«»7<). \ (.e.mapliical \tlas d' World Weed-. John Wi- . -oiisenal mi, prioril i, - Nature 10.,; I'M-} ;.>{}. 

lev & Sons. New York. Olson. I). \1. * K. |)in.-M<-m l«*K5 Hi.- 2(K 

I VI \(,K learn. 2001. The |\1 \(,K 2.2 Implementation ol representation approach to rons.-miif- thr KarthV . 

iheHJKS Scenarios: \con,p,dic,,s,w anaKs.s.donns- lnuln^„,,IK valuable ocoro^,o„s. I „„-,,„ t,„|. 

sions climate elian-e and imi.acls In ihe •> 1st ce \ 

.,- .,,, | .... , Olson. . |. S... |. \. Walls X I.. J. Allison. I '){«. ( iarbn 

" ' " " | . ,, , ' ' , " , ", ' l.iv \ ion ol Maim Woilil I'eosvslems. '" 

\. Ferreira. J. M. liankin-de Ylerona X S. (,. 
. 1998. Rain forest (rau.nenlation and tin- dv- 

v.-p-lahon. Wildlil. l.'.-M-airh (.*.,». .it n... 2. I >. I Mi and \\ . I. it if. 

i. Soil Pol- Service Washington. DC. 

THUG. 2002. Tropical Kam forest l.ilornialinn Cenlei 

Ohlcn. /. Available at: <liltp://vvu\v.hsrsi.msn.cdii/lrlic/} (acccsse. 

Zh.i. I.. >a.iu K .1. W. Men I, ,„l. 2( D.NclopnuMii ol % <>" 2i \ <><tober 2002). 

a d.,1,,1 land , , l,:„ :,,,,M ,. s .I.Hal,:., and lU'.l' V '" ^' k ''"- "' f 2 A °°° Shjub inVM.ons o| ^nl,_\,nc, 

l "-;-;:,^ l -" lU,ll;l! ' ""■'■"• s " v^X^T^ 

'"-- l: ;"■; '"■ > "- , , UaKuii. !'«'.(.. Ilinn.mappropn, , ol ihe prodnc.s o 

.owe. >.. M. Mnmiie e> s. Ilnudjelas. 2(M)I. I(K) ol ll.c pholosw.lhesis. ISioseience 3(r. 368-373. 

Worlds Worsl liua-iM \l,c„ >•,„■>. bnasne Specs v 111|t| „. | V ^ \ V s . |. ( , M „| tllI1 |. |.,.,„. I,,,,,,,,., „| ;l |„, 

Specialist (.roup. II CN/SSC (ilSI'. |ililllK .1 Kim.. , anican. s. Wed led I. 10 


lut. 76: 133-161. 

: H.. M. C. Reed. J. K Brown, 

REVISION OF ARDISIA Jon M. Ricketson- and 



\ taxoiiomic revision of the [Neotropical Anlisia subg. Auriailanlisia (Lundell) Hickelsoti & Pipoly is presented, and 
a key to the Neotropical subgenera of \rdisiti is provided. The y,ni|i comprises lava iormeilv placed by Lundell in his 
segregate genera Auneulardisia, Amallama. and \<denoanthus. Ardisia sulm. \tu " uSaniism s defined bv lis unique 
calyx lobes, which arc asv mmelrical. usually nolched just below ihc apex, and always auriculate nasally. Six sections 
are recognized within \rdisia subg. Auricidardisia. comprising 75 laxa: seel. \iuirulardisio (I.uudell) Ricketson <\ 
Pipolv ( t species), sect. Pleuroholryae Ricketson iN Pipolv I I sp< ( ics). s, , i / . « Ox Pipolv |7 species), 

sect. Wedelia Hicketson X Pipolv |i! species), sect. \m,ithimn il.undell] Pi. -kelson vN Pipolv (3 species, including 6 
subspecies), and sect. Palmanae Hicketson iK Pipoly I 17 species, including 2 subspecies). Thirteen new species and 
two new subspecies are described and illuslrated. I he new combination^ \rdisoi pellueula Oerst. subsp. pectinatu 
(l)onn. Sm.| Kickelson cK Pipoly. \rdisia lichmanmi Oeisl. M il.sp. lulaprr.sis il.undell) Ricketson «N Pipolv. and Ardisia 
croalii Lundell subsp. correae (Lundell) Ricketson iN Pipolv are proposed, \rdisia carcluana and \rdisia :akii are 
transferred to (»i^,n- • id, II) ICckHson vN Pipolv and (,eissautlius :„l u (Pipolv) Rick- 

Pipolv v una . I e | | u •nbgeneios ti 

segregados por Lundell en s M s geueios \mit nioh.' I 1 • - >- \'.in tdatdisia s, define 

auri<ailados en la base. Se ie( onoi en scis s, , , i,,ii I i . ul .,i pi. iiden 7 > laxnncs. 

listados a continuacion: \;au ,•'"•<■ Inn. II II ! i - n\ Pipolv i I . s, ■- ) ~, < I I'lcnoholnae Ricketson & 

Pipolv (1 especie). sect. /u lfi erlindia Rickelson X Pipolv i 7 especies). sect. Wedelia Rickelson t N Pipolv (« especies). 
sect. Amatlania (Lundell) Rickelson & Pipoly (3 especies. incluyendo () subespecies) v sect. I'almanae Ricketson <N 
Pipoly (17 especies. iuclliv endo 2 slibcspe. ics). Se descriheii c llnslran Irece especies V .Ins subespecies COmO nuevas 

para laciencia. Se proponen ties coiiibiuaMoti. s ma \as I,,'-,- ■ - I 1 1 i. Nil.) Rickelson 

vN Pipolv \rdisia lahmaiuui Oeisl subs, , I i I, ! i . u o, Pip,, I v \idma , toatii Lundell subsp. 

correae (Lundell) Ricketson & Pipolv. Se Iransfieren \rdisia carcluana v Ardisia zakii al genero Geissantlius. como 
Geissanlhus carchianus (Lundell) Rickelson & Pipolv v Geissanlhus zakii (Pipolv) Ricketson & Pipoly. 

I'lic p.uil topic al \rdisia Svv. is the largest genus circumscription has been problematic owing to a 
in the family Myrsinaceae. containing perhaps as lack of comprehensive In atnieiil since that of Mez 
many as 500 species (Chen <\ Pipolv. 199o). Its (1902) in Kngh -i"- I ho I'fhuu-mrrirh a century ago. 

1 We thank the Missouri Botanical Carden and the Flora Ylesoamericaiia Project, as well as ihe Kairchild Tropical 
Garden, for their support to bring l.'i I ' - Il'i| »l- l«v lll «' 1 •> «,.ik on llr pr. \> i \\> _u ilelullv acknowledge the 

material Lor assistance with the scanning electron micrographs we thank George Taylor of the Llectron Microscope 
Laboratory at the Florida International University. Jack Fisher of Fairchild Topical Garden generously assisted with 
the plates and reviewed the morphologv section, for which we are wrv graleful. We also v.rv much appreciate the 
review of the taxoiiomic concepts seclinn bv Walter Judd of I he I niversilv of Florida. Scoll Zona, dim Utteridge. and 
Gerrit Davidse generously gave of their time to provide thorough reviews ,,1 the entire manuscript. Special^ 
due to Rov Gereau for his critical review 
Glen and Janet Hahn, Dean and Gail I 
I'inallv. the assistance of Jen Kir I I „ 

' Missouri liolanical Carden. P.O. Nov 2«)<). St. Louis Missouri 63166-0299. U.S.A. 
' Fairchild Tropical Garden. I 1935 Old Culler lU.d. Coral Cables (Miami), Florida 33156-1212, U.S.A. 

Ann. Missouri Bot. Card. 90: 179-317. 2003. 


-ill.lCI'Ilf llV \ 

^s (Mez. 

I«>02). I'ipolv ; ,n,| Kickels >08a). himrvrr. 

louiid lh;il I Ik* stamens m all Ardisia are actually 
connate I >;l-:i 1 1 \ l>\ then filaments in lorm ,i hvaline. 
inconspicuous lube, hnl thai (he tube is live from 
(lie corolla. While onlv a leu -roups within \nlisi,i 
s.l. have hern segregated as separate genera in ihc 
I'aleolropies (Sadiria Mr/ |I'>02|. \Jrardisia Mr/ 
(1002). Telrardisia Mez (1002). Hymenandru A. 
DC. (1834). Farardisia Nayar ci Ciri (1986)). there 
has hern an enormous increase in separation of 
specie-, -roup-, from \rdisia to new genera in the 
Nootropics, starling with Auhlel's (177-")) dose -rip- 
lion ol Iracoiva (against which Ardisia is conserved) 
followed by Alphonse de Candolle (Dill). Ducke 
(1930). and finally I I II (1963. 1961. |08la. 

1081b. |98lc. 1981,1. 1982). l.undell's contribu- 
tion was clearly the most extensive, consisting of 
the segregation of Amallania l.iindell ( = Ardisia 
-uhu. \nrii ulardisia sect. Amallania (l.uuo in l.'n 
kelson K Kipolvi. \nri<-idaidisia I uiidell ( = Ardisia 
suhg. \urindanhsia 1 1 .iindell I llickel-on vK I'ipolv ). 
Chonlalrsia l.undell (= Hymrnandra A. DC. p.p.). 
)iinrl,-cria l.undell (- Clcnard/sia Ducke). (irnllra 
l.undell. Craphardisia l.undell ( = \rdisia suhg. 
Craphardisia Me/.. I'ipolv «\ Kicketson. 1 998a). 
Ihaitrti l.undell i \nli\ia -uh-. \rdisia). Oerste- 
dianthus l.undell. Synardisia (Mez) l.undell. \ah- 
rioanlhus l.undell (= Ardisia suhg. Aiirirulardisia 
seel. Auricitlardisia (l.undell) Kicketson & I'ipoly). 
and Ziinilia l.undell (= Ardisia suhg. 6;,. ;, . 
Me/.. Pipoly & Kicketson. 10<>0|>). Because (he 
group is comprised of o\er 8.00 names, it will he 
some tune he-lore each species has heen reviewed 


il, I 

Ctrnanlisiu. Hvnirnandra. Genllca. and 
i). l.undell ( 198 Id: 342) separated his ge- 
idardisia from \rdisia s.l. using ihc lol- 

cies as a part of ihc huger g.-mis \idi\ia in. hides 

■ hei go.iip. \iii,iiiama. s.-grcgah-d from Ardisia 

In l.undell (1082). l.undell slated thai \matlania 
resemhlcs anolher segregate genus. (>>::■:■ ■I'.m'.hus 
l.undell. heeaiise hoi h h,i\e filament- vmMi ■.•laud- 
tipped hairs |a character I'ipolv and Kicketson 
( I '>«>«:.,_ |« >00|,) observed in Ardisia suhg. Craphar- 
disia\. hut differs nolahly in the nature of the in- 

\Ve also discovered that Lundells (1082) genus 

\alcriaanllni'. helonged to the \uriridaidis,a group. 
based on its auriculatc. asymmetric calyx lobes 
with suhapical notches, l.undell (1982) defined \a- 
lerioanthus b\ its ,oai><- i.-d hirsute iiidumcni. with 
Irichomes up to 2 mm long. During the course of 
this study, we noted with interest thai the species 
l.undell (1082) included in the group, including 
\alerioanthus ni-ramanmi (Standi.) l.undell (= Ar- 
disia ncrcrmannii Standi, in sect. Fa^nhndia\. I. 
iirsimis (l.undell) Lund. -II ( Ardisia ursina l.undell 

dell) l.undell (which we ha\e placed in synonymy 
under \rdisia ursina) have strikingly distinct ves- 
titure that in no vvav serves as a uniting character 
slate. As defined in this treatment, \nhsia suhg. 

disia species in the Neolrnpios and i- comparable 

\<-raidisia. which ranges From Southeast Asia and 
the Pacific to Mesoamerica and the Caribbean. 

While preparing a treatment of Ardisia suhg. Au- 
rirulaidisia lor inclusion in the Myrsinaceae for 
Flora Mvsoamcricana. we saw the need to present 
a full account of the group with detail not possible 
in the lormat of Flora Mcsoanwricana. and as a 
precursor to our eventual Flora \roltopna Ireal- 
menl. Herein we formally propose a new la\onoinv 
lor \rdisia suhg. Ann, ulardisia to include six sec- 
tions, ol which four. seel. Fa^crlindia. sect. I'/cii- 
robolrya. sect. Wedelia, and sect. Falmanae. are 
new; we reduce the genus Amallania lo a section 

rictl'v racemose or suhsp'ica.e ' ' vvilhi " ll "' subgenus. I nlorluna.ely. it was 

Ardisia «'l)l«' lo conduct a phylogenetic analysis, sc 

musly seal), furfuraceous or rarely tent to which each section is monophy lelic 

■ asymmetrical, depressed-orbieu- rently unknown, 

broadly ovate, and er ^^ . ',■„„•„,/„„,;,,•„ se( ,. .,„,„,,„, 

I inHu 

. \\r Whelhel ill, hah 

have determined (see VIoi phologv. Vesliture) that 
llir vostiUue 1(11111(1 among llu- suite of species is 
actually composed <>l papillae or papillae mixed 
will, uniseriale. mull ic-l lulai glandular villous tri- 
chomas (Fig- M). Other features that define the 

the leaf margins and the dense yellow ghmdular- 
granules within the corolla tube and on the fila- 
ments. Recognition of this group of species as a 
separate genus would leave \rdisia undefined. In 
all species of \mallania. the stamens are united 
kisalh h> theii lilaincnl- l<> form an incon^picuou- 
tuhe free from the corolla tube. clearly placing 
them within Ardisia. In addition, all species in 
Amallania have asymmetric calyx lobes with a sub- 
apical notch and an (albeit more imperleclb 
formed) auricle on one side, which also indicates 
that this group of species belongs to Ardisia subg. 
\ ■ . \\i Ii n I hi i. di . . I< > , , ,, U in 

generic to sectional status within \rdisia subg. \n- 

\nlisia subg. \iiik -idardisia sect. 1.'.. 
eonlaining lour species, i- defined by the combi- 
1 branches 
i i ■ouue-led. glomerate cor\ nibs, each 
of which is subtended by a persistent, enlarged in- 
florescence branch bract, and finally, the individual 
flowers subtended by a persistent, enlarged floral 
bract as long as the flowers. This section 
the rare and verv interesting specie^ I" 
which has unique red to pink floral bracts and is 
visited by hummingbirds. I'he section includes two 
species placed ill the genus [ii/criuatllll //s b\ l.un- 
dell (1982) based on liis misinterpretation of their 

along with A. dud^ei. I. iilomcrala. and 1. nervo- 

-ubg. [uricidardisia seel. I'd 
contains se\en species and is defined by exhibition 
of Kagerlind's Architectural Model (see Morpholo- 
gy), combined with terminal, pendent inflorescenc- 
es. Because the vegetative and reproductive shoot 
leaves are dimorphic, main herbarium specimens 
are incomplete, containing the flowering shoot only, 
or with a vegetative shoot flowering precociously. 
This has led to taxonomie over-description and a 
fundamental lack of understanding of the devel- 
opmental dynamics of these plants. Members of 
section h'diicrlindid are normally inconspicuous el- 
ements of the understory. 

\rdisia subg. \iiririddidisid seel. /'almanac is 
clearly the largest section in the subgenus, with 47 
species. The section is admittedly artificial and is 
defined bv the 

guishing character -^lalc i- unknown at this time 
and will require more rigorous study. 
Ardisia subg. \nricidardisia seel /' • 

9nly Ardisia plcurobotrya 

long naked peduncles to 6.2 cm 

long, dense, overlapping chocolate brown lepidole 
scales on mosl plant pails, and pendent flowers on 
long, usually sigmoid pedicels. The lateral inflores- 
cences and chocolate blown lepidole scales are 
uni(|iie within subgenus \itricitlardisid. and within 
flic genus as a whole are otherwise frequent onh 
in subgenus Akosmos Mez of Asia. 

\rdisia subg. \uri< idardisia seel. Wcde/ia is de- 
fined by a subshrubby to small arborescent habit, 
and terminal, columnar lo siib-cohuiinar and often 
pendent inflorescences, on peduncles at least l A flic 
length of the inflorescence and subtended bv large, 
loliaceoiis brads. Like spe. ie- of section h'aiscriin- 
did. these plants are inconspicuous members ol lh. 
forest understory. The most salient feature of the 

cences. similar to the branches found in the basal I v 
blanched, pvi.tmidal inflorescence of [rdi^ia pscn- 
■ ■-/ i se ( i I'dlmanae). 

hides Morphology 

Species of Ardisia subg. \iiricnlaidisia are ter- 
restrial, small trees ;IM( | shrubs, rarely subshrubs 
(A. pellucida). The vast majority of laxa (sect. Ainal- 
lania. sect. Auriculardisia. sect, /'almanac, and 
sect. Wedelia) display Scarrone's Architectural 
Model (Halle et al.. 1 978; Bell, 1991). which is 
defined by an orthotopic monopodial. rhythmically 
growing principal axis (stem or trunk) differentiated 
from the branches, which are produced in pseu- 
doverticillale litis. I he branches are composed of 
orthotopic shoots thai produce terminal inflores- 
cences: thus branching is svnipodial by substitu- 
tion. In a mature individual, there is often a dense 
canopy crown whose ultimate units may appear pla- 
giotropic by gradual bending as ihey develop, a 
phenomenon documented bv Kisher and Stevenson 

Species ol \rdisia subg. \nricnlardisia seel. 
I'i , ... «,.■-■ exhibit Hand's \iclnte. t ur ill Model, 
which is i haiai leri/cd bv a monopodial. rhythmi- 
cally growing, readily distinguishable trunk that de- 
velops pscudovei'licillale Ik i- of branches morpho- 
genetically identical to itself. All branches arc 

The most unusual mod. 'I we found was thai of 

\nif.t<i -Mihg. \urii nlanlisii/ -.el. I'ulk i Imdui. de- 
scribed lu-rcin. which exhibits Fagerlind's Archi- 
tectural Model (Halle et al.. 1978). Fagerlind's 

the apex gradually coalesce into a poor 
resinous area during secondary xylem 
.mil lirauclilel thickening. The bark 
smooth, or is rare Is exfoliating i \rdt>ia i 

gross in- hunk, ssith tiers ol modular, plagiolropic 
[.ranches thai branch sympodiallv. The modular 
liranches form ss llepti<-alls .iml -s mpodialls. often 
by apposilion. and ssilh spiral phvllolaxv. First, a 
mink ("vegetative shoot") develops rhythmically. 
producing -uccessisc pseud. >\ ci I iccls of "vegeta- 
tive" leaves until the apex loses dominance and 
latent lateral ("axillary") liuds are released, pro- 
ducing successive, monopodial pseudoverlieels of 
shools. each of which is sylleptic. ssilh a long hypo- 

P"d i and exhibilmg i . 1 1 . i < I extension grosvlh. The 

sylleptic brunch shoots produce lirst a prophyll. 
then |)seudoverlieels of "reproductive shoot" leaves 
(similar to the leaves of (he "segetalive shoot" bill 
notably smaller), and llien may either lose domi- 
nance and once again produce "segelalise shoot" 
leaves, branch ss mpodialls by apposition growth 
(repealing the i lule) without llowering. or pro- 
duce a terminal or pseiidolcrmiiial (acting as a ler- 
minall inllore-ceiiee. followed by -vmpndial 
branching bs substitution. I'he pscudolerminal in- 


- of the stems or 



and peduncle i 

i (In n 

a given species. Much < 

i lh<' 


e regarding I he 

branchlets. the pel- 

scales of the leases had not been thoroughly inves- 
ligaled. \\ bile a study of the covering scales for the 
•■•'I in- family is clearly warranted, we here define 
llie principal types of scales and other veslilure 
found In \rdisiti subg. \iinnii,udisiu. 

The most striking type of trichomc lomid ill the 
subgenus is the cupuliform scale, which is sessile 
to stibsessile. with small lobes to arms, composed 

form a deep cup or a small cup with radiating arms 
(Fig. FA-ID). These trichomes arc quite large, 
ranging in size from 200 to 300 pm in height. The 

cupiililorm -.calo in. is be mixed ssilh much smaller 

. While Ihe lern.inal bud i 

is interesting that all 

ral genus Hymenandra 
del and fused anthers 

sect. Fagerlindia may 

negatively geolropic shoot is termed a "trunk" if it 
attains a diameter greater than or equal to 2.5 cm 
diameter at breast height (DBH); if il does not. it 
is termed a "stem." Branchlels are defined as llie 
upper 10 cm of any branch. Branchlets max be 
stout or slender, straight or llexuous. terete, suble- 
rele. somewhat ungulate ssith longitudinal ridges, or 
bearing well-developed interpeliolar ridges. The 
hraiichlels always bruise blue-purple when cut. due 
lo the less lo numerous Is so-schizogenous resin ca- 
nals present in the secondary phloem. The resin 

elv furfuraceous- 
12 cap cells and 
one stalk cell, approximately .'iO-50 n_m in diam- 
eter, shown in the middle left area of Figure 1C 
and throughout in Figure ID. 

Furfuraceoiis-lepidote scales are those that do 
not form a discernible cup and which may be entire 
to subentire but never with radiating arms, and 
range in size from 2(10 lo M)() pui in diameter (Fig. 
2B). The apparent thickness of the cap cells, owing 
lo their respective I iginli. al ion. varies go-alls, from 
extremely thin, almost translucent (Fig. 2B) lo ap- 
pearing opaque (Fig. 2A, 2D). 

In mans laxa of the subgenus, van ing qiiaulilile- 
ol hsdiopoles may be present, mostly on llie abaxial 
leaf surface (Fig. 2 \). They are often confused with 

lepid scales, but their obvious subsidiary eells 

and unusual cap morphology easily distinguish 
them. A good example is seen on the left side of 
Figure 2 A. found in Ardisiu puiiamrnsis Fundell. 
Hvdropotes ("water drinkers") were described bv 
Mavr (10|f>). then Griiss (1927a. 1027b). and fi- 
nally. Gessner and Volz (1951). all of whom re- 
ported these structures from studies of submerged 
aquatic plants. They were fust reported for the 
Myrsinaccac by Pipoly (1987) in Cyhmnthiis Mart. 
subg. Cnimimidrnid (Benth.) I'ipolv. but are now 

Figure 1. Cupuliform scales in Ardisia subji. Aurit 
reclined habit. — C. Close-up of cupuliform scales. A. 
|). Cu|)ulifonn scale- id above lmfuraccou-lc|,idole -ale- if). I. fiirfiinn ,;,. Nolo that cupuliform scales reseml. 
those of 1. brencsii. I.ul an- more -tipilale. I \. I! Iron, II. Ilahr, <v 11. Zuchmrski 'U.V, |\H»: C from K //after & I 
Zuchowski 9128 (MO): 1) from «. «m/e <£ (,. Co,// '/;_' (FTC).) Scale ha,-: \. C: _><IH p.,,,: II. I): 50 o, m . 

known for most subgenera of Cvhianlhus (Pipoly, 
1998). Hvmenandra (A. DC.) A. DC. ex Spaeh (Pi- 
poly & Ricketson. 1999a). and H.,//em"a Sw. (Pan- 
fet. 2(X)2). These structures may easily be distin- 
guished from lepidote scales b\ their 5 to 7 
subsidiary cells, a central foot cell, a hasal cell, a 

stalk cell, and up to 12 cap cells. In earh -rn\. 

a imicilaemun- -uh>tanee accumulates around the 
hasal cell. Later the cap breaks off. leaving a mu- 
cilaginous ring around the basal cell, w hereupon 
they function as hulathodes in water regulation, 
based on elegant ultra-l ruelural and chemical stud- 
ies hv Liitlge (19641 and l.iittge and Krapf (1972). 

port minerals into the leaf lli - H 
substance secreted around the stalk. The mucilag- 
inous substance stains easiK with Sudan IV (Pi- 
poly, 1987). Given the 
vial tropical life zo 
Myrsinaceae, we expect to find that these structi 
are ubiquitous within the family. Ilydropotes 

ea>il\ distinguished hum I raii-lucent lepidote 
scales h\ their subsidiary cells. 

\rdisia ncKTtihimin Standi, is u tii< |i it ■ uilhin 1/ 
disia subg. \uriadardisia sect. Fa^rrlnuiia be- 
cause of its ferrngineous hirtellons \estiture. com- 
posed ol stiff Iriehomes less than 2 mm long. These 
deiiseb |i.k ked hau- arc apparent b unicellular, 
with subulate apic- ,,nd bulbous bases (Fig. 2C). 
Lundell (1982) did not adequately distinguish the 
hirlellous loinenlum of Ardisia > 

inhabited by most 

slipitale-stellale loi nenlel Ions. The arms frequenth 
break off, leaving an apparent glandular-\ illou- lii- 
choinc. indistinguishable from those found on the 
branehlels and inflorescences of section \ii«ilhin'm. 
such as Ardisia prllurida (Fig. 3A). While we can- 


7wmt B 

* :,'^fel '^ffl^ 

}. Wj 

vv "^-^ '* \-i -j^ 1 )7^§S5fc ^ 

1 ^^^ 




ol (li-iin-- ilu- i .. »--i I »i I H \ 1 lli,- glandular-villoiis 
icliomes of section \matlania (fig. '.\ \) arc de- 
\<-d ll'oiu the glandular sli|)||.ilc-sli-|lalc Iricliomes 
)iiii(l in set-lion I'ulmanur (fig. M). we have nol 
)iiiid any remnant arms on the villous trichomes 
(any member of section \mallania. At times, the 
lipitate-stellale trichomes form a layer ol vestiturc 
l>o\e I hi- mimile. Mibx'-silc I r in-ln, . -ril scales ill 

1983). are epiden 


ularly notable for the well-developed glandular- 
pillae (Fig. 3A). which in Ardisia pellucida 
hsp. pellucida are also accompanied by glandular 
lows trichomes. These trichomes preserve poorly 
d detach easily during the pressing and drying 

ment with lime and handling. The papillae are 
iy similar to those found in Cybianthiis subgenera 
nmmadenin (Pipoly. 1987) and Microconomorphu 

•cies ol \rdisia siibg. 1//. 

The leaves of ; 
culardisia are alt 

slipul.ile. and -impl,-. ( dven thai all hiau, lung i- 
sylleplic. rapid extension growth of each sliool is 
followed h\ production of one or two prophylls. fol- 
lowed by fully formed leaves. In those species ex- 
hibiting fag, -Hind'-. Model, the easily distinguish- 
able "vegetative -hool leave-,"" are much Lugcr ill, in 
the "reproductive shoot leaves" and have a smaller 
length-to-vvidlh ratio. While the leaves of the prin- 
cipal avis and ihe blanches have spiral phyllotaxis. 
the leaves of the branches aie leoiu-uled in a plane 
lilor-ivcnlral) ibrough a twisting of ihe internodes 
rather than by distichous phyllotaxis. The blades 
are most I v eh; 

ar-villous |»\ I lin - - ■■■•'•>•■ ' ~uh-.|>. jivllui idn. This is trpiv- 

. ursina. showing the glandular. -I i f »i tal.-->lrl laic tricliomes. —I). Ii 
-stellate Iriehomes. 1. ursina. (\ from C. Martinez 1071 (KT<;): I? I 
■.«>/» 14(142 [Y'\r,).) Seal.' liars: -\-C: 201) urn: I): 50 jxm. 

>tic. oblanceolate or oblong, fences on long peduncles terminating in con nibs, 

apically acute to lonu-acimiinatc. ba-alh cimcale. The panicles arc normalb p\ ramidal and much 

to obtuse or rounded, or rareb auriculatc broader than long. However, an exception is found 

The blades ma\ also be deciirrcnl on the petiole in section Wcdc/in. where the panicles are columnar 

or not. conspicuously or inconspicuously (\isible (much longer than broad I lo -ub-columnar (longer 

but not raised) or prominentb (raised) punctate or than broad). The rachis ma\ be straight, (lexuous. 

punetatedineale. variously f'urfuraceous- or eupu- or rareb geniculate. The inllorescenee bract is nor- 

liform-lepidote. hirtellous-lomentose. \illous glan- ma ||y early caducous; the inflorescence branch 

dular tomentose. or >tipilale->te||ale lomeutose. The bracts may be (bliaceous. large, and often envelop 

margins may be entire, crenate. serrate, pectinate. t he terminal flower cluster, at least in early ontog- 

deulate. flat, or re\olute. The petioles ma\ be ob- env 

solete or subobsolete. canaliculate or marginate. The fi owers can | )e 4 _. 5% or 6-merous. with 

asymmetric cahx lobe- thai have a subapical notch 

IM I olilx.l \<.l> \\D H.owERS an( j a p rom i, u . n t auricle on one side, a feature that 

Within Ardisia subg. Aurindardisia. the most tlefiurs the subgenus. The enroll is are usually 
common inflorescence is a panicle, with flowers ar- white to pink. The corolla lobes may be ovate, el- 
ranged in corymbs or glomerules at the end of the liptic. or oblong, symmetric to asymmetric, appar- 
branches. rareb appearing suhracemose. The inflo- ently epunctate. pellucid, black or orange punctate, 
rescence is a terminal panicle except in section and/or lineale. The stamens are fused basal ly by 
Pleurobotryae, which ha- oh\ ioush lateral inflores- their filaments to form a hyaline, inconspicuous 

aighl or graduallv broader lowaid the base. The work has thus far produced onb preliminary clad- 

thers mav bo o\al.'. lanceolate. linear-lanceolate, ograms. and that all homologies have not been 

oblong. . i p i < . 1 1 1 s apiculale. subulate, caudate. worked out. I low ever, we believe it is more appro- 

eordate to priate to formulate a prolimiunrv pin logenetie h\- 

subapical pothesjs r .il In i Mian ba\e no working framework 

>. with the until the dual nnabsis is done. liv definition we 

uoiisly or advocate the use of sectional rank only within for- 

prominently punctate dorsally. The pistil is most mal subgenera. 

often obturbinate but may be obnapiform or ellip- Our concept of subspecies follows that of Pipoly 
soil th( ovu\ is globost Igl b < II | I (19»7: 46). who defined a subspecies as: "groups 
ovoid, or oblong. The placenta may be globose, sub- "' populations within a single lineage of ancestor- 
globose, or ellipsoid, with apparent!) uniseriate descendant populations that show variation by 
(few ovules in a high anthotacli. spiral), biseriate uni(|ue combuialions of plesiomorphies, or homo- 
or pluriseriate. few to nianv ovules. plasic apomoi phi.-, correlated with bmgeogiapliv 
and/or ecology This rank i- prunaiib used to con- 
i . ,., || vey information regarding variation in the life his- 
tories of these populations and character stale dif- 
The fruit is a one-seeded drupe that may be glo- f rrni( . es hypothesized to be the result of this 
bose to subglobose. inconspicuously (pellucid) variation. The subspecific rank in no way attempts 
punctate, or proinineulb black punctate and/or liu- tn predict speeialion e\enls." 

I.indlev (UHI!) and Pipoly (|<)K7. 1992) for the in- 
florescence, racliis pedicels, and Moral parts. De- 
scription of leaf morphologv follows lliekev (IWI). 
The kevs are artificial and designed to expedite ir "*"™ ^*™\«™ <<>''«- TI,e„l,aM et al. (1984), 
en.ilication of herbarium specimens. An attempt ""[J^" ' "' "^ * mwm]o & foU ° W MH " 
is been maile lo emplia-i/e \ei;«-lali\e characters '''.' .'' ';,„,,, „„ ... . , , 

t , ■ f, ... .-, . , Lundell 1981b. I9J50 published 0« new names 

increase their iiselulue-s with sterile material. . . . . ,. . '.' ... A ,. . 

ml comhin.i ■ i- I- m \nhsia and 

Quantitative and qualitative data presented 
and descriptions for Moral parts anil brae 

taken from organs rehvdialed from ln-rbariu 

imens by boding in water. Measurements from 

Myrsine L. In both papers, he indie 
opposed to placement of the taxa i 

J provided in anticipation of future eircuniserip- 

iaii-e Iroin 10', i.i I .)' / greater than those .■ ,, , ,. 

, iL . „ II,. ,1 r i- , tions of the genera. We agree with Morales (1997) 

that in doing so. I.undcll inadvertently invalidated 
his new names ami combinations, according lo the 
International Code of Botanical \oinenrlalnre. \r- 
ticle 34.1 «;.euler et al.. 2000). lor Anlisiu spe- 
cies, we (Pipoly cK Hicketson. 19915b) corrected this 

names. ,,| which 1-1 belong to members ol \nli>ia 

taken directly from dried 
Data regarding stem diameters, inflorescence ni- 
duses, pedicels, and leaf and fruit shape and size 
were taken from dried hcibaruim specimens. \ List 
of Species and Subspecies and an Index to Kxsic- 
catae are provided in \ppendices I and 2. respec- ^j|, 1;1 |j , 
lively. The numbering system of \ppendix I has names , 
been used for the species and subspecies in the su ) )lr ^ 

text lo help the reader locate taxa within the article. |, llm V||- s mv;1 | i( | t .„ m | ,„ ,,, Ils s |, K for the pu, numbers are also used m the sectional kevs. |)()srs <)f ( .|. m(v ,,;,,.„ ,| |e |)n . |H)ll( |crance of n, 
Any correlations with phylogenetic relationships menelatural s\ non> ins complicated by the fact th; 
are coincidental. |M( , same e pj t | lets [ um . | )een llS( . ( | f or .\nlisia a 

Our concept of a subgenus is here defined as: a w ,.|| as f or manv species in the segregate genen 
group of species that comprises a major basal pin- \\ |,j|,. W1 . aiv r „gnizant of the fact that these Lm 
logenelic subdivision within a mouophvletic genus. ( |,.|l names have no nomenclatural stains, we fell 
whose formal recognition enables better under- was important lo summarize all relevant nomench 
standing of the phylogenv of the entire genus. A tural matters in on*- place to avoid confusion, 
corollary to that is our concept of the section 
fined as: a group of species within a formally 

ognized subgenus that appears lo share a common Anlisia Sw.. Prodr. A: M 
ancestry, based on the fact that they share uni(|iie Anlisia liml»lia Sw. 

Taxonomic Treatment 


herbs, h-ares simple, exstipulate. alternate or 
pseudovcrlicillale. normal l\ punctate or pi 1 1 M -| ;i 1 1 - 
lineate. peliolate or sessile. Flowers bisexual. 1- to 
.">( to <i)-m< toiis; p. -n, 1 1 ill i wild imlirieate or (|uincun- 
, ial aestivation (either i lext mr-e or sinislrorscl. 1 1 m - 
lobes shortly eonnale: stamens included, the Ma- 
free from the corolla tube, the anthers free, often '.\ 
times longer than wide, dehiscent b\ apical or sub- 
apical pore-, -nliapieal pores opening into longi- 
tudinal slits or simple longitudinal slits, rarely 
trans\ersel\ septate: |)istil obturbinate. the ovary 
ovoid or oblong, the ovules lew to numerous on a 
basal placenta, at limes appearing uniseriate. some- 
times biseriale. mosl commonly |)luriseriate. the 
-Isle elongate, normallv exserted at anllwsr-. Fruits 
drupaceous. 1 -seeded, punctate or punclate-lin- 
eale. sometime- longil udi na llv eoslale. with a some 
what fleshy exoearp and crusty or slightly bon> en- 
docarp; seeds covered by membranous remnants of 

Distribution. About 400 to 500 species, pantro- 
picalb distributed but with the highest diversities 
in central Malesia and southern Mesoamenca. fot 
Mesoainerica. we expect that of the nearly 800 
names for the group, we will recognize approxi- 
mately 120 species. 

For groups within Ardisia native to or naturalized 
in Mesoainerica. we present the following key to 
the subgenera: 

Kki to thi: Si iti.iM n \ ni \ui)isi\ i\ nil- \i;<>ri;on< - 

vascularized nodule in the sinus of each crena- 

hut never vvilli nodules in the sinus of each cre- 

2a. Calyx and corolla lobes dextrorselv imhri- 

le-s lliaii V'i ;is long a- the anthers. 

.'5a. hllloreseence a short-peduncled. termi- 
nal panicle; anthers widely ovate, de- 
hiscent hy wide longitudinal slit- sub- 
eontluenl apieallv at antllesis 

Sllhg. \< nir<lisiu 

3b. Inllore-cenee a long-|)eduneled lateral 

sul.g. mils 

cate or ()uineuncial: apically free portions of 
the filaments longer than W a- long a- the 

la. Anthers yellow or while, rarely light 
pink, not concolorous, dehiscent hy lon- 

gitudinal slits or subapic 
5a. Inflorescence brand 

margin, asymmetrical, auriculate 

basally suhg. \uri< ulunli^iu 

Hi. Anthers concolorous (diving gray I. de-- 

brous or rarely sparsely 

isually clasping the developing 

inflorescence and floral bracts 
mostly orange, rarely black, punc- 
tate, the perianth mostly orange, or 

Treatments of Ardisia subg. Cmphardisiu were 
published by Pipoly and Ricketson ( 1 998a, 1999b) 

ami ol \rdisiti subg. \rrardisia In Pipoly and Kie- 
ketson (2000). Treatment of the other subgenera 
(subg. \rdisia. subg. Icacoreu. subg. Tinus, and 
subg. Crispardisia) are now under way (Hicketson 
X Pipoly. in prep.). 

Ar<lisia Sw. subg. Wiculai disia (l.undell) Ric- 
ketson & Pipoly, comb, et slat. nov. Basionym: 
Auriculardisia Lundell, Phytologia 49: 341. 
1981. TYPE: Ardisia glomerata l.undell. 

muiuui iStaiull.l Lundell). 
Subshrubs or small trees. Stems or branehlets 
and leares furfuraceous-lepidote (the scales lawny, 
or rarely chocolate brown in e..lm. oppressed or flat. 

amis, stiff vet lian-liiciil 

visible), mixed lepidote, hirtellous-lomei 

hairs simple, apparently one-celled, long a 

what stifO rufous stell 

on multicellular, uniseriale stalks (the star section 

frequently breaking oil and the hairs then seeming 

fous). or villous-lomentose (indistinguishable I mm 
slipilate-stellalc with -tar aims broken off). Inflo- 
rescences terminal or raids ( 1. pleurobolrya) lateral 
panicles that are pyramidal, or rarely (sect. Wede- 
Ini) « > In, mi. ii to sub-columnar, the branches bear- 


caducous. IKMI-Iol i, it roii> I 

- llol subtended by ;i foliaceous inflorescence 

. Inflorescences lateral (axillary) 

sir I. 1'leuroho, 

. Inflorescences strictly lerniiual i.r |isoudo- 

bale. the apical free portions of the filaments mostly 
longer than W anther length, at limes subeqiialing 
to slightly longer lluui the anthers. ihe anthers nar- 
rowly ovoid or lanceoloid. apically apieulate. cau- 

sagillale. deliisrenl b\ suhnpical p.. res opening into 

pislil oblurbinale. the ovary globose lo siib.nlobo-e 
to oblongoid, the style slender, al tunes piinclale 
and/oi piuiclaledineale. llie placeula globo-e or 
ovoid, the ovules few lo numerous, in 2 or more 

s. wtien lew, at tunes appearing uniseriale. 
'.s subglobose to globose, conspicuously or 
liuenlly piinclale and piinctate-liiieale. smooth 

Distribution. Anlisia subg. 
known from San Luis Potosi. Hidalgo. Pueblo. Ve- 
racruz, and Oaxaca. Mexico, soulh throughout Me- 
soainerica to (.olombia. Kciiador. and in pails of 
Peru and Venezuela. Il grows from sea level to .'WOO 

Ecology. \nlisia subg. Aiirieulanlisia occurs in 
a wide variety of moisl or humid, wet and pluvial 

turbed areas, oak-pine woodlands, premonlane. 
montane, cloud, and elfin forests. 

The subgenus contains 70 species, including 75 
taxa. and is defined by asymmetric calyx lobes llial 
are usually siibapicalb notched anil always hasallv 

not -.periali/ed I 

the same general 
la. Peduncles at 

5a. Pedicels and inllorescence ra- 
chis minutely rufous glandular 
|i.i|iill.iU or mixed mill scat- 

licellular stalks bel 
late group of arms, ll 

ing an apparently < 

I unci 

Vnli-in ~ul .«_■. 

(Lundell) Kicketson & Pipoly. comb, et stai 
no\. basiouvm: \inotlania l.uiidell. Wnglili 
7: 38. 10«2. TYPK: Anlisia liebmannii Oersl 
|= Amatlania liebmannii (Oersl.) Pundell). 

Shrubs or small trees. Hranchlets slender, te 
rele. densely and minutely rufous glandular papil 

glaudulai villous hails, olleii d.ibieseeiil with ag( 
I raves mononiorphic: blades membranous lo con 
aeeous. elliptic |o oblong or obovate lo oblanceo 
late, moslly inconspicuously ,,i conspicuously liu 
often prominently piinclale and piiiiclale-linealc 
petioles slender, terete. Inflorescences terminal 

erect. I)i|>innalel\ to quadri-piniiateb paniculate, 
pvrnmidal. shorter or longer than the leaves, usu- 
ally loosely congested corymbs: inflorescence 
bracts foliaccoiis. usually caducous: inlloif sc.-tirc 
hianch bracts and floral bracts often caducous, 
margins \ario i I loo I 1 i i I idi i-i ul i 

finely serrate ■•pectinate" teeth, or regularly dentate 
or serrate, or rarcb entire: pedicels .slender, terete. 
Floirrrs 5-meroiis. while, light pink, light purple, 
red. or blue-violet: calyx lobes essentially five. 

basally aiiriculale: corolla membranous, the lobes 
ovate to lanceolate, inconspicuously to conspicu- 

OllsK ailtl lis lalK p:'OM:ll '-l,||-. p.uieialc minI puiic 

tate-lineale. densely yellow glandular-granulose 

inenls apieally lice, connale basally into an elobale 
lube, free from the corolla lube, epunclale. densely 
yellow glaudular-graniil(»e. ihe anthers ovoid to 
lanceoloid. apieally apiculale. basally sagittate l<> 
. ordale. dehiscent b\ siibupieal pores, opening into 
wide, longitudinal slits, the connective punctate: 
pi.-lil glabrous, the ovary oblong, the style slender, 
erect, epunctate to inconspicuous!) punctate, the 
oxides pluriseriale. h'ruils globose-. inconspicuously 
or conspicuously, often prominenlly punctate and 

Ih 7 to 12 per cm. ••pectinate" 
-- An 

Vestilure of the braiichlets. infloies 
chis, branches, and pedicels with ; 
of densely minutely rufous glandu 
lose and scattered erect uniseriale 
hilar glandular \ illous hairs; calw 
to densely and prominently punt 
3a. Calyx lohes ovate. 1.7-1.9 X 

I).') I. 

thers ovoid, 1.2-1.4 mm long 

5. Ardisia peUucida subsp. /W/m idn 

3h. Calvx lobes ovate to lanceolate. 3.1-3.2 
X 1.5-1.7 mm; corolla lohes 1.3-1.5 
mm long; anthers lanceoloid. 3-3.1 mm 

--- 3. Ardisia pellttcidu suhsp. I<inc<-iill,-nsis 
. Vesliture of the hranchlels. inflorescence ra- 
chis. branches, and pedicels densely mi- 
nutely papillose only: calw sparsely, ineoii- 

: 2.8-2.9 m 
X 0.8-1 r 

style 3.2-: 

Distribution. Ardisia subg. Auricitlardisia seel. 
Amallania is found from San Luis Polos i', Hidalgo, 
Pueblo, Veracruz, and ()a\aea. throughout the \1e- 
soainerican region. In South \merica it is found 
from Venezuela. Colombia. Kcuador. and Peru. It 
grows from 10 to 1700 m in elevation. 

Ecology. Members of Ardisia subg. \uricnlar- 
disia sect. Amallania grow in a variety of humid 
habitats, including tall moist, humid, wet and plu- 
vial forests, and lowland, premontane, and mon- 
tane elevations. There are a few species that occur 
in gallery forest in seasonally dry areas as well. 
They generalb have broad ecological tolerance for 

Ardisia subg. Auricitlardisia 
defined by (1) its vesliture of sparse to dense mi- 
iiule papillae or lliose papillae mixed Willi dense |o 
sparse uniscriate. mull icellular villous hairs; (2) 
leaf margins variously toothed, usually with irreg- 
ularly finely serrate "pectinate" teeth, or regularly 
dentate or serrate, or rarely entire; (3) dense yellow 
gluudiihu granules within the corolla tube and on 
the filaments, similar to those found in Ardisia 
subg. Graphardisia. 

if blade margins entire to regularly denlale In 

V.»mlla lohes o.O-o.!', X 2.0-2.9 mm: an- 
thers lanceoloid. apii alb -ubiif ile-api. u- 
lale. 2.5-2.0 mm long; styles 4.6-4.7 mm 

long 7. Ardisia schippii 

Corolla lobes 2.7-3.2 X l.n-2 mm; anthers 
linear-laneeoloi.l. apicilb apiculate. 1 .9- 
2.2 mm long; styles 3.1-3.6 mm long 

cence branches glabra 

ose below: . 
1.0 2 mm: ai 

Missouri Botanical Garden 

\nli>ia In I. ma.mii Oerst.. \ idensk. Meddel. 
Dansk Naturhist. F0ren Kj0l)enha\ n 1861: 
129. 1862. Icacorea liebmannii (Oerst.) 
Standi.. Contr. U.S. Natl. Herb. 23: 1110. 
1924. Amatlania liebmannii. (Oerst.) Lundell. 
Wrightia 7: 40. 1982. TYPE: Mexico. Vera- 
cruz: prope Amatlan [de los Reyes]. July 1812 
(II). F Liebmann 7A (holotype. C!). 

Shrubs 0.3-6.1 m tall. Hranchlets 1-4 mm diam.. 
glahrescenl. or of scattered to denselv and ininulelv 
ru 1'ous glandular papillose, or with a mixture of 
densely ;m<l miiiiilelv ruloiis glandular papillose 
and scattered to denselv erect uniscrialc inullicel- 
lular glandular villous trichomes. Leaves with 
hlades chartaceous. oblong to elliptic. 4.8-21.5 X 

with an acumen 0.6-1.2 cm long. basallv acute to 
cllllealc. deciurcnl on the petiole. iiiconspiciioilsh 
punctate and puiictate-liiicale. usuallv 

alleied miiiutelv papillose, the indu 

• to serrate, flat: petioles slender. 

orescences erect, hi- to tripinnately paniculate. 
-14 X 2-13 cm. pyramidal, shorter to longer 
n the leaf hlades. the rachis with vcstiture like 

hranchlets. the [tranches terminating in 4- to 
flowered corymbs: peduncles 0.1-3.1 cm long. 

lower hranches often suhtended hy leaves: in- 
escence hracts unknown; inflorescence hranch 
its memhranous. ohlong. 1.2-7.5 X 0.2-3 mm. 

ally acute, hasally sessile, niidvein obscure to 
iressed above, prominulous below, prominently 
ictate and punctate-lineate. papillate and/or vil- 

* as in the leaves, the margins regularly entire, 
mtely erose, hyaline, sparsolv glandular cilio- 
: floral hracts similar to the inflorescence 
rich brads, but ovate or deltate. 0.6-2 X 0.2- 

mm, usually glabrous above or sparseh and 

leied 1 1 1 1 1 1 1 1 1 * papillae: petiole-, -.lender, terete. 2.8- 

') mm long, hicoiispieiioiish punctate and puiH late- 

. papillate and/or \illoiis like the hranchlets 

pink to light purple or red: calvx 

.4-1.9 X 0.0-1.3 mm. 

■ inconspK iimish ti 
prominently punctate and pimctale-lmeale. gla- 
brous adaxiallv. -pai-el\ and niinuteh papillatt 
abaxialh. the margins .■rose. h\aline. sparseh glan- 
dular ciliolato: corolla membranous. 3.6-1 mm 
long, the lube 0.4-0.6 mm long, the lobes ovate to 
lanceolate. 2.7-3.2 X 1.6-2 mm. apicalh acute, 
lew inconspicuous and proiiiineiitb punctate and 

pillielate-liueale. glabrous 1 1 in.llgl lonl . bill de||sel\ 

yellow glandiilar-granulose adaxialh. apicalh 
abo\c staminal tube as well as between th.- junction 
of the corolla lube and the lobes, the margins cn- 

1.6-2 mm long, the tube 0.3-0.0 mm long, the api- 
calh free portions 1-1.7 mm long, epunctale. 
dense!) yellow glandtilar-granulose. the anthers 
lanceoloid. 1.9-2.2 X 0.6-0.8 mm. apicalh long 

live epunclale to lew conspicuous and proiuuieiitlv 
punctalc: pistil 1-1,5 mm long, the ovarv oblong. 
0.8-1 mm long, the style 3.1-3.6 mm long, epunc- 
lale. the ovules 8. biseriale. Fruits globose. 3.6-5 
nun diam.. usualh apicalh iiicoiispieuouslv punc- 
tate and punctate-lineate. prominently costate. 

Distribution. \rdisia liebiiiaiiim is found from 
Hidalgo. Pueblo. Oaxaca. Veracruz, and Chiapas. 
Mexico, growing from 550 to 17(H) n. in elevation. 

Ecology. \rdisia liebmannii occurs in primarv 
and secondary lower montane forests, montane rain 
forests, and evergreen wet forests. 

Within Ardisia subg. \uriculardisia sect, \mat- 
Idnid. \n//<Hi /ii'bniatinii appear- to be most closclv 
related to \. schippii based on its entire l.< i.--nlarlv 
dentate or serrate (not pectinate) leaf margins. 
However. 4. liebmannii ma\ be . J i-t i ii-ii i-ln-< I from 
1. schippii bv its smaller corolla lobes to 3.2 X 2 
mm. shorter liiiear-lanceoloid anthers to 2.2 mm 
long with apiculate apices, and the shorter styles 

During the 1930s. J. F. Mac-bride of the Field 
Museum of Natural History photographed Neotrop- 
ical types and '■authentic specimens" of Furopean 
herbaria. I nfortunatclv. the widelv circulated pho- 
tograph of \rdisiu liebmannii. F neg. 22951. is from 
F Liebmann 7 (C). not of the type specimen. F 


dell) Ricketson & Fipoly. comb, et stat. nov. 

Basionym: Ardisia jalapensis Lundell. Wrightia 

6: 104. 1980. Amatlania jalapensis (Lundell) 

Lundell, Wrightia 7: 40. 1982. TYPE: Mexico. 

Veracruz: Mpio. de Xalapa [Jalapa], km 7 carr. 

San Andresito. 1320 m. 4 Aug. 1976 (fr). M. 

Zola B. 610 (holotype, LL!; isotypes, F!, 

MEXU!, XAI.!). Figure 4. 
Shrubs 1-6 m tall. Branchlets 1-3 mm diam., 
with a mixture of densely and minuteb minus glan- 
dular papillose and densely erect uniseriate mul- 
ticellular glandular villous 'trichnmes. Uares with 
blades 5.4-15.8 X 2.5-6 cm, with an acumen 0.6- 
2.3 cm long, usually glabrous above or sparsely and 
itiiimleb papillose especially along the midrib. gl.i- 
I .Km- below or sparseK and minutelv papillose, the 
imlument of the midrib with a mixture of densely 
and minutely papillose and densely simple multi- 
cellular hairs; petioles 0.(5-1.7 mm long, \estilure 
as in the leaves. Inflorescences 3-10.5 X 3-12 cm. 
vestiture as in the branchlets. the branches termi- 
nating in 4- to 8-flowered comnb-: peduncle-. 0.8 
1.7 cm long: inflorescence branch bracts 2-6 X 
0.3-1.4 mm. denseb glandular papillose and/or vil- 
lous; floral bracts 0.8-1.3 X 0.2-0.4 mm; petioles 
4-4.3 mm long. Floicers pink to \iolel: calyx lobes 
1.4-1.5 X 0.9-1 mm; corolla 3.6-3.8 mm long, the 
tube 0.5-0.6 mm long, the lobes 3.1-3.2 X 1.6- 
1.9 mm: stamens 3.4-3.5 mm long, the filaments 
1.6-1.7 mm long, the stamina] lube 0.5-0.6 mm km 
long, the apically free portions 1-1.2 mm long, the 19? 
anthers lanceoloid. 1.9-2 X 0.6-0.7 mm: pistil 
4.4-4.5 mm long, the ovary 0.9-1 mm long, the 2. 
style 3.5-3.6 mm long. Fruits 4.2-1.8 mm diam. 

Distribution. \nl I. -p. pdapen- 

sis is found in Hidalgo. Pueblo, and Veracruz, Mex- 
ico, growing from 800 to 1700 m in elevation. 
Ecology and conservation status. Ardisia lieb- 

and pine forests. It apparently has some ecological Ard 
tolerance because it has been collected in second- 
ary forests, and data suggest that it is currently un- 

Within Ardisia subg. Uiriadurdisia sect. Amat- 
lania. Ardisia liebmannii subsp. jalapensis is most Am 
closely related to 1. liebmannii subsp. liebmannii 
by its leaf blade margins regularly dentate to ser- 
rate, corolla lobes to 3.2 X 2 mm. linear-lanceoloid 
anthers to 2.2 mm long with apiculate apices, and 
styles to onlv 3.0 mm long, \rdisia liebmannii 
subsp. jalapensis may be easily distinguished from 

the leaves usually glabrous above or sparsely and dia 

brous below or sparseh and miuuteb papilln-e. the 
indument of the midrib with a mixture of denseb 
and nuniileb papillose and of densely simple mill- 

g tnduL 

to 3.2 x 

anil (he 

• slvle 

, 3.6 i 

linutely rufous glandular 
l uiu-.'iiate multicellular 
. the longer corolla lobes 

Specimens examined. MLMCO. Hidalgo: Mpio. I'e- 
lianu il« Dnri.l. 5 kin \K ol lenaugo (le Dona, toward 
llueh.iella. 5 Julv 1979 (fr). It Hernandez \l. .iiili (MO); 
Mpio. Jenangn de Dona. 0.5 km \ of Tenango along ma. I 
lo San Karlolo Tutotepec. 9 \,,g. 1987 (II. IV). S. Koch 
11772 (LI.). Pueblo: Canada del Mm \pulco \ of Zaca- 
poaxlla. 21 Juti.- 1077 ill). M. Marline: e, ,d. ^ i M I A I . 
MO): Mpio. de Palmatlan. \opa,iapa fi km al S\\ de Da- 
lmatian. 22 Max 1980 (II). l\ Tenon,, I.. & C. Romero de 
I. //5V5 (MLAl . MOi: Mpio. Uiuacallan. path lo /apo- 
lillan. \gua Dulee. I km SK ol M.uacal Ian. 2 Julv 1987 
(II). G. Tori: 1. .■/ al. .Vll (.TO. \1K\l ). Veracruz: Mpio. 
Iliialn-eo. trail lo lepampa. I km from the lluatiiMo- 
CoM-omatepee llvvv.. 1 \ug. 1070(11). .S. [rendano R. ix 
./. I.alzadn lO.i ( K I.I.I: Mpio. Ilualliseo. Cern, l.e.ea.i... 
Kfo Seco. along I luatuseo-Coseomatepec Hwy.. 7 Nov. 
1070 (fr). .S. \rendano R. 7,7,6 if. LL): Mpio. .In. Iii<|im de 
I'errer. Cerrn de \ .11 . Idea. I. Max 1981 (III. 6. Castillo 
C. el al. 1704 (K 1,1.. \AL): Mpi<, .|uelu,|„e feirei. be- 
tween I .as llavas ,x Pena Colorada. 21 June 1072 (II). R. 
Hernandez \1. ,v /. Cdzada 1 77U> ( I. \1K\l I: Mpio. J.i- 
«-lii(|ue Lerrer. La (una. Las llavas. 21 June 1072 (II). R. 
Hernandez \l. 17,07, I L MIAl i: Mpio. Tlapaeovan. ra. 
km hv air S of llapa. '..van on ,oa.l lo \llotonga. I I JuK 

loi!2 (fr). U. \eecl-t,. Di^s 2 n::::i if). 

"via crempelala Me/., in LngL I'llanzenr. IV. 236 (He 
9): 91. 19()2. Svn. nov. leacorea crenipetala (Me; 
Standi.. Contr. U.S. Natl. Herb. 23: 1110. 192' 

sia rekoi Lundell. in Sehultes. Bot. Mus. Leah 1 . 9: 185. 
1911. Sw. nov I MM Mexico. Oaxaca: Dtto. Teo- 

chitlan. I8°2I'N. <)96°45'\\r I KM) m. 24 JuK 1938 
(fl), R. Sehultes & B. Reko27.'i (holotvpe. MICH!. LL 
neg. 1971-12!; isotype, GH!). 
itlania elliptica Lundell. Phytologia 56: 19. 1984. Syn. 
nov. Ardisia elliplifolia Lundell. Phytologia ol: 63. 
1986, non Ardisia elliptica Thunb., Nov. Gen. PI. 8: 
119. 1798. TYI'K: Mexico. Oaxaca: Mpio. de Co- 
maltepec. km 149 carr. Tuxtepec, Sierra Juarez, 
Puerto Kligio. 8<M)m. 17 June I 906 (fr). 6. Marltnez- 


(holotvpe. XAL 

ubs 0.3-6.1 m tall. Branchlets 1.5-4 i 
, glabrescenl. or of scattered to densely ; 


< glandular |>;i- 

minutely rufous glanduL 
ture of densely and min 

piling- ainl -rallnvil erect iini^naU' multicellular 
ulatuliilar villous tricliomes. I saves with blades 
4.8-21.5 X 2.5-9.1 cm. with an acumen 1-4.2 cm 
long. usually glabrous above, scattered minutely 
papillose below; petioles ().(>-2.7 mm long, vesti- 
ture as in the leaves. Inflorescences 2.5-14 X 2-13 
cm. with the vesliture like the branchlets. the 
branches terminating in 1- to 10-llowered corymbs: 
peduncles 0.1-3.1 cm long; inflorescence branch 
brads membranous. 1.2-7.5 X 0.2-3 mm. with 
scattered minute papillae; floral bracts 0.6-2 X 
0.3-0.8 mm: petioles 2.8-0 nun long, papillose or 
villous as in the branchlets. Flowers lavender or 
pink lo light purple or red; calyx lobes 1.5-1.9 X 
1-1.3 mm; corolla 3.9-4 mm long, the lube 0.4- 
0.5 mm long, the lobes 2.7-2.8 X 1.9-2 mm; sta- 
mens 3.9-4 mm long, the filaments 1.9-2 mm long, 
the staminal tube 0.3-0,4 nun long, the apically 
.5-1.7 mm long, the anthers 2.1-2.2 
-0.8 mm; pistil 1—1.1 mm long, the ovary 
long, the slvle 3.1-3.2 nun long, I rails 

Ardisia liebmannil subsp. lieb- 
cted to Chiapas, Oaxaea, and Ve- 
growing from 550 to 1680 m in 

is glahreM-ent. or ..I scattered lo densely and mi- 
nutely rufou> glandular papillose, or v\ilh a inixliu. 
of densely and minutely rufous gl.mdiilai papill. isl- 
and scattered erect uniseriale multicellular glan- 
dular villous tricliomes. the shorter corolla lobes to 
2.8 X 2 mm. the longer anthers to 2.2 mm long, 
and the shorter styles to 3.2 mm long. 

Populations corresponding to the type of Ardisia 
crenipelala are unique only for the more scattered 
and slightK longer miilticcllului -Ian. hilar \illou- 

\\iil-u, \niisia subg. \ariealardisia sect, \mal- 

ImiKi \rilisia ;, ii mail is most 

closely related lo 1. liehmannii subsp. .>,!,:•,< i,^\ 
(see under that subspecies loi a discussion). \nlisia 
liehmannii subsp. liehmannii may be easily distin- 
guished from subspecies jalapensis because the 
vesliture of the leaves is usually glabrous above, 
with scattered minute papillae below, and the ves- 

ubsp. liehmannii in all other re- 
spects. In Mez's (1902) original description of Ar- 
disia creni/>elala. he listed three synlypes. fiolteri 
146. Ml. and Conzalli 169, from GDC, \\ and GH. 
Slandley (1921) did not lisl any specimens in his 
treatment for Mexico. I.undell (1982), however, in- 
directly lectolypified the Bolleri 146 collection at 
G as the lectotype in his Irealmenl of \mallama. 

Ardisia rekoi was originally described as being 
related to Ardisia ni^rescens Oersl. However, study 
of the type shows that it has auriculate calyx lobes. 
which clearly place il in \nlisia siibg. \nricidar- 
disia. I lu type of \mallania elliplica was collected 
in fruit and represent- populations unique for their 
slightly wider leaves with shallow dentate margins. 

MEXICO. Chiapas: Mpio. of 

Kotol le, 16 July 1965(11). I>. Hreedlore 
lighlamls. 6 km NK. I.v road, of 
lloehll mi hwv. 10,1. 10 \„j.. pir,;, |l,|. k. Roc rial. 1143 
(Ii. Ouvaea: Upm. ,,| San Felipe I sila. Cerro \enle. - 
km al \\K de San h-lipe I sila. 30 Sep. 1002 (Ir). C. 
Iharra M. el ,d. :i72» |\1<M: Mira.loi. 1813 (II). I: l.ieh- 
mann 7 (C. V ncg. 22o:, I ): \1pio. ..I \ ,1 1-- \a.ional. Cerro 
Ylira.lor. |.» km WW de \alle \arional. 16 Oct. 1992 
(Ir). ./. Mean- del Caslilh, el al. 17,11 (VIO). \l,„... \ngel 
\lbino Corzo. slope neai l.ancho Vicjo of the fui.a I'rn- 
sia. 23 Jan. !«>(,» (I,), t. Sluhm, T. i5<>7 (f, LL NY). 
Veracruz: Cerro .1.1 liom-o. O.I. IJJ56 (II). I/. Halle,! 

Ocampo', 19 June 0506(11). C. Conzalli IW (( -I I ). I hi/al.a. 
lf!55 III). 1/. Cnminfi >.n. (C): Mpio. Misantla. Kl Mava. 

(MEXU) Orizaba \lav lour, (II). f . /'„,,,„> 1212 it 

(41. \1(). \U II Paso Meiero between \l/ .,l„l II- 

surface. — C. Detail of basal area of s 
ahaxial surface. — C. Stamen, lateral m 
Zola B. 610 (LL); C. K-ll from M. \h 
Avendafio R. 556 (F).) 

leal surface. —I). Detail of in 
— II. Stamen, a.laxial s.irlac.-. 
e, al. W (V1FXU); D from R. 

ahaxial surface. 

i 7A (C): I) from K Liehmann 7 (C): K-H from S. I 

dowering branch. —15. Detail of margin of ahaxial 
. Detail of inflorescence. — K. flower. — F. Stamen. 
1 surface. —I. fruit. (A-C drawn from the holotype. 
ler 9517 (F); I from A. Shilom T. .1567 (F).) 

<>M 7 |K Ml A I 


Vnlisi;i |.< Hik ida <»( i>l . \i,lciisk. Yleddel. I);msk 
Naturhist. F0ren KJ0benhavn 1861: 150. 
1862. Icocorea pelhicidu (Oersl.) Standi. , 
Contr. U.S. Natl. Herb. 23: 1110. 1924. Amat- 
lania pellucida (Oerst.) Uundell, Wrightia 7: 
40. 1982. TYPE: Mexico. Veracruz: prope Pi- 
tal, (II). F Liebmann 29C (holotype, C!, F neg. 

Shrubs or subshrubs ().3-5(-10) m tall. 2.5-4) cm 
diarn. Brancblets 2- ( > mm diam.. scattered to 
densely and minutely rufous glandular papillose, or 
with a mixture of densely minutely rulbus glandular 

lanceolate. 0.5-61.2 X 5-20.0 cm. apically acute 
to acuminate, with an acumen 0. 1-1.8 cm long. ba- 
sally obtuse to cuneale or oblique or auriculale. 
usually deeurrent on the petiole. inconspicuously 
or conspicuously and usually pr..niiueiill\ punctate 
and punctale-lineate. glabrous above, sparsely and 
minutely papillose, denser along (lie midveins and 
secondary \eius. the midrib impressed above, con- 
spicuously raised below, the secondary veins 12 to 
59 pairs, prominulous above and below, the mar- 
gins irregularly finely serrate, the teeth 7 to 12 per 
cm. "pectinate." flat; petioles slendei. maigiiiale. 
0.6-5.6 cm long, glabrous above, sparsely to dense- 
ly and minutely papillose. hi/b,rcscences erect, bi- 
lo (|uadri-pinnately paniculate. 2.8-28.2 X 2.6- 
22.7 em. pyramidal, shorter or longer than the 

lets, the branches terminating m •>- to 16-11, ,wered 
corymbs: peduncles obsolete to 2.1 em long, the 
lower branches often -ublendcl by leaves: inflores- 
cence brads obsolete: inflorescence branch bracts 
membranous, linear or oblong to elliptic. 2.8-12/) 
• 6.2 2.'» mm. apicallv acute to attenuate, sparsely 
to densely and conspicuously oi prominently piinc- 
lale and piiiiclate-lineale. the margins entire except 
minutely erose apically. hyaline, sparsely glandulai 
ciliolate or irregularly finally serrate "pectinate."' 
the midvi'iu iv to impies-ed above ami raised 
below, the secondary veins obscure. Moral bracts 
membranous, linear. 0.6-3.5 X 0.2-0.8 mm. api- 
cally acute, basally sessile, the veins obsolete. 
sparsely to densely inconspicuously or conspicu- 
ously or prominently punctate ami punetale-lmeale. 
the margins entire except minutely erose apically. 
hyaline, sparsely glandular ciliolate: pedicels slen- 

puiiclate and pun. lale-l ineale. -parse I y to densely 
minutely papillose. Flowers 5- or 6-meroiis. lav- 
ender to red-violet or blue-violet to deep purple: 
calyx lobes essentially free, membranous to char- 
laceous. ovate to lanceolate. 1.7-5.2 X 0.9-1.7 
mm, apically acute to long acuminate, notched be- 
low the apex, rarely lacking a notch, basally auric 
ulate. sparsely to densely and inconspicuous or 
conspicuously punctate or piinclate-lineale. and of- 
ten prominently punctate and punctate -liueale. gla- 
brous adaxiallv. sparsely minutely papillose, the 
margins erose. hyaline, sparsely glandular ciliolate: 
corolla membranous. 5.5-0.1 mm long, the tube 
0.7-1.4 mm long, the lobes ovate to lanceolate. 
2.2-5.2 X 1.8-2.9 mm. apically acute, epunetate 
to bearing a few inconspicuous or conspicuous 
lions and punctale-linealioiis. -.fibrous idu 
xially. or densely yellow glandular-granulose api- 
cally above the slainiiial lube as well as between 
the junction of the corolla tube and lobe, glabrous 
outside or sparsely minutely papillose medially, the 
margins entire, hyaline: stamens 2.5-1.8 mm long, 
the filaments 1.5-2.;) mm long, the stammal lube 
0.2 0.8 mm long, the apically live portions 1-1.8 
mm long, epunetate lo slightly pun, tale, -fibrous 
to densely yellow glandular-giaiuilose. the anthers 
ovoid to lanceoloid. 0.0-5.1 X 0.6-1 mm. apically 
apiculate. basally cordate lo sagittate, the connec- 
tive epunetate lo inconspicuously punctate; pistil 
2.5-1.8 mm long, the ovary oblong. 0.9-1.5 mm 
long, epunetate or conspicuously punctate and 
punctale-lineate. the sly le 1 .5-5.5 mm long, epune- 
tate or inconspicuously punctate, the ovules 9 to 
57. Fruits globose. 5.9-5.8 mm diam.. inconspic- 
uously punctate and punetate-lineate. eostate. 

Distribution. \rdisia i«-llu< i,l,i is found from 
San Luis Potosf. Pueblo. Veracruz. Oaxaca. Tabas- 
co, and Chiapas, Mexico, and throughout Ylesoa- 
merica. In South America it is loiuul I com Vene- 
zuela. Colombia. I* mi, b ... an. I I'eni. growing from 
35 to 1650 m in elevation. 

Ecology. Ardisia pellucida occurs in primary or 
secondary humid, moist, wet. or pluvial liopical for- 
ests from the lowlands lo premonlane levels, rarely 
in cloud forests where it occurs on low. isolated 
mountain peaks. 

Ardisia pellucida is a highly variable species 
within Ardisia subg. Auriculardisia seel, \niatlania. 
from the other species in section \matlania. it may 
be recognized by its leaf blade margin irregularly 
Itiiely serrate with teeth 7 lo 12 per cm. "pecti- 
nate.'" whereas the leal blade margins in \rdisia 
schippii and I. liebwannii are regularly dentate to 

Valley, near Tela and the Janlfn Botanico Lancetilla 
Ardisia pellncida >ubsp. Iniu -elillrnsis Kir- Biological lb-serve in \llanlida. I londuras. growing 

i Pipoly, subsp. nov. TYPE: 
Atlantida: hills above Lancetilla, 1500 ft. [457 
n.|. 15 Jul> 1934 (fl). T. Yunckei 4588 (holo- cida s 
type, MO!; isotypes. A!. F!, NY!). Figure 6. mary i 

Subspecies haec a subsp. jiellucida. )• >l>u I i - cal 

to 4500 mm annual precipitate 

3.1-3.2 (nee 1.7-1.9) nun l.mgis 1.5-1.7 (nee 0.9-1,1) search was conductc<l to look for this subspecies. 

mum lull-. lobuli- < uiullim- 1.3 I.., iik.ii 2.2-2.7) nun Ion- without success, so we may assume the subspecies 

gis. .lenique antheris 3.0-3.1 (non 1.2-1.4) mm longis js ,| lreate ned. 

I" a '' l: " r mstal - Etymology. The subspecies is named for the 
Shrubs 1-3 m tall. Branchlets 5-9 mm diam.. Lancetilla Valley region, near Tela. Atlantida. on 
sparsely to densely minutely rufous glandular pa- the north coast of Honduras. We are very grateful 
pillose. or with a mixture of rufous glandular pa- to lug. Ciro \a\arro. Director of the Janlfn Botanico 
pillae and uniseriate multicellular glandular villous de Lancetilla, who brought this subspecies to our 
trichomas. U-aves with blades oblanceolate, 13.7- attention and helped Pipoly look for it in the field. 
64.2 X 3.9-20.6 cm. with an acumen 0.7-3.6 cm Within Ardisia siibg. \uriculardisia sect. Amat- 
long, the secondary veins 21 to 29 pairs: petioles limia. Ardisia pelhuida subsp. lancelillensis shares 
0.9-3.6 cm long. Inflorescences hi- to tripinnatelv with sul»peeie> j,cih,cida ,i -miliar vestiture with a 
paniculate. 8.-5-28.2 X 5.6-22.7 cm. shorter or mixture of densely minutely papillose and scat- 
longer than the leaves, the peduncle, vestiture of tered, erect, long, simple multicellular hairs on the 
the rachis, and branches similar to the branchlets. upper branchlets and throughout the inflorescence 
the branches terminating in 11- to 19-flowered cor- branches, but subspecies lancelillensis may be dis- 
ymbs: peduncles obsolete to 4.8 cm long: indoles- tinguished by its larger ovate to lanceolate calyx 
cence branch bracts linear. 3.8-12.9 X 0.8-2.9 lobes to 3.2 mm long, larger corolla lobes to 4.5 
mm, densely and prominently punctate and puiic- mm long, and laigei lance.. |,. id anthers to 3.1 mm 
tate-lineate. the veins obscure, llie margin- minute- long. 

ly erose hyaline sparsely glandular ciliolate: floral HONDURAS. Atlantida: Res™ del jar- 
bracts 1.5-3.3 X 0.2-0.6 mm. densely and prom- ||m |?ol . mro (|| , mrr| . ||;i _ S|( ,,,-„ ,„, „_ ,„,„„-, , 
inently punctate and punclate-lineale: pedicels /- (m ^ i, ail( . ( .,jl| ;1 \ a llev. Lancetilla, 22 June-27 July 
3.1-12.5 mm long, inconspicuously, punctate and 1020,11,. 1. Chi, /..,-/._■ 200 il'i: Lancetilla Valley, ca. 10 
punctate-bneate densely minutely papillose. Flo, mi. SK ^ m IWm ,,n>.n ^**%^^ 
ers 5-merous. dae blue-volet, v ,ole.-purple to (M0): a|ong nw<1 for 1IluIliH ^ wa|er SU|(|lK <riv , 4UjI1 , 
deep purple: calvx I'd*.- ovale to lanceolate. ...I- ....,,)!;, (.anlc,-. ,,„ ,,,,,1 ,a. 2 mi. WSW of Tela 
3.2 X 1.5-1.7 mm. apicallv acute to long acumi- a ,„l S of man, Inw.. " I .1, I'CT ,|, ,. T. Croat & I), //an- 
nate, without notch below the apex, sparsely to nan hl(,:\7 <\1< »,: I,,-,,; i„ from „l ..Id dam, 21 Aug. 1999 

prominently pun,, „ I , „ - II , ," * < ; ^ J',-" , ■ ,/ „ ' '^ ^^ '/ ^ 

5.4-5.5 mm long, the tube 0.9-1 mm 

, 12 Nov. 1978 (fr), I). Hazlel, 2' >r,n WWW I: I 
cetilla, 10 Aug. 1978 (fl). I). Hazleii 5058 (HJBL): i. 

lobes 4.3-4.5 X 2.2-2.3 mm. -par-elv but con- ( |am. Lmcetilla \.ill.-y. 2'< Jub 10 \ug. 1951 (fl), R. How- 

spicuously punctate and punctate-lineate. densely ard el al. 500 (A): Tela, lancetilla Valley, above Experi- 

yellow gl and ular-granu lose adaxiallv. apicallv »"■<" >'•"">"■ aim,, -man, ami dupes stream but 

• , , ■ 1,1 II I, ,1 h"l>'« the dam. I W I<«!0llil. ./. Uarlhmgaleial. 5170 

above the s.ammal tube, as well as between the ^ _ ^ m , _ m| ^ „ 

corolla tube and lobe junction. -p;u-elv and mi- m iiXlu p,<>| ,r r)< <;. \,,/. v „„ / 7.125 ( I IJM.): vicinity of San 

nutelv papillose medially outside: stamens 3.8-4 Alejo. at base <.f hills S <,l San \lejo near Rio San Alejo 

mm long, the filaments 1.3-1.6 mm long, the sta- [<a. 15 km >\\ ..I I. la . 22 27 \,„ 1917 (ster.), P. Stand- 

„ „;.„. o ■> ,) -, mm I,,.,. ,| lt . ,,,„., |K p.,. .,,„, Am 7720 ,|- ,. banc, nil;, \allev. nca, Tela. 6 Dec. 1927- 

minal tube 0.2-0.5 mm long, the apually por ^ ^ ^ p S/(/n(//fV - 2f)92 {V , S) s:j ,- 7 (K 

tions 1.1-1.2 mm. ..-conspicuously punctate. L S) _ 5 ,„-^ (K , S|; Llll , ,,,11., \allcv. <a. 3 mi. S of Tela. 

densely yellow glandulur-granulose. the anthers ; > () Jim p„i2 ,|| L ,,. Hrhsirr et al. 12681 (F, LL): Lila 

lanceoloid, 3-3.1 X 0.9-1 mm: pistil 3.1-3.2 mm River 2 mi. above Puerto Sierra. 18 Jan. 1903 (IV). l\ 

long, the ovary 0.9-1 mm long, conspicuously Wilson 79 (NY, US). 
punctate and punctate-lineate. the style 2.1-2.2 

mm long, inconspicuously punctate, the ovules 35 4. A 

to 37. fruits 5..5-5.8 mm diam. * 


Gaz. 12: 132. 1887. Icacorea pectinata (Doiin. 
Sm.) Standi., Contr. U.S. Natl. Herb. 23: 1110. 
1924. Ardisia pellucida Oerst. var. pectinatu 
(Dorin. Sm.) Lundell, Wrightia 3: 99. 1964. 
Amallania pellucida (Oerst.) Lundell var. pec- 
/.t/m/.« (Dorm. Sm.) Lundell, Wrightia 7: 40. 
1982. Amatlania pectinata (Donn. Sm.) Lun- 
dell, Phytologia 55: 235. 1984. TYPE: Gua- 
temala. Alta Verapaz: Pansamala. 3800 |)|) 
|1158 m|. June 1886 (fl), H. von Tiirckheun 
942 (holotype, US!, US neg. 2381!: isotype. 
GH!). Figure 7. 
Shrubs 3-5 in tall. Branchlets 4-7.5 mm diam., 
scattered to densely rufous glandular papillose. 
Leaves with blades oblanceolate. 6.5-41 X 3-12.3 
cm, with an acumen 0.7-4.2 cm long, the second- 
ary veins 12 to 31 pairs; petioles 0.9-2.1 cm long. 

1 Hi i 11 nil I i il. |v | i.ltlK III lie. 

6.2-20.1 X 6.1-22.5 cm, shorter or longer than the 
leaves, vesliture of the rachis. branchlets. abaxial 
bract surfaces, and pedicels similar to the branch- 
lets, the branches terminating in 5- to 16-llovvered 
corvmbs: peduncles obsolete to 2.1 cm long: inflo- 
rescence branch bracts linear. 3.2-4.9 X 0.2-0.8 
mm. sparseb and inconspicuously punctate and 
piinelale-lineale. llie \eins obsolete. the margins 
minutely erose. hyaline, sparsely glandular cilio- 
lale: floral bracts 1.5-2.6 X 0.2-0.6 mm. sparsely 
and inconspicuously punclale and piuiclate-linente; 
|)edicels 6.2-7.8 mm long, inconspicuously punc- 
tate and punclate-lineale. Flowers 5- or 6-merous, 
lavender to red-\iolet: calyx lobes ovale. 2.1-2.2 X 
1.4-1.5 mm, apieally acute, sparsely inconspicu- 
ously punclale and punelale-lineate; corolla 5.9- 
6.1 mm long, the tube 0.7-1.1 mm long, the lobes 
5-5.2 X 2.8-2.9 mm. mostly epunctate, rarely or 
sparsely with glandular-grariiilo-c papillae around 
the top of the corolla tube inside, sparsely and mi- 
nutely papillo-e medially oiilside: stamens 1. 3—1.8 
mm long, the lilamenls 1.(3-2.2 mm long, the sta- 
minal tube 0.5-0.7 mm long, rarely or sparsely with 
'Jaiidiilnr-graniilose papillae on the slaiiunal lube 
and lilamenls. llie apieally free portions 1.1-1.7 

mm long, epimctate. llie anthers lauceoloid. 3-3.1 
X 0.8-1 mm, basally subcordale; |)istil 4.74.8 mm 
long, the ovary 1.4-1.5 mm long, epunctate. the 
styles 3.2-3.3 mm long, epunctate, the ovules 21 
to' 22. Fruits 5.5-5.8 mm diam. 

Distribution. Ardisia pellucida subsp. pectinata 
is found from Mpio. bason in Chiapas. Mexico, to 
ISnj.i \eiapaz. and Izabal. Guatemala, growing at 
(50-)l 158-1851 m elevation. 

Ecology and conservation status. Ardisia pellu- 
cida subsp. pectinata occur-- in primary premunlaiie 
and montane wet forests, because of its restricted 
distribution, it should be considered threatened. 

■j\. The specific epithet was derived 
from the latin, meaning "with narrow closely-set 
divisions, hki- a comb" relet ring to the teeth on the 
leal blade margins. 

Within Ardisia -nl^ \mi<-ui,irilisia sect, \iuatla- 
,,;,, : ..- dimes a similar 

yeslititre will) subspecies tliomascroalii. sparsely to 
densely, minutely, nilous glandular papillose through- 
out. However, Ardisia f>ellucida subsp. pectinata may 
be distinguished by its oblanceolate leaves, smaller 
but wider ovate calyx lol>es to 2.2 X 1 .5 mm, larger 
corolla lobes to 5.2 X 2.9 mm that are sparsely and 
minutely papillose medially outside and sparsely 
<_l udul -miiulose inside around llie corolla tube 
and filaments, the larger lanceoloid anthers to 3.1 X 
1 mm, the larger pistil to 1.8 mm long, and the longer 

style to 3.3 r 

Specimens examined. MFXICO. Chiapas: Mpio. Kay- 
o.i. near I'ucrto del \ lento. <> mi. \ \\ <»f Pueblo \iievn 
Solistahuacan aloiin mail !•> Tapinla. 2(1 \u». !'»<.;> ill). />. 
IU,e,l!,„r HW.'i ll. I I. Mi: N-ha Vgra 10 km ah-ve 
Ray6n mezealapa along road to Jilotol. 10 Jan. I ( >!!| lltt. 
I). II,,-,;!!,,, , ,y /; Kelt,-, !'>:<, / 7 ( 1 .1 .. MO. NY), (il \TK- 
MAbA. Baja Verapaz: along dirt toad 4 mi. NE of Pu- 
,„||,,,. IT.Iiilv I077.ll). y; Cr,,,,, li:L'.-,'.\Uh. l/ Km. 
.luvama. SK ol" Cliexenne. about 17)1.11. S\\ ol I Sana ... i a. 
!>, \pr. I'M.) ill. Ir)../. Si, ^e, mark Will 1 1- | inlloi. m-,,,,-, 

5. Ardisia pellucid;! -ad.sp. pellucida. figure 8. 

I ' -' . ' ' •■ ",,. M.I. II 

Figure (left I. ■\rdi.siu pellucida s.... ,.. 
— C. Detail of inflorescence. —I). Flower. — F. Stamen, ahaxial surface. — F. Stamen, adaxial surface. — ^. auuiieii, 
lateral margin. —II. Fruit. I \ I intloies, eneei X I drawn In.m isolvpc. T. )uneker IStltUCU): \ deaf) & B from isotype, 
7.- )u„,ke, 177,',!', (F): l)-<; from C. Webster el al. 126X1 (1,1.); U from P. Slandley 5M22 (F).) 

Figure 7 (right). Ardisia pellucida sul.sp. pectinata. —A. Flowering branch. —15. Detail of abaxial leaf surface. — 
C |)Hail of inlloreseeiice. —I). Flower. F. Stamen, ahaxial surface. — F. Stamen, adaxial surlace. — (7 Stamen. 

lateral mamii —II Fruit (A. 15 drawn from hololxpe. //. ran Tiirekheim "12 il <l: < In ...type. //. von Tiirekhein, 

01- ((3D- I) (i from I). Iheedtove IIWH (F): II from I). Hreedlove & B. Keller 49317 (NY).) 

Volume 90, Number 2 

iSl ... i! i I mi I I 1 i lih ) ,,i i. mi lot, st. < s|h < i. 

1««!2 nil I'm' n < 

gaps, and at the forest edge along 

. • ! V, 1925 {Iffisto^^V^ Can t0lera f te ™ d T a t t0 ^^ Shade ' bUt ^ T' 
US!, US neg. 2376!) paction of soil. At this lime we see no immediate 

threat to this subspecies. 
Subshrubs 0.3-3( -10) m tall. 2.5-6 em (ham. Common N(imes . "Tapacajete" (L. Williams 

Branvhlets 2-6.5 mm diarn., sparsely or d '■ 

■at h. 

fous glandular papillose, or with a mixture of rufous wuhin AnUsul sl| i )g Aunt ulanhsia sect. Amat- 
glandular papillae and uniseriate multicellular /(//;m ^.^ peUui , lda subsp . peUucida appears to 
glandular villous hairs, /^m.s will, blades 8.7-42.9 ^ mQ ^ doselv re|ale(| to subspecies l an cetillensis 
X 3.7-14.6 cm, with an acumen 0.4-4.8 cm long. lh . m ((( ^ ()th ' r slU)spe( . ies because of the mixed 
the secondary veins 14 to 29 pairs; peholes 1.1- ^ |||im , ()I , ,,„. vr . r ,alivc parts consisting of mi- 
3.1 cm long. Inflorescences hipinnatelv paniculate. nu{e papillae and viUous tric homes. However, sub- 
2.8-15.2 X 2.6-10.2 cm. shorter than the leaves. ^.^ ,„.„,„.,■,,„ m;|V ,„. ( | islmgll ; s he,d from sub- 
vestiture of the rachis, branches, bract sin ^.^ lanceiMensis | n its snia |l el ova te calyx 
faces, and pedicels similar to the branchlets. the u ^ ^ , () mm , (|| ^ smaHei . (>oroUa lobes tQ 2 . 7 
branches terminating in 9- to In " ■ . -■ Mm j (1II „. am u ma || rr ovo ,d anther, to 1.4 mm long. 
peduncles obsolete to 2 cm long: mlloreseeiiee yhe tvpe of 4rr/i.sia myrtof/onia is in fruit; how- 
branch bracts linear or oblong to elliptic. 2.8-1 1 .2 evr] . ||( ; ( . ;ms) , ()( - i(s Mmil;ir u . Milure and smaI1 

X 0.6-2.6 mm. midvein obscine to imputed ^^ ^^ u ^ ]{ ^^^ , [M , Ullcida subsp . 

' ' " ' pmunneullv /W/ , /r/fA/ in a |[ respects . 

ly erose, hyaline. sparselv glandular ciliolale or ir- Specimen* examined. MEXICO. Chiapas: Mpio. Oco- 

regularh finally serrate ••pectinate": floral bracts 1- zocoautla de Kspinosa. M kin N of Ocozocoautla along 

32 X 0.2-0.8 mm. densely anil pronunenlK mail I,. I' .-- I'M '■ I I'"'' <!., />/.' ',-.. \P 

. . , , , , Jill \lp.., I'.il.i.qn. (>-12kmSofPalen- 

punctate and pun, . ,1, I.... .1. |,.,l..-,-U ,._ I... ,_,,„, f) /w/o|v 

mm long, inconspi, i . i . and punctate- 9 6 . 5// (LL) . Oaxaca: Mpio. Santa Ma.fa Cl.imalapa. ra. 

lineate. Flowers lavender, red-violet to deep purple: 5 km S \\ of Sania Ma,, a along i-.miI to La/am Cardenas. 

calyx lobes ovate, 1.7-1.9 X 0.9-1.4 mm. apicallv 17 Juh 1984 (fl), H. Herndnd. (, 211 MOi On., In 

acute. S l ig l«l y nolehed below ,he apex, p^inenUy «,,,,,,. , ,,,., .i^U*. 

|)Uiietal« m<l pun. ' • , . I)( , |„..- hl( \]„ llim . : \ u ~ \ (| | ( ,. Vil \H. MO), 

long, the tube 0.7-1.3 mm long, the lobes 2.2-2., |» m .i,| a \l, . , ,|, - „, l'i. ... M.n. I'),| if,) //. Hum, //. 

X 1.8-2.2 mm. willi lew but conspicuous puncta- 268 (MEXU); Mpio. Piedras Negras, El Salto, 9 km NE 

pun.. '[aii--'l 

allv. apicallv 

r La Ceiba. 25 Feb. 1987 (fr), A 
M IX I 1. San l.ui- l'..lo*i: \l|.i... \ililla. Poblado Xum- 

, . , , , ehiaio, 12 Aug. 1976 (fr), ./. Calzada 2572 (MKAU): 10 

minal tube as well as between, 0,0 b. ,u,l I..I, ^ ( , , , |uK ,„ |( , „„ ( „,„,„,„,, 

junction, sparselv ami nunutclv papillate imdiallv A/ Shinhl ,„ „., ,,, \ |.,|,., M -„ \1,,„, leapa. vi< 
outside; stamens 2.3-2.5 mm long, the filaments „l l,- ; ,pa. : mad loin,,,, leapa and h.rntalpa. .,. I 
1.3-1.5 mm long the staminal tube 0.3-0.4 mm ^gg^ ST^S^^It^ 

long, the apicallv free port,,,,,. I .1 mm. d,,i~, v ^ ^ ^^ ,1,1,1- 

vrdlow glandular-granul. , li 1 1 s ovoid. 1.2- |M1K m j,^ P)71db: \bsantla. JuU I'M J dl li I. A" /A / 

ndndez M. 1227 (F), C. Purpiis 5959 (BM. F, GI"" 

MO); Mpio. Hidalgotitlan, La Escuadra, 

and ,-,-, Ihe ? ,e !..>-,,„,„, ,„.,, ^^^^^^2^^^. 

epunctate or mconspicuoiislv punctate, the ovule- . >( ^ „ m r|h u. >v/n>/> -S-559 (F). GUATEMALA. 

lactic. 5 Apr. 1939 (ster.). P. Slandley 

lintla: barranco of Rio Hun ion. NK of 

1941 (fl. fr). P. Slandb-y IMM) (F|. 

J), • .. \ , ltu'<i . -I I >- ■ • < ■ ' 

is found from San Luis Potosf. Pueblo. Veracruz. 
Oaxaca, Tabasco, and Chiapas. Mexico, through 
Belize. Guatemala. Honduras. Nicaragua. Costa 
Rica, and Panama. In South America it is found 

from Venezuela. ,:„ ,i«. K.-.U •. I IV™. Il ,',,.,;,' ^•,:," I ".',V- I "£ Roan,' 

grows from 35 to 1650 m in elevation. trems 6348 {U | :5 | ): p., c,,,,,),,,.. | km K on Kfo Purula 

5. Ardisia pcltu 

mtafia del Mic 
38745 (F, LL). Peten 

Ill I. 1/ < inn urn A C. Ilcm-i 
Til;..;,,,. 10 31 ,| ;m . 1926 (fr). 

V,, ', ] u] ( "\" ' '!",,' ""VVi j " '•' '"''' M '»- Vnlisia p.dluridhi subsp. tlioiiia*croatii Ric- 

/'/'!<</ ,'\l'<> u'V'liV.i.". ( „ | ,':,,-,' || ,'|"; /',, kelson & Ripolv. subsp. nov. TYPK: Costa 

\l;n. I'»;;i) llrl. II. >/./,„, !<>/,.->/ |\1()|. |{i >:) >: Hacienda ,vi,a - SaM ■h> s, ' : about I mi. beyond divide be- 

Lalima. Montana La Victoria. N i.l ilu- l„„i.e <,| Hacienda tween San Isidro del (ieneral and coastal town 

|, ;i l ,'"'^ SV V;' , | S " , '"' l ' l k ' S ', 1 '' l " !; - ""• ' S "'" 1 " 1 " '' "■' <)f ' n <>">ineal. 900 in. 22 Mav 1976 (fl). T. Croat 

!,,. 2 lIsXZnI;; ;;.■; ;;!, l:i /" !•:;; ■;"";'/.'; ; ^ 7 ° ( M< »^ mod. f^ * 

(MO. M >; Mpio. ,lc Uaspam. Kescxa Rosaxxas. Lelxxeen Propter ramulos. rhachides intlnreseentiamm. pedicclos 

nil. 20 Jan. P*X,(lr). A*. Kurd,, <■/ „/. W.-/ (MO). « (M'\ ,>i, ,„ ,„, ,ti. ,-t I,!, ,■.„,,-,„'„'', uV. , ' ll |,lV.'/"vm"!i,7'i- 

UICA. Alajiifla: uhoul ., km S ,,l (nnalelc near llic liio milans. sc,l ah ca l.,|,ulis calxcm. 2.:, 2.(>lnon 2 I 2 2) 

/apale and along ihc mch road to I pal.,. 12 \„x. P>7:, „„„ l,o,-i> 0.9-1.0 (nee I.l l.:» mm latis. loluilis coro- 

l ','•>• "• ""V" * '«'• / ;" / " "><>»"• l I';:, ,lc >,,„ ||„„s .'x.(, :}.,'! (,„,■ 5.0 :..2| mm longis. anteri. oxo„le,s 

(.arlos, 21 I'ch. 1900 (Irl. 1. !/,<////,/ A', n „l. I7(>:i!i (L). ad apices acutis (mm laneeolatis a,l apices caudatis) per- 

<:«••!:.».. lim,,n l, man, l„,» .-I < \- facile recognoscilur. 

TIK.Turrialha. .iOJnlx P«;7> (II,. 1/ CnnumXI', ll,un- 

m<-/ 5754 (MO). (Miaiiarasu- \ oll{,o Las I lores. ,a I Shrubs I in tall. Hranclilcts 4.5-5 mm diam., 

'ii.;:.i :l Moult •/Lima. 21 Jan. I <>;;."> scattered to densely i i|.|\ minus glandular pa- 

pillose. hums ■ with blades elliptic. I ."J. 5-28.5 X 5- 
I I em. with an acumen 1.5-1.7 em long, the sec- 
ion on the Rfo I'licrio \„io |i, -i I „| ik jimr'ii,,,, Villi il.'c <>"«l«ry veins 20 to 25 pairs; petioles 1.2-1.7 em 
!i<> Sarapi,|iii. along W Kixer Road. I June I97R (II). 1/. long. Inflorescciicrs 1 1 ipiunalelx paniculate. 12-12.5 
•rnuun n.l'i ,l»l KL). 7 \pr. N,'!.) „|). //. //„„„„,/ «/«// X | |_| 1.5 rm . toiler llum „„. |,, 1Vr> . u .„ lilln ,. „,- 

11. nv ',. i'"i ■\ , 1 ! , ."'i!;';.'v ' /; :";,;';"' ' ",'; t ;';;v ,l,r mthis a,,<l l,,a,i<1 ^ simiia| - "■ ^ i» a »<i<i^- 

Miilon <l, •''I'alamai'ica. llVerx'a VidiCiV J'alamanca" Vail." l ' U " man( ' ,1(>s terminating ill 5- to 7-flowered eor- 

e Talamanca. Vmnhi i. 17 .Inn, 199 | (I,). ,,'. ,,„- vmbs: peduncles 1.6-1.7 em long: iidloreseenee 

unto 256 (Fm. MO). I'.iiiiarenas: almig Oiiehrada Mo- branch bracts linear, 5.5-7.5 X 0.4-0.6 mm, in- 

'_''"""''''' | ,n ''" ' Ml '" '■'"■■"' ! • ' -"'■ c(.ii>pieu,)ii-l\ pniiclalc .iiid pini, lale-lineate, below 

l' ()| ' 1;| ■•) (| |' ,V"|„ „| ,|{ Vita ' I ' " ,' s l ,;lls, ' ,x ilM<l mirmleb ■papillose, the veins obsolete. 

OillHll). l/.(,„ ( ,™,„ ,/ .,*MI 10|2|) \-\\\\]\ lIx-HKii-i-iiiseiilire. miimlrlx .-rose. hxaliiK-. spars,- 

olon: Vchiote. 2 June P»75 ill). K. Ihvsslr, III).! (MO). I> glandular eiliolate; lloral bracts 0.6-5.2 X 0.2- 

►arien: vicinilv „l T..rti. IMU, ,„,. I ,,l Ka>a.,o Dan. 0.1 ,„„,, ineonspicuoiislv piuu-tate anil punctate- 

';"^ ;::■;.,::: ,::: l ;,',::;:.:;:r,^;, 1 ";:™;r; ""'v- • : ■ • >■■•- •r-'r" 

entry, -lot. 2H55I (\l(ll: unniH ol Lax a. Uio I'axa. Hall l ,lll,,l;lU ' all « l piin<lat<-lu,«-ate. sparsely and mi- 

ctween I'aya el I'axila. lO.liine l«).V>(ll). It. Sirru vi al. "iihl\ papillose, l-lnucrs 5-merous. lavender: calyx 

y5(MO) Can;,,,,;, /one. I'.a. ,„ ( :,,l,,ra,lo Island. I<»be- lanceolate. 2.5-2.6 X 0.9-1 mm. apically 

U 'l' lr ' '" ■''''■ wil1 ' V||II " 1H - ' M"- L),0 (II). '/: C„, t ,i | n|| ,, aciiiiniiale without a notch below the ipe\ 

Vl kla: xiauirN?,;".', m'jV!! ." ". ':"!,. 'i ,', V. rVw.' ^'" """^ '""" , " 1 " ■"" l P — ' a '<-'i- a .e: cV 

i de Perija. along mlla [i] -^ ,nl11 lon ^ l,l< ' l,ll,r ■ — ■ - ""*■ ">'" '""S- 'll< J 

./. I- rum till,-. 



Mpio. I'ereua. 

Hacienda Malabar, 

1 km I.,.,,, 

'creira Hwv.. 27 Nov. 1 W) (fr). 



.5709 (MO). ECUAI 

tera Hollin-Ix)reto, t 

• ■ \V., .. 

x. !««!«> (fr). 

76 W |\1()|; \\ 

V, Nuevo RooafuerU 

la pohlacion en 

film del |{i„ 

serio Ouichua, 2 

'. LI.. 1 '»;; | ', |, ). J. Jaramillo & 

/ '.'' 

42.U (OCA). /274 (OCA). PKRU. 

Mailrr de 

e Kas.nera. Biological I«>Im-s 5.6-5.8 X 2.4-2/ 

la. alSVV „j Macl,i,|,„s. densolx. xellow glandular-graniilos,- adaxiallv. api 
tally above the slaminal tube, as well as betweei 
the corolla lube and the lobe junction, glabrous out 
side: stamens 5.2 5.5 mm long, the lilamenls 2.1 
2.5 mm long, the slaminal tube 0.7-0.8 mm long 
the apically free portions 1.7-1.8 mm long, epune 
late, densely yellow glaiidular-granulose. the an 
thers ovoid. 0.9-1 X 0.6-0.7 mm: pistil 5.8-5/ 
mm long, the ovary 0.9-1 mm long, conspicuously 

Distribution. Ardisia pvllucida subsp. lliomas- 

Volume 90, Number 2 

Ecology and conservation status. Ardisia pellu- 

forest. Unfortunately, no lurlliei ecological dala arc 
ayadahle I mm the label, Given the an essibi lit\ of 
the site, and lack of collections, this subspecies 
should be considered threatened. 

Etymology. It is an honor to dedicate this taxon 
to Thomas 15. Croat ol' die Missouri Botanical Car- 
den, scholar, gentleman, and preeminent authority 
on the systematies and ecology ol \eotro|)ical \ra- 
ceae and the genus Anthurium in particular. 

i ( i , , id I ism sect. Amal- 

i ! [ I 1111 

a similar vestiture with subspecies peclinala. 
sparsely to densely, minutely, rufous glandular pa- 
pillose throughout. However, subspecies thiuwis- 
croatu may be distinguished by its elliptic leaves. 
longer but thinner lanceolate calyx lobes to 2.0 X 
1 mm. smaller corolla lobes to 3.8 X 2.6 mm that 
arc glabrous outside and deuseb glandiilai-gruuu 
lose inside around the corolla lube and lilam.-nls. 
the smaller ovoid anthers to I X 0.7 mm, the small- 
er pistil to 3.9 mm long, and the shorter style to 

7. Ardisia sehippii Standi.. Field Mus. Nat. Hist.. 
Bot. Ser. 12: 412. 1936. Ainatlania sehippii 
(Standi.) Funded. Wrightia 7: 40. 1982. TA I 'F: 
Belize. Toledo: Temash River, 200 ft. [61 m|. 
8 Aug. 1935 (fl. fr). W: Schipp 1365 (l.ololvpc. 
F!, F neg. 68249!, LL neg. 1971-99!; isotypes, 
A!, BM!, G!, GH!, K!, MICH!, MO!. NY!). Fig- 
ure 10. 

Ardisia izuhalunu Lumlcll, Wrigl.lia 5: 88. 1975. Syii. nov. 

Ainatlania ' > : I I H) I ' '« " ' ' > 

53: 413. 198.'!. TM'I : (malcmala. l/iibal: Kl F-lor. 
6 km S. 30 Jan. 1975 (fr). C. Lundell & K. Contreras 

ularly dentate or serrate, the teeth 3 to 5 per em 
when present. Hal: petioles slender, marginale. 1.2 
2.8 em long, glabrous abo\e. papillose below. In- 
florescences erect, tripiimately paniculate, 7.5-24 X 
9-23 cm, pyramidal. mostly longer than the leaf 

Mi 1 i inn I. . b he-, and peilieels milillleK 

papillose, the branches terminating in 5- to 10- 
llowered corymbs; peduncles 1.1-3.3 cm long, the 
lower branches often subtended by leases: inllores 
eenee bracts unknown: inlloieseeiiee branch bracts 
membranous, oblong, 2.1-2.4 X 0.9-1.1 mm, api- 
cally acute, basally sessile, scattered, inconspicu- 
niisK punctate and punctated iueate. the veins ob- 

glandulai eiliolale: Moral bracts membranous, ovale 
to debate, 1.8-1.9 X 0.9-1.1 mm, apically acute, 

scattered inconspicuously punctate and | late 

lineale. the veins obscure, the margins minutely 
emse. hyaline, sparsely glandular eiliolale: pedicels 
slender, terete. 3.8-6.3 mm long, inconspicuously 
punctate and punelate-lineate. Eloicers 5-merous, 
lilac to purple; calv x lobes chartaceous. ovale. 1.7- 
1.8 X 1.2-1.3 mm. apically acute to acuminate, few 
prominently punctate and punctate lineale. gla- 
brous adaxially. sparsely and minutely papillose, 
the margins erose. hyaline, sparsely glandular eil- 
iolale: corolla membranous. 6.(>-(>.8 mm long, the 
tube 1-1.2 mm long, the lobes ovate to lanceolate. 
5.6-5.8 X 2.6-2.9 mm. apically acute, prominently 
punctate and punctate lineale. glabrous throughout, 
but densely yellow glandular-granulose adaxially. 
apically ahoyc staminal tube, as well as between 
the corolla tube and the lobe junction, the margins 

u, nails 2.5 2.6 mm long, the staminal lube 0.8- 
0.9 mm long. I he apically free portions 1.6-1.8 mm, 
cpunclale. densely yellow glandular-granulose. the 
anthers lanceoloid. 2.5 2.6 X 0.7-0.8 mm, apically 
long siibiilale-apiculale. basally cordate, the con- 
nective inconspicuously punctate; |)islil 5.9 <> nun 
long, the ovary oblong, 1.3-1.4 mm long, promi- 

S7//-//A.S or small trees 1-10.7 m tall. 5-26 em 
diam. Hranchlets 3-1.5 mm diam.. scattered to 
densely minutely rufous glandular papillose, often 
glabrescent with age. I raves with blades coria- 
ceous, elliptic, 7.8-25.9 X 2.6-9.1 cm. apically 

coti-picuoiisly punctate and punctate-lineate. gla- 
brous above, sparsely and minutely papillate along 
the midrib, often on the blade, usually glabrous, 
the midrib impressed above, conspicuously raised 
below, the secondary veins II to 22 pairs, promi- 
niilous ahoyc ,\)u\ below, the margins entire to reg- 

Distribution. Ardisia schippii is found in Cayo 
and Toledo, Belize, and Alia Verapaz. I/.abal. and 
I'eten. Cuatemala. mowing from <>l to 19<X) m in 

Ecology and conservation status. Ardisia schij, 
pil occurs in primary rain forests. While il is eel 

lamb nol eoi u. at ibis lime there are no dal 

to suggest that the species is threatened. 

Within Ardisia subg. Auriculardisia sect. Amat- 

lania. Ardisia srhippii appears to be most closely 
related to A liebniannii (sec under that species for 
similarilies). However. 1. srhippii may he distin- 
guished from I. liebniannii In it- hugei coiolla 
lobes to 5.8 X 2.9 mm. longer lariceoloid aullieis 
to 2.6 mm long with subulate-apiculatc apices, and 
the longer styles to 4.7 mm long. 

The populations corresponding to the type of Ar- 
disia i:uhalana are unique only for their narrow 
leaf blades with more entire margins. The vestiture 
is sparser than in A. srhippii, but otherwise they 
match perleclb. The Ivpe of \rdisin izabalaua is in 

Specimens examined. BFLIZF. Cavo: \icinilv of Cue- 
x.i- < ol Milhoiiario. 2' '-MO Mas I'C! (II). /.' Cfal 2.'.5i'.7 
(F. FIT,. I.I.. MO): Valentin. June Jub IWO (hi. C. I.un- 
,lrl! (tf>l if. I.I..N1I. Ifh. C. Landed 6277 (F. OH. I.I.. 
\V TFA). 01 VILMM.v. Altu Yerapuz: Rul ' 

. 2') .1 


(in. <;. , 

:• a: <.Vm/re 

,.« /o.>r6il.l.|2|. M- i l/ : . 

,, on new l/ahal lioad. about 
2 5 km from village. I 7 \u«. I"*,') (IV). /.. Ciwl/rm.i .'WW 
(LI |2[. TK\l: Ca.lenas. kn, S\\. ') J..l> P>70 (fl). E. 
Cnnnems HUM (1)1 KL. I.l.|:t|. NM- <•"•->. /o>/6/(L. LI. 

|.',| M ); Puerto Mendey. on Bin Dulee lioad ahoul 10 
k,,.. (.Sep. I')7I)HV|. /■:. Cnnlnras H>2I7\\)\ KL. I.l.|:i|. 
M |: LI Li,,,. 2.', Mar. 1<>72 llrl. E. (,mi,enis ll-WI ( I L 
[2 | Peten: Jlllec. km ol of road between Santo lonhio 
and Sanl.i \na. 20 J.ilx DOl (II. (r). E. Conlreras 2675 
(1)1 KL. I.l.|2|. MAr. LaCumbre. \\ of km 1 42/ 1 43. 500 
m from the road. 10 Sep. P)7o (hi. C. Landed X E. C„n- 
irems /<«.«( I ,L 1 2], MO). 

krdisia suhg. Auriculanlisia sect. \iirirulanli- 

Subshrubs or small trees to 20 m tall. 10 cm 
iiam. Bninchlets stout, terete. denselv furfura- 
eoiis-lepidole. < U| >ul iform-lepiclote (the scales with 
nrving number of lobes or arms, at lime- ol two 
izes). mixed fuifuraceous- and cupuliforiu d. -pi- 
tar often breaking off. leaving an ap|.arcntly vil- 
ous tricboiue. the indument normally persistent. 

I,r,iirs iiKHiomorpliie: blade- membranaceous to co- 
riaceous, elliptic, obovate to oblanceolate. promi- 
nently, inconspicuous (pellucid) or conspicuously 
pmielale and |iimctale-liiieale. the margins entire, 
flat: petioles at limes obsolete, or slender or stout 
and much longer, then canaliculate or maigiiialc 
inflorescences terminal, erect, pinnately to tripin- 
natelv paniculate, pyramidal, -holler than the 
leaves, the flowers on the secondary or tertiary 
branchi-s jjomcrnte . c..rv mbose oi ol coiige-l<-d cor- 
ymbs, rarely pscudorae.-mo-e: inflorescence bracts, 
inflorescence branch bracts, and floral bracts mem- 
branaceous to coriaceous, usuallv persist. -nl. Moral 
bracts white, light pink to red. 3.7-9.7 X 1-8.5 
mm: pedicels stout, terete, often recurved and ac- 
crescent or inerassate with maturity. Fhurcrs 5- or 
6-merous. green to white or pink to purple in hues: 
ealv\ lobes nienibranaceous to coriaceous, the mar- 

riaceous. the lobes variously connate, ovate to 
lanceolate, the margins entire, hyaline: stamens 
with obvious filaments, connate basally into an elo- 
bate incoiispieiiou-ls hvaliue tube, the anthers live, 
narrowly ovoid to linear-lanceoloid. lanceoloid. to 
ovoid, apically apiculate. basally cordate to sub- 
cordate, dehiscent by suba 
wide, longitudinal slits, tl 
uouslv to conspicuously 

inconspicuously or eonspicuousK. or prominently 
punctate, the ovules few 
merous verlicels. Fruits globose to depress ed-glo- 
bose. often conspicuously to prominently punctate 

Distribution. Four species, from Puntarenas 
and San Jose in Costa Rica, southward to Code. 

Panama, and San bias in Panama, lo \ulioqiua and 
Vaupes in Colombia. They grow in habitats from 5 
to 1400 m in elevation. 

F.roh>i±\ and runsemitinn slot lis. Members of 
section Auriculardisia occur in primary pivmonta- 

)■■,,,„,■ 10 (left). \rdisiu srhippii. —\. Idou.-n,,^ branch. —15. Detail of abax,al leal surface. -(.. Detail ol ml —I) Llower — K. Stamen, ahaxial surface. — K Stamen, lateral margin. O. stamen, adaxial smlaeo. 
H Lruil (A-C drawn from holotvpe. IL Sehipp \U>5 (L): D-0 from C. Lundell 6277 [CAW. H from E. Contreras 102 

Figure I I (right). Ardisia dod^ei. —A. flowering branch. — B. Detail of aba 
rescence, inlloresceiH •«■ branch brad, anil lloral bracts. — D. Flower and floral hi 
F. Stamen, adaxial surface. — 0. Stamen, lateral margin. — IL Fruit. (A, I? drawl 
9HHI (MICH): C-0 from /•] Ones,,,!,, 5.10 (Li: II from R. \guihu 554 (MO).) 

s defined by die «<»inhi nal ion of ils terminal inflo- above Km Paquita. 5-50 m. 15 Aug. 1036 ( 

eseenee with branches terminating in congested. C. Dodge S: V. Coerger 9««/ 0iololype. VI, 

lomerate corymbs, each of which is subtended by neg. 68147!. PL neg. 1971-45!: isolvpes. CI 

persistent, enlarged inll, .rcsceucc branch brad. (dl!. MICII!. I.I. iicn. 1071-15!. \1<>!. I.I. ii, 

nd finally, the individual flowers subtended by a 1971-40!. N> !. I'll!. I C!. IS!. PS neg. 2571 

ersislent, enlarged floral bract, as long as the flow- Figure I I . 

Shrubs or small trees to 20 in tall. 10 cm dia 

• n m i ,„•; Tax a uv A«n,sn si ,,;. Anaa unmm "ranchlets 5-7 mm diam.. densely lomentose w, 

ECT. AURICULARDISIA a mixture of lui luraccoiis-lcpidote and cupulifoi 

a Plants wlf i\| f t I I scales and with scattered stellate and slipitate-sl< 

lucent scales. ||„. scales oll.-n earl? cat I u'ous. Ia,r ln( ' ,mm «7 h '"'^ with blades chartaceous 

12.7-25.2 X 5.2- 
inate, with ; 

long. hasalb acute to nbtus, 

■ trichomes, the stellate 
uniseriale. multicellular stalks to 1.5 mm long, 
the apical stellate arms often caducous, ill,- hairs 
lli'-n appealing glaihlulai ■-villous: culv\ lobes 
narrowly lanceolate. 7.2-8.2 mm Ion-.' apicallv !,ll(l |H"iclalc-lin,-ate. glabralc above 

'.' I, , ,n| " ,' M '"" r> ' '"' iMonnuclb ,,„■,! h.-lou. ||„- M-condan vein. 

i "II. lie Irichomes. when present the , A . ' . , 

'') 01 pan-. |iroiiiiiuiloiis above and below: pcli< 

stout, marginalr. I 12 mm | ( , nil . moslb glabralc 

scales: eahx lobes male, ohlale or orbicular to above, below with a mixture of dense lurfuraceous- 

- 1 ' 1 "" 1, :; '■■ m ; ,,l,,,,L ''' v""'"d«'d lepidotc ami cupuliform scales and wall, 

lo Irillieale. ollen >h-,lilK eniaimiiah . . ,, , . . , , * 

2, floral laae.s red or pink: calyx lohescori- *"'"""' i " ,<l -"P'l<<«— tellale Inflores- 

aceous. 5.-1-5.7 X 1,8-5.2 mm: corolla <''''<''« bipimialclv or Iripiiuialcb paniculate. 2-7.5 

lobes 8.3-8.5 X 3.2-5.5 mm: anthers linear- X 2.5-7 cm. the rachises densely lomentose with 

lanceoloid, 4.5-5.7 X 0.9-1.1 mm. basally a mixture of lurluraceous-lepidote and cupuliform 

2,, :t:;;ti, „,„, „, ,*,„ ,1^7^ sr t s vvi,h ™ h ™ 1 si " ll; "" ami ^^^- 

calyx lobes membranaceous to cliarlaceous. Inchoincs. the branches Icriiiinaling in 5- to 10- 

l<"> 2.7 1.9 I , mm: corolla lobes 4.5- (lowered glomerate corymbs: peduncle 0.4-1.4 cm 

5.7 X 2.1-2.7 mm long; anthers ovoid to long: inflorescence bract caducous, membranous 

!onir.e m"sa 2 ^a.e. * ^^ m '"' ^^ ^^ ° VaU " '" " blon * M - 2 - 6 X l) ^ ] - h «''»• a l»" 

3a. Petioles ""canaliculate, 3-1.5 cm long- fally lollM, ' ( ' ( l lo truncate, prominently punctate 

leaf blades Hal above, the secondary illul puiidalc lui.alc -labial.' above, lurfuraeeous- 

vcins prominuloiis above and below: lepidotc below, the midrib inconspicuous, the mar- 

gins entire. Hal: inflorescence branch brads similar 
to the inflorescence bracts, but persistent, ovale. 
0.8-2.1 X 0.0-1.6 cm; floral brads similar to the 
inflorescence bracts, but red to light pink, usually 
persistent. 8-0.7 X 5.2-8.5 mm. enveloping the 
llower: pedicels 3.6-1.2 mm long, incoiispi, uousb 
pmiclatc-luicale. with mallei. ■<! lurfuraceous-lepi- 
dolc scales, /''lowers 5-nicrous. white or hghl pink: 
calyx lobes coriaceous, very widely ovate to sub- 
orbicular. 5.4-5.7 X 1.8-5.2 mm. apicallv rounded 
or truncate, conspicuously and somewhat promi- 
nently punctate and punctate-lineate. glabrous ada- 
'> hroadened \ially. scattered furfuraceous-lepidole abaxially; co- 

10. Ardisia nervosissinu, m ||., chartaceous. 10.3-10.5 mm long, the tube 2- 
2.2 mm long, the lobes ovale. 8.3-8.5 X 3.2-3.5 
, Publ. Field Mus. Nat. mm. apicallv obtuse lo rounded. 
'<•. 1058. \uneid(irdisi,i piuidale and piuiclalc-liiiealc. glal 
II. Phvlologia 49: 343. sparsely furfuraceous-lepidole aba 


h blue-gray or lav 

emler; calvx" 

cliarlaceous, oblate 

. 2.:; 2.7 < 

...... long. Il.c 3.8-1.1 

iTe 1 '" 

ig, ihclilamen.siV 

\rdi.siri liloiiienila 


s n.arginale to near 

: lei bUlo 



low; perianth lighl green to pink; calyx 

. 1.8-2.2 X 1.0-2.1 

<..;: 7 i 

urn long, the lobes 

1.5 I.J! .,., „ 

long: a 

nthers 2.0-2.8 mm 

long. Ill, 1,1 

Volume 90, Number 2 

9.6-9.8 mm Ion*:, the filaments >.2 -.). I mm long. 
the slaminal tube 2.5-2.6 mm long, the apieally 
free portions 2.7-2.H mm long, pellucid punctate, 
the anthers linear-lanceoloid. 4.5-5.7 X 0.9-1.1 
mm, hasally siihconlale. the connective conspicu- 
ously punctate: pistil 7-8.1 mm long, the ovary 
0.9-1 mm long, the style 6.1-7.4 mm long, con- 
spicuously punctate ami punctale-lineate. the 
ovules 56 to 59. h'niils depressed-globose. 6.8-7.2 
mm diam., conspicuously punctate. 

•»::■;, tlauion. \rdisia dodgci was considered en- 
demic to the area of the Osa Peninsula in San Jose 
and Puntarenas. Costa Kica. until Pipoly (1991a) 
reported disjunct populations from Antioquia and 
Yaupes. Colombia, growing I ti nil 5 to 700 m in el- 

Kcology and conservation status. \rdisi, 
occurs in primary wet or pluvial forests. Because 
of its restricted distribution, it should he considered 

Etymology. Ardisia dodgci was named in honor 
of the late Carroll \\ illiam Dodge (18.95-1988). li- 
chenologist. mvcologist. and slafl mcinher at the 
Missouri Botanical Carden from 1951 to 1965. 

Within Ardisia suhg. Aiiriciilardisia sect. Auri- 
culardisia. \rdisia dodgci is distinguished hy its 
red or pink persistent Moral bracts to 9.7 mm long 
and to 8.5 mm wide, the large corolla to 10.5 mm 
lung with lobes In 8.5 nun long, and linear-lanceo- 
loid anthers to 5.7 mm long. 

Specimens examined. COSTA KICA. Punlarenas: 

Cjml.iii deO>a. Ucsrna hueslal (.oil.. I'uhe. I • - Mogos. 
Collil... 5 Sep. 1991 (IV). R. \guilar.2rl (CI!. IM!. Mil!: ile Osa. Ouel.rada Banegas. 25 Sep. 1001 Hrk It 
\»uil,ir IW (Clt. MO): Cuenca Terraba-Si.-,,,,-. llalua 
Chal. La Parcela. 21 Sep. I90(, Mcr.l. II. X^uilnr K,W 
|IMJ. MO): 25 kin \\ ile Cliacarita l.v road. between Kin- 
,-,„, and Chaearila. 29 .1 1 1 1 h ■ I « >* > I (III. /,'. Ihnu.ncl X M. 
•ij MCI! I TO IM! MOi < ..1.-. I...s M.,- 
uos. headwaters ol Ouel.rada 21 Nov. 1991 (II). 
<; timer,, /'MMCIi. ITI ;. IMS. M< >): Peninsula de Osa. 
IJ.meho Ouemado. SL Seelor. If. Sep. I '>'<2 1 1 . ). ./. Marin 
X l>. Mam, 52M< I!. I TO. IMS. MO): Canton de Oolfho. 
Par<|lie Yieiorial Coreov ado. Peninsula ile Osa. Corcnva- 
,|„. 20 Max 1905 111). /■;. Mjhru 21" (CI!- IMS. MO): V>- 
quinas Forest, area l.etvv. Ui'o bsquinas anil Palmar. 11 
Jub 1951 (II). /' Mien f,2.-,.'i (CVS. I |2|. (51): n.a.l to 
|- u ,,lo.|,n,e„e/.. Osa. 19 km \\ ol I. A. roule 2. sal. (fl). /, 
Come: I'. I<> /Oft ( I T |2 |. M()|: Canton lie Osa. Sirrpe. I ...s 
Mo-os. upper drainage of Km Sierpe. affluent of Ouebrada 
lahoga. 15 Dec. 1900 ill). C. Ilenera 1T»I (OIL I TC 
INIC MO): Parotic \aeional Corcovado. I pper Ollas Trail, 
27 June lo:;;; ,||,. ,;. Reman X I' l'hilii,,s 629 (CR, MO); 

Canton de Osa. Kaneho Ouemado. 5 Julv 1901 ((]). K 
Onesada .',21) (CR. K f'TC. IM!. MO): Canton de Oolfilo. 
Kelugio Naeional de launa Nhes„M.„ll,lo. 21! lei,. 1001 
11,1. (,. Hurra a al. 2227 (CIC k I: l.etw. C;«,lfo Dulce and 
KfoTerraba. Dee. 19 17 (II). -1. Skulch MW ( K. MICH |2|. 
| S): Canton de Osa. ICmeon de Osa. Kntranee to (Jioeua- 

eo. 27 Aug. 1992 (fr). A. Zamora et al. IHUl (CI,'. IM'.. 

I\1P,I\. \nlioqiiia l.igliwav lo the sea. near 

\ ill., \ileaga.0 Dee. 1 II! |ll ). /' h>,„ : X M. Sanchc: M. 

2,H (COL. IS). Vaupes: 19'.. Kananari and Cerro de Isi- 

1951 (II). //. Carcia-Harriga I2766A (COL). 

9. Ardisia glomerata Lundell, Amer. Midi. Nat- 
uralist 29: 486. 1945. \uneul ardisia glomc- 
rata (Lundell) Lundell, Phytologia 49: 344. 
1981. TYPE: Panama. Code: hills N of hi Va- 
lle de Anton, trail to La Mesa. ca. KMK) m, 2 
Sep. 1941 (fl), P. Allen 2741 (holotype, M0!, 
LL neg. 1971-51!; isotypes, A!, US!). Figure 
Small trees to 8.5 m tall. Branchlets 6.5-11 mm 

raeeous-lepi.lote and . upiilale scales and -lipilale- 
stellate lomenlellous mi iniill iccllular stalks to 0.7 
nun lonu. appearing glandular-v illous as in A. ur- 
sina Willi the stcllale portion broken off. I /-ares with 
blades membranous to cliartaceous. elliptic. 51.5- 
39.5 X 9.6-16.2 cm. apieally acute to acuminate, 
with an acumen 0.7-2.5 cm long, hasally acute or 
acuminate, decurrenl on tin- petiole, mosllv incon- 
spicuously [)iinclate. but with a few punetale-lin- 
eations above and below, glabrate above, below 
with a mixture of dense furfuraeeous-lepidole and 
scattered copulate scales, except densely so along 
midrib, the midrib impressed above, prominent be- 
low, the secondary veins 25 lo 75 pairs, prominu- 
lous above and below; petioles stout, canaliculate. 
3-4.5 cm long, glabrous above, below with a mix- 
ture of dense furfuraceous-lepidote and scattered 
cupulate scales. Inflorescences bipinnatelv or tripin- 
nately paniculate.' 12-24 X 6-16 cm, the rachis 
densely tomentose as in the branchlets. the branch- 
es terminating in 5- lo 9-flowered glomerate cor- 
ymbs; peduncle 1.7-5.8 cm long; inflorescence 
bract unknown: inlloreseenec branch bracts often 

persistent, inembn us. oblong. (1.9-1.2 X 0.4-€.7 

cm, apieally acute, inconspicuously (pellucid) 
punctate and punctatc-lineate. glabrate above, fur- 
d stellate-tomentose below. 

onspicuoiislv punclate-lineate, furfura- 

.Tous-lcpidole and slellnl.-loinentose. Flowers 5- 
or 6-merous, blue-gray or lavender: calyx lobes 
chartaceous, oblate, 2.5-2.7 X 4.5-1.7 mm, api- 

calU hum ale and -lighllv emarginale. eonspieu- 
ouslv punctate and punetalc-liucale. glabiou- ada- 

xially, furfuraccou-depirfole abaxialb: corolla 
coriaceous. 8-8.4 mm long, tin- tube 2.7-2.8 mm 
long, the lobes ovate. 5.3-5.7 X 2.4-2.7 mm. api- 
callv obtuse to rounded, eucullate. conspicuous!) 
punctate and punclate-lineate. glabrous adaxially. 
sparsely furfuraceous-lepidote abaxiallv; stamens 
6.6-6.9 mm long, the filaments 3-3.2 mm long, ha- 
sallv widened above the junction with the tube, the 
staminal tube 1.4-1.6 mm long, the apicallv free 
portions 1.4-1.6 mm long, the anthers lanceoloid. 
3.8-4.1 X 1.3-1.5 mm. basally cordate, the con- 
nective conspicuously punctate dorsally; pistil 5. 1 
6.2 mm long, the ovary 0.9-1.1 mm long, the style 
4.5-5.1 mm long, conspicuously punctate and 
punctate-lineate. the ovules 19 to 21. Fruits de- 
pressed-globose. 6.2-8.2 mm diam.. densely and 
conspicuously punctate. 

l)!,;nLsi>!nn. \nlisia glomernln is endemic to 
Cerro Pilon and adjacent areas near El Valle in 
Code. Panama, from 6(K) to 1000 m in elevation. 

E,-i>lntx\ and eunsemiliau status. \nlisia ir/o- 
merata is found in cloud forests, a life zone that is 
rapidly disappearing in Panama. Because of its re- 
sin, led distribiilh.ii. il should be considcicd threat- 

Ftuiioloitv- The specific epithet was derived 
from the Latin glomer, meaning to form or collect 
closcb together into a sphere, and refers to the con- 
gested flowers of the inflorescence. 

Within Ardisia subg. Aurieulardisiu sect. Auri- 
eultirdisia. Ardisia glomerala is most closely related 

orbicular to suborbicular calyx lobes, which are 
apicallv rounded to truncate, and chartaceous to 
coriaceous corollas. However. 4. glomcrala can 
casilv be separated from 1. nerrosissima because of 
its canaliculate petioles to 4.5 cm long, flat leaf 
I. hide-, chartaceous. oblate calyx lobes to 2.7 mm 

-lepidote abaxiallv.; 


Specimens examine,/. I'WWIV Code: belu. Cam 
I'M,,,, a, „l Kl \allc. K> \ng. PW.7II1)../. Ihl.rX J. Ih.u; 
!;•»,! ill.. \1()i: foothill- (ilCrmi I'ilon near Kl \allc. 5 
Oct. I')f.7(ll). ./. Du/.rX- I/. Cone,, 1. IIW2W .1.1: (,-.„. 
Pil.iii. II \allc. I Jan. I'*..", (hi../. I hike & B. Ldlatlun 
I l<WI (FTC. I.I. MO): Kl \alle «lc \nloii ill the loot of 
Ccrro Pilon. 15 Aug. P>o7 (fl). ./. l)wver& \l. Corn;, 1. 
;'/.;/; (IK,. IK. MOi: I,. Mesa. Cem, I'ilon area. P) Jan. 
!<)(,;; ill). ./. Duxer 8321 (MO); woods adjacent to ehicken 

- Kl Valle. 


with lobe- to .>.7 mm long a 

. 2.7 . 

10. Ardisia iiervosissima Kundell. Wrightia 4: 
62. 1968. Auriculardisia nervosissima (Lun- 
dell) Lundell, Phytologia 49: 345. 1981. 
TYPE: Panama. Code: El Valle. 800-1000 m. 
28 June 1967 (fl), J. Duke 13150 (holotype, 
LL!, F neg. 55645!; isotypes, GH!, LP!). Figure 
Shrubs or small trees to 10 m tall. Branehh-ts 3.5- 

8.5 mm diam., densely furfuraceous-lepidote. the 
scales persistent, h-ares with blades chartaceous. 
elliptic, obovate or oblanceolate, 9.2-26.5 X 3.2- 

9.6 cm, apically acuminate, with an acumen 0.3- 

1.7 cm long, basally acute to nearly auriculale. de- 
current on the petiole. prominently punctate and 
punctate-lineate. sparselv furfuraceous-lepidote 
above, more denseh so below along midrib and to- 
ward base, strongly bullate above, the midrib im- 
pressed above, prominent below, the secondary 
veins 22 to 35 pairs, deeply impressed above. 

auriculate. subsessile to 3 nun long, densely fur- 
furaceous-lepidol<-. In/loivsreixvs pinnately to bi- 
pinnalely paniculate. 0.2-12.5 X 4—8.5 cm. to- 
nienlose with rufous Iraiishicent cupuliform scales. 
at times with arms, the scales 0.1-0.6 mm long. 
and with flat furluruceou — lepidote scales below, the 
branches terminating in 3- to 7(or 9)-flowered glom- 
erate corymbs: peduncle 0.5-1.4 cm long: inllore- 

F.gure 12 llef.l. Udisia plomeraU,. \. flowering (.ranch. -IK Detail ol abaxial leaf surface. -(.. Inchon, 
types, sessile slellale il.-lli. -I qdale-avllale on a multicellular stalk with the ra\s lermmallv rotate leenlerl. ami 
sessile, flat scale (rislil). D. Detail o! inlloreM ■ence. -how ing iiill..n->r.-iiee I, ran- I. brad and floral bracts —F. Howe 
showini; floral brad K "stamen, ahaxial surface. — G. Stamen, adaxial surface. -II. Manual, lateral margin. — 

Frail. (A-C & K-H drawn hum holotvpe. I'. Mien 2741 (MO): I) from isotype. /' Mien 2711 (US): I from J. Du<yt 

figure l.i | right I 

i«. —A. Flowering branch. IK Detail of ahaxial leaf surf; 
• nmllicellular stipitate (eculerl scales, the scales termim 
, II. ,1 (riglil) -ales. — D. Detail of infloi 

neeolale lo elliptic, 0.5-4. 1 X 0.3-1.4 cm, acn 
ally smaller, apically acute to obtuse, densely a 
eminently punctate and punctatedineate, tom< 

nder that species for 



spicuously impressed above and prominent and 
conspicuous below, the secondary veins inconspic- 
uous, the margins entire, sparsely glandular-cilio- 
late; floral hracls similar to the inflorescence 

4.8-5.7 X 3.4-4.3 mm; pedicels 0.5-5.1 mm long, 
lomentose like the inflorescence racliis. Flowers 5- 
merous. light green to pink; calyx lobes membra- 
nous, orbicular to suborhicular. 1 .fi-2.2 X 1.9-2.1 
mm. apically broadly rounded lo truncate, promi- 
nently punctate and punclale-lineate. glabrous ada- 
xiallv. sparseb lurliiiaceoiis-lepidole abaxialb: Co- 
rel hi coriaceous, (,.8-7 mm long, the tube 2.2-2.3 
mm long, the lobes ovale, 4.5-4.8 X 2.1-2.3 mm. 
apically flat, rounded to obtuse, prominently [mnc- 

lale and punelalc dim-ale. glabrous lluougl ; sla 

mens 5.8-0 mm long, the lilamenls 3.7-3.9 mm 
long, not widened above the tube junction, the sta- 

1 "I '"be 1.3-1.5 mm long. I In ■ apicalb live pt.i 

tions 2.3-2.5 mm long. 0.4-0.6 mm diam.. epunc- 
tate, the anthers ovoid. 2.6-2.8 X 1.2-1.4 mm, 
basalb sagittate, the coiineclive inconspicuously 
punctate dorsal ly; pistil 5.3-5.8 mm long, the ovary 
0.9-1.1 mm |„„g, die style 4.4-4.7 mm long, prom- 
inently punctate, the ovules 19 to 21. Fruits glo- 
bose. 0.1-7.2 mm diam.. densely and in 

1 ■ 
lated to 4. ^lomerulu <->. 
similarities). However, 1. 
separated from 4 ^/onierala because of its subses- 
sile. marginate to auriculate petioles lo 3 mm long, 
bullalc leaf blades, membranous and orbicular lo 
siiboibieular calyx lobes to 2.2 mm long and to 2.1 
mm wide, and coriaceous corolla with lobes to 4.8 
mm long and lo 2.3 mm wide and glahmiis aba- 
xially. and anther to 2.8 mm long. 

Specimens examine*/. PAN AVI A. Code: Cerro I'ilnn. 
Kl Niillc I Jan. L968 (fr), /, Duke & B, Lallathin 1 1968 

il.L MO). I tun i (Li.. \k )):(:,,„,«:,,.„■,„, I. i-. | :1 „. i<).„, 

(ID../. Duke & J. /W, /.,/.>':> (I. I.. M<M: < >,,„ I'il 

S'|. l"o;;,||i../. IhxerX- It La/lath, n 11077 (VTC.MO): 
^ «»l I'l \iill--. I.Hucc, (,■,,„ Cai. ic,„al ami ( vrru Oailal. 
'■"■ .!•■!> l ,, o- ) HI). \ Kn, w ei „/. otiMl (II. MO): C, m , 
I'iloii. Sep. l')o,'l (II. Irl. /,'. 1,,/lallun 11-2 1 1 I < . |2|. MO); 

Panama, in Code (Cerro Caracoial. Cerro I'ilon. La 
Mesa, and Kl Valle) and Panama (Cerro Campana 
and Cerro Trinidad), growing from 800 to 1 KM) m 

is. Ardisia i 
in elfin and cloud forests, 
ed distribution, il should be . 

M(H: Cm, Campana. 27, JuK 
nl. IK).', (|,|.. VIO. \\ ); (; rlT , 
(11. III. C \1el'he,snn 7W/II.I 


. Ardisia ursina Lundell. Wrightia 0: 92. 1979. 
Yulerioanthus ursinus (Lundell) Lundell. 
Wrightia 7: 50. 1982. Auriculanlisia ursina 
(Lundell) Lundell, Phytologia 57: 450. 1985. 
TYPE: Panama. Panama: Kl Llano-Cart f 

ad, 10 km from 

, MIAt !. MO!. M 


ii- It.hi.-1i. 15. Detail ..I a 

scales, ill.- seal.-- often earlv ea< Iim •< m-. and densely 
lomentose or \ i 1 1 < n i - In hirsute rufous slipitatc-slcl- 
late trichomes throughout, the stellate branches on 
uniseriale. multicellular stalks to 1.5 mm long, the 
apical stellate arms often caducous, the hairs then 
appeal oil.' glaudular-villous. Iran's with Uades 

X 2.2-6.9 cm, apically 

3- Id nini long, hasally 
iole. prominently punctate and punctatc-lineale. 
ahove and helow with a mixture ol sparse to scat- 
tered, minute, sessile translucent scales, the scales 
often earl\ caducous, and with scattered villous or 
hirsute ru foils stipilale-slellatc Irichomes. denser 
alnnu imtKein, otherwise as in the hraiichlets. the 
midril) impressed above. prominently raised below, 
the secondary veins 21 to 35 pairs, prominulous 
al>o\e and below; petioles -dernier, canalu ulate. 2 
mm long, with a mixture of scattered minute, ses- 
sile translucent scales, the scales often early ca- 
ducous, and scattered to densely lomentose or vil- 
lous to hirsute rufous stipilate-stellate trichomes 
above and below, Iiiflorrscciu-rs bipimialeb panic- 
ulate, 4.2-8.4 X 2.3-4.7 cm, with a mixture of 
scattered minute, sessile translucent scales, the 
scales often early caducous, and densely lomentose 
or villous to hirsute rufous slipilale-slellate tri- 
chomes. otherwise like the hranclilels. the branches 
terminating in 5- In '>-llowercd glomerate . mwiil.-; 
peduncles ().2-l.."> cm long: inflorescence bracts 
early caducous, unknown: inflorescence branch 
bracts persistent, membranous, ovate to elliptic. 
0.6-4.2 X 0.4-1 cm. apically and basally acute, 
conspicuously and prominently punctate and punc- 
tate-lineate. with a mixture ol scattered minute. 
sessile translucent scales, the scales often early ca 
ducous. and scattered villous or hirsute rufous slip- 

chis. the midrib impressed above, prominently 

rolla membranous. 0-0. I mm long, the tube 2.1- 

2.7 mm long, the lob.-s ovate. 0.7-0.0 X 3.0-L2 

stamens 0.0-6.8 mm long, the filaments 3.9-4.1 
mm long, the slanuiial lube 1.3-1.7 nun long, the 
apical free portion 2.1-2.0 X 1-1.4 mm. the an- 
thers ovoid. 3.8-4 X 1.5-1.7 mm, basally subcor- 
date. the connective inconspicuously punctate dor- 
sally; pistil 7,1-7.7 mm long, the ovary 1.7-1.0 mm 

long, the style 5.5-5.0 mm long, in -piciion-K 

punctate, the ovules 21 to 20. Fruits globose. 5- 

5.8 mm diam.. prominently puu< laic an. I punctale- 

Distribution. Ardisic 
ama, Cocl6, and San R 
300 to L400 ... in eleva 

Panama, growing I 

Etymology. The specific epithet, meaning " 
the hear," refers to the rufous slipitale-stellate i 
dument of the branchlets, leaves, and inflorescem 

that resembles the fur coat of a bear. 

Ardisia ursina is unique within Ardisia subg. A 
riculardisia sect. Auriculardisia because its branc 
lets, leaves, and inflorescence branches have an . 

liale. itiullicclhilar stalks In 1.5 mm long, lli.- apical 
stellate arms often caducous, the hairs then ap- 
pearing glandular-v illiiu-. Udmuidi the branchlets 
and inlloresecnce branches ol \rdisiu dodgt'i, A. 

ilale-slellale trichomes. they are considerably 
smaller, to only 0.4 mm long, 

long, iiicoiispicimi, I! pan i and puiiclate-lin- 
eale. veslilure as in the iiillorescence raehis. Flm,- 

meinbranoiis. narrowly lanceolate. 7.2-8.2 X 2-2.3 
mm. apically attenuate, prominently punctate and 



scattered v illo. 

;E of road, 8 Sep. 1980 (fl). <.. Mrl'h.-ts,,,, 
I 1012 [V. ITC. Ml At . MO. I'M \). San Itlas: Cerm Ol.u. 
25 June I01S0 (fl). (,'. dr \rrrrs rt «/. H0U, (I.I.. MO): 
Nasagandi. Sender.. Wedar. 10 July I Wo (fl). J. \h - 

Dmu^lir, at. 21)2 (MO): Cm. II. ,l,„. l,.,il I, ISmSidi... 

20 Dec. IOi',0 if,-). K. Ssismu <>/ ,,/. 2770 (I.I.. MO). 

Quoad sepala asy mmetrica ad bases auriculala atque 
ramulos furfuraceo-lepidotos ad Anli.siam subg. Aiiricular- 
disiam perlinel. Al> aliis seetionibus subgenei is lulu- di- 
niMi'pliin-. jhiii, pLmla exemplar architecturalem Fager- 
lindii exhibente praeclare distat. 

Small suh.shriihs or lives exhibiting h'u»<-i-| i nelV 
Architectural Model (Halle et al.. 1978), to 7 m tall. 
3 em diam. Trunk, re get at ire shoots, and reproduc- 

Isave.s dimorphic; 

with cupuliform 
live shoot leaves 
glabrous above. 

usually similar to the shoots below; petioles stout. 
obsolete to petiolale; repr. m In, Iiv <• -boot leaves with 
blade- -mill, it In die vegetative shoot leaf blades, 
but usually smaller. Inflorescences terminal or pseu- 
dotermiiial. pendent, pinnatelv lo tripirinatelv pa- 
niculate, pyramidal, usually longer than the leaves, 
the branches terminating with flowers in loosely 
congested corymbs; inflorescence bracts usually 
persistent, loliaoeoiis; inllorescciice branch bracts 
and floral bracts caducous, the floral bracts much 
smaller than the flowers; pedicels slender, terete. 
Hon, is ."j-nieroiis. white, i • in i i ii i i ifill. 

red; calyx lobes essentially free, membranous to 
cbartaceous. ovate to suborbicular. basally auricu- 
late; corolla membranous, the lobes ovate to lan- 
ceolate, conspicuously and often prominently punc- 

lilamenls apieally live, connate basally into an elo- 
bate tube, free from the corolla tube, epimelate. the 
anther- ovoid or narrowly ovoid to lanceoloid or 
linear-laneeoloid. apieally apiculate. cuspidate, su- 
bulate, mucronatc or caudate, basally sagittate or 
cordate, dehiscent by subapical pores, opening into 
wide, longitudinal slits, the connective punctate: 
pistil ulabrou.-. the ovarv oblong, the style slender, 
erect, inconspicuously or conspicuously, rarely 

Distribution. Seven species, including one pop- 
ulation from \eia< in/, \le\ieo. and one population 
in Izabal, Guatemala, then southward from Nica- 

ragua (lim San .liiau) thiouid I ( osta Hi. a lo l!o- 

cas del Torn. Panama, with a disjunct population 
in the Choeo ol Colombia. 1'liev grow between sea 
level and 1500 in in elevation. 

primary, secondary. 
I and pluvial forests 
and in evergreen Liquidambar-Quercus forests. 
Etymology. The name "fagerlindia" is derived 

from the arcliitecluial model exhibited by the mem- 
ber- of this section. 

I.' •';:•: -ubg. \urii ulanlisia sect. Fagerlindia is 
defined bv the species' exhibiting Fagerlind's \r- 
chitectural Model (Halle et al.. 1978), and its ter- 
minal, pendent inflorescences. One of the outstand- 
ing features exhibited by Fagerlind's Model i- that 
the llnwering -boot- bear leaves markedly different 
in shape and si/e from those of the vegetative 

:r. Fagerlindia 

Trunks and shout-, leaf blades, and inllore 
racliises hirtellous-tomentose. the Irielion 
parenlly unicellular. O.ii 1.8 nun long 

..-- 1 \., b 

2b. Indument of lower leaf surface of strictly In 
furaceous-lepidote scales: corolla lobes I..'! 

Leal blades broadly elliptic, 
ward the ha-e: petiole- con- 

corolla lobes pellucid pum 

13. Ardisia bastonalei 

I narrowly elliptic to ob- 
<■< ►lut<- or spatliulate. gradually 

florescence bracts persist 

punctate-lineate; anthers 

cronate. merelv apiculate 

5a. Leaf blades crenate t 

peduncles 0.4^1.7 

calyx lobes 1-1.1 > 

1.2-1.3 mm; anther 
mm long, on apieally 
ments 0.6-0.7 mm 1< 

15. Ardisia gordonii 

. Vegetative shool leaf blades 25.2-12.6 

x 9.4-11.] cm. 

fta. Branohlets 7.5-H.5 mm ilium.: veg- 

pelioles of reproductive sliool 
leaves sessile to 0.3 cm long; inflo- 
rescences 12.4-12.8 cm long; calyx 

ftb. Branchlets 1.5-5.4 mm diam.; veg- 
etative si t leaf blades 25.2-34.ft 

■ ■m I - - 1 1 ;_■ : apicallv abruptly acumi- 
nate: petioles of reproductive shoot 
leaves 0.2-0.7 cm long; inflores- 
cences 9.2-11.5 cm long: calvx 
lobes 1.3-1.5 mm long 

. Ardisia apoda Slandl. & Slt-yerm.. Publ. 
Field Mus. Nat. Hist.. Hot. Ser. 23: 219. 1947. 
Icacorea apoda (Standi. & Steyerm.) Lundell, 
Phytologia 49: 347. 1981. Auriculardisia apo- 
da (Standi. & Steyerm.) Lundell, Wrightia 7: 
266. 1984. TYPE: Guatemala. Izabal: Cerro 
San Gil, 300-900 m. 25 Dee. 1941 (fr), J. Ste- 
yermark 41923 (holotype, F!, F neg. 68130!. 
LLneg. 1971-18!). Figure 15. 
mall trees to 6.1 m tall. Trunk and vegetative 
■its 7.5-8.5 mm diam.. densely furfuraeeous-lep- 
:e; reproductive shoots similar, hut 3-4.2 mm 
n. I saves dimorphic: vegetative shoot leaves 
i the hlades membranous, elliptic to slightly oh- 
•eolate, 39.8-42.6 X 9.4-9.8 cm, apically long 

. inconspicuously punctate and punctate-lin- 
•. glahrous above, sparsely furfuraeeous-lepidote 

slighth crcnulate. flat or revolute: petioles stout 
maiginale. (ilisolele lo 0.3 cm long, furluiaceniis- 
lepidnte: reproductive shoot leave-, Willi ill. hlades 
similar lo the vegetative ones hut 0.3-12.0 X 3.4- 
3.6 cm. the secondary veins 9 lo 21 pan-: petioles 
similar to the vegetative ones. Inflorescences pin- 
nately or bipmnately paniculate. 12.4-12.8 cm 
long, longer than the leaves, the rachis. hranches. 

branches terminating in 7- In 12-llowered corymbs: 
peduncle 3.1-6.5 cm long: inflorescence bracts 
persistent, oblong. I. 1-1.0 X 1.8-5.1 mm. apicallv 
acute to rounded, midrib prominulous above and 
below, prominently punctate and punclale-lineale. 
the margins entire. Hal: inllorescence loan. Ii brads 
similar In tin inllorescence brads, but 1.7-2.2 X 
0.4-0.9 mm: floral bracts unknown, early caducous; 
pedicel 7.4-9.1 mm long, prominently punctate 
and punclale-lineate. Flower color unknown: calyx 
lobes chaitaceous. suborbicular to oblate, 1-1.2 X 
1.1-1.3 mm. apicallv acute, prominently punctate 
and punclate-lineale. sparsely fiirluraceous-lepi- 
e. hyaline, -paiselv 


known. Fruits 7.4-45.5 

punctate and punctate-lineate, slightly eostate. 

Distribution. Ardisia apoda is only known from 
Cerro San Gil. Izabal. Guatemala, growing liom 300 
to 900 m in elevation. 

Ecology and conservation status. Ardisia apoda 
is found on damp, forested slopes and ravines, be- 
cause it is only known from the type collection, it 
should be considered threatened. 

Etymology. The specific epithet refers to the 
sessile leaves. 

Because this species is only known from the 
type, the relationships of \rdisia apoda are uncer- 
tain. However, within Ardisia subg. Auriculardisia 
-eel. Fagcilindia. \nlisia apoda may lie most easily 
■ •■■iilused willi \rdisia lortitgnerensis by virtue of its 
large vegetative shoot leaves to 12.0 X I 1.1 cm. 
gradually tapering to the base, the relatively thick 

I). II,,,, 

i Imstoiudensis. — \. Flowering branch. It. Detail of abaxial leafs 
— E. Stamen, abaxial surface. — K Stamen, lateral margin. —C. Ski 
i drawn from hololvpe. Ii. (.',■,/„'/„ /.' A (,. Higareda 2HW (MKXU).) 

gure 17 (right). Ardisia brenesii. —A. flowering branch. — B. Detail of abaxial leaf su 

( \I(M 1 ' 


13. \rdisia bastonalensu Ricketson X I'ipolv. 
sp. nov. TYPK: Mexico. Veracruz: Mpio. De 
Catemaco. Rancho La Chingada. 10 km al SK 
de Tebanea, camino a Bastonal, 22 Nov. 1984 
(fl), /?. Cedilla T. & G. Higareda 2890 (holo- 
type, MEXU!; isotype. US!). Figure 16. 

[] tall. Trunk and regelatire dioots 
3-7 mm diam., sparsely and minutely furluraceoiis- 
lepidole: reproducing shoots similar, hut 1-2 mm 

3.8-7.6 cm. apicallv aciiminale. with an acumen 
5-9 mm long, basally acute slightly decurrent on 
the petiole, conspicuously punctate and punctate- 
lineale. glabrous above, sparsely lurfuraceous-lep- 
idote below, glahreseent. the midrib impressed 
above. pronurienlK raised below, (he secondary 
veins 40 to 45 pairs, slightly raised above and be- 
low, the margins entire. Hal: petiole slender, mai 
ginate. 12-27 mm long, glabrous above, sparsely 
linliiraceous-lepidole below, glahrescent: reproduc- 
tive shoot leaves with the blades similar to the veg- 
etative ones, but 3.2-0.4 X 2.2-5.1 cm. the sec- 
ondary veins 23 to 20 pairs: petioles similar lo the 
vegetative ones, but 4-12 mm long. Inflorescences 
pinnate, 1.6-4 X 0.8-5 cm. shorter than the leaves, 
the rachis and pedicels nidi densely fnrluraceoiis- 
lepi<lote scales, the blanches termiiialing in I I- jo 

, 1-0 i 

lloieseenee bracts and branch I 

ly caducous): biacl- membranous, oblong. 2- 

2.0 X (). |—().(, in,,,. ,||ne;i||\ acute. illeoilspielloi|s|\ 

punctate and punctate-lineatc. the margins minute- 
ly erose. apieallv hyaline. spar>elv glandular eili- 
>late: pedicels 5.2 I 1.2 mm long, inconspicuously 

lense minute furfuraceoiis-lepidolc scales. Flowers 
rtdute; calyx lobes membranous to chartaceoiis. 
uiborbicular lo widely ovate. 1.4-1.6 X 1.1-1.3 
urn. apicallv acute lo rounded, prominently pinn- 
ate and punclate-lineate. pellucid, glabrous a. la- 

tely erose. hvaline. -par-elv glan- 
dular ciliolale: corolla 5. 1-5.0 mm long, ihe lube 
1.2-1.4 mm long, the lobes narrow Iv ovale lo ovale. 
4-4.2 X 2.2-2.4 mm. apicallv acute, prominently 
punctate and piinclale-liiieale. pellucid, glabrous 
throughout, the man-jus entire, hvaline: stamens 
4.3-4.4 mm long, ihe lilamenls 1.4-1.6 mm long, 
the staminal lube 1.2-1.3 mm long, the apicallv 
free portions 0.2-0.3 mm long, epiinclalc. the an- 
thers lanceoloid. 2.0-3.1 X 1.2-1.6 mm. apicallv 
inucronate-apiciilalc. ba>allv -.ig:i I t;il, . ihe eonncc- 
livc inconspieuoiislv punctate: pistil 5.6-5.9 mm 
long, the ovary 1.6-1.9 nun long, the styles 3.8-4 
mm long, epunctate. the ovules 10 to 12. Fruits 

Distribution. Ardisia bastonalensis is endemic 
to the area around Bastonal. in the Mpio. de Ca- 
lemaco. Veracru/. \le\iro. growing al around 500 
m in elevation. 

Ecology and conservation status. Ardisia bas- 
lonolensis occurs in tall evergreen f.iuuidambai 
Qiiercns lb res Is on clay soils. Because of its re- 
stricted distribution, il should he considered 
lli" '. I 

Etymology. The specific epithet refers to the 
type locality, around Bastonal. Veracruz. Mexico. 

Within Ardisia subg. Auriculardisia seel. Fager- 
lindia. \nlisia bastonalensis is most closelv related 
to 1. tilaranensis anil 1. gordanii bv then smaller 
vegetative shoot leaves less ihau 21.5 cm long and 
7.6 cm wide. However, the elliptic leaves with 
acute bases, conspicuous petioles lo 2.7 cm long in 
ihe vegetative shoot leave- and long peduncles |,, 
9 mm long, ihe pellucid (not black) piuiclate-liii- 

lale anthers clearly set il apart from these oilier two 

Paratype. MLXICt 
Cumbres de Bastonal, 
(MKXU, US). 

1 I, Ardisia brenesii Standi.. I'ubl. Field Mil- 
Nat. Hist.. Bot. Ser. 18: 855. 1938. \uricula, 
disia brenesii (Standi.) Lundell. Phylologia 4* 
342. 1981. TYPE: Costa Rica. Alajuela: Ca 
taratas (bos Angeles) de San Ramon, vicinit 
of San Ramon. 17 Apr. 1935 (fl). 4. Brent 
20537 (holotype, F!, F neg. 68135!, LL nej 
1971-23!: isotypes, NY! |2|). Figure 17. 

i 6: 79. 1979. Syi 

(Lundell) Lundel 
I'hvlologia I'): .ill, \*M\. T\ PL: Cosla Rica. Limor 

Volume 90, Number 2 

15 Mar. 1978 III. ID. (, Davids,,,,. I. kaminer. I, W tne staminal tube 0.3-0.7 nun long, the apical 

eeoloid. 3.1-3.1 X 1. 1-1.2 nun. api- 

55677!; isotypes, F!, F 

Shrubs or small trees 2-7 m tall, 1-3 cm diam. call\ apiculate. basally sagittate. 

7nm/r «m/ vegetative shoots 4—11.5 mm diam.. in- punctate; pistil 2.9-4 mm long, the ovary 1-1.2 mm 

dument with a mixture of dense ferrugineous-lepi- long, the style 1.7-2.8 mm long, inconspicuously 

dote and scattered erect cupuliiorm scales. 0.1 -0.9 punctate, the ovules 12 to 14. Fruits ( 

inin long. Hie scale margins loned or with z 
arms; reproductive shoots similar, hut 1.2-3.' 
diam. Leaves dimorphic; vegetative shoot 1. 
with the blades membranous, elliptic to obla 
late, 22.5-53.7 X 5.4-22.6 cm, apieally a< 
nate, with an acumen 1.3-3.2 cm long, b; 
auriculate, prominently punctate and punct£ 
neate, glabrous above, indument with a mixti 
dense ferrugineous-lepidote and scattered ere< 
puliform scales, 0.1-0.9 mm tall, the scale rm 
lobed or with 2 to 

, prominently punctate and puriclate-liiieate. 

Distribution. Ardisia brenesii occurs from Rio 
San Juan, Nicaragua, through Alajuela, and Here- 
dia to Limon, Costa Rica, with a disjunct popula- 
tion in San Bias, Panama, growing at to 1500 m 
in elevation. This is the first report of the species 
from Panama. 

Ecology and conservation status. Ardisia bre- 
nesii occurs in priinarv. secondary, and remnant 

-,.,,, , in. moil I I"!" 

«o .»* • i- i i j j u ■■ " ■■ ' '■ ' - ; " ' ■>-■' ■'' ''''' ~ :! '"' " " " 

. , . ■ ". ■ ii at a densitv ol less than a dozen . i- 

I, • gularh i re- - . 

nate, flat or revol tile I .ward I l»— ; pel loles stout, u ( ;| .; . § '. , 

marginate. 0.3-0.8 cm long, indumenl similar to 
the vegetative shoots: reproductive shoot leaves 
wilh the I. lade- similar to the vegetative shoots but 
oblong, 4.9-23.9 X 0.7-8.8 mm, the secondary 
veins 3 to 30 pair-: petiole- similar to the vegeta- 
tive ones but 0.1-9.5 cm long. Inflores, ences bipin- 
, 9.2-35.5 cm long. 

While it is < 

time there are no data to suggest that the species 
is threatened. 

Etymology. This species was named in honor 
of Alberto VI. Brenes of the Museo Nacional de 
Costa Rica. 

Ardisia brenesii is unique within Ardisia subg. 
Fagerlindia because 

longer than the leaves, indument as in shoots, the ' ,.,.'.' , -, i it 

uJLu^ .„_;„..,;..„;..-,.. KJ_fl I -v.»lw miMlin ' ol b'rmgineous-lepidote and cupuliiorm 

ered corymb; 

peduncle 3.1-19.4 cm long; inflorescence bract; 
early caducous, oblong. 1.. 5-8.6 X 0.5-2.4 cm. api 
callv acute to rounded. prominentlv pun. tale am 
punclalc-lineale. glabrous abo\e. iiidiimeiil belov 

nd below, tb 

Hat or revolute toward the base; 

ferrugi neou s-lepidote 
present throughout the plant, although it 

belong- to llie section because of the di- 

n-productive shoot-, \rdisia brenesii mav be most 

, n|> easily confused with 4. tortuguerensis because of 

1 and below, the n,ar,,ns entire. "^ '"'^ s.nkmglv d.mnrplnc lea blades to 53.7 

cm long, those ol the vegetative shoots elliptic to 

oblanceolate wilh obvious acumen to 3.2 em long. 

blanch bra. t- -imcu I I! ■-< ence „.^.„, 

but 1.1-3.3 X 0.5-14.9 mm: floral bract- similar and gradually tapering toward the base. However. 

le inflorescence bracts, but 0.8-1.2 X 0.2-0.5 
; pedicels 5.4-10.2 mm long, pellucid punctate 

elalbe shoot-. Fionas pink, purple, or red: calyx 

lobes membranous to charlaceous. ovate to subor- 

bicular, 1.2-1.4 X 0.9-1.2 mm. apieally acute, 

prominent!) punctate and puiiclate-lineate. with 

sparsely ferrugineous-lepidote or cupuliiorm scales. 

the margins minutely erose apieally, hyaline, 

-par-elv glandular-ciliolate; corolla 5-5.3 mm long, 

the tube 0.5-0.7 mm long, the lobes widely ovate, 

4.5-1.8 X 2.2-2.4 mm. apieally acute, prominently Specimens examined. NICARAGUA. Rio San Juan: 

punctate and puiictate-lineatc. sparsely furfura- near |{ fo San Juan at "Kl Mel. is." .a. midpoint between 

ceous-lepidote abaxiallv. the margins entire: sla- Kl (.a-lill,. I I Mia de San Juan. 23 Mar. 1961 (fl), G. 

Ardisia brenesii, with its mixed vestiture. apieiilate- 
subulale anthers, much larger calyx lobes to 1.4 
mm long, and corolla lobe- lurluraeeon-depidol.- 
abaxially, is easily recognized. 

The t\ pe of irdisia Umonensis is identii aJ to pop- 
ulation- of I. brenesii. except it is more robust. It 
is important to note that the MO holotype of A. 
linionensis is made ol two vegetative shoot leaves 
and an inflorescence, and both the F and LL iso- 
types are made up ol reproductive shoot leaves with 
an attached inflorescence. 

or Cafio (I.- Oro. loward Cerro el Oigai He. IT, S< M ». 1. 11 2.0 mm tongas 1.0-1.1 mm lal 

. h. Martinez S. 2 I .><) (\1K\1 I; Km San Juan. Fl tie arch- al'tini-. sed ah ea liahitu Irulicoso I nun suffruti- 

. 14 Sep. |9»2 (IV). R. Riviere 358 (MO); along com.) lol.ulis cabciiiis O.R 1.0 limn 1. 1-1.3) mm lalis. 

Top lions,. Creek „f Km |„,li„. 5 J,,b anllieris I.!! -2.0 Inon 2.9-.H ) ,„,„ longis 1.0-1.1 (nee 

. R. Rueda et at. I7W (III II. MO): liesena 1.2 1.0) run, lahs. donique ,>isii llo 4.2-4.4 (non 5.6-5.9) 

i/. slopes lacing moulli o| Rio San Carlos, on (lie longo facile distinguilur. 
Juan, 11 Feb. 1996 (fl), R. Rueda et at. 4054 

WO); Mpio. San Juan del Vrte. Resena Indio- Small trees to 4 m tall. Trunk and vegetative 

no Kl I dgan.e. 5 k„, del «„, San Juan. 21 Sep. ,,,„„,, 3.4.5 mm dia ,^ (|ww ,., f urfll raceoiis-lepi- 

. A. hue, la el a/. ...... (J | :\> i I 0STA i , ,, , a , ,.., j ■ i 

ajuela: on Caribbean slope between San Lorenzo ,° " llw wa,,,H llaU l ,all,(l ' reproductive shoots sim- 

Angeles de San Ramon, above the |{fo San I ,o- ll; " tu ""' M'^'lalm- shoots, but 1 .5-2.5 mm diam. 

' Sep. 1978 (fr), W. Hurler <t T. \nhmio 111X3 leaves dimorphic; vegetative shoot leaves with the 

mi de San Ramon. I{esci\a I'oieslal >an Raumu. blades ineiiibranoiis. narrow l\ elliptic lo oblnnceo- 

'/'''J/'it'i; i'tc' "|\|; \| " ' \ , i ","'i l;,,< ' '" s l ,a,hllla,< "- ' 1-^-21 .5 X 3.2-5.6 cm, api- 

dwr&E. Cruz hW7 \(\\\. IMC MO. I S); Resena gmdualb tapering to an auriculate, amplexicaul 
!•' San Ramon, ea. 10 kin \\ of Lagilos along base, prominently pimelale and piiuclatedineale 
abo\e ami below, glabrous above. densely lui lui a- 

mnlrib impressed ab..\e. promnieiil l\ raised below, 
the secondary veins 37 lo 45 pairs, promiiiulous 
abo\e and below, the margins entire. Hal; pelioles 
obsolete; rej)roduelive shool leaves with llie blades 
similar to the vegetative shoot leaf blades, but 6. 1- 
8.8 ' I ■> 8.1 'in. In//iiieu , -,,, < .«, bipimiateb panic- 
ulate, l()-22 X 5.5-8 em, longer than the leaves, 
denselv furfuraceous-lepidote. the branches termi- 
nating in 5- to 12-flovvered corymbs; peduncles 
Cliirripocilo and Kin S.udina "sanlinal" on Chirripo •">-">- I 5.5 cm long; inflorescence bracts similar to 
miHu,|i,a ( |, angle. II [\„ v . 1088(11). M. Crayum et al. the vegetative shoot leaf blades, but oblaneeolate 
'' " |- |{ - M,li '""«!'> K.YF.S liana del ,„ ( ,blo,,g 1.8 1 ! 1-0.8 cm; inllorescence 

',,"," " h|| , i l "' u I'".'"', h bia, Is persistent, memb.anous. oblong, 1.5- 

I Car.ari. mi il„- l,„,„ „| Mario Chavama I'.! I.: Sep! 3() X ' • V ~ 2 - ' "»^ apically acute, the veins absent. 
I (II). k. Thnmsen /62/|C. MO). I'\N\M \ San Itlas: prominently punelale and pimclalc-lineale. gla- 
IVom Kio Ksaili lo Cerro Ranega. 21 Dee. 1 9R5 (II. brous above, sparseb fiirluraceoiis-lepulote below. 
C. de Never* & H. Herrera 66-/6 (MO |2|). , h( . .....^j,, milllllr | y „„,,, , lyu|ill ,.. || ()r;| | ,„,„.,, 

similar to the inflorescence branch bracts, but 0.7- 
. Vnlisia iionloiiii Picketson & Pi poly. sp. 2.1 X 0.4-0.8 mm; pedicels 0.6-1.1 cm long. 
nov. TYPE: Panama. Pooas del I oro: above prominently punctate and punclate-liiieale, sparse- 
Chiriqui Grande. 08 55'0|"\. 082 'lO'O l"\\. Iv Flour,, pink: <alvx lobes 
300 m, 26 Dec. I08(. (fl). 6. MrPherson <£ J. membranous, ovate. 1.3-1.5 X 0.8-1 mm. apieally 
Aranda I01M (holotype. MO!; isolype. 1,1.!). a<aite. prominently punctate and pimclate-linealc. 
Figure '**• glabrous adaxiallv, sparsely lurluracooiis-lenidole 

i holotype, C. MrPherson & ./. Aranda HUM (MO); C-(i from C. I 

et al. 15261 (1,1, 

!], WO 



San Ramon, 

n Midge. II 

ino Trail. 20 Apr. 

. /2(CI{. 1. 

le. Kstaeion 

din's. 2 Oct. 1990 

(III. \. i 

/ (CR. INK. 

); Los Angeles d« 

1081 (IV). ./■ 

oly 7113 (CR, N' 



ipiquf, Cuenca del Sarap 

lills. ea. I.", 

N of Puerto Viejo, 

II 1 ... 


15. MO). Limon: 1 
a in the Kio Coloi 

ii lb .-,.., 




oi I :..,,., ,|. 

.ra.lo. L>-I0 Sep. 

1 '),",(, ( 

H. hi 

. (,. 

Dandse & C. Her- 

. Flowering branch. — R. Detail of ahaxial leal surface. 
, stiff, mostly erect modified hairs. —I). Detail of inflorescence. — K. Flower. — F. Stamen, al 
nen, adaxial surface. II. Stamen, lal, r.,1 margin. I lanil. ( \. R drawn from M. Crayum et a 
. C. Rarbasa 650/ (MO); I from /,. I'nreda I. et al. 4174 (¥).) 

long. the tube 1-1.4 nun long, the lobes ovate. 3.2- 
3.4 X 1.9-2 mm. apieallv acute, prominently hla< k 
punctate, glabrous ihrnughoul. llie margin.- enlire. 
hyaline: stamens 2.5-2.8 mm long, the filaments 
1.7-1.8 mm long, the stamina! tube 0.9-1 mm long, 
the apieallv free portions 0.7-0.9 mm long, epiuic- 
tate, the anthers ovoid. 1.8-2 X 1-1.1 mm, apieallv 
apioulate. hasally deeply cordate, the connective 
inconspicuously punelate; pistil 1.2-1,1 mm long. 
the ovary 1-1. 1 mm long, the styles 3.1-3.4 mm 
long. inconspicuously pnncl.ilc. the ovules 8 to II. 
Fruits 7-7.5 mm diam.. inennspiciiouslv pellucid- 

Distribution. Ardisia gordonii is endemic to the 
area above Chiriqui Grande around the area of Cer- 
ro Pila de Arroz, in Bocas del Torn, Panama, grow- 
ing at 300-500 m in elevation. 

Fcology and conservation stains. \rditia g;>v- 
donii occurs on slopes ol premontane wet forest. 

1926(H),/? Standler&J. Valeria ■ 
lotype, US!, LL neg. 1971-73!, US n 
Figure 19. 

Because of its restricted distribution il should be 
considered threatened. 

Ft\niolo»\. Il is an honor to dedicate this spe- 
cies to Cordon MoPherson. a curator at the Mis- 
souri Botanical <. anlm. Gordon i- an indefatigable 
collector, prodigious lloristician. a noted specialist 
in African and Central \merican lloras. and above 
all, a scholar and uonlleiiian. 

Within Ardisia subg. Auriculardisia sect. Fager- 
lindia. \rdisia gordonii appears to be most closely 
related to 4. tilaranensis b\ the narrow l\ elli|itic lo 
oblanceolale or spalulale leaf blades gradually ta- 
pering to an aurieulate base, the very short or ob- 
solete petioles, persistent inflorescence bract, black 
punelate corolla lobes, and apioulate anthers. How- 
ever. Ardisia gordonii can easily be separated from 
1. Id, in, tin, sis bv the enliie leal blades, calyx lobes 
to 1.5 X 1 mm. mm li laiget. bla< k punctate corolla 
lobes to 3.4 X 2 mm. larger anthers to 2 X 1.1 mm 
on longer apieallv free filaments to 0.9 mm long, 
and larger non-coslate fruit lo 7.5 mm in diameter. 

I'umtyfte. PAIN A MA. Boca* del Toro: along mail lo 
Chiriquf Grande. 10 mad mi. Imni ( Oivide and 
'2 mi. along |»i|., lui. ace,-- mad I ,,l Imv.. on (run Pila 
de Arm/. 10 Mar. PM«, (II. Ir). (,. Ucl'lu-rson H7.i; ill. 

. Ardisia nevernianiiii Standi., J. Wash. Acad. 
Sci. 17: 524. 1927. Valenoanthus nevermannii 
(Standi.) Lundell, Wrightia 7: 50. 1982. Auri- 
culardisia nevermannii (Standi.) Lundell, 
Phytologia 57: 450. 1985. TYPE: Costa Rica. 
Lim6n: Finea Monteeristo, on the Rio Reven- 
tazon, below Fl Cairo, ca. 25 m, 18-19 Feb. 

Shrubs or subshrubs I 
Ive shoots 6-9 mm dian 
richomes apparently ui 
lm long; reproductive 

tall. Think and vegeta- 

llular, the hairs 1.4-1.8 
ns similar, but 2.5-4.5 

uilli llie blade- membranous, narrow Iv oblong lo 
oblanceolale, 34.2-42.5 X 10.2-13.9 cm, apieallv 
acuminate, with an acumen 1.3-2.9 cm long, ha- 
sally obtuse or slightly aurieulate. prominently 
punelate and punclale-lmeale. lui tellous-lonientose 
above and below. I he hairs 0.8-1.8 mm long, much 
denser along llie midribs, the midrib impressed 
above, prominently raised below, the secondary 
veins 2(> lo 12 pairs, slightly impressed above, pro- 
minulous below, the margins entire. pdmlcs 
stout, canaliculate, 0.7-1.8 cm long, hirtellous-to- 
mentose above and below, the hairs 1.4-1.8 mm 
long; reproductive shool leaves with llie blades sim- 
ilar to the vegetal ive ones but 5.5-27.5 X 2.7-8.9 
mm, the secondary veins 13 to 31 pairs; petioles 
similar to the vegetative ones but 0.2-1.2 em long. 

Inflorescences pinualelv <n bi| jldv paim ulale. 

1 1-22.5 cm long, longer than (he leaves, hirtellous- 
lomenlose. llie hairs I. I 1.8 mm long, llie branches 
lerininating in .'>- lo 8-llovvered corymb-; pediuude 
4.1-12.7 cm long; inflorescence brad unknown; in- 
florescence branch bracts earlv caducous, oblong. 
0.5-2.6 X 1.1-8.2 mm. apieallv acute to rounded. 

Ilous-tomenlose, the hairs 0.8-1.8 mm long, the 
midrib prominulous above and below, the second- 
ary veins inconspicuous, the margins enlire: floral 
bracts similar lo the inflorescence branch brads 
but 0.4-1.6 X 0.2-0.4 mm. hasally sessile: pedicel 
6. 2-7.6 mm long, eoiispieuiuisb puudale. hirle- 
hairs 1.2-1.8 mm long. Flou- 

eliarlaceoiis. ovale lo siiborbicular. 1-1.2 X 1.1- 
1.4 mm, apieallv aeule. prominently punctate and 
piinelale-luieale. sparsely hirtellous alu\ullv. llie 
hair- 0.3 0.8 mm long, gl.ibion- adavially. the mar- 
gins irregular, minutely erose. hyaline, sparsely 
glandulai. iholate: corolla 4.3-1,6 mm long, the 
tube 0.3-0.7 mm long, the lobes widely ovate, 3.9- 
4.3 X 2.1-2.3 mm. apieallv acute, prominently 
punctate and pmiclate-liiieale. glabrous lliroiighoul. 
Il margins enlire: slamens 3.1-3.3 mm long, the 
lilamenls 0.9-1 mm long, (he staniinal tube 0.3- 
0.7 mm long, the apieallv live portions 0.3-0.6 mm 
long, the anthers narrow l\ ovoid to lain, ol,, id. 2. I 
2.7 X 0.9-1.1 mm, apically caudate, hasally sag- 

ittate. the connective punctate dorsally: pistil 
3.3 mm long, the ovary 0.5-0.6 mm long, the 
2.5-2.7 mm long, inconspicuously punctate 
ovules 7 to 9. Fruits 7.2-7.9 mm diam., p. 
nently punctate and punctate-lineate, costate. 

Distribution. ArdL 
Jose and Limon, Cost 
lomhiu. Il is not known from Panan 
he expected in the Darien, growing 

ind the Choeo of Co- 
Panama, hut should 

Ftymohtfu. Slaudley |l<>27: 521) stated llial 
"the species is named for Mr. Ferdinand \ever- 
maim. a keen sludcnt of Costa Rican Coleoptcra 

Ardisia nevermannii is unique among all thi 

iiit'iiilicr- of \rtlisin -iibg. \iirn ulardisia sect. Fa 
gerlindia hecause of its hirtellous-tomentose vesti 
lure ihroimhoul the planl and llie caudate a[)ices o 

mined. COSTA RICA. 

(,. (Aijodonl, 6tf.5 (W): Zona Protectors Barhilla. W side 
ill plateau M-paratin:: headwater- of \ fork of Rio Dantas 
from lieadualer- ol Ouelnada \arreal. Rio Karl.illn drain- 
age. SK of I I Jan. I9H7 (II). M. Gmyum el at. 
7<L>:> 111. \1(). TK\); Cerro Muelulla. Kila Matama. ( !or- 
,|,||,, a de lalaman, a. (. \pr. !««!«> (frl. A', fiuhles X 1. 
CI,,,,,-,,, !'/„'!.; KJflU,. I\li. \10|:|{ioReve..l : ./ 

Ft I 'J If): flora \lonleen-ln. on llie Rio Revellla/oii. be 

low Cairo. 05-16 Kel». 1926 (II). P. Standby & J. Valeria 
//;//;/ It Si: Hamburg tinea, on llie Rio Rev enla/on. he 
low Cairo. 16 Kek |02h ill). I> Standi^ <v ./. l«A-/7» 
ir.-.ll |1 Si. //;/;i'/ ll N. SanJ.w': Carrillo Station, liran- 
lio Carrillo. 16 \,.r. I'WJl ilrl. /. Cnmez el at. 2/ /.7" 1 1 I .. 
\lOi: l'ar,|iie Naeional IJ.aulio « .arrillo. Kslaeibn Car, illo. 
27 Nov. I<)!i(, (IV). /, /\»m/« 1. el at. 1171 1 1 I: l.agnu.i 
Ho Ca- 
rci 56/ 

(MO). COLOMBIA. Choeo: Vlpio. de \uqiii. Corregi- 
inienli. Te, males. \ of Ouebrada 1'iedra. 1'iedra. !*. Sep. 
1661 |fr). /? -Wifdn-Rdfiz. el at. h,H75 (Fl'l,. I S): had on 
Morn, de \1iro to the seeme lookout. along " ." 

„/ , , \i " i'!«>n ,ii. i,, < [•;»'• • ' k n- - 

FTC |2|. \10|2|): \\\ ol Alio Curiehe. 26 Max l<»67ilrl. 
J. Duke & J. Idrobo 11248 (1.1,1: Mpi... I'.alna do -olano. 

I'm, la Hi,-,, and Caala (.ocalilo. Hi \pr. 1 660 (fr). ./. 
Ks,,inu el at. 3639 (CHOCO, MO); Wpio. Balua .le 
I'anjiie Naeional Natural Knsenaila de f Iria. trail between 


) III I. ./. r.s/emi 

(Standi.) Lundell, Phytologia 49: 345. 1981. 
TYPE: Costa Rica. Guanacaste: Quebrada Se- 
rena, SE of TilanSn, ca. 700 m, 27 Jan. 1926 
(fr), P. Standley & J. Valeria 46169 (holotype, 
US!, US neg. 2389!, LL neg. 71-1 15!). Figure 

Subshrubs to 3 m tall. Trunk and vegetative shoots 
2.5-5 mm diam.. il. -ns. Iv nppres>ed iiiiluraee.m-- 
lcpidote, reproductive shoots similar to the vegeta- 
tive shoots, hut 1.5-3.5 mm diam. h-ures dimor- 
phic: vegetative shoot leaves with the blades 
membranous, narrowly elliptic. 7.5-17.2 X 1.9-4.2 
cm. apically attcnuale I., a. auiiinale. with an acu- 
men 0.7-1.4 cm long, gradually tapering to an au- 
ri( ulale base, prominently puiielalc ami | nmolale- 
lineale. essentially glabrous above, sparsely 
appre-M'd luil'iiraceous-lepidole below, the midrib 
impressed above, proniinently raised below, the 
secondary veins 21 to 31 pairs, slightly depressed 
above, prominulous below, the margin crenale to 
dentate, flat: petioles -hml. marginal, subobsolele 
to 2.5 mm long, glabrous above. < b r i~<l \ lurfiua- 
ceous-lepidole below: reproductive shoot leaves 
with llie blades similar to the vegetative ones, but 
3.5-14.7 X 1.1-4.3 cm. Inflorescences bipinnately 
to tripinnately paniculate, 5.4-9.2 X 2.-5-5.6 cm. 
longer than the leaves, the rachis and pedicels 
densely furfuraceoiis-lepidolo. llie branches termi- 
nating in 7- lo 13-flowered corymbs: peduncles 
0.4-1.7 cm long: inflorescence bracts persistent, 
membranous, ovate to lanceolate, 0.4-3.7 X re- 
productive shoot leaf blades: inflorescence branch 
brail- caducou-. membranous, ovate. 1.5-3.2 X 
1.4-2.6 mm. apicallv acule. basallv auriculate. 
promineiilly pimcljli and punctate-lineate. essen- 
tiallv -labt)iii> abov- . -par-cly furfuraeeoii--|epi- 
dole below, the midrib inipi.-- -ed above, slightlv 
prominent below: pedicels 4.5-6.2 mm long, prom- 
inenllv punel.ite and punclate-liiieale. densely fur- 
luraceous-lepidote. Flmvers pink: calyx lobes mem- 
branous lo chartaceous, ovate. 1-1.1 X 0.6-0.7 
mm. apicallv acute. prominently punctate, sparselv 
furfuraceoiis-lepidoli'. the margins minutely erose. 
hyaline, sparselv glandular-ciliolate; corolla 2.8- 
2.9 mm long, the tube 1-1.1 mm long, the lobes 
ovate. 1.6-1.9 X 1.2-1.3 mm, apically acute, 
prominently black puiietale and punclale-liueale. 
glabrous throughout. lh«' margins entire: stamens 
2.1-2.2 mm long, the filaments 1-1.1 mm long, the 
staminal tube 0.4-0.0 mm long, the apicallv Tree 
portions 0.0 0.7 nun long, epunetale. the anthers 
ovoid, 1.2-1.3 X 0.4-0.5 mm. apicallv api< ailat. •. 
basallv cordate, the connective inconspicuously 

nun Ion-;, the style 1.8-2.1 mm long, prominently 
.iiiil conspicuously punctate, llu- ovules 10 to 12. 
Fruits <)..") iiini diam.. prominently punctate, 
slightly costate. 

Distribution. Ardisia tilaranensis is endemic to 

ami Alajuela. Costa Riea. growing from 600 to 
1400 m in elevation. 

Ecology and conservation status. Ardisia tila- 
ranensis occurs in preniontane humid forests. We 
believe that the protected status of the areas where 
this species has been recorded is enough to prole. I 
it for the time being. 

Etymology. The specific epithet referred to the 
type locality, along the Cordillera de Tilaran in 
Canton Ciianacasle. Costa Rica. 

By virtue of its subsessile leaf blades gradually 
tapering to an aurioulate base and apiciilate an- 
thers, Ardisia tilaranensis is most closely related 
to \. "-on/onii w illiin \nlisia >ubg. \uricidardisia 
sect. Fagerlindia. However. \rdisia tilaranensis 
can be easily separated from 1. gordonii by il> 
crenale io denlale leaf blade margins, the smaller 
calyx to 1.1 mm Ion- d\u\ corolla lobes to 1.9 mm 
long, shorter, apically tree portions of the fila- 
ments Io 0.7 mm long, and the smaller Iruits to 
6.5 mm in diameter. 

Specimens examined. COSTA KICA. Alnjuela: l!es,i- 
\a i'.inldmca Mool.v.i.l, . I5(,s.|iie Flevno de los Psitlos. 
Ouehrada UuaCala. 27, Jan. I <>'H) I IV). K. Hello C. It'Jl) 
«:i{. FTC. I\H. MO): forest trail Iron. Maradamu \ill.i..- 
Io summit of Cevm Chato. :H)et. |W| (IV). I. Funk el al. 
UlOIll it Si: I.Vsena Mn.ilewnle. I Wo Sol. 1.5 kin S I ...- 

;' :i{. IMS. MO): Canton Upala, Bijagua, Kl Hrlim. slope. 
nfCemi Monle/.uma. 2.'> Jub l<>".'. ill). C. Ilerrera 6.7 1 1 
(CB, F): S slope of Cerm Chain. 2.. Feb I W> (IV). c. 

-una de \renal. I.'! .|ul\ l'»<>2 ill). C. Hroien ITUn\ (K 
I -I l: Ou.-I.rada Cramle. Tilaran. Fsperan, Nuhes 
Trail, on Continental Divide. 21 Feb. 1 0<<!7 (II. IV). 11. //,/- 

cional Tortuguero, Lomas de Sierpe, 4 km NE 
of gate at Parque Nacional, along the Rio Sier- 
pe, 10°24'N, 083°33'W. 100 m. 15 Aug. 1988 
(fl. fr), R. Robles, G. Herrera, L Flores & M. 
Rojas 2052 (holotype, MO!; isotype, CR!). Fig- 
ure 21. 

longos cum ad A. apoda 
abea ramulis 4.5-5.4 
ais foliaribus 25.2-34.6 

12.1 12.!!) cm Inngis, .leni(]ue lohnlis calyciiiis 1.3-1.5 

Small shrubs, height unknown. Trunk and vege- 
tative shoots 1.5-5.4 mm diam., dense lurfura- 
eeoiis-lepidnte to short cupuliform scales, the 
scales sessile or slightly stalked. 0.1-0.2 mm tall, 
lobed or with 2 to 8 arms, very similar to those of 
I. hrencsii bill of one size inslead of I wo: reproduc- 
tive shoots similar, but l..>-2.3 mm diam. Ira res 
dimorphic: vegetative shoot leaves with the blades 
membranous, obla.iceolale, 25.2-31,6 X 9.6-11.1 
cm. apically abruptly acuminate, with an acumen 
0.3-0.5 cm long. ba>allv graduallv lapering Io an 
am mmI. ile base, prominently punctate and pune- 
tate-lineate. furfuraceous-lepidote, (he scales 
sparse above, denser below, much denser basallv 
below and along I In midrib, the midrib impressed 
above, prominently raised below, (he secondary 
veins 28 to 37 pan-, slightb depressed above, pro- 

pclioles sloul. niaiguiale. subobsolele to 0.3 cm 
long, detiselv furfuraceous-lepidole: reproduclive 
shoot leaves with the blades similar to the vegeta- 
tive ones but 10.3-19.4 X 4-7.4 cm. the secondary 
veins 25 to 32 pairs; petioles similar to the vege- 
tative ones but 0.2-0.7 cm long. Inflorescences bi- 
pinnately or Iripumatelv paniculate. 9.2-11.5 X 
6.1-8.5 cm. longer than the leaves, indument as in 
the vegetative shoots, the branches terminating in 
5- to 9-flowered corymbs: peduncles 0.5-1.5 cm 
long; iullorex-eiiee brads and branch brads early 
caducous, unknown; Moral brads charlaceous. ob- 
long. 1.4-1.7 X 0.5-0.6 mm. apically acute, prom- 

inllorescence. — I), Flouei 

Figure 21 (right). Ardi 
inflorescence. — D. Flowei 
— H. Fruil. (A-H drawn fi 

i W. Haher & E. Hello C. (ulO (MO).) 

tortuguerensis. — A. Flnueriim braucli. -I>. Detail of abaxial leal s 
— F. Stamen, ahaxial surface. — F. Stamen, adaxial surface. — 0. St 

aboye. densely furfi 1 1 ;ii «-< n I -- 1 « ■ | > i . 1 1 M t ■ Ittl. .u. midrib 
mid secondary veins inconspicuous, the margins 
entile, hyaline; pedicels I..UI nun Ion*;, ennspic- 

shoots. Elowers light purple: calyx lohes membra- 
nous to chartaceous. ovate to subnrbicular. 1.3-1.5 
X 1.1-1.3 nun. apicallv acute or rounded, promi- 
nently punctate and punclatedineate. sparsely I'ur- 
luia<-eoiis lepidote. the margin- entire. minutely 
erose. hyaline, sparseh glandular-ciliolalc: corolla 
3.1-3.4 nun long, the tuhe 0.5-0.0 mm long, the 
lohes ovate. 2.0-2.8 X 1 .0-1.0 nun. apicalh acute. 

hrous adaxialh. I'urfiiraceoiis-lepidote abaxialh. the 
margin entire; stamens 2.9-3 nun long, the fila- 
ments 1.1-1.3 mm long, the staminal tuhe 0.0-0.8 
mm long, the apicallv live portions 0.5-0.7 mm 
long, epunctate. the anther- narrow Is os < >■< I to lan- 
ceoloid. 1.9-2.1 X 0.7-0.8 mm. apicallv upiculate- 
( uspidate. hasalls sagittate, the connective punc- 
tate; pistil 3.2-1 nun long, glabrous, the ovary 1- 
1.2 mm long, the style 2-2.4 mm long. 

ipieihus suhmcises ad liases aun< ulal< 
,ubg. Auriciilanlisiani |>ertinet. Ah aliis s 

(sersus lireyes) insidenlihus [lerfacile rngnosi itnr. 

Shrubs or small trees. Branchlets straight or fit 
uous. slender to stout, terete, siihlerelc. with fi 
longitudinal ridges, or angulate, with densely fi 
furaceous-lepidote and/or cupuliform scales, ran 
stipitate-stellate tomentellous (A. liesneri), the v« 
titure mostly persistent, hut at times glahrvsce 
Leaves monomorphic. with blades membranous 
coriaceous, elliptic to oblong or obovale to obla 

ceolale. al tunes UK «»n>pic i>lv pun, tale and; 

punctate-lineate. the margins flat or revolule. Infi 

erect, pinnately to tripinnately paniculate, pyran 
dal. or obpyramidal. mostly shorter than the leave 
the branches loosely to tightly congested into c< 
ymhs; peduncle short to obsolete, inflorescen 
bracts usually early caducous, inflorescence brain 
bracts and floral brads often -mall and carls c 

(conspicuously punctate 

, the ovules 8 to 10. Im- 
■ haiu.. prominently pun. 
, appearing non-costate. 

Distribution. Ardisia 

lly from the type in I.ii 

lortuguerensis is known 
non. Costa Mica, grow ing 

•alvx lobes es- 

Evology and conservation status. Ardisia lortu- 
guerensis occurs on fairly sleep, densely Ibresled 
slopes with well-drained soils. because it is only 
known Irom the type, it should be considered 

Etymology. The specific epithet refers to the 
area in which it is found. Par<|ue Naeional Toitu- 
guero. along the east coast of Costa Pica in Limon. 

Within Ardisia suhg. ■\uriculardisia sect. Eager- 
lindia. \nlisia lortnguerensis most closely resem- 
bles. 1. apoda, because of the large \egelative shoot 
leaves 25.2-42.6 X 9.4-1 1.1 cm. Ardisia lortngue- 
rensis is easily separated Irom I. a pod a by the nar- 
rower branchlets to only 5. 1 mm in diameter, small- 

abruptk acuminate apices, shorter petioles of re- 
inflorescences to 11.5 cm long, and smaller calyx 

Taxonomic Treatment of Ardisia 

to short. Flowers 5- or 
sentially free, membrar 

metric, widely ovale, oblate to ovale, subapicall 
notched, basallv aurietilate: corolla menibraiious U 
coriaceous, the lobes narrowly ovale, lanceolate, o 
oblong, inconspicuously |o conspicuously or prom 
inently punctate and/or piinclale-lineate: stamen 

bate tube, free from the corolla tube, epunctate 
glahrou-. the anthers oyoid to lanceoloid. ha-alb 
lobate or cordate, dehiscent by longitudinal slits 

oyoid. the style slender, erect, epunclate or pun 

laic ami .a punclalc-liueale. the osules pluriseriat 
Eruils globose to depressed globose, ilicollspici 

Distribution. Forty-seven species from bYli/i 
and (aialemala. southward ihrough Mesoamerica I. 
the Choco Floristic Province of Panama. Colombia 
and western Fcuador. from sea level to 22011 „, ,, 

\r.lisia -nbg. \iiiiculai(lisi ; , -eel. Palmaiia 
Rieketson & Pipoly, sect. nov. TYPE here de: 
ignated: Ardisia palmana Donn. Sin. 

rarely in tall, lowland wet forest. 

Ardisia subg. Auriculardisia sect. Palmanae 

defined by the terminal, sessile to subsessile inff 
<<■ ol phs logciiclic sludie 

■ if this character 

lonophyly o 

2a. Infloreseenct 

5b. Calyx lobes 1.4-1.7 mm wide: corolla lobes ovate, apicalh acute. 2.3-2.5 mm wide; stamer 
5.8^6 mm loii": anthers 2-2.2 X 0.8-1.1 mm: -hies 1.5-5.2 nun 15. \nli*i,i lu,L 
Abaxial leaf surface with < > r i I \ one lype of scale, either of densely cnpulilorni or liirfuraceoiis-lepidole scale 
or occasionally with glandular stipitale-stellate hairs. 

7a. Petioles slender. 1-3 mm diam. 

8a. Calvx lobes as wide as lon» or wider than long, deeply notched below the apex. 
9a. Pedicels 4.5 mm or longer. 

long: styles 1.1-1,7 mm long: plants of Costa Itica 

18a. CaKx lol„- 3.5 3.7 ,„„, wide: corolla lobes 5.3 :,.(. mm g: anther- 3.0 

3.8 m... long; styles 6.9-7.1 mm long 17. |„/m a ftojfwi 

18!.. CaKx 1.8 3 mm whIc: corolla Initio 1.5 5.2 nun Ion- anil,.-,. 2.2 
.'..2 mm Ion- styles 5-5.7 mill long. 

19a. Leaf blades 5.2- 1 2.6 X 1 . 1-3.2 cm: |>««li.«ls 0-1 .2 mm long; rorolla 
lobes 1.8-2 mm w.d, anthers 2.2-2.1 X 0.9-1 „„„ 

26. \rdisia hlrpharod, 
191). Leaf blades 12.6-31.8 X 1 6.7 em: pedicels 1.8-3.5 mm long: co- 
rolla lobes 2.3-3 nun wide; anthers 2.6-3.2 X 1. 1-1.5 mm. 
20a. Trees 3-25 in tall: blades inconspicuously punctate and pune- 
lale-lineate on the upper -ml. ice: Mower- white or light yellow: 
ealy \ lobes coriaceous. 2.2-3 mm w ide: anthers 2.6-2.6 X 1 .3- 
1.3 mm ...... 42. Ardisia fimbrillifer 

lineal.' on the uppe. surface: Mowers while |o light pink: caKx 
lobes chartaceous. 1.8-2 mm wide: anthers 3.1-3.2 X 1.1-1.2 
inn. 57. Ardisia pscudoracemijlar 

CaKx lobes longer than wide, rarch deeply notched below the apex. 

21a. Calyx lobes 3-5 mm long. 

22a. Calyx lobes 1.8-5 X 3.9-1.1 mm; styles 5.9-6,1 mm long 37. \rdisia da 

22b. Calyx lobes 3-3.7 X 2.6-3.1 mm; styles 5-5.6 mm long. 

23a. Leal blades 16.1-32.8 X 1. 1-9.8 cm: 8.2-32. 1 X 4.8- 
•'1.3 cm: Mower- ino-ilv 5-merou-. rareK (i-merou-: caKx lobe- 3 
3. 1 X 2.0- 3.2 mm: corolla lobes 5.8-6 • 2.8 3.2 mm: anthers 3.2 
3.5 X 1.3-1.5 mm; shies 5-5.2 mm long 27. \rdis,a cupilrliah 

23b. Leaf blades 1,8-11.5 X 1.7-1.2 cm; 4.8-7.2 X 3-1,5 
cm; Mowers all 5-merous: calyx lobes 3.1-3.7 X 3.2-3.1 mm: corolla 

5.4-5.6 mm long '.. 44. Ardisia 'zwnilnisi 

21b. CaKx lobes 1.7-2.9 mm long. 

Leaf blades basallv obti 
25a. Calyx lobes 2.7-2 



25b. Calyx lobes 1.7-2,1 X 1.2-2 mm; fruits 5-9 mn 

-2.8 mm long; corolla 1. 

.bes 5.4-5.5 

mm long; anthers 3 

CaKx lobes 2.6 

styles 5.4-5.6 n 

Calyx lobes 2.1-2.6 mm long; corolla lobes 4-1.5 mm long: anthers 2.3-3 „ 

2-5.1 mm long. 

27a. Leal blades longer than 51 cm long: l.ranchlets 15-20 mm diam.; pe 

cm long: corolla lobes 1.5-1.7 mm wide __. 19. 

27b. Leaf blades shorter than 51 cm long: bran, hid- 5 10.5 mm diam.: pe. 

-cssile to A.7, cm long: corolla lobe- 1.6 2.0 mm wide. 

28a. CaKx lobes ovate. 2.5-2.6 X 1 .8- 1 .9 mm: anthers 2.3-2.6 X 1.3- 
2-2.2 mm long; fruits 8-9.8 mm diam. J 

1.1 nun: styles 4.6-5.1 mm long; fruits 1,5-6 mm diam. 

29a. Leal blades coriaceous, mloiis tin liu.i. above a 
ioles 1-5 mm diam.: caKv lobes 2.8-3.1 mn, wide: corolla 
X 2.2-2.3 mm: styles .5-5.1 mm long 3 

29b. Leaf blades me. .1 in , , „■ ,,|„,\e and below: pell 

28 cm; calvx lobes 3.3-3.0 mm wide 15. \rdisia [>i 

s 1-6 mm diam.: leaf blades 2-10.0 x 0.0-5.7 cm: inllorescences 3-24 X 2- 0.7-3.2 mm wide. 

Calyx lobes 1-1.4 mm long. 

34a. Branchlets with interpetiolar lidge- loniini^ up In 5 angles. 3-5 mm dian 
calyx lobes 1.2-1.1 mm long: corolla 2..°. -3 X 1.0-1.7 mm; anthe 
0.7-0.0 mm wide; styles 3-3.1 mm long; IVuils 3-5 mm diam. 

- - 01. Ardisia t, 

34b. Branchlets terete or angled, hut without interpetiolar ridges. l-3(-3.5) m 
diam.: calvx lobes 0.0 1.2 mm long: corolla lobes 2.3-2.0 X 1.2-1.5 mi 
anthers 0.5-0.7 mm wide: styles 1.3-2.0 mm long: Iruils 1.3-8 mm dial 
35a. Branchlets angled, sparsely anil mimil.K femigineou- fin furaec... 

1.1 mm: anthers 1.8-1 .') mm long: -hie, 1.3-1.5 mm long: fruits 

cupuliform and furfuraceous-lepidolo; ca- 
de: corolla 2.3-2.1 X 1.4-1.5 mm: 
.tyles 2.0-2.0 mm long: fruits 1.3-1.7 mm 
03. \,<l,sia I. 

30a. Branehlcl- 3 mm diam.: pelmle- 1.1-2. I mm long: calvx 1.0-1.0 

mm wide 30. Ardisia Colorado, 

301.. Branchlets 1-3 mm diam.: petioles 4-7 mm long; calyx lobes 0.8-1 mm 

37a. Branchlets 2-3 mm diam.; pedicels 3.0^1.8 mm long: corolla lobes 
2.1-3.1 X 1.7-1.8 mm: anthers 1.0-1.7 X 0.0-0.7 mm: styles 3.1- 
3.3 mm long .--- 41. Ardisia eucune 

37b. Branchlets 1-2 mm diam.: pedicels ,.2 I0.3 mm long: corolla lol.e- 
1.7-1.8 X 1.2 1.3 mm: anlbers 1.8- 1.0 X 0.8-0.0 mm: styles 1.0- 

32b. Corolla lobes 1.0-3.3 mm wide: anther- 2.2 1.2 mi 

38a. Branchlets 1-3.5 mm diam.: pedicels 12-15 . 

38b. Branchlets (2-) 1-5 mm diam.; pedicels 4-9 n 

39a. Leaf blades coriaceous; pedicels 4-5 i 

oblate. 2.8-3.2 mm wide, apieallv rounc 

X 2.3-3.3 mm: anthers 3.2-1.2 X 1.2-1,1 mm: styles 4.9-5 mm long: 

fruits 8-9.8 mm diam. 40. Ardisia dinrri 

39b. Leaf blades membranous; pedicel- 0-0 mm long: calyx lobes membranous 
to charlaccous. 1.1-1.2 mm wide, apie.ilb acute: corolla lobes membra- 
nous. 3.3-3.4 X 1.0-2 mm: anthers 2.3-2.1 X 0.8-0.0 mm; styles 3.4- 

Pedicels 2.5 mm long or shorter. 

40a. Calvx lobes as long as or much longer than wide. 
41a. Leaf blades more than 18 cm long. 

42a. Leaf blade- 2 1. 5 51 <).2 I 5. I cm: . ,iK x lobe- 1 . 1-1 .0 X 1 .3- L5 mm: corolla 

49. Ardisia knappii 

12b. Leaf blade- 18,5-20 < 0.7 7.9e,„:oabx lobes 1.2-1.3 X 0.7-0.8 mm; corolla 
lobes 1.7-1.8 X 1-1.2 mm: anthers 1.1-1.2 X 0.0-0.7 mm: styles 1.2-1.4 mm 

long 58. Ardisia indrrrnlriiia 

lib. Leaf blades less than 15 cm long. 

43a. Branchlets 3-5 mm diam.: cabx lobes 0.0-1 mm wide: corolla lobes 3-3.2 X 
2-2.2 mm: anthers 1.7-1.8 mm long: styles 3.1-3.2 mm long 

20. Ardisia albisrpala 

43b. Branchlets 5-8 mm diam.; calyx lobes 1.1-1.0 mm wide; corolla lobes 2.9-3 X 

1.0-1.9 mm: anthers 2.2-2.1 mm long: styles 1.9-2 mm long 04. Ardisia lysonii 

40b. Calyx lobes wider than long. 
44a. Leaf blades coriaceous. 

45a. Branchlets horizontal!) checking and exlnhaling: leaf blades 23-23.5 X 5-5.0 

to oblate. 0.8-1 X 0.0-1. 1 mm 00. Ardisia smurfilana 

45b. Branchlets smooth: leaf blades 1 1.3-1 1.9 X 2.9-1.4 cm: petioles slender, can- 
aliculate. 3-0 mm long: calyx lobes charlaccous. ovale. 1.3-1.0 X 1.3-1.8 mm 

44b. Leaf blades r 

For illustration 

Pipolv (1995:419, fig. 1). 

Small shrubs <■; 

i. 3 .. 

i lall. Hraiichlcls stout, terete. 

15-20 mm diam. 

. dm 

seK appressed 

rulous 1 ii rf ii- 



e.s with blade? 

, chartaceous. 

>blanceolate, 51.5-67 

X 12-1 7.5 mi. 

. apiealb aru- 

ninate, with an i 

■n 0.5-1.2 cm 

louii. basallv 

iuneate, decurrei 

.1 on 

the petiole, im 


I .a. Calv\ lobes I. 1-1.5 nun long: corolla lohes 1. 1- 1. 1 mm long: anthers 2.8- 

2.9 mm long; styles 5.3-5.4 mm long 21. Anlisia anchicavuna 

Mi. Calvv lobes 1.1-1,1 .mm long; corolla 2.7-3.7 mm long: anthers 1.9- 

2.5 mm long; shies 3.5-3.0 mm long 55. Ardisia palmana 

19. Ardfeia aguirrt'uiiu Pipolv. Caldasia 17: 119. eate. glal tl ugl U tl i i r r tire, hya- 

1995. TYPK: Colombia. Choeo: area of Baud... line; stamen 1,9-5 mm long: the filament 2.9-3 mm 

on left bank of river Baudo. about 1.5 km up- long, the skiminal lube 1.1-1,1 mm long, the api- 

stream of estuary, practically opposite the K- rally free portions 1.6-1.8 mm long, the anthers 

most houses of Puerto Pi/arm. about 20 m in- laneeoloid. 2.1-2.6 X 0.8-1 mm. apicalK broad I \ 

land from the shoreline of the bay in front of apimlate. basallv lobate. the eonneclive conspiou- 

Kstero del medio, ca. 5 m above shoreline. |0- ouslv punctate: pistil 1.8-5 mm long, glabrous, the 

100 m|, I 1 Keb. 1967 (fl. ft), //. Fuchs & L ovary ovoid. 1-1.1 mm long, styles 3.7-4 mm long, 

Zane/la (holotype. COL!; isotypes. F!. conspieuously punctate, the ovules 24 to 30. Fruits 

US! |2|). (immature) globose. 5 7 mm diam.. conspicuously 
punctate and punctate-lineale. inconspicuously 

Distribution. Anlisia aguirreana is found in 
Choeo in the area of Baudo and the Paripie Nacio- 
ual Natural de I liia. Colombia, growing from 5 to 
100 m in elevation. 

Ecology and conserralion status. Ardisia agui- 
punetate and punctate-lineale. glabrous above. rrcana occurs in secondary vegetation on relatively 
deiiseb liirluracroiisdepidote below, the midrib im- wel. sliglith swampv ground. Because of it-, re- 
pressed above, proniineiitlv raised below, the see- stricted distribution, it should be considered threat- 
ondary veins 75 to B5 pairs, nilid above, promi- cned. 

nently raised below, the margins entire, revolute: Etymology This species was named in honor 

petiole stout, 3-5.5 cm long. 5-8 mm diam.. gla- of Jaime \guirrc. e\-direetor of the Institute) de 
brous above, furfural -eoii>-h pidole below. In/lores- Ciencias Naturales. Museo de Historia Natural. 
cenccs erect, tripinnalelv paniculate. 12-22 X 6- I'niversidad Nacional de Colombia. 
12 cm, pyramidal, shorter than the leaves, the Common \ame. Arrayan (Fuchs it Zanella 

rachis. branches, ahaxial bract surfaces, and ped- 21851). 

icels furfuraceous-lepidote. the branches loosely Within \rdisia subg. \uriculardisia sect. Pal- 

congested into 6- to I I -flowered corymbs; peduii- manac. Ardisia agiiirrcana is closely related to 4. 
c-les 4-6.3 cm long: inflorescence bracts unknown: megistophyllu. A. cogolloi. A. crassipes. and 4. car- 
inllorescence branch bracts caducous, membra- tagoana because of its long calyx lobes with thick 
nous, ovale to lanceolate. 1.5-1,2 < 0.6-1.5 mm. petiole-,, \rdisia u»uiiTeanu i- separated Irom 1. 
ipically acute, coiispicuously punctale and punc- megistophyllu bv its shorter cab \ lobes to 2.6 mm 
:ate-lineate. glabrous above, furfuraceous-lepidote long, corolla lobes to 1.1 mm long, anthers to 2.6 
aelow, the margins entire, sparse glandular ciho- mm long and styles to I mm long. Ardisia agui- 
late; floral bracts similar to the inflorescence rreana is easily distinguished from the other related 
>ranch bracts, but 2-2.5 X 0.2-0.3 mm; pedicels taxa by its longer leaf blades to 67 cm long, larger 
•itout. subobsolele to 2.5 mm long. conspicuously bianrhlel- In 20 mm in diameter, longer peduncles 
Hinctate and punrlale-lineale. Iinliiraceous-lepi- to 6.3 cm long, and narrower corolla lobes to 1.7 
lole. Flowers 5-merous. light violet gray; calyx mm wide, 
ohes chartaceous. ovale. 24-2.6 X 1.2-1.4 mm. Specunens examined. COI.OMMA. Choeo: Mpio. ,le 

ipically acute, conspicuously pum-fnlc and punc- O m M,,. !'.,,, ,,„• \a,„,nal \. al <lr I In'a. along Ouebra- 

ale-bneate. glabrous adaxiallv. furliira. a-ous-lepi- da I .a \guada. 21 \pr. I<><><) (III. ./. E.spimt ,-/ „/. .»T5/ 

sparse glandula. ciliolale: corolla chartaceous. 6- ^ ,uuh ,l ;' llu ' '"' "ul.I...,, hi o ,1, la- ball, na. 

., I ,l . l l n •-> o I iii - June ' ° ! " ' > ' 

).2 mm long, the lube 1.9 2.2. mm long, the lobes (;o |. T(; V]()| 

larrowly lanceolate. 1—1.1 X 1.5-1.7 mm. apically 

icule. conspicuously punctate and puiictate-lin- 20. Ardisia alhiscpala (Lundell) Pipoly & Bic- 

Volume 90, Number 2 

ketson. Sida 18: 511. 1998. \uriculardisia ul 
bisepala Lundell. Wrightia 7: 266. 1984. Ar- 
disia albisepala (Lundell) Lundell, Phytologia 
61: 62. 1986. num. inval. TYPE: Panama. Ve- 
raguas: trail on ridge to summit of Cerro Tute. 
Cordillera Tute, 1 km past Kseuela Agrfcola 
Altos de Piedras, W of Santa Fe, 08°36'N, 
081°06'W. 950-1250 m, 15 Dee. 1981 (fl). S. 
Knapp & A. Sytsma 2548 (holotype. LL!: iso- 
types, MO!, NY!). Figure 22. 

Trees 6-8 m tall. Branchlets slender, terete. 3-5 
mm diam.. densely cupuliform lepidote. h'aies with 
blades elliptic. 7.2-10.7 X 2.2-3.9 
cm. apicalb long acuminate, with an acumen 6-1 1 
mm long. basally ol.tusc. decurrent on the petiole, 
prominently punctate above and Itelow. glabrous 
aho\e. furluraceous-lepidote below except mixed 
Willi < 1 1 1 • l i 1 1 1 • > t 1 1 1 lepidote scales along lllidl ib. the 
midrib impressed abo\e. ])rominently raised below, 
the secondary \eins 15 to 51 pairs. piomnnil.Mi- 
above and below, the margins entire. Hat to slightly 
inrolled: petioles slender, canaliculate. 8-13 mm 
long. glabrous above mixed lepidote below. Inflo- 
rescences erect, bi- to tripinnatcly paniculate. 5.2- 
13.7 X 4.5-11.7 cm. pyramidal, longer than the 
leases, cupuliform lepidote. the blanches loose I \ 
, Miiiie~led into 3- to 7-llowi-reil corymbs: peduncles 
obsolete to 1.4-1.7 cm long: inflorescence bracts 
unknown: inlloreseeiiee branch bracts unknown: 
floral bracts caducous, membranous, lanc<'olale. 
1.2-1.4 X 0.4-0.5 mm. apicalb acute, prominently 
punctate, glabrous adaxially. furfuraeeoiisdepidote. 
the margins irregular, minutely erose. hyaline, 
sparsely glandular ciliolate: pedicels stout, obsolete 
to 1.5 mm long, prominently punctate, furfura- 
ceousdepidote. Flowers 5-merous. light pink: calyx 
lobes membranous, ovate. 1.2-1.3 X 0.9-1 mm. 
prominently punctate and punctnle-lineate. gla- 
brous adaxialb. scattered furfuraceous-lepidole 
abaxially. the margins irregular, minutely erose. hy- 
aline, sparsely glandular ciliolate: corolla membra- 
nous, 4.4-4.5 mm long, the tube 1.2-1.5 mm long, 
the lobes lanceolate. 3-3.2 X 2-2.2 mm. apicalb 
acute, prominently punctate and puuelale-lineale. 
idahrous throughout, the margins entire, hyaline: 
stamens 3.1-3.2 mm long, the filaments 1.9-2.1 
mm long, the staminal tube 0.6-0.7 mm long, the 
apically free portions 1.3-1.4 mm long, the anthers 
narrowly ovoid to lanceoloid. 1.7-1.8 X 0.8-0.9 

live conspicuously punctate: pistil 1-1,2 mm long, 
glabrous, ovary oblongoid. 1-1.1 mm long, the style 
3.1-3.2 mm lorn;, uicoiispieuoiisly punctate, the 
ovules 18 to 20. Fruits unknown. 

Cerro Tute in \eragua>. I'an.i 

ma. growing from 950 

to 1250 m in elevation. 

Frolt>»\ Olid conservation 

status. Ardisia albi- 

sepala occurs in lower mon 

tane rain forests and 

clou<l forests, Because of its 

restricted distribution. 

it should be considered 1 h r . -. 


Ftvmology. The specific 

epithet was derived 


i dull x 

""sepala"" i« fii i i ng to the light-colored calyx lobes. 
Within Ardisia subg. Auriculardisia sect. Pal- 
nuuuie. \rdisia albisepala is most closely related to 
1. tvsonii because of its short calyx lobes that an- 
as long as or much longer than wide short pedieeb. 
leaf blades less than 15 cm long and less than 6.5 
cm wide, and long petioles. However. I. alhisep,da 
can easily be separated from 4. tvsonii by its thin- 
ner branchlets to 5 mm in diameter, narrower cab x 

to 3.2 X 2.2 mm. shorter anthers to 1 .8 mm long, 
and longer styles to 3.2 mm long. 

Sperimens examined. I'WWIV Yrrajfua*: "Cerm 
Tute" ridj-r up from former Kscllela \gncnla. Santa I e. 
20 Id,. \'>',V.>, ill). (.. Hamilton X: H. Dossier :>,!<») (LI.. 

21. Ardisia an. I.i.ay ana bicketsoii & Pipoly. 
sp. nov. TYPE: Colombia. Valle del Cauca: 
Alto Yunda, Rio Anchicaya. 1000 m. July 
1072 (II). S. I lilt v Jv-28 (holotype, US!; iso- 
type, ARIZ!). Figure 23. 

ceas pedic-ellos brevio- 
, sed ah en lobulis caly- 
l longis, lobulis enroll i- 

Trees to 15 m tall. Branchlets stout. terete. 5-6 
mm diam.. bark exlolinling. Iiirluraoeous-lepidole. 
Leares with blades membranous to chartaceous. el- 
liptic to oblong. 20-36 X 6.9-9.2 cm. apicalb 
acute, with an acumen unknown, basally acute, de- 
current on the petiole, inconspicuously punctate 
and punctate-lineate. glabrous above, luifuraceous- 
lepidote. (he midrib impre—ed abo\e. prominently 
raised below, the secondary veins 75 to 85 pairs, 
promimilous above and below, the margins entire, 
inrolled: petioles stout, canaliculate. 3-7 mm long. 
glabrnii- above. I'm iuraceoiis-lepidotc below. Inflo- 
rescences likely pendent, tripinnatcly paniculate. 
23.5-27.5 X 10-15 cm. pyramidal, shorter than 
the leaves, densely furluraceous-lepidote. the 
branches loosely congested into 5- to 7-flowered 
eorvmbs: peduncles obsolete, the lower branch sub- 

i i 

florescence branch bracts caducous, membranous 
ovate to oblong. 1.5-2 X 0.4-0.6 mm, apically 
acute, prominently punctate and punclate-lincale. 
glabrous abo\e. lurfuraceous-lepidote below, the 
margin.- irregular, minulelv crosc. hyaline, sparsely 
glandular ciliolate: Moral bracts similar to the inflo- 
rescence branch bracts, but 1.2-1.7 X 0.4-0.6 
mm: pedicels stout. 1-1.5 mm long, inconspicu- 
ously punctate and punctale-lineate. furfuraecous- 
Icpidotc. Fionas 5-merous. while: calyx lobes 
membranous, ovate, 1.4-1.5 X 1.8-2 mm, apically 
acute to rounded, proiihnenllv punctate and punc- 
tate-lineate, glabrous adaxially. sparsely lurfura- 
ceous-lepidote. I he margins irregular, minutely 
cruse, hyaline, sparsely glandular ciliolate: corol la 
membranous. 5.4-5.0 mm long, the tube 1.2-1.5 
mm long, the lobes narrowly ovate, 4.1-1.4 X 2.4- 
2.5 mm, apically acute, prominently punctate and 
punctale-lineate. glabrous throughout, the margins 

ciilne. Ii\:il : slameus 5.1 5.5 mm long, ihe lil- 

amenls 2.6-2.8 mm long, the slaminal tube 0.5- 
0.7 mm long, the api< alb live portions 1.9-2.5 mm 
long. Ihe anthers narrowly ovoid, 245-2.9 X 1.1- 
1.2 mm. apicalb |.n i l„i ■• basally lobale. the con- 
nective conspicuously punctate; pistil 6.4-6.7 mm 
long, glabious the ovarv oblongoid. 1.1-1.5 mm 
long, the style 5.5-5.1 mm long, epiuiclale. the 
ovules 18 in' 24. Fruits unknown. 

Distribution. Ardisia anchicayana is known 
only from the type collection from Valle del Cauca. 
Colombia, growing at 1060 m in elevation. 

Ecology and conservation status. Ardisia anchi- 
cayana occurs in premonlane pluvial forest, on the 
western slopes of the western Andean cordillera. 
This area of the I Toed I Lrislic IYo\ ince is known 
for annual precipitation of over 8000 mm. and is 
also known for very large individuals of commer- 
cially valuable limber species. Therefore, all the 
forests of ihe area are under serious threat. 

Etymology. The specific epithel comes from the 
type collection along ihe llfo \nchicaya in Valle de 
Cauca, Colombia. 

Within \nlisia subg. \uriculardisia sect. Pal- 
munue. \rdisia anchicayana is most similar to A. 
palma/m because of the calyx lobes wider than 
long, membianoiis lo elurluei-ous leaf blades, and 
short pedicels. However, \rdisia anchicayana may 
be separated from 1. palmana by ihe longer calyx 
lobes to 1.5 mm long, longer corolla lobes to 4.4 
mm long, longer anthers to 2.6 mm long, and longer 

, Ricketson & Pipoly, 

sp. nov. TYPE: Costa Rica. Puntarenas: Can- 
ton de Osa, Fila Costena, Fila Cruces, head- 
waters of Rio Piedras Blancas, Cerro Angucia- 
na, western slopes, 08°48'56"N, 083°10'37"W, 
14(H)- 1 600 m. 10 Dec. 1993 (fr), B. Hammel 
19280 (holotype, MO!; isotype, INK not seen). 
Figure 24. 

i, », i II: .! .nli.s caKcitms longioivs quam latioivsovalos 
.holi- .aUnni- 2.7 2.<» (noii 1.7-2,1) inin longis. 2.1- 


ill. Hranchlcts slender, terete. 1-5 
mm diam.. densely and iniiiuleh appressed rufous 
luifuraceous-lepidole. often glabrate with age. 
traces with blade- nieinbrai s lo charlaceous. el- 
liptic to narrowly elliptic. 28,.8-33.3 X 9.9-11.6 
cm. apically acuminate, with an acumen 10-12 mm 
long, basallv acute, decurrent on the petiole, prom- 
inently raised below, prominently punctate and 
punctale-lineate above and below, glabrous above, 
densely and minutely appressed rulou- lurlura- 
, c,ii>-lepidol.-. ihe midrib unpiessed above. prom- 
iiiently raised below, ihe secondary veins 42 lo 48 
pans. |)roiiunulous above, the margins entire, flat; 
petioles slender, canaliculate. I 1-13 mm long. 2- 
3 mm diam.. glabrous abo\c sparsely furfurai-eous- 
lepidote below. Inflorescences erect, bipinnatelv pa- 
niculate. 29.8-30.5 X 17.5-18.2 cm. pyramidal. 

(this, bn 

, ! I, 

pedicels densely and minutely apprised ruloii- 
furfuraceous-lepidote. the branches loosely con- 
gested into 5- lo 7-llowered corymbs: peduncles 
2.7-2.9 cm long; inllorescence bracts and branch 
bracts unknown; floral bracts unknown; pedicels 
stout, 2-2.8 mm long, inconspicuously punctate 
and piuictale-liiiealc. Fionas 5-merous. calyx 

bicular. 2.7-2.9 X 2. 1-2.7 mm. apically acute to 
rounded, proinuieiilb punclale and punclate-linea- 
te, glabrous adaxiallv. furluraeeoiis-lepidote aba- 
xially, the margins uiegular. ininulely erose. hya- 
line, sparsely glandular ciliolate; corolla, stamens, 
and pistil unknown. I- mils ! immature), globose. I- 
5 mm diam.. prominently punctate and punclate- 
lineate, glabrous. 

Distribution. Ardisia angucianensis is known 
only from the type collection in Puntarenas. Costa 
Rica, growing from 1400 to 1600 m in elevation. 

Ecolog\ and conserration status. Ardisia an- 
tis occurs 
limestone. i 
it should be considered threatened. 

subg. Auriculardisia 
manae, Ardisia angucianensis is 
nigropunctata due to its calyx 

23. \rdisia aliupuipii. ca Lundell. I'hv lolo-ia 
48: 134. 1981. Auriculardisia atropurpurea 
(Lundell) Lundell, Phytologia 49: 342. 1981. 
TYPE: Panama. Panama: from Torti to the Pi- 
lota del Toro, the mountain overlooking Torti 
Arriba, 400-700 m, 27 Aug. 1977 (fl), J. Fol- 
som, G. Alonzo de Monte & relatives 4998 (ho- 
lotype, LL!; isotypes, FTG!, MO!). Figure 25. 

Shrubs 1-1 in tall, Hranchlets slender, terete, 1- 
2 mm diam.. densely lurliiraeeous-lopidote. h-ares 
with bla<les membranous, elliptic. 2.8-7.1 X 0.8- 

mm long. basallv acute, decurrent on the petiole, 
inconspicuous!) pimclalc and piuiolale-liiieale. gla- 
brous above. dcli-rb, liiTma< < in,- j-i.::'«« < belo.v. 
tlie midrib im|ircssed above. prominently raised be- 
low, the secondary veins 12 to 20 pairs, obscure to 
prominulous above and below, the mar-ins entire. 
Hat; petioles slender, canaliculate. 1-6 mm long, 
glabrous above, densely furfuraceous-lepidole be- 
low. Inflorescences erect, pirinaleb to bipinnately 
paniculate. 5-7 X 2-7 cm, pyramidal, longer than 
the leaves, denselv fiirluraeeoiis-lepidnle. the 
branches loosely congested into 5- to 7-flowered 
corymbs; peduncle 1-1 mm long, the lower branch- 
es subtended by leaves: inflorescence bracts un- 
known; inflorescence branch bracts caducous, 
membranous, ovate to oblong. 1 .3-2.3 X 0.5-0.8 
nun. apualK acute, prominent!) punctate and 
punctate-lineate. glabrous adaxially, mixed cupu- 
lilorm and lurluraceous-lepidote, the margins irreg- 
ular, minutely erose. hyaline, sparsely glandular 
ciliolate: Moral bracts similar to the inlloreseetice 
branch bracts, but 1-1.3 X 0.3-0.6 mm; pedicels 
slender. 7.2-16.3 mm lony. prominently punctate 
and punclale-lineale. mixed furfuraceous- and cu- 
pnliloiin lepidole. Flmrcrs 5-meruiis. deep purple: 
calyx lobes chartaceous, ovate, 1.5-1.6 X 0.9-1 
mm. apically acute, prominently punctate and 
punctate-lineate. glabrous ada\iall>. furluiaeeoii- 
lepulote abaxiallv. the margins irregular, niiniitely 
erose. hyaline, sparsely glandular eiliolale: corolla 

mens 2.3-2.1 mm long, the filaments 0.9-1 
long, the stamina] lube 0.2-0.3 mm long, the api- 
cally free portions 0.<> <».<". mm long, the anthers 
ovoid to narrowly ovoid, 1.8-1.9 X 0.8-0.9 mm. 
apieallv apiculatc. basallv cordate, the connective 
conspicuous^ punctate: pistil 2.(>-2.7 mm long, 
glabrous, the ovary ovoid. 0.9-1 mm long, the style 
1.6-1.8 mm long, prominently punctate and punc- 
tate-lineate. the ovules 12 to 14. Fruits globose. 4- 
7 mm diam.. prominently punctate. 

endemic to t 
Panama, growing from 400 to 700 t 

Ecology and 
purpurea occurs ii 
is so poorly know 

Flvmologv. The specific epithet comes from the 
Lit in ""alio" meaning dark and "purpurea" meaning 

purple, and refers to the dark purple flower color. 
Within Ardisia subg. Auriculardisia sect. Pal- 
immue, \rdisia alropiirpniea is most closely related 
to A. eucuncula because of the small calyx lobes, 
long pedicels, narrow corolla lobes, and short an- 
thers. How ever. Ardisia atropurpurea is separated 

mm in diameter, longer pedicels to 16.3 mm long. 

Specimen examined. I'WWU. Panama: a..:. >,„. 
rounding Rancho Chorrn. mountains above Torti Arriba. 
(,„u,a- mountain chain. * \)ee. 1077 (II. fr). J. Folsom 
el al. 66/7(1,1., MO). 

24. Ardisia aurieulata Donn. Sm.. Bot. Gaz. 24: 

395. 1897. Auriculardisia aurieulata (Donn. 

Sm.) Lundell. Phytologia 54: 285. 1983. 

TYPE: Costa Pica. Limon: forests of Suerte 

[Guppies|. Llanura de Santa Clara. 900 ft. 

[274 m|. Keb. 1890 (fr). J. Donnell Smith UhlU 

(holotype, US!, US neg. 2367!, LL neg. 1971- 

19!). Figure 26. 

Shrubs or small trees 1-8 m tall. 2-7 cm diam. 

lirunchlets slender, terete, the bark longitudinally 

ridged, 3-7.5 mm diam.. densely and minutely fer- 

riigineous furfuraceoiis-lepidote. the scales early 

caducous. I. cure with blades membranous to char- 

ovate. 12..'?- 16.7 X 3.4-17.7 cm. apically acumi- 
nate, with an acumen 2-13 mm long. basal h 
aurieulate. prominently pun. laic and punetale-lin- 
calc. glabrous above, scattered minutely ferriigin- 
eons furluraeeous-lepidole below, tlie midril) im- 
pressed above, prominently raised below, the 
secondary veins 32 to 1!'. pairs, ptomimiloiis above, 
slightly to promineiilly raised below, the margins 
entire. Hal: petioles slonl. marginule. subobsolclc to 
5 mm long. 2-3 mm diam.. glabrous above, furfu- 
raceoiis-lcpidote below. Inflorescences erect. hipin- 
nate to tripimuilely paniculate, 13-10 X 10-30 cm. 
pyramidal, longer than the leaves, the rachis dense- 
ly and minutely ferrugineous furfuraceous-lepidote. 
the branches loosely congested into 3- to 9-flowered 
corymbs, sparsely furfuraceous-lepidote; peduncle 
nearK obsolete to 8.6 cm long, densely lurlura- 
c-oiis-l.-pidole. the lower branches subtended by 
leaves: inflorescence bracts unknown: inflorescence 
branch bracts caducous, membranous, oblong. 10- 
18 X 2-4.5 mm. apically acute, inconspicuously 
punctate and puiictale-lineale. glabrous above, lur- 
I'uraceous-lepidote below, the margins irregular, mi- 
nutelv erose. hyaline, sparsely glandular oiholale: 
floral bracts similar to the mllorescencc branch 
bracts, but 1.1-1.8 X 0.5-0.8 mm; pedicels slen- 
der. 6-10.2 mm long, inconspicuously punctate and 
punctatc-lineate. very sparsely lurluraeeous-lopi- 
dote. glabrescent. Flowers 5-merous. greenish, 
white, pink, purple: calyx lobes membranous, 
ovate. 2.3-2.7 X 1.4-1.6 mm. apically acute to 

cat--, glabrous throughout, the margins irregular, 
minutely erose. hyaline sparsely glamlulai < ■iliolale: 
corolla membranous. 5.6-6.2 mm long, the tube 
0.4-1.2 mm long, the lobes ovale. 4.8-5.4 X 2.8- 
3 mm. apically acute, prominently punctate and 
punctatc-lineate. glabrous throughout, the margins 
entire, hyaline; stamens 4.7-5.1 mm long, the fil- 
aments 2.2-2.5 mm long, the slaminal tube 1-1,4 
mm long, the apically free portions 0.0-1.1 mm 
long, the anthers ovoid. 2.5-3.1 X 1.3-1.5 mm. 
apically api< ulate-mucronate. basally cordate, the 
connective proininenlK punctate: pistil 6.2-6. 1 mm 
long, glabrous, the ovarv oblong. 1.3-1. 1 mm long, 
the style 1.6-1,7 mm long, epunctate. the ovules 
19 to 21. Fruits globose, 5.5-7 mm diam.. promi- 
nently punctate and punclate-lineatc. 

Distribution. Ardisin aurictilala is found on the 
Atlantic Slope of Nicaragua in Jinotcga, Km San 
Juan, and Zelaya, throughout Costa Rica (except 
Cartago and Puntarenas). and Panama in Bocas del 
Tom. Colon. Code. Panama. San bias, and Vera- 

premontane wet and pluvial forests. It apparently 

has some tolerance lot disturbance and has a broad 
hahilal range. Therefore, we do not believe it is 
threatened at this time. 

Etymology. The specific epithet was derived 
from the Crook, "auricle." or ear. and refers to the 
lobed leaf bases. 

Within Ardisin subg. Auriculnrdisia sect. Pal- 
mniiue. \rdisia uuriculntu is closely related to both 
4. angucianensis and I. nigropunclntu because- ol 
its large calyx lobes to 2.0 mm long, which are 
I,, tiger than wide. a\u\ -lender petioles, \rdisiu nu- 

ll,,, >an ,„mi Ciiio Chontalefio. 20 km \K of I'd 
. ; ;1 >iillo. Win Iii.Ih, ualershed. 7-0 Mar. I«>7H (fr). I), W/ 
.'««/ (HNMN. MO); Mpio. Castillo. Hcserva 

y-,. | 


.level to 12011, 


Oil II,. MO). Zclava: \tlanla. Cafio el ligrillo, ha l'iea- 
,1a V, Nov. 10!!2 Ifll. I. hiuunu I Ml (MO): Km I'unta 
(„,„la. \llanta. "l.a liichard" 2(10 in SK. \\\ Nov. I9HI 
ill! /' \h.renn X /. Sundim, 12077 0IW1Y MO): Cam, 
\l„,,lec,,sto. inoiill, ol Cano Kl Con-nclo. 7 I eh. PK'.L'(lr). 
!' \lmvnn /of/o/lMOl: K ol Nueva \ llanla. 20 f'eh. I 00 I 
In /,' Runhi r»U (III II.. MOI: Mom. .1,- II., nan/a. I.'e- 
senabosaua- Ce, m < ,,la Ulan, a. 2 June I 007 I -In. ,. /,'. 
Hneda & I. Cnnmndn o/O. (III I.K. MO|. COSTA MICA. 

(II). V. Funk vl ,d. Hl')L>:i IMO 
Monleverde. Ouehrada Cerro N, 
lil.meas. 20 Mar. 1 0i (7 ilrl. U. , 
(MO); Canlon de I |.ala. Dos I 
righl hank Urn I'izole. 2!! Oel. 
(OK. MO): I, a hutuna. San Carl 
Calarala Km lortuna. 7 Nov. I 
FAhnndo 740 (CM. f I'C INI!. ^ 

slrcam on the Km I'nerlo \ ie,o. .Vd Jan. 1007 (IV). II. 
Ihuuc S: (.. Mmn, I . 1231 (OK. Dl KK. K C. (41). I A- 
mmi: near Km loro Amarillo. (a.apiles. 10 Feb. 1005 
(IV), /.'. f,W//.-» <><>7,M2 I MO. II i.eii.) San Jose: I'arque 

(.Mi.A.l.v in.l tributaries. 5 Oct. 100.5 (III. ./. Monde, el 

til. mr,5 «:i{. no. im$. m<». i>\\ \\i \. it,,,,,* ,i,.| 

Ton.: along mad In ( :hiri.,ui « -iaii.1. . L><> Oct. lO'C. (ill. 

(.. \ui>hc,su„ 7:1:12 (\io). <:,„.|,'. : c ;.«-i«-.-itt» mad. .-leva- 

i-i.lftr i-.|.. I-M.m. I'»i;r, n,i. C ,/,. \,-„7> ,-/„-'. /„-;/: I I. 
MO). Colo.,: walking upstream In,,,, bridge over Km 
Ouamhe. 10 Jan. I0f!<) (IV). 7.' l„,, w ,„ .7.T, / HI. Ml »,: 

along Id., Cuancla- ,-;,. .{-f, , n j. J.,I:,,mI. 3 Mlg. I <>7 1 (f]|. 

/. Cm,,/ _'o/,W (LI.. MO). Panama: headwaters ul Km 
< :ii = t^r< — _ Km L-peran/.a an.l Km IVdra>. 17 Oel. |0J!| 
(II). C. </<■ \e,ers el ,,l. M7f>\ MO). San lila>: I .,„„.,,, ., , lc 
S.-m l!la>. I'fnon Chi,.,,. Km, I kupsriii. along Km I knp- 

seni. Imm Camp Nelia Duiiiiiial to tin- Talk :'{ i. |'«)| 

III). //. Henna el ul. 1027 (HO. INK. MO,. Yera^.as: 
vicinity ul Lscmla \gi i, nil m .i Ml,, IV, Ira near Santa IV- . 
I hour walk along r„;nl bevnnd >,|„„,|. | |),, . |07<» ,|] L 
'/: \nlonio :UH)1 (MO). 

1 & Pipolv. s|t. nov. 
TYPE: Ecuador. Esmeraldas: San Lorenzo de 
Canton, Reserva Indfgena Awa. Parroquia Ri- 
caurte, comiiiiidad Ibdsarefio, Rfo Palabf, 
01°09'N, 078°31'W, 100 m, 15-29 Apr. 1991 
(II). 1). Kubio & C. Quelal 1453 (holotype, 
MO!; isotypes, FIX;!. QCNE not seen). Figure 

Trees 8-25 ,„ tall. 10-15 cm diam. / 
ender, terete. 4.5-6.5 nun diam., densel 
itely apprcssed rufous lurluracenus 

pressed above, prominently raised below, the see- 
tale and punetate-lineale above and below, glabrous 
above, densely and minutely appressed rufous fur- 
luraeeous-lepidote. inconspicuously raised above 

diam.. glabrous ah,,ve. lurliuaoeous-lepidole below. 
Iiijinicsmirrs erect. I ripinnalely panieulale. 8-21 
X 10-18 fin. pyramidal, usually longer than the 
leaves, peduncle, the rachis. hranchlels. abaxial 
bracl surfaces and pedicels densely cupulilonn and 
furfuracenus-lepidote. the branches loosely con- 
gested into 3- to 7-flowered corymbs; peduncle 
nearly obsolete to 1.7 em long: inflorescence bracts 
unknown; inflorescence branch bracts unknown; 
II nil hit membranous, ovate. 1-1.4 X 1-1.3 



punclate-liiicale. glabrous adaxially. lurluraceous- 
lepulole abaxiallv. the veins unknown, the margins 
irregular, minutely .-rose, hyaline, sparsely glan- 
dular ciliolale: pedicels stout. 1 .5-3 mm long, in- 
conspicuously punctate and punetate-lineate. 
sparsely furfuraceous-lepidote. Flowers 5-merous, 
while: calyx lobes coriaceous, suborbicular. 2-2.5 
X 2-2.5 mm. apieallv acute to rounded, promi- 
nently punctate and punclate-lincale. glubious ada 
xiallv. -parscb liirliuaceous-lopidole abaxiallv. ob- 
solete- Hal scales with the margins entire or with 
small teelh. the margins irregular, minutely erose. 
hyaline, sparsely glandular ciliolale: corolla char- 
laceous. 5-0.2 nun long, the lube 1.4-2 mm long, 
the lobes lanceolate. 3.9-4.5 X 1.7-2.1 mm. api- 
eallv acute, prominently punctate and punetate-lin- 
eale. Jahiou- tin, .ii-h,, ul. I In- margins entire, hy- 
aline; stamens 3.7-4.2 mm long, the filaments 1.5- 
2 mm long, the staminal tube 0.5-0.7 mm long, the 
apieallv free portions 1-1.3 mm long, the anthers 
ovoid. 1.9-2.2 X 1. 1-1.5 mm. apieallv apiculate. 
basally deeply cordate, the connective conspicu- 
ously black punctate; pistil 5.2-5.0 mm long, gla- 
brous, the ovary oblong. 1.2-1.4 mm long, the style 
4-4.2 mm long, prominently punctate, the ovules 
46 to 48. Fruits red. then burgundy, then black at 
maturity, globose. 8-9.2 mm diam.. prominently 
black punctate. 

Distribution. \rdisiu oicarum is endemic to the 
Reserva Indfgena Awa in San Lorenzo. Ecuador. We 
would anticipate thai its range extends In the south- 
ern portion ol Narino. Colombia, on the western 
slopes ,,f the Wesicrn Cordillera, growing between 
(80-)200 and 600(-1000) m in elevation. 

rum occurs in primary, pluvial premonlaiie forests, 
apparently along the forest margins. Hecause the 
majority ol the coll.-clions are from protected areas. 
it does not appear that the species is threatened. 

Etymology. The- specific epithet is in honor of 
the Awa. a group ol indigenous people found along 
the western slopes of the western Andean Cordil- 
lera, from Fsmeraldas. Ecuador, north to central- 
western Narino. Colombia. 

Within Ardisia subg. Aitrieulardisia seel. I\d- 

lup'hiihi l.ecau-eol ii- leal blades coriaceous, calyx 
coriaceous, corolla ehartaeeous. the tube glabrous 
outside, shorter anthers to 2.2 mm long, and shorter 
slyles to 4.2 mm long. 

Paralvpes. LCI M)Oli. Carclii: Tulcan Canton. IV 

i"'"i|iiia se,|,,i Cualpi medm. Keserva Indfgena 
\ua. Iiail I,, Nin Marco- \ ,,l the <■, .mmiimlv , rule,. 2'.\ 
27 May 1002 (|||. <;. 'fipuz n ul. 10211 (III.. MO). IV 

rva film 

70 Ir TO. MO. OCNK. IS); Reserva f 
( |iiia Malar. <-rntm Malajr. 21 Sr|.. \<><>2 (III. C. \uleslia 
el al. 177 (FTO. MO. (KAK): II.xtu r'.lni.-a \«a. I'a- 
m.t|iiia Malajr. Ceiilm Malaje. 21 Sep. I<><>2 (IV). C h/- 
leslhtelal. r,l!i\H,. MIAl . MO. \VO(\l l: I'am.iiuia 

\||„ Taml.o. >,Tl..r Kl Cristal. 13 Apr. I<><<2 1 1. 1. (,. Tip.,: 
el al. 777 {WW.. MOl: lieserva Inilrgciia \»;i. hVaiirte. 

10-21 O.I. 1<><)2 ifrl. -A 7//,«-- <•/ al. 2044 (FTC. MO. 

26. Ardisia blepharodes Lund. II. Wrightia 4: 

55. 1968. Auriculardisia blepharodes (Lundell) 
Lundell. Phytologia 49: 542. 1981. TYPE: 
Costa Rica. Cartago: Kl Mufieeo on the Rio 
Navarro. 1490-1 5(M) ,n, 6-7 Mar. 1926 (fr), P. 
Standley & R. Torres R. 51266 (holotype, US!, 
LI. neg. 1971-22!: isotype, GH!). Figure 28. 

Small trees 3-5 m tall. Branchlets slender, terete, 
2-5.5 mm diam., densely and minuloK appressed 
ml.. us liirfuraceous-lepidote. Irenes with blades 

rowh elliplie In ol.laneeolate. 5.2-12.6 X 1.4-3.2 
cm. apicalK abruplK a. umu, ale. with an acumen 
6-12 mm long, basal ly acute, decurrent on the pet- 
iole, prominently punctate and punctale-luieale 
above and below, glabrous above, densely and mi- 
nutely appressed rufous lurlura. eous-lepidote. the 
innlnl, impressed above, proiiiineiilh raised below. 
the secondary veins 17 to 29 pairs, prominulous 
above and below, the margins entire. Hat: petioles 
slender, marginate. 6-11 mm long. 0.5-1.5 mm 
diam.. glabrous above lur fiua< eoiis-lepi.lole below. 
Inflorescences .Meet, hi- to iripim lately paniculate. 
9.2-16.7 X 8.1-12.5 cm. pyramidal, longer than 
the leaves, the mollis, branchlets. abaxial bract sur- 
faces and pedicels furfuraeeous-lepidote, the 
branches looseK congested into 3- to 6-Howercd 
corymbs: peduncle 0.8-2.6 cm long, the lower 
branches subtended l>\ leaves; inflorescence bracts 
unknown; inflorescence branch bracts caducous, 
membranous, ovate. 0.8-2.2 X 0.4-1.2 mm, api- 
calK acute, promin. ailK punciale and puuclal.-li- 
iieale. glabrous above. f urluraceoii — le| )ldo|. ■ below. 
the margins 

sparselv : 

ular to ovate. 2.2-2.8 X 2.1-2.8 mm. apicalK 
obtuse to rounded. prominenlK pun. late and punc- 
tate-lineate. glabrous adaxially, furfuraeeous-lepi- 
dote abaviallv. the margins irregular, minutely 

liolate: corolla 
. 6.9-7.1 mm long, the tube 2.2-2.4 
mm long, the lobes narrowly ovate to lanceolate. 
4.5-4.9 X 1.8-2 mm, apicalK acute. proiiiinentK 
punctate and puuctate-liiieate. glabrous adaxiallv. 
furfuraeeous-lepidote abaxially. the margins entire, 
hyaline; stamens 6.5.-6.9 mm long, the filaments 
4.5-4.7 mm long, the staminal tube 0.8-0.9 mm 
long, the apicalK free portions 3.6-3.9 mm long, 
tin- anthers narrowly ovoid. 2.2-2.4 X 0.9-1 mm. 
apicalK api.ulale. kisalK deepK cordate, the con- 
nective conspi. iioii-Iv punctate: pistil 0.2-6.6 mm 
Ion-, glabrous, the ovarv oblong. 1.1-1.3 mm long, 
the stvle 5.1-5.3 mm long, prominently punctate, 
the ovules 15 to 21. Fruits globose. 5-6 mm diam.. 
inconspicuously punctate. 

Distribution. Ardisia blepharodes is endemic to 
the Cordillera Talamanca. in Union and Cartago. 
Costa Rica, growing from 1200 to 1500 m in ele- 

Ecology and conservation status. Ardisia blep- 
harodes occurs in moist premonlane forests. Be- 
cause it is known from so lew ooll.-eliotis. no further 
informalioii is available aboi 

Etymology. The specific epithet was derived 

from the Creek "Blepha-" meaning relating to eve- 
lashes or eyelids, referring to margins fringed with 
hairs or ciliated, and "odes" or "oides" meaning 
like or resembling, specificalh referring to the 
glandular eiliolatc calyx lobes. 

Within Ardisia subg. Auriculardisia sect. Pal- 
manae. Ardisia blepharodes is easily distinguished 
from 1. hagenii bv its smaller leaves to 12.6 cm 
" 5es to 2.8 i 
long, shor 

mm long, and shorter styles to 5.3 mm long, \rdisia 
blepharodes is separated from both 1. (imbrillifera 
and 1. pseiidorarcnnfloru bv it- -mailer and narrow- 
er leaf blades to 12.6 X 3.2 cm. shorter pedicels 
obsolete lo 1.2 mm long, narrower corolla lobes to 
2 mm wide and shorter and narrower anthers to 2. 1 

Speeimens examined. COM \ till \ < ar...^. 
\\ of Casa Blanca. lapanli. 2d I >.•.-. PWSI (ill. 
1/ C.nnum ei ,,l. lOVUVW,. I.I.. MO). 22.1,,,,. !<>;;.,, III. 
\1. (innum & It Warner 7,1.11 (CI!. ITO. LI- MK\l. 

MOl. I.i ,!„„„, i.,;i, l,.-| W r.-i. «:.-m. Cliinui and Cerm 

Malania. 30 \|-r. PKi.') (In. /.. C»t»e; el al. 1 1 TO. 
MO, NY). 

27. Ardisia capitellata Lundell. Wrightia 6: 67. 
1979. Auriculardisia capitellata (Lundell} I in 
dell, Phytologia 49: 342. 1981. TYPE: Costa 
Rica. Puntarenas: above coffee fincas along 

Rio Colo Brus, near Cotan, 23 km N Fa Union, 
on Panama border, 9 Aug. 1974 (fl), T. Croat 
26678 (holotype, FF!. F neg. 55613!; isotypes. 
MO!. NY!). Figure 29. 

cidunlisia latisepala l.undell. Wrightia 7: 269. 1984. 
Syn. mov. \rdisia lalisqxila (l.undell) l.undell. Pliyto- 
logia 61: 65. 1986, nom. inval. Ardisia latisepala 
(l.undell) Pipolv & Kiekelson, Sida 18: 513. 1998. 
TYI'K: Costa Rica. Puntarenas: on and around Wil- 
son's f.nea. (, km S of San Vito de Java. ea. 4(KK) ft. 
[1219 m], 19 Aug. 1967 (fl). I\ Karen 2/6553 (ho- 
loivpe. F!. K neg. 68324!; isotype, MO!). 


lie ni.iigins entire, opaque: 
stamens 3.9-4.2 nun long, the filaments 2.6-2.8 
mm long, the staminal lube 0.7-0.8 mm long, the 
apically free portions 1.8-2.1 mm long, the anthers 
lanceoloid. 3.2-3.5 X 1 .3-1 .5 mm. apically broad- 
ly apieulate. basallv deepb cordate, the connective 
conspicuously punctate; pistil 0-0.2 mm long, gla- 
brous, the ovary ovoid 1-1.1 mm long, styles 5-5.2 
mm long, epunetate. the ovules 22 to 24. Fruits 
depressed globose. 7-9 mm diam., inconspicuously 
punctate and punctate-lineate. inconspicuously 


diam. Uraiichlels slender, terete. 3-7.5 mm diam.. 
denselv and minutely appressed rufous furl lira - 
ceoiis-lcpidole. Iran's with blades chartaceoiis to 
coriaceous, elliptic- to narrowly elliptic. 10.4-32.8 
X 4. 1 9.8 cm. apically acuminate, with an acumen 
0.5-1.7 cm long, basally acute, decurrent on the 
petiole, inconspicuously punctate and punctate-lin- 
eate above and below, mostly glabrous above, 
denselv and minutely appressed rufous lurlura- 
eeoii-lepidole below, deii-er along the midrib and 
secondary veins, the midrib impressed above, 
prominently raised below, the secondary veins 39 
to 57 pairs. prominulou> above and below, the mar- 
gins entire, flat to slightly revolute; petioles slender. 
6-18 inni long. 1-3 mm diam., glabrous above, lur- 
luraeenus-lepidote below. Inflorescences erect, bi- to 
tripinnately paniculate. 8.2-32,1 X 4.8-31.5 cm. 
py ramidal. longer than or as long as the leaves, the 
peduncle, rachis. branches, and pedicels densely 
and minutely appressed rufous furfuraceoiis-lepi- 
dote. the branches loosely congested into 4- to 9- 
flowered corymbs: peduncles 0.0-3.0 cm long, the 
lower branches subtended by leaves; inflorescence 
bracts and branch bracts unknown; floral bracts 
usually persistent, membranous, ovate. 2.5-4.5 X 
0.8-3.2 nun. apically rounded to acute, inconspic- 
uously punctate and punctate-lineate. glabrous 
above and below, the midrib unknown, the second- 
ary veins unknown, the margins irregular, minutely 
erose. hyaline, sparsely glandular eiliolate; pedicels 
stout. 0.9-2. 5(- I) mm long, inconspicuously punc- 
tate and punctate-lineate. Flowers 5-merous. rarely 
6-merous. white to pink or light purple: calyx lobes 
chartaceoiis to coriaceous, orbicular to ovate. 3-3.4 
X 2.6-3.2 mm. apically rounded, conspicuously 
red punctate and punclatedineale. sparsely furfii- 
raceous-lepidote abaxially. glabrous adaxiallv. the 
margins entire, sparsely glandular eiliolate: corolla 
-7.6 mm long, the tube 1.1-1.0 

Distribution. Ardisia capitellata is endemic to 
ie southwestern portion of the Cordillera Talaman- 
i. Puntarenas. Costa Pica, growing from 1200 to 
!>>()<) m in elevation. 

Fcology and conservation status. Ardisia capi- 
■llala occurs in pi unary and secondary lower uion- 
me wet forests, occasionally in disturbed areas. 
realise it is known from so few specimens, its cur- 

Etymology. The specific epithet refers to the 

compact, capitulate appealing corymbs of the inflo- 
rescence branches. 

Within Ardisia subg. Auriculardisia sect. Pal- 
manae. Ardisia capitellata is very similar to 1. ge- 
neialensis. bill is d i-l i ngiiislied from it by the short- 
er and narrower leaf blades to 32.8 X 9.8 cm. the 

32.1 - 31.5 , 

mm long, the lobes narrowly o\ 
5.8-6 X 2.8-3.2 mm. apically t 
inconspicuously red punctate am 

Populations corresponding to the type of Auri- 
culardisia latisepala Lunclell match the type of Ar- 
disia capilellata l.undell in all respects except for 

•dighllv larger flowers. 

Sep. 1981 (II). C. Daridse el al. 
I'roteetora I .as 'labia- t. Marion li 
ralelo Trail. 31 Aug. I'»)2 (fl). 
(C|{. KTO. IMC MO): Canton .1. 

(INIC MO): Canton de Colo Urns. Cuenca lenalia-Sierj 
Kslacion Biologica Las Altura-. trail to < I < : 
(), l 1007,111. II. Camlxm H /*-MMl \U ' I: al< ....■■ I r. 
between I .as Cruees liotanieal Oarden and Rfo Java, I 
3.5 km sf ol San \i(o de Colo Krus. 12 Sep. 1985 ( 

\l. Cnnun, rl „/. .><>,-,' I IK, LI.. VIO); las Mimas. 2(. Ul |„. 1 ._> •> ...... | ,..„ ., , , . ,_, ., 

An- M)7I (II). /! !/„„* A /.'. !/,./„ //<>_' <|-,. ( ; ; „ ,| ( . ,,..,_ " . ' "" '* m , *' '~ 

<: Urns. Pai<|,ic lti,li.u<n.i I., \niisla<l. «;,.,-. li I U-r., lala " ' """' ;, l'"' ;lll > '" lllr - prominently puiicl 

nnmca. Pill,, i. Puii.r Trail. I \i, u . P>»r> Uli /■ al " 1 P»"<'lalc4 mcale. glabrous I h 101 i-l ,. >i . I . ill,' in 

\ararn, 170 |C|{. IT.;. IMI. VIO,; WiWs I'a.m. h k„, gins entire, hyaline; stamens 4.8-5 mm lom> 

s ... s,,n \ ,,„ i, | „ „. W ;■,,; ,„, /- ,,„„, ,, //t?r f|lam( . |lls 2 . V2 ., mm |ong> tli( , slamiiia| in|jr () 

L, ,a,i,. ,,,,,11:';,"; ,i ;„:,;„ , ;: ', , ;i; , ; mi ,,;;,',," «»- ...... i.... B ..i,,.a,,i«-aiK in-,,.,,- i «, !.»„ 

na-ia. l'J! J.m I «<> , . I , A' I ,, „, | |< .. |'\ |{."\1( )' : l»"K- H.«- anlli.-i-. lancetiloid. 2.9-3 X 1-1.1 ,,, 

Caiilnn^h- C Bins. Pai(|iir liuhVriia I .a \inisiad. On- apicallv apiculale. I>asall\ deeply cordate, (lit- c< 

""''■ Im '"'">" I'm"'.. 12 June PM5 (II). A'. nective conspicuously punctate: j>istil 6.4-6.5 n 

long, glabrous, the ovary oblong, 1.8-1.6 tnm loi 

the style 4.6-4,7 mm long, epunclate. the ovu 

2«. Ardisia rartagounu l.undell. Wrighlia 6: 68. l6 l() 21. Fruits (immature) globose. 4.5-5.2 n 

l ( >7«>. \nricidardisia carlagoana (l.uii(lell) long, proinineiitb punctate and piiuclaledinealc 

l.undell. Phvlologia 46 : 343. 1681. TYPK: /v ... , .. . 

Costa Hi,,, Carla,,,: aim,, road |, ,„ .|„an ,. ' "'j;'"'""''- '"'«<' '«»,,*,,,,,,, is endemic 

Vinas & lurrialba. 7 km \\ lurrialba. along ° S,U . , " ""' hus , *?" M 'T" 1 "f '" ,l " 1 ( <"« l 

ditch. I July 1676 ,11). 7. ,.„„„ .;„/;// ,| l( ,|„- 'j'' 1 ,!",''""' ""'" "' m '" mi,l,u ,( - i,,l!,,a " a 

Ivpe. MO!. K neg. 55671!: iso.ype. I.I.!). figure "' hl,i ; .'''' Ma,amit *' " ,e , «" i « ,w «'^ "»' ""■ » 

• 5() • ' ' h RoVCI ll.lllK- 

>7irii/« or .v#i,«// /rec.v 1-8 m tall. Hranchlets stout. £co/o£ V 

terete. 5-10 mm diam.. densely rufous furlura- ><'&>«">' 

ceous-lepidole. h-arcs with blades membranous. l)erM lollml 

oblanceolate. 15.4-32.6 X 5.1-45.7 cm. apicalK Ht ' l(,s an,j ,mi * is ' 

acuminate, with an acumen 6-10 mm long, basally m " <<>nsidered I hut 

acute, decurrent on the petiole, inconspicuously Etymology. The specific epithet 

and conspicuously punctate and punctatedineate. Province of Carlago. Costa Kica. where the type was 

glabrous, the midrib impressed abo\e. prominently l 'ollected. 

raised below, the secondary veins 40 to 51 pairs. Within Ardisia subg. \uricidardisia sect. />,//- 

nitid above, prominnloiis below, the margins entire. """"»'• \rdisia carlagoana is closely related to 4. 

flat to slightly inrolh-d: petioles stout, marginal*-. "X'^i^ophy/lo. A. agiiirmuui. A. cogolloi, and A. 

nearly obsolete to 1.5 cm long. 3-1 mm diam.. -la- frussipes because of its long calyx lobes with thick 

brous. hi/lori'scrnct's civet, bi- to Iripinnalelv pa- P*'"""'™- However. 4. carlagoana is separated from 

nicidate. 13.1-28.4 X 6.7-18.8 cm. pyramidal. 4 " f"t'gi"l'>/>hyUa h\ its shorter calyx lobes to 2.3 

usually longer than the leaves, mixed minutely cu- Inin ' >- s,l, "'l<''' <«'rolla lobes to 4.2 mm long. 

I 'I I ' nil lurfura lepidote, the branches sll< >'"ter anthers to 3 mm long, and shorter styles to 

congested to loosely congested into 5- to 6-flowered ''' ,m " ,() "^- ,/v/ ' v "' mrlagoaiia is easily distin- 

racemes with flowers clustered in a pseudocorv ml. Pushed from I. aguirmma by its shorter leaf 

at apex; peduncles 1.5-3.5 cm long, the lower ,,la(l, ' s to :i2 - (> <"' long, smaller branchlets to 10 

branches subtended by leaves; inflorescence and """ '" ,li; '""'l'''- shorter peduncles to 3.5 cm long. 

branch bracts unknown: floral bracts caducous. aml ui(l, ' r <orolIa lobes to 2.7 mm wide. Ardisia 

membranous. 1.7-2 X 1 .4-1.8 mm. apically acute. <<"l<igo<m<t is separated from 4. cogolloi by its 

conspicuously punctate and punclatc-liiieale. gla- small, 'r and wider orbicular, not ovate calxx lobes 

brous adaxiallv. fuiluiaceous-lcpidot.. abaxially. the 2A X 2 " ' " 1,n - longer and narrower anthers 3 X 

margins irregular, minutely erose. hyaline sparseb '•' " lin - l,, "f?<' 1 ' st >h's to 1.7 mm long, and smaller 

glandular ciliolale: pedicels slout. nbsnlclc to 2 mm lniil ^ '" "'^ l,im '" <liaineler. Ardisia carlagoana 

long, inconspicuously punctate and punctale-lin- (,if '«'rs from 4. crassipes by its membranous not co- 

«'alc. furfiiraceoiis-lepidote. Fionas 5-nierou-,. '"laceous leaf blades, which arc glabrous aho\e and 

white to pink or violet: ealyx lobes chartaceous, helovv - thiimt-r petioles to 4 mm in diameter, nar- 

orbicular to oblate. 2-2.3 X 2.1-2.4 mm. apicallv rower l ' al - vx lo,)< ' s to 2A lllm wide, shorter corolla 

rounded, inconspicuously punctate and punctate- lo,u ' s lo ''■- """ 'o' 1 ^- am ' shorter styles 5.2 mm 

M.icate to nearly epunclate. glabrous adaxialk. fur- l(MI - 

ltt<;i<;-<.,<- I. -pi.lnlc abaxially. ihc margins irregular. ., . . , 

;-; ■> ;:■>"- Y™- '^ sla Iar ' " *'S:'^ ■'■■■ ' "-'"'..llirr^ 

lat< ' ; Hi > I » I i n the deChirripo. 1(1 <),l. P><)2 (II. I,), f. Cam^s .u H\\{. VW,. 


IMS. MO|. _"• June IW.5 (II). I\ Campos 7h (l-'TO. |\|{ i , in . , 

MO). Uinon: Canton dr I, im,,n. <;,,,,!, II, T aTalamanea N , '*' M> '' ' "" ' . _ ,u " n< ' rl ' V( ' conspicuc. 

Hank „l I'iL, de Uatama al 1,,-adual, ,- of |,'„. IJmei. 17 '> P lm,lal <' : I"- 1 ' 1 •»--•••«". nun long, glabrous. 

Au S . 1995 (fl), M. Grayum 11046 (INB, MO). "vary oblong. 1.6-1.7 nun l.-n-. the style 2-2.2 i 
long, epunelate. the ovules 19 lo 27. Fruits glob, 

29. Ardisia eogolloi Pipoly, Caldasia 16: 277. 8 ~ 9 ' 8 llim <liam - l»«»miiiriitly punctate. 

1 991. TV I'K: Colombia. Anti.ujuia: Mpio. de Distribution. Anlisia cogolloi is endemic to 

I -n.o. Pa,,,ue Naconal Natural "Las On,,,,- (; ( . n |,||en, ( >ee,de„lal of Colombia. ,-,■„«,„, I, 

(leas. Sector Venados. right hank of upper Rio {;()<) t( , 17-^) , n ;,, elevation 
Venados, 00=31Tv 070 |9<W. | 150-1300 ,„. hnll) ^ ,„„, rnllsrrr(ltinn „,„„, ,,„,„,„ 

26 July 1988 (fl) .4 <;,*„//,, ./. A>„„„W, & (). ,/„/ oeenrs in prcnonlane pluvial forests.' when 

/IWez .«98 (holotvpe. .1 Al VI!: isolvpes. is eeolngiealb ^nduunl. Ire (| ue,„b alia a 

COM. FMH!. HIM!. MO!). Figure 31. a.neter a,, ,,,, .5 ,,,, be-anse , 

7Wv.s .VI 5 in tall. 9.5-2< I , Hr fit ma i oril > of •'"' collections are from protecl<-d a 

(..:» Id., , 

furfuraceous-lepidote. /^/ms with blades charta- £'.vmo/o#v. This species was named in I 

reous. elliptic. 14,1-50.2 X 5.9-17.6 cm. apically ° f ,Ji, ' ,L A,vam (:<, « ,,,,n '^checo. Director o 
obtuse will, an abrupt acute acumen, acumen {).(*- * van ' ] \ ; " lllr ''""daci,,,, Jardin botanic,,. "Jo; 

, decurrent on the pel- 

• Uribe." in Medellm. Colombia. 

iole. prominently and meonspieuousb punctate and NX l,hl " A " hsi " Sllh - ^'iculardisiu seel. Fal 

punctato-lineatc above and below, glabrous above """""' W ' s '" n '#" lloi is rl,,s, ' l > , '«' lat <-< 1 "' *• '"'" 
densely rufous furfuraceous-lepidote I.elou. the ^ ls " , l ,liyll < 1 - ■'*• <W"rreunu, A. crassipes. and A. car- 
midrib impressed above, prominently raised belou. '"^"""" demise of its long calyx lobes with thick 
the secondary veins 38 lo 15 pairs, nilid above P<' llo,< ' s - H«»we\er. rofjn/loi is separated 
raised below, the margins entire, subrovoluto peli- ,mm l "»'A''*'<V>>n-//<i '»> its shorter calyx t„ 
oles stout, canaliculate. 1.5-3.5 cm long. 1-7 mm 2<) """ l,,n - sh,, '"' r (omlla l,,l,( ' s "> 4S »»" '""K- 
diam., glabrous above, furfuraccous-lepidnlc below. slm,l,T m,ll,,,, ' s lo 2(> mm l,,M ^ a " ,l sl,( "f>- styles 
Inflorescences erect, bipinnatelv paniculate. 15,1- ,c> 2 ' 2 """ l,)M ^ W,w " '"A">AW i* «'asily distin- 

24.6 X 4.5-13.6 cm, pyramidal, shorter than the ' !:ms,ln, ,mm l «H'"™""' '»> its shorter leal 
leaves, the branchlets. abaxial bracl surlac- UUi,{v * '" MV2 <m lo "^ s '" al '<''- branchlets to 10.5 
es, and pedicels densely furfuraceous-lepidote. ||,c """ '" ,liam, ' l, ' r - ■ s| ""' t '' peduncles to I cm long, 

branches loosely congested into 2- to 7-flowered ;m<l U " l< ' 1 ' nM ' ,>lla lo,)( ' s lo 2() mm wi<lt '- ^disia 

corymbs; peduncle 0.2-1 cm long; infloresc.-n.-e " w " // "' '^ ^'Parated from both 1. crassipes and I. 

bracts and branch bracts unknown: floral bracts ca- r " r "W""'« '>> lls '"'^r and narrower ovate, not 

ducous. membranous, ovale, 1-1.5 X 0.5-0.8 cm or,,lt ' l,li,r <' a, .vx l»»>«"s to 2.6 X 1.9 mm. shorter and 

apically acute. ,,romin, punctate and pundale- wi,,, ' r un,,M ' w '" 2i > X ] ^ "»»■ «l""1er styles lo 

lineale, glabrous above, h.rluraceousdepi.h.le be- ~ 2 '" m lu "^- a,,<l lai W' r l,llils "' 9 -« """ '" <H- 

sparsely glandular ciliolat<- ; pedi«,-ls sloul. obsohMe N /w ,,„,,,,, ,,,,„„■„,,/. C()|.()MMIA Anlio,,„ia M,„u 

to 0.7 mm long. proniinenlK punctate and punc- <le l'rra». Pan|,i,- \a, -io.ial Natural "Lis 0,(|u,«l,-as." \,- 

tatedineate. furfuraceousdepidole. F/oicers 5-mc- r,,,la ( -allcs. n»ln hank <.l i;„. ( all... |, r ,an-e N\\ „C 

rous. white; calyx lobes c.hartaceous. ovate. 2.5-2.0 ^■^ \^\\ ™ n ^{ J™ In,,!' pTe'v " / ' 

X 1.8-1.9 mm. apicalU obtuse, prominently ■punc- n .,| Natural -| a (i,,,,,,,),,,! " \,\ ,| , j" ,ll,~ '( ),„ |,!' I( '| 

late and punclatcdinealc. glabrous adaxialb. fur- Honda, on the ranjie N\\ uf ( ahan 1 ( all, ■ :;li',' i'l'V' 

furaceous-lepidotc abaxiallv. the margins irregular. i-i'-'-i- .A I'tf-h .■;.,!. /^,:'",J\l \|. \|,),. H,7,'L>(\\\ VI 

minutely Iiyalinc. sparsely glandular cilio- Xl " ! ' I \l M. \1(H: \1pi... ,le 

late; corolla tiiembi. us. (..2 ,...> mill |,,im. if, , (j| >( . , l ,' ( ' i | l ' 1 "" ' '" ,',"'.' 

tube 2-2.3 mm long, the lobes narn.wly oblong. \L.*\<m ,',,',■." !/!' 'I'ipJi n "/'. /"/:;', |"lo! I \\\\ 

4.2-4.o X 2,4-2.6 mm. apicallv acute. pmmimnlK \l(»: M|.io. d, I nan. Panpie Na.mual Natural "Lis Or- 

pintclale and |Minclalcdincale. glabrous adaxialb M'l'dra-."" \civda (.all.--, ri-ht bank „l Km Call,-. '> I),-, 

furfuraceousdcpidnl, abaxiallv. ih. „,.,„,„, r ,i P^< <Hi. .A /'»/W. ,/ „/. UUMin ,\ I (,. J \[ \1. \K) ); \\ 

|„ ,|; , -•)-,. ' 1 1 ,-. " ' lr lll,lllll ">- \creda Venados. Pairiue Na<-„.„al Nalnril 

I wh ,e. stamens ,.2 .,9 mm long, .he filaments ,, s ()t , imi|( . a> . M|| „ , „ „, m ,',,,,. ,„„; / ," 

• U. .J.o mm long. Ihe -luminal lube |- |.| ,,1111 long. /«./t rial. l!i:U)7 |.| VI M. MO), 
the apically free portions 2.5-2.9 nun long, the an- 
thers ovoid, 2.3-2.6 X 1.3-1.5 mm. apically apic- 30. Ardisia eoloradoana l.undell. Wriditia 6: 

Volume 90, Number 2 

69. 1979. Auriculardisia colomdoana (Lnii- 
dell) Lundell, Phytologia 49: 343. 1981. 
TYPE: Panama. Chiriqiri: Genu Colorado. 31— 
35.6 km above Rio San Felix, 13-14.6 km 
above turnoff to Escopeta, 1390-1410 m, 15 
July 1976 (fl), T. Croat 37235 (holotype, MO!, 
F neg. 55684!). Figure 32. 

Treelets or trees 2-6 m lull. Branchlets slender, 

terete. 3-6 nun ill, km . d< -i- i u I li> u hi ' I m 

liliaceous -lepidote. Leans will] I >lad< 's eliai tare, in- 
to coriaceous, elliptic lo oblong, 3.4-10.8 X 1.1- 
3.6 cm. apieally long acuminate, with an acumen 
0-12 mm long. basallv acule or cuiicale. deeurrenl 
mi llic petiole. inconspicuously | hit nl .it « - and pune- 
tale-lineate. nearly glabrous above, densely ap- 
pressed rufous furfur.ieeous-lepidole. the midrib 
impressed above, prominently raised below, llie 
secondary veins 23 to 2') pairs, prominulous above 
and below, llie margins entire. Hat. at limes drying 
i rolled: . i in., c- sli nd< . e, mil i< i laic. I . I -2. I cm 
ong. j ill. inns above, densely appressed furliim- 
eeous-lepidote below. Inflorescences erect, bi- to tri- 
piiinaleU paniciilale. 5-2 1 X 1.5-1 1 cm. pyrami- 
dal, longer than llie leaves, peduncle, the racliis. 
branches and pe«l ,rb densely eupuliform and lur- 
liiraceous-lepirloic. the branches looselv eougc-led 
into 3- to 7-flowered con nibs: peduncles nearly ob- 
solete to 3.8 cm long, the lower branches subtended 
by leaves; inflorescence brads unknown: inllores- 
cence branch l.rael- cadiicou-. membranous, ob- 
long. 2.5 2.'> X 1 -1.3 mm. apioallv acule. promi- 
nently punclale and punctate-liiieate. glabrous 
above, appressed furfuraeeous-lepidole below, the 
margins irregular, minutely erose. hyaline, sparsely 
glandular cdiolale; floral bracts similar to the inflo- 
rescence branch bracts, but 2.8-4.1 X 0.6-0.9 
mm; pedicels slender. 3.5-5.1 mm long, inconspic- 
uously punctate and piuiclate-lineate. mixed cu- 
puliform and furfuraceous-lepidote. Fionas 5-me- 
rous. white or cream; calyx lobes ohartaceous. 
orbicular lo oblate, 1.4-1.6 X 1.6-1 .9 mm. apieally 
acute to rounded, conspicuously punctate and 
punctate-lineate. glabrous adaxially. furluraeeous- 
lepidotc abaxiallv. the margins irregular, minuleb hyaline, sparsely glandular eiliolate; corolla 
membranous to chartaceous. 3.9-4.1 mm long, the 
lube 1-1.4 mm long, the lobes lanceolate. 2.7-2.9 
X 1.6-1.8 mm. apieally acule. inconspicuously 

punclale am pin i I i In i ■ i i e li u 

the margins eiilire. hyaline: si. miens 2.9-3 mm 
long, the filaments l.l-l.() mm long, llie slaiuinal 
tube <).<> (1.7 mm long, the apieally free portions 

tale: pistil 3.8-4 mm long, glabrous, the ovary ob- 
long, 1.1-1.5 mm long, llie style 2.5-2.7 mm long, 
nearly epimctatc. the ovules 19 to 23. Funis glo- 
bose. 0.5-7.2 nun diam.. inconspicuously punclale 

Distribution. Ardisia colomdoana is endemic to 

Oeiro Colorado ami Cerro Hornilo along I hi- Con- 
tinental Divide, at the junction of Hocas del Torn 
and l.hiiii|iii. Panama, growing at 1200 to 1950 m 
in elevation. 

Ecology and conservation status. Ardisia colo- 
mdoana occurs in cloud forcsls. where il is locally 

Etymology. The name refers lo llie typ< 
tion on Cerro Colorado. 

Within Ardisia subg. Auriculardisia sect 

maaae. \rdisia < <>h>i adoana is most closely related 
to I. alropntpniea because of llie small calyx lobes. 
between 1.4 and 1.7 mm long, long pedicels, nar- 
row corolla lobes, and short anthers. Howe' 
colomdoana is easily distinguished I mm 1. 
neala and 1. alro/mmurea hv its thicker biaiichlets 
to 6 mm in diameter, shorter petioles lo 2.4 cm 
long, and wider calyx lobes lo 1.9 mm wide. 

Ardisia colomdoana is most easily confused with 
1. plciitidiotrva (seel. Fleumbolnae) based on llie 
leaf size and shape. However, it is easily separated 
by its terminal inlloresoeiiee and shorl pedicels lo 

■1,8 . 


pedicels to 1(1.5 mm long. 

. plea robot rya. 


Specimens examined. PANAMA. Boras «lil Ion.; 12 
mi. beyond Campamento Charm. 12 + 12 mi. Com l.'m 
San helix. 20 June I9o0 (fl). It. ir\rc\ HW, (II. YIOl: 

vide. 7 mi. from Charm Camp. 12 Apr. I 986 (ff), G. Me- 
Fhersnn HfM2{\.\.. MOl. Itonler oNiiM-a* ilel 'loro iiii.l 
Chi.i.,...': Cerro Colorado, road along top. 13 Aug. 1977 
Or). ./. Folsom el „l. 1710 |MO). < ho ,oo, rorlm.a Dam 

rid-eoi ualeish, ,1. 17. Ian. I0„0(lr). (,. Mel'hrrsnn I .KM, 

(FTG, MO, PMA). 

31. Ardisia couglomerata I irdell. \\ right 1,1 (r. 
71. 1979. Auriculardisia conglomerata (Lun- 
dell) Lundell, Phytologia 49: 343. 1981. 
TYPE: Panama. Veraguas: NW of Santa Fe, 2 
km from Fseuela Agrfcola Alto de Piedra, on 
ridge top below summit Cerro Tute, 28 Mar. 
1975 (fr). S. Mori & J. Kallunki 5275 (holo- 
type, MO!, F neg. 55675!). Figure 33. 

1.5-1.7 X 0.7-0.5 

. apiealb , 

la 1 1. I Ira art dels 

mixture of densely cupuliform ai 
1 1 - 1 > i < ! • > I < -cales. I. curs with blade 
liptic, 9.6-16.5 X 3.6-6.8 cm, apieallv acuminate, 
with an acumen 0.6-0.9 cm long, hasally ohluse. 
(leciirrent on the petiole, inconspicuously punctate 
and |)iinctate-lincalc. glabrous above at least with 
age. below with a mixture ol densely ciipulilnrni 
and lurfuraceous-lcpidote scaler, especially mi 
along the midrib, the midrib impressed abo\e. 
prominently raised below, the secondary veins 22 
to 21! pairs, pi'orni inil<>n> above and below, the mar- 
gins entire, revolulo: petiole ~1< ml. ttiar^iiiale. ()..'; 
1 .2 cm long, glabrous above at least with ;iuc below 
with a mixture ol densely cupuliform and lurfura- 

nalclv paniculate. 3-3.9 X 4-4.8 cm. obp\ ramidal. 



Iiranchlels. and pedicels with a mixture of densely 
cupuliform and furfuraceousdepidolc scales, the 
branches congested into 2- to 8-llowcred corymbs; 
peduncles to mm long; inflorescence bracts per- 
sistent, chartaceous. oblong. 1.2-1.4 X 0.6-0.8 cm. 
apieallv acute, scattered inconspicuously punctate 
and puiK late-linealc glabrou- above, below with a 
mixture ol scattered cupuliform ami lui -furaeeou- 
lepidole scale-, the midrib inconspicuous, the sir- 

i the inflores- 
cence bracts, but 4.5-6.7 X 2.5-5.5 mm; floral 
brads persistent, membranous, ovate. 0.6-1.2 X 
0.7-0.9 mm. apieallv acute, scattered inconspicu- 
ously punctate and punctate-lincatc. glabrous 
above, sparsely fulfil raceous-lepidote below, the 
mar-iu- irregular, minutely erose. hyaline, sparsely 
glandular ciliolate; pedicels stout, obsolete to 1.2 
mm long, inconspicuously punctate, with a mixture 

of scattered cupuliform and furfuract - Icpidd.- 

-cales. Fionas (measurements I mm a single, par- 
tial, old flower] 5-merous, color unknown; calyx 
lobes membranous to chartaceous. orbicular to very 
widely ovate, 0.9-1.2 X 0.9-1.2 mm. apieallv ob- 
tuse to rounded, prominently punctate and punc- 
late-lineale. glabrous throughout, at least with a-c 
the margins irregular, minutely erose. hyaline. 
sparsely glandular ciliolate; corolla membranous. 
2.4-2.6 mm long, the tube 0.5-0.7 mm long, the 
lobes ovate to narrowly ovate, 1.9-2.1 X 0.8-1 mm. 
apieallv acute, prominently punctate and pimclalc- 
hiieale. glabrous throughout, the margins entire, 
hyaline: stamens ca. 2.8 mm long, the filament- in 

the apieallv free portions ca. 1.2 mm long, the an- 
thers laneeoloid. ca. 1.0 X 0.0 imn. apically apic- 
ulatc. basallv subcordate. the connective ineon- 

glabrous. the ovarv oblong to globose. 1-1.2 nun 
diam.. the style 3-3.2 nun long, conspicuously 
punctate, the ovules unknown. Fruits globose. 1- 

ri.5 mm diam.. pmiuincnl b punclale. 

Distribution. Ardisia conglomerate, is known 

only from the Ivpe collection on Ocrro Tute in Ve- 
raguas. Panama, growing at 700 to 800 in in ele- 

Ecology i 

glnmerata occurs in cloud forests near the summit 
of Cerro Tute. because of il- restricted di-lribiilion. 
it should be considered threatened. 

Etymology. The specific epithet meaning clus- 
tered, often spherically so. referred to the small 
compact corymbs of the inflorescence. 

Within Ardisia subg. Auriculardisia sect. Pal- 
manae. \nlisia conglomerala is one of a number of 
specie- that have Iiranchlels covered with a mixture 
of dense cupuliform and furfuraceoiis I. pidole 
scales, \rdisia conglomerala can be confused with 
I. lundelliana owing to the coriaceous, elliptic 
leaves wild obtuse to rounded bases and the rela- 
tively stout, terete braiiehlds. However, \rdisia 
conglomerala is easily separated from 1. lundclliu- 

-horler p. -die 


bin Ardisia 

fused with Ardisia conglomer 
which also has Iiranchlels covered with a mixture 
of dens*- cupuliform and furfuraceous-lepidole 
scales and straight Iiranchlels. However. 1. m//g/o- 
merata differs by the braiiehlds terete, with large 
petiole scars, inlloi ■,>,■, nee- hipiunatelv paniculate 
ami obpvramidal. ihe pedicels obsolete to 1.2 nun 
long, the orbicular lo widely ovale wider ealv x lobes 
to 1.2 mm wide, the smaller corolla lobes to 2.1 X 
I mm. and the smaller anthers to 1.0 X 0.0 mm. 

Ardisia ronglomerata is known only from the ho- 
lolvpe. winch is in young fruit. A fragment ol the 
hololvpe is located at 1,1. and contains pieces of a 
-ingle Mower consisting of a corolla lube with two 
attached lobes ami two filament- with a -ingle loose 
anther. The floral measurements in the above <U - 
sciiplioii come from this material. It appears thai 
this corolla had failed lo completely fall from the 
expanding ovary. 

32. Ardisia crassipedicellata Lundell, Wrightia 
6: 73. t. 140. 1979. Auriculardisia crassipedi- 
cellata (Lundell) Lundell, Phytologia 49: 343. 
1981. TYPE: Panama. Veraguas: NW of Santa 

Missouri Botanical ( 

Fe. 4.2 km from Fscuela Agrfcola Alto de Pie- 
dra, 25 Fel). 1975 (fr). S. Mori & J. Kallunki 
4H:U) (holotype, \\\. V neg. 55643!; isotype, 
MO!). Figure 34. 

lepidote. Leaves with blades 
14.2-32.4 X 4.7-9.8 cm, apically acuminata, with 
an acumen I- 13 mm long. ba>alk acute, deourreut 
on the petiole, inconspicuous, ineoiispn ■uoiislv 
punctate an. I puuelale-lineate. essentially glabrous 
alune. Iiiiluraceoiis-lepidole below. I lit midnh mi 

long. 2.5-3 mm <H mi glahi »i - ib..\e. denselv fur- 
furaceous-lepidote below. Inflorescences erect, hi- to 
tripinnately paniculate. 6.8-32.4 X 6.2-26.6 em. 
pyramidal. usuallv longer than the leaves, the ra- 
chis. hranches. and pedicels densely fill I lira, colls 
lepidote. the hranches looselv congested into 3- to 
7-flowered corymbs; peduncles 0.4-2.8 cm long, 
the lower branches often subtended by leaves: in- 
florescence bract unknown: inflorescence branch 
bracts caducous, eharlaeeous. oblong. 7-8.2 X 2- 
2.3 mm. apicallv acute, prominenllv punctate and 
punclale-lineale. e»entiall\ glabrous aho\e. lurlii- 
raceous-le|.idole below, the margins entire, miiiule- 
Iv erose. livalme. > parse I \ glandular ciliolale: floral 
bracts similar to the inflorescence branch bi.nK. 
but 0.6-1.3 X 0.6-1.4 mm; pedicels stout. 7-12 
mm long, inconspicuously punctate and punclate- 
lineate, furfuraceous-lepidole. Fionas 5-merous. 
pink; calyx lobes coriaceous, oblate. 2.4-3 X 4- 
1.5 mm. apicallv rounded. uioonspiciioiisb pune- 
lale and punclate-lineale. glabrous adaxialK. fur- 
furaceous-lepidote abaxiallv. the margins irregular. 
ininuIeK erose. hvaluie. -paisdv glaiidulai eilio 
late: corolla coriaceous. 9.6-9.8 mm long, the tube 
3.7-3.9 mm long, the lobes lanceolate. 6.1-6.3 X 
2.8-3. I mm. apicallv acute. pioimii.nlh punctate 
and punel.ilediueate. glabrous throughout, the mai 
gins entire, hyaline; stamens 7-7.1 mm long, the 
filaments 4-1,2 mm long, the slaminal tube 2-2.1 
mm long, the apicallv free portions 2 2.1 mm long. 

X 1.2-1.4 mm. apicallv apiciilale. basallv lobate. 
the connective conspicuous!) punctate: pistil 0. ( >- 
7.3 mm long, glabrous, the ovary oblong. 1.8-2 mm 
long, the sl\ |e 5. I 5.3 mm long, prominenllv punc- 
tate and punctale-lmcate. the ovules 37 to 45. 
Fruits globose. 6.8-7.5 mm diam.. prominentlv 
punctate and punctate-lineate. 

Distribution. Ardisia crassipediccllata is en- 
demic to the Cordillera Central near and along the 
Continental Divide in Bocas Del Torn. Code, and 
\ciagius. Panama, growing at 350 to 900 m in el- 

Fcoli>it\ and . 
sipedicellata is found i 
cause it is relatively i 
sidered threatened. 

Flvmo/o^Y. The specific epithet was derived 
from the Latin "crassi" or thick and "pedieellata" 
meaning pedicelled. referring lo the thick pedicels. 

Within Ardisia subg. Auriculardisiu sect. Pal- 
manae, Ardisia crassipediccllata has flowers that are 
most similar to those of I. lirsncn because of their 
lame calw lobes || u | arc as wide as or wider than 
long, long and slender pedicels, relatively long co- 
rolla lobes for [be section. and lanceoloid anthers. 
However. 1. crassipediccllata is easilv separated 
bom I. Iicuicn bv lis Jens. I\ furfnraee.iiis lepidote 
scales, coriaceous leal blades, larger oblate calvx 
lobes to 3 X 4.5 mm. longer corolla lobes lo 6.3 
mm long, shorter and wider anthers to 3.8 X 1.4 
nun. and shorter slvles to 5.3 mm long. 

The type of Auric ul 'ardisia roseiflora Lundell is 
in voung bud. and without mature pedicels. While 

S/ici imcn.s 

along road ah 

\lrl%;s»,, ;: 

V s„. 7 km 

(11)../. Folsom 


Lundell. Wrighlia 1: . r 
1968. Auriculardisiu crassipes (I.undell) l.u 
dell. Phytologia 49: 343. 1981. TYPF: Pa. 
ma. Bocas del Torn: Bobalo Trail. N slopes 
Cerro Horqueta. 60(M)-7(MM) ft. 1 1829-2 1 
ml 5-7 Aug. 1947 (fr). P. Allen 4991 (ho 
type. MO!. I.I. neg. 1971-29!; isotvpes. C! |: 
OH!, LL!). Figure 35. 

Missouri Botanical Garden 

Plivlologia 19: .HI. 19151. T\ PK: Panama. Cliiriqur. 
rail to Cerro Horqueta, 6000-6500 It. | 1829-1981 
jii|. 15 Mav 1971 (fr). C. I'roctor :il<)25 (hololvpe. 
LL!, F neg. 55668!; isotvpes, K!, K n«- 61 UM, lon & lhe s, y ,e S" 5 

long, glabrous, the ovary oblong. 

TEX!). ovules 20 to 25. Frails globus,-. 5-6 mm diam.. 

Ardisia floritleiita I nn.l.-ll. \\ i iphli.-i <>: 75. |07*». Sui. n<>\. epunt'tate. 
\tiri< uhn<li<i<i Jlorulenta (Ptindell) l.uiulell, Phytolo- 
gia I0 : :m. I0JW. H PK: Panama, (".hiriqi.i: Cerro Distribution. Ardisia crassipes is endemic to 

I'an.l i the CiiMliii.-ntal .in, I the Panama- western Panama, in bocas del Toro and Cluriqui. 

Co-la !{„., l„„,l,-r. r;,. |(, km \ \\ ,.| II Halo ,1. o lmvm „ •„ |<> L ><) ,,, 22.5 I II, in elevation. 

\olr.,n. 2000 2I,".2 .... 21 .li.K P>75 Hh. S. l/„„ ,v ' , .... 

i /.w,.,, r:;n .i,,,,,,,,,. i i <. , „,, .v,u,H: is,,- '" w '*' v "'"' <—""" *"""*■ >"''-' <"- 

, ype vi()!). s '/"' s occurs alongside watercourses and footpaths 

in wet montane and cloud forests. It is locally in- 

Trees 8-20 m tall. Hranehlels stout, terete. 7-0.5 frequent, but appear- to be able to withstand mod- 

mm 'bam.. densely rufous lurfuraccous-lepidote. crate disturbance. I>e< ause of this observed resil- 

elliptic. 12.7-20.8 X 1,7-10.2 cm. apically acu- time. 

ruinate, with an acumen 0-16 nun lout;. basallv A7 w//o/og\. The specific mean- thick, 

acute, decurrent on the ))etiole. inconspicuously referring to the thick inflorescence rachises. 
punctate and punetate-lineate above and below. Within \rdisia -ubg. \iiih idardisia sect. I'al- 

sparsely to densely rufous lurhii'acectus-lepidole manae. Ardisia crassipes is closely related to A. me- 

abo\c and below, the midrib impressed above. tfislophvllu. 1. ugiitrreanu. \. cogolloi. and I. cur- 

prominently raised below, the secondary veins 15 lugnana because of its long calyx lobes with thick 

to 65 pairs, proiiiinulous above and below, the mar- petioles, \rdisia crassipes i- separated from 1. rar 

gin entire. Hut to iiuollod; petioles stout, marginal,-, ^istopliyl/a by its shorter calyx lobes to 2..? nun 

1-1.7 cm long. 1-5 mm diam.. furfuraceous-lepi- long, shorter corolla lobes to 1.5 mm long, shorter 

dote above and below. Inflorescences erect, hi- |<> anthers to 3 hum long, and -horlei -Ivies in 5 1 mm 

tripinnately paniculate. 16.5-20.6 X I 1-31.2 cm. long. However. Ardisia crassipes is easily distin- 

pyramidal. longer than the leaves, mixed, densely guished from 1. aguirreunu by its shorter leaf 

eupuliform and lurfmaeeoii-lepidole. the branches blades to 20.8 cm long, smaller branehlets to 0.5 

congested into 5- to 10-flowered pseudos|)icate cor- »»" '') diameter, shorter peduncles to 2 cm long. 

ymbs; peduncle obsolete to 2 cm long, the lower '"id wider corolla lobes to 2.3 nun wide. However, 

branches subtended by leaves: inflorescence bracts \rdisia crassipes is separated from 1. co^olloi by its 

and branch brads unknown; floral bracts caducous. smaller arrd wider orbicular, n vale calvx lobes 

membranous, ovate, 0.0-1.8 X 0.0-1.3 mm. api- 2.1 X 2.3 mm. longer and narrower anthers to 3 X 

cally acute, inconspicuously punctate and punc- ' mm - 'o' 1 ^' 1 " «tyl«-« to 5.1 mm long, anil smaller 

tate-lineale. glabrous adaxially. mixed lepidote aba- (mils lo ( > mi " «" tliameter. Ardisia crassipes differs 

xially. the margins irregular, minutely erose. ,mm '■ <<irta^mna by its coriaceous not membra- 

hyaline, sparsely glandular ciliolate: pedicels, when nmis l, ' ilf ' l,la,l «" s - whi«-h are tomentose above and 

distinguishable, stout, obsolete to 0.5 mm long, in- »»« glabrous^ below, thicker petioles to 5 i 
conspicuously punctate and punetate-lineate. 
mixed lepidote hirluiaeemi- lepidote. Floieers 5- 
ineroiis. 1 1 Li 1 1 1 pink: calvx lobe- ehartaceous. orbic- 
ular. 2.1-2.3 X 2.8-3.1 mm. apically rounded, 
prominently punctate, glabrous adaxially. furlura- 

nulely erose. hyaline, sparsely glandular ciliolate: 
corolla membranous, 6-6.1 mm long, the tube 1.5- 
1.6 mm long, the lobes ovate. 1.1-1.5 X 2.2-2.3 

aller flowers, but i 

■, | 

apically acute, prominently punctate and Specimens examined. PANAMA. Chiriqui: Ori 

, glabrous Ih.oilghoul. the margins f' carrelera liaeia Ml,. ( .1,,, |..e.o. 20 Dec. 1 007, ( 

entire, hvaline: stamens 5.6-5.7 mm long, the 1,1- "»»"■>»*>» /l'/5 (MO. STHh: Disirii,, <le |{ „i 

'*r a, i ■ •,,,,. Horqueta. 255 June lOtUlfr). /. f'inuh 7'" ' v ' 

aments 4.5-4.6 mm long .1, tarmnal tul l-l.l ll( . a A :t . n , ( p aUl (l( . Maelu, 7 Jan. 1085 , 

mm long, the aprcallv bee portions 3.^-3.6 mm & I). Rnuhick /.;.>'/ lITO. PI. M0). 
long, the anthers lanceoloid, 2.8-3 X 0.0-1 mm. 

apically apiculate-subulate, basally cordate, the 34. Ardisia erassiramea Standi. 

Mus. Nat. Hist.. But. Scr. 18: 887. 1938. Au- 
riculardisia crassiramea (Standi.) Luntlell, 
Phytologia 49: 343. 1981. TYPE: Costa Rica. 
Alajuela: Cerros de San Antonio de San Ra- 
mon, 1200-1245 m, 26 July 1927 (fl), A. Bre- 
nes 5652 (244) (holotype, F!, F neg. 68143!, 
LL neg. 1971-30!). Figure 36. 

Ardisia solomonii Lundell. Phytologia 18: 135. 1081. 

Wiidih.i 7: 2ir.. I 1 ';;.",, n IT: Costa Him. Pun 
nas: Monteverde, cloud forest at edge of Contin 
Divide and on Pacific side of slope, 10°5 
081° 18' W. < a. 1300 m. 18 Aug. 1976 (fl), J. Soh 
5M2 (holotvpe. 1.1.!. NY neg. 11078!, 11078A! 

. \HI , 

Syn. no, 

': 269. 1984. 

1818 |1819], non Ardisia mirranlha 
Donn. Sn... Hot. Gaz. 14: 27. 1889, nom. illeg. Ar- 
disia dryeri Eundell, Phytologia 61: 63. 1986, nom. 
inval. Ardisia dryeri Pipnlv K Uickelson, Sida 18: 
512. 1998. TYPE: Costa Rica. Alajuela. Puntarenas. 
or Guanacaslc: Finca de Monteverde, Cordillera Ti- 
laran. along the Window Path. Continental Divide. 
1560-1580 ,n, 15 Aug. 1976 (fl), V. Dryer 577 (ho- 
lotvpe. MO!: isolvpc. K!. V 68322! ). 
eulardisia spalhulala l.undcll. Phytologia 63: 75. 
1987. Syn. nov. Ardisia spalhulala (Lundell) l.undcll. 
Phytologia 63: 463. 1987. TYPE: Costa Rica. Gua- 
nacaste: pasture adjoining main road, and Chomogo 


h d 


Shrubs or trees to 10 m tall. 8-15 cm diam. 
Branchlels straight, slmil. angulate. longitudinally 
ridged, the petiole Mais small, inconspicuous. 5 
I I inin diam.. w ilh a mixture of densely cupuliform 
and furfuraceous-lepidote scales. Leaves with 
1. 1. ides rliarlaeeous. elliptic or oblanceolalc. rarely 
ohovate. 5.8-25.2 X 1.8-5.8 cm. apically acumi- 
nate, with an acumen 0.5-2 cm long, basally acute, 
decurrent on the petiole. inconspicuously punctate 
and punctatc-lincalc. essentially glabrous al»o\e. 
with a mixture of densely cupuliform and liuhira- 
,>.i - epidnic -cales. moiv densely so along the 
midril). the midrib impre-sed above. prominently 
raised below, the secondary veins 17 to 23 pairs, 
promiiiulous above and below, the margins entire, 
flat: petiole slender, canaliculate. 0.5-3.5 cm long, 
essentially glabrous above, below with a mixture of 
densely cupuliform ami In rl 'uraceous- lepidote 
scales, fiijlorrsci-nees erect. Iripiimalely paniculate. 
9-28.5 X 6-22 cm, pyramidal, usually longer than 
the leaves, the peduncle, the raehis. secondary 

el- .•. ill 

cupuliform and (urfui 
branches loosely congested in 
corymbs; peduncles subobsoleti 

i mixture of densely 
lepidote scales, the 
ito 7- to 9-flowered 

lowei branches subtended b\ leases: inflorescence 
brads unknown: inflorescence branch bracts ca- 
ducous, membranous, oblong, 0.5-3.5 X 0.9-3 
mm. apically acute prominently punctate and 
piniclale-lmeale. glabmu- above, the margins irreg- 
ular, minutely erose. hvaluie. glandular 
ciliolale: floral brad- similar to the inlloresccncc 
branch bracts, but 0.6-1.2 X 0.4-0.8 mm; pedicels 
stout, 2-4 nun long, incoii-pii unii-lv pum lale and 
punctate-lineale. with a mixture of densely cupu- 
liform and furfuraceous-lepidote scales. Flowers 5- 
iiiemiis. while to red or purple after pollination (Pi- 
polv. pers. obs.): calyx lobes membranous to 
chartaceous. ovate. 1-1.2 X 1.3-1.4 mm. apically 
acute, prominently punctate and pun. lale -linealc 
glabrous adaxially, -par-el\ furfuracenu- l.pi.loie 
abaxially. minutely erose. hyaline, sparsely glan- 
dular ciliolale: corolla membranous. 4.4-4.7 mm 
long, the tube 0.9-1 mm long, the lobes elliptic to 
ovate. 3.5-3.7 X 1.9-2.2 mm. apically acute, 
prominently punctate and punclale-luieate. gla- 
brous throughout, the margin- entire, hyaline: sta- 
mens 3.3-3.5 mm long, the filaments 1.6-1.7 mm 
long, llie slaminal tube 0.5-0.8 mm long, the api- 
cally free portions 0.9-1.1 mm long, the anthers 
lanceoloid. 2.1-2.4 X 0.6-0.8 mm. apically apic- 
ulate. basally cordate, the connective inconspicu- 
ously punctate: pistil 1.2-1,8 mm long, glabrous. 
the ovary oblong. 1-1.4 mm long, the style 3-3.0 
mm long, inconspicuously punctate and punctate- 
lineale. the ovules 17 to 19. Fruits globose. 4.5- 
5..". mm diam.. prominently punctate ami pum lale 

Distribution. Ardisia t 
from the Cordillera Guana, aste |\olcan Cacao) in 
Guanacaste. and from the Cordillera Tilaran in the 
vicinity of the Monteverde Piological Reserve, at 
the junction of Guanacaste, Alajuela. and I'unl.i- 
renas. Costa Rica, growing at ll(K) to 1700 m in 

Ecology and c 

siramea occurs on windy, exposed ridges in dim 

rain forests. Fieldwork by Pipoly has shown that 
the species occurs in light gap- and along lon-sl 
margins. While it is locally common, and the areas 
in which it occurs arc somewhat protected. I he spe- 
cie-should be considered threatened because of il- 
very restricted distribution. 

Etymology. The specific epithet was derived 
from the l.alin. "crassi-" meaning thick and "ra- 
mea." for branch, referring to the thick inflores- 

Within Ardisia subg. Auriculardisia sect. Pal- 

-.,.«',',,;■„ lli,!l aave branch els covered vi4 a mixture 
of dense oupuliform and furfuraoeous-lepidote 
- I i , ., i In inn -i < ,i-i]\ « nii- 

fused with A. conglomcrata, with their straight 

ealw lobes to onlv 1.2 inni wide, and shorter co- 
rolla lohes to 3.7 mm long. However, the ungulate 
brauchli ts will riii il r "•:!»< i ll< 

idpueialeh ;>;inicida'e aed \ >\ ramidal. the ovate 
longer calyx lohes to 1.4 mm wide, the larger co- 
rolla lohes to 3.7 X 2.2 mm. and the larger anthers 
to 2. 1 < 0.8 mni distinguish \. < rassa unr, i. i 
A. conglomerate,. 

The t\pe ol !;■'';• •-, •./,..»'-.,;.; ,\;|s collected in 
bud, and evidently Lundell (1981a: 136) never 
i mi, ii. i! n ii i ■ i i.' >l I ,,' .-< ,<" •„■ Ih i in 

,,s,s ol 

, lad. I .midell's 

lated to the genu I • > ;,„> • •'>,• i ' - < • , ! 
However, the type of A. solomonii perfectly matches 

The type of Auriciilardisia. micrantha is unique 

onlv in the more open inllorescenees and slightly 
larger flowers with what Lundell (1984a: 269-270) 
described as "small dentate, dcuseh punclale lo- 
liaceoiis bracts." However, our studies of the ma- 
terial show that the "dentate"" brads are siinpb an 
artifact of the diving process, ami the brads aie 
actually entire. The slightly larger flowers and more 
ii ,i 

< in be. ausi ol i iii i i s|)c« imeu. ( urn 

pared to the inflorescences of the Ivpe ol \rdisiu 
crassiramea. which is congested because the llovv- 

The type of Auriciilardisia spalhulala was sepa- 
rated by the ••distinctive spallmlato leaves" (bun- 
dell, 1987: 463). However, .loser examination 

show s il'.i die apices ol the leaves are all deform, ii, 
which gave the appearance of wider apices. The 
popul .ns represented bv the Ivpes ol \rdisia 

\ ,' • , >■ i<-r •'•,,, , ■ > , • 
ha, and 1. s/xilliulala are all from the same area 

;!• d i-i- b i: ■•.-■' ' |; "-'! e \ i . . > Moii- 

teverde. and all are similar in appearance and 

ample, light exposure and ; 

Palma de San R 

(451) (F, NY); Cord] 

Ramon, 8 Aug. 192 

I ((!). 

/ Pipoly 7094 

(CI!. MO. NY). 7696 (CIS. K NY). C.ana.asK-: I'a,,|ue 
Naeional Ouanacaste. Lsta.aon Cacao. 23 Nov. 1990 if. I. 
/.'. Kspinoza 66 (INIi. MO): Canton (le Liberia. I'arque 
Naeioual ( ,uatia. -asle. Cordilleia Ouaiiaeasle. \olc;in ( ia- 
ra„. 12.lub 1996 (IV). J. Morales el al. 7W.W I ll .. INIi. 

I „ I >,".> . .(I ' ' . I I ' 

I.I . \|(»: Caiilon .le I'luilarenas. Ooiddlera Tilaian. San 
Luis. Cerro Banquet.-, trail In Surluhal. \1..nlev erde. 22 
Mar. IWI (II). /. Fuenles 71)1 (Id!. ITC. INI!. Mill: I!.- 
>rrvil Moiilevevde. 2(> \ug. 1985 ill). I!. Habei & E. Bello 
( '/_■/: iVIO): Keserva Biologic-a Monteveid, l!,illa„i, 
Trad. Omiineiilal Divide. 13 Mar. 1989 (fr). It. Ilaher & 
ll. Zucluwski 0121! (CK. LT0. IMC MO): Keserva Bio- 
logic-a Monteverde 

- !,, d - 

I ; i < \ 

i (CR. LTC. INI!. MO): 

all. II- llle CIliimoiM) n ) . \L> l< >.",:!" < .1 .11.1 l-nresl |{, 
srnr. 2 1 Aug. 1975 (II). C. Hartshorn I Till) I K MO): .-a. 
2 km SL of Moiileveide. Pacific «ale,s|„,|. .! I Julv 1980 
(II). R. IaiwIoi, I2IU |I9: VI, .I, lev erde. 20 Ma,. 1959 ill;. 
C. I'ohne, //.) (N\): Monlev erde. ( limning.. Trail. Ilrsrivi 
limlogiea. I! \ug. 197.-, Isteh. L. I'oreda 1. 1 1 If, (L. MO). 
lii.nler ol" \l..|m I i Onaii n .-I. iikI I'.iol itii:i> 
Keserva flora de Monteveid.'. Continental Divide. Hri- 
llante Trail. II Dec. 1976 (IV). I. Diver 107!', ll. MOi. 
Border lietween Alajmla and Puiitareims: vvindsuepl 

, : ont 

levenle. C 20 \ 

ll-l: M.mlevenle lieserve. Cordillera Tilarau. Cerro Cen- 
linela. 19 Aug. 1981 (II. IV). M. Cnnvm &■ l> Sleeper HM7 
ILL. \|()):s,.||lemenl al Monlev erde in N L seelion of, He- 
serve. 5 Mav 1975 ((I). ./. Vtley & k. llle, 2.' I <11 (L |2|). 

lisia croatii Lundell, Wrightia 5: 281. 1976. 
Auriciilardisia croatii (Lundell) Lundell. 
Phytologia 49: 343. 1981. TYPE: Panama. Ve- 
raguas: valley of Rio Tercero Iba/.o beyond Ks- 
cuela Agricola, Alto Piedra above Santa be. 
along road, 29 Aug. 1974 (fl), T. Croat 27327 
(hololype. LL!: isolypes. V1LNI !. MO!). 

irubi . 

1 20 i 

, 4.2-16 . 

V... OHM: • - 

and miiiutclv appressed rufous lurfuraceous-lepi- 
dole. Iran's vvilh blades membranous to charta- 
eemis. ■ > I > I ■ ■ n ■_• _ elliptic, or i.uclv uarrovvlv oblanceu- 
late. 4.1-48.5 X 1.4-13.9 cm, apically acuminate, 
with an acumen 0.4-1.7 cm long, hasallv asym- 
metric, acute. deeurrenl on the petiole, pionuin nib 
punctate and punctate-lineale. glabrous above, 
scattered rufous fur Im.i. eoiis-lepidole below, 
denselv so along the midrib, the midrib impressed 
above, prominentlv raised below, the secondarv 
veins I.", h. ii.'i pairs. impressed aho\e. prominulous 
below, the margins entire. Hat: petioles slender, 
canaliculate. 4-24 mm long. 1-3 mm diam.. gla- 
brous above, rufous lurfuraeeous-lepidnle below. 

6.2-32.8 X 4.1-21.2 cm. pyramidal, usually longer 
than the leaves, the raehis. braiichlels. abaxial 
bract surfaces, and petioles densely and niiuiilely 
appre--. d rulous f'urfuraceous-lepidote. the blanch- 
es looselv congested into 3- In 0-|loweied corvmbs: 
peduncles obsolete to 3.7 cm long, the lower 
branches subtended bv leaves; iullnrescence bracts 
iinkiiowu: inflorescence branch bracls caducous. 
membranous, oblong. 5.5-23.5 X 2.5-15.2 mm, api- 
eallv acute. |>rominently punctate and punctate Im- 
eale. glabioiis above. iiilou- I iirlu raceous-lepidote 
below, the margins entire, flat: floral bra. I- similar 
to the inflorescence branch bracls. but 3.2-5.9 X 
1.8-2.3 mm; pedicels slender. 8-16 mm long, in- 
conspicuously punctate and piinclate-lineate. ru- 
fous lurfiiraeeoiis-lepi.ini. . Fionas 5-merous. 
while, cream, pink to light purple: calv \ lobe- ehar- 
taceous to coriaceous, oblate. 2-2.9 X 2.5-3 mm. 
apieallv broadlv rounded to Iruncale. inconspicu- 
ouslv punclale and punclale-lineatc. glabrous ada- 
xiallv. rulous hirlura. eou--lepido|e abaxiallv. the 
margins irregular, mimilelv .-rose, hvaline. sparselv 
glandular ciliolale: corolla chartaeeous. 6.7-84 
mm long, the lube I .(>-3. 1 mm long, the lobes 
ovate. 4.8-5.2 X 3-3.6 mm. apically acute, incon- 
spicuouslv punctate and punclalc-lineale. glabrous 

through the margins entire. hyaline; stamens 

4.7-6.5 mm long, the filaments 2.5-4.1 nun long. 
the slaminal tube 0.9-1.2 mm long, the apically 
free portions 1.3-3 mm long, the anthers narrowly 
ovoid to lanceoloid. 3-3. I X 1.1-1.7 nun. apieallv 
apiculale. hasallv deeply cordate, the connective 
eotispi. uoiisK punclale: pistil 1.2 -8.5 mm long, 
glabrous, ihe ovary oblong. 1.7-2 mm long, the 
style 2.5-7 mm long, inconspicuously punctate and 
punctate-lineate. the ovules 35 to 49. Fruits glo- 
bose. 7.7-10.8 mm diam.. prominentlv punclale 

fhoma- II. Crnnt. the P. 
Missouri Botanical Garden. 

Within Ardisia subg. Auriculardisia sect. I'al- 

from 1. inlunilulosoinar^inala bv its huge, oblale 
calyx lobes to 2.9 mm. with broadlv rounded to 
truncate apices. Iighllv appre--<d rulon- 1 1 ■■ 1 1 1 1 ;i 
ceous lepidote scales on the hianchlels. mllores 
cence parts, and pedicels, the secondarv veins of 
ihe leal blade-, impressed above and proiiunenllv 
rai-ed below, the larger corolla- to 8.1 mm long, 
and larger anthers to 3.4 X 1.7 mm. Ardisia cmtttii 

nized subspecies. 

35. Ardisia croatti subsp. correae (Lundell) 

Ricketson & Pipoly, comb, et slat. nov. Ardisia 
correae Lundell, Wrightia 6: 72. 1979. Auri- 
culardisia correae (Lundell) Lundell, Phytolo- 
gia 49: 343. 1981. TYPE: Panama. Chiriquf: 
NK of Fortuna Dam cam[). Hardwood, trail In- 
ward Lima Landau, 08°45'N, 082° 1 5' W. 
1000-1200 m, 21 Sep. 1976 (fl). M. Correa A.. 
R. Dressier, /V. Salazar, J. Mediety. C. Garibal- 
di, F. Ham urn & T. Bailees 2690 (holotype, 
MO!, F neg. 55681!: isotype, PMA!). Figure 

.1 from Patricia Pilar to 24 
a, 2 Oct. 1979 (fl), C. Dod- 
mpp 8677 (holotype, MO!; 

Tives 2-20 in lall. 1.2-16 cm diam. Iliunchlcis 
1-6 mm diam. traces with blades 17.5-18.5 X 
5.1-13.9 cm. with an acumen 0.9-1.7 cm long, the 
secondarv veins U* In 03 pairs; petioles 15 2 I mm 
long. 2-3 mm diam. Inflorescences 19.4-32.8 X 
8.3-21.2 cm. 5- to 9-flowered corymbs; pedicels 
1 1-10 mm long. Fionas with calvx 2.0-2.8 X 2.7- 
2.9 mm; corolla 8-8.1 mm long.' the tube 2.9-3.1 
mm long, the lobes 5-5.2 X 3-3.2 mm: stamens 
0.3-0.5 mm long, ihe filaments 3.9-1.1 mm long, 
the slaminal tube 0.9-1.1 mm long, ihe apieallv 
free portions 2.8-3 mm long, the anthers 3.3-3. I 
X 1.4-1.5 nun; pistil 8-8.5 mm long, the ovarv 
1.5-2.2 mm long, ihe style 5.8-7 mm long, the 
ovules 35 to 43. Fruits 8.2-10.8 mm diam. 

Distribution. Ardisia croatii subsp. correae is 
known Iroiu lliree disjunct areas, hum ncai the For- 
tuna Dam. Boras del loro. Panama, and in Lena- 

cloud I 

r is most frequently found in I 
and upper premonlano forests, 
■stricted (listril)iition. it sliould he 

sidered threatened. 

Etymology. This suhspeeies was nam 
or of Mireya I). Correa. eurator of the P; 
National I lerharium. 

\\ illiiii Ardisia s.d.g. !».'•■ 
manor. \rdi-.ia iroatil suhsp. corrcor is 
from suhspeeies croatii hy a series of q 
characters, ineludini; lis longer petioles 
long, longer calw lohes 2.8 mm long, na 
rolla lohes to ."1.2 mm wide, and longer 
3.4 mm long. Despite the similarity l>< 
suhspeeies in ahsolule altitudinal range, i 
correac is found more often in eloud and, 
extent, premonlane or n 


The type of \rdisia dodsonii is unique only for 
its slighth longer calw lohes and pedicels: other- 
wise it matches \rdisia croatii suhsp. correac in all 
other aspects. 

Specimens examined. I'\N\M\. I'.oea- del lor.. 

hetwren lil.ena \ i-la Cnlle linea and Crrrn I'llon.on Chi- 
,-i,,,ii Trail. 17 \pr. !'«»;; ih). ./. KidhrideX ./. />«/„• o,'0' 
(\1(). M I: 8.5 nia,l nn. Iron, bridge near Fortuna Dam ..n 
load Inwards Chiriqui Orande. 1.3 mi. N of Continental 
Divide. 10 Mill. I9i!5 (In. C \h I'hcrson <u7,2 (FTO. IT. 
MO. I'M A). (>7r,i (\1()i: alomr, ,.i,..-lim- road in area of 
Fortuna Dam. near end ol mad. .-a. <". Mar. !<).",(, (In. ... 
Mcl'hersnn H7()h (FTO. IT. MO. I'MU F<T \DOK. Ks- 
m.ral.las: O.iuunde Canton. I'.ilsa IS.olog.eal Nation. 
Mach<§ Mountains. 35 km \\ of Ouuiinde. 5 km \\ of 
Santa Isabel. Dogala 30 Oct. IWI (Ir). ./. Clark cV' 
/>. \eill 2:il (MO. OONKi: Invader Had. II Nov. IWI 
0.1. /. Clark a al. 2.V, ( I- 10. MO. (KINK). 20 Nov. 1905 
I IV). ./. C7„,/, n al. 1077 1 1- TO. MO. OCNF). (slcr.i. / 
CAW. _'/// I FTO. MO. OIAI ): Maehe-Clundul Feological 
|{eser\e. Milsa ISiolom. al Slation. Maehe Mountains. 35 
km \\ nl Oiiininde. 30 V '" 

I I I nn I I li 1 1 
lOiil. lYI'F: Panama. \n;i:ii;h: Ion-ted 
\\ of Alto de Piedras. \\ of Santa Fe. 
[975-1707 rn|. » Sep. 1070 (II). It. 
Hummel 4599 (holotvpe. MO!. F 55625!: iso- 
type. LL!). 

Shrubs or trees 2-12 m tall, to 10 cm diam. 
Branchlets 2-1.5 mm diam. Lanes with blades 4.1- 
25.2 X 1.4-6.7 cm, with an acumen 0.4-1.2 cm 
long, the secondary wins 18 to 33 pairs; petioles 
4-13 mm long. 1-2 mm diam. hi/lorescciaes 0.2- 
20.1 X 4.1-26.3 cm. 3- to 7-fIowered corymhs: 
pedicels 8-14 mm long. Flowers with calyx 2-2.5 
X 2.5-3 mm; corolla 6.7-6.0 mm long, the tube 
1.6-2 mm long, the lohes 4.8-5.1 X 3.4-3.6 mm: 
stamens 4.7-5 mm long, the filaments 2.5-2.7 mm 
long, the stamina] tube 1-1 .2 mm long, the apiealh 
free portions 1.3-1.1 mm long, the anthers 3-3.1 
X 1.5-1.7 mm: pistil 4.2-0 mm long, the ovary 
1.7-1.8 mm long, the style 2.5-1.2 mm long, the 
ovules 43 to 49. Fruits 7.7-9 mm long. 

I Panama, in Code. Yeraguas. 

1 believe it is not t 

characters, including its -I i- •! -1 petioles to 13 m 
Ion-, u ider calw lohes 3.0 mm Ions-. shorter eon.l 
lobes to 5.1 mm long, and shorter anthers to 1 
mm long. See under suhspeeies carina- foi 

The type of Ardisia retusa is unique only for i 
rounded, somewhat reluse apex of the calw lohe 
and pedicels longer than the median for the sp< 
, ies. Otherwise it I. craalii siil.-p. croai 

I'ANAMA. Code: Fl Potroso. 

Feb. 1977 (fr), J. Folsom & L. Collins 

si at hase of Can, I'llon abme Fl \alle. 
1. (renin .Ho 19 (].|.. MO, NY); sawmill 

A. Pitman & J. CI 
1 1 lull 1 1 t 

36. Ardisia eroatii suhsp. croatii. Figure 38. 
\rdisia relusa Funded. Vv right iu (r. 88. 1979. Syn. no 

Volume 90, Number 2 

Vrraijiias; Cerro Tut.', trail past agricultural school near 
Santa h. IT N-,. !<>.«> Hi. / \ni..n„, I!U<) il I. M< i. 

M I .'. lr.,.1 I |.ul( liile.OOel. |079<|||. / \, ,..,,;„ 

I'M! (IT. MO). 3 \|)i. 19C0 <IV). T. \nlonin :>,>>!!.-> (IT. 
Mill. |||) / \nlnn i III) I .n I l!i. Ire- Pia/o- 

9 km from Santa le. 24 JiiU 1971 ((I). /.' Cr»„l I'.i.V'.l 
(CAS. IT. MO. M I: along road belueen Kseuela \gncola 
:in.l Ml.. Piedra. Simla f e. a.r.l K lo Dos boeas ea. 
5-8 km from Ksn.rla. 20 Julv 1971 (III. /.' Coat 2501)7 
(I.I. (K neg. 55W.ll. MO. NM: along dirt roa.l between 
Santa be ami Ki'i. San Luis, bevond Kseuela Circle Mlo 
.1.- Piedra. 5.9 mi. N of school. 28 June 1987 lilt. /. Cat 
MUM (ITO. 1. 1.. MO. I'M \): I rail on ridge to summit of 
Cerro Int.'. Cordillera Tute. I km pasl Kscuela \gviola 

Altos de Piedvas. W ..(Santa le, 15 Dee. I9{-,1 (IV). S. 
Knap/, <£ K. Svlsma 2193 (IT. MO. W): along trail to 
summit to Cerro Tile. ea. 3 km above Ksetiela \gvicullma 

Mlo P.edva near Santa Le. I Jan. 1981 (fl). K. Svlsma X 
7. \„„„,in .SOU (II.. MO!. (IV). K. SMsmn & T. \ntanio 
3015 (LL, MO). 

» Lundell. Wriglilia I: 58. 
1968. Auriculardisia darienensis (Lundell) 
Lundell. Phytologia 49: 343. 1981. TYPF: 
Panama. Darien: Dtto. de Chepigana. crest. 
Cana-Cuasi Trad. 5500 ft. |I676 nt|. 15 Mar. 
1940 (fr), M. Terry & R. Terry 1563 (holotype, 
F!. F neg. 68145!; isotypes, A!. MO!, LL neg. 
1971-41!). Figure 39. 

Shrubs or trees to 10 in tall. 14 em diam. Branch- 
lets slender, terete. 3.5-0.5 mm diam.. densely and 
minutely appressed rufous furfiiraceous-lcpidote. 
Leaves with Made-;. elliptic lo ..(.ovale 
1 1.2-28.5 X 3.9-9.8 em. apieallv subacuminate to 
i 4-12 . 

i long, hasally 

spieuously punctate and punctate-line; 
and below, furfuraceous-lepidote above a 
the midrib impressed above, prominently raised be- 
low, the secondary veins 55 I.. 70 pairs, prominu- 
lous above and below, llie margins entire, flat; pet- 
ioles slender, marginal*-. 8-28 mm long, 2-3 mm 
diam., furfuraceous-lepidott- above and below. In- 
florescences erect, bipinnately paniculate, 8-20 X 
4-14 em, pyramidal, usually shorter than the 
leaves, the peduncle, raehis. branehlets. abaxial 
bract surfaces, and pedicels densely furfurnceoiis- 
lepidole. th<- branches h.oselv congested into 4- to 
7-llowered corvnih-: peduiiele- nearly obsolete to 
2.8 em long, the lower branches subtended by 
leaves; iiillorescene*' braels unknown: inflorescence 
branch bracts caducous, membranous, oval*-. 4.2- 
6.5 X 3.3-5.2 mm. apieallv acute, prominently 
punctate and punctate-lineate. the margins entire. 
hvalme; floral braels similar to the inflorescence 
branch bracts, but 3.9-6.2 X 2.8-1,1 mm: pedicels 

long. 1,8 5 x .',.<) I.I ,„,„. apieallv a*ut<- lo round- 
ed, the subapieal notch almost medial, so as to 
make llie lobe ap«-\ appear einarginale. conspieu- 
ouslv piuielale and puuelale-lin*-ale. glabrous ,nla 
\iallv. furfurac<«lol<' abaxially. tin- margins 
irregular, minutely erose. hyaline, sparsely glan- 
dular ciliolale; corolla cliarlaceous and coriaceous. 
9.1-9.4 mm long, the tub*- 3.8-4 mm long, the 
lobes narrowly oval*-. 5.1-5.6 X 2.6-2.0 mm. api- 
eallv acute, prominently punctate and puuclale-liii- 
eal. . -lain. .us ihroughoul. the margins entire, hy- 
aline: stamens 7.3-7.5 turn long, the filaments 4.2- 
4.4 mm long, the staminal tub.- 2-2.2 mm long, the 
apieallv fret- portions 2.2-2.4 mm long, the anthers 
lanceoloid. 3.4-3.6 X 1.2-1.4 mm. apieallv apie- 
ulate. basally lobate. the connective coii>pieuou-lv 
punctate: pistil 7.8-8.1 mm long, glabrous, the ova- 
ry oblong. 1.5-1.7 mm long, the style 5.9-6.4 mm 
long, inconspicuously punctate and punctate-lin- 
eate. the ovules 15 to 18. Fruits globose. 7.2-8.4 
mm diam.. inconspicuously punctate and punetale- 

Distribution. Ardisia darienensis is endemic to 

Darien and San lila.-. Panama, grow ing from 610 to 
1981 m in elevation. 

Ecology and conservation status. \rdi<ia daiie- 
nensis is a common uiiderstory tree in cloud and 
elfin forests, on steep slopes and ridge-. I!eeau-e 
of its restricted distribution, il should be considered 

F.lytuolo!t\. I'he -peeilic epithet referred to the 
loealilv in the llarien vv here the type was collected. 

Within Ardisia subg. \uriculardisia sect. Pal- 
manae, [rdisia darienensis is most similar to I. _gc- 
neralensU due lo its elliptic or ol.ovale leal blade-,, 
bipinnate panich-s. and deiis.-lv corvmbo-e ll.meis. 
However, it is easily separat*-d by the longer and 
thicker marginal.- petioles to 28 X 3 mm, charta- 
(toii- leaves wilh moie numerous secondary veins 
to 70 pairs, and the apieallv emarginate caly\ 

Specimens examined. PANAMA. Darien: Cerro Pirn 1 , 
11 Apr. 1967 llv). V llnslan 59.7 (MO): on C.-rro Pine. 
I f Dee. 1902 (IV). J. Ihike 650/ (IT. MO): Cerro Pine. 
9-10 Aug. 1907 (II), ./. Duke & T. Klias 13750 ((,11. IT. 
MO. t S): vi.lge lop area N of ( l.-rn. Pine, between Cerro 


'. !,ii! : 1 \ 

.48. Vidisia ,luki-i Luiidell. Wrighlia 1: 15. 1008. 
Icacorea dukci (I, urn Id I) Lundell, Phytologia 
49: 348. 1981. Auriculardisia dukci (Lundell) 
Lundell, Wrightia 7: 267. 1984. TYPE: Pan- 
ama. Darien: peak between Rio Hales & Rio 
Aretf at their confluence, ca. 300 It. [91 in], 
13 Sep. 1906 (II). J. l)„kc 8741 (hnlctvpe. 
MO!: isutypes, LL!. US!). Figure 40. 
Trees with height unknown. 10-10.5 cm diam. 

densely lurluraceous-lepidole and cupuliform 
scales. Leaves wilh blades memltranous. « 1 1 ipl it to 
widely ol .ovale. 20.9-31/) X 8. 1-1 1.4 cm. apicallv 
acuminate, with an acumen 7-1 I mm long, hasally 
cuneate. decurrenl on the petiole, inconspicuously 
punctate above and helow. fiirfuraceousdepidoto 
above, with a mixture of fiirluraceous-lepidote and 
cupuliform scales below, the midrib impressed 
above, prominently raised helow. the secondary 
veins 00 to 08 pans, proininiiloiis abo\e and below. 

niarginale. 0-21 mm long, fiirluraceous-lepidote 
above and below, Injlorcs, em es eiccl. |»i- to Iripin- 
natelv paniculate. 27.8-28.9 X 12-21 cm. pv ra- 
midal. longer than the lea\es. Inrl u i a<«<>us-b-| >i- 
dote. the branches looseb congested into 3- to 
8-(lowered corymbs: peduncles obsolete: inflores- 
cence bracts unknown: inflorescence branch bracts 
unknown: floral bracts usually persistent, membra- 
nous, ovate. 0.9-1.1 X 0.7-0.0 mm. apicalb acute, 
prominently punctate and piinclale-bneate. gla- 
brous above, furluraceoiis-lfpidole below, the niai- 

glandular cibolale: pedicels slender. 0.0-1.5 mm 
long, inconspicuously punctate and punctate-lin- 
eate. indumenl as in the branehlels. Flowers 5-ine- 
rous. cream: calyx lobes inemb 
ceous. ovate. 0.9-1.1 X L. 1-1.3 mm. apical!) 
acute, prominently punctate and piiiielate-liiieale. 
glabiou- a d ax i ally, furfuraceous-lepidotc abaxialb. 
the margins irregular, muiiiteb erose. hyaline, 
sparsely glandular cibolale; corolla and stamens 
unknown: pistil in vouug fruit 3.2-5.0 mm diam.. 
glabrous, the ovary in young fruit globose. 1.2-2 
mm long, the sl\ les in \oung Iruil 1-3 mm long, 
prominently punctate, the ovules (according to Lun- 
dell. 1968) 22 to 21. Fruits unknown. 

nail peak in Darien. Panama. 

anil was collected 

: ca. 91 m in elevation. 

Ecology and conservation sta 

tus. This species 

known from one of the most 

remote but highly 

vploite.l areas of the Darien. si 

> its rarity is prob- 

>lv a realitv and not a collecting artifact. There- 

ire, we expect that the species 

is at least threat- 

led if not in danger of extincli. 

FimiioIo^. This species was named in honor 

who has dedicated his life to the understands 
economic and medicinal lro|)ical plants arouix 

\\ ithiu \rdisia subg. Auriculardisia sect. 
rnanae. Ardisia dukci is most similar to A. at 
cayana and A. palmana because of its short ( 
lobes that are wider than long, menibranou 
ehartaceous leaf blades, and short pedicels. I 
ever. Ardisia dukci may be separated from but 
anchicayana and A. palmana by the shorter c 
lobes to 1.1 X 1.3 mm. and shorter styles to 3 

piano (P. H. Allen) Lundell. Phvlologia 10 
343. 1981. TYPE: Costa Rica. Punlarenas 
(.anion de Osa. vicinity of Ks(|uinas Fxperi 
menl Station, sea level. 16 Apr. 1949 (fl), F 
Allen 5274 (holotype. US!. LL neg. 1071-48! 
isotypes. F!. F neg. 68148!, MO!. NY!. IS!) 
Figure 41. 

Shrubs or trees 2-20 m tall, 8-10 cm diam 
llramldels slender, terete. 3-o mm diam.. denselv 
and miiiulelv appresscil furfuraceous-lepidotc 
h-ares with blades membranous, elliptic. 7.5-181 
X 3.1-5.0 cm. apicallv acuminate, with an acuniei 
6-21 mm long, hasally acute, decurreiil on the pet- 
iole, prominently punctate and puiielale-lineale 
glabrous above, furfuraceous-lepidote below, the 

ini'hib impre — cl .ibo\e. | . r . . r i 1 1 1 1 \ raised below. 

secondaiy veins 23 to 31 pairs. | i I 

harta- above and below, the 

He, 8-1 3 mm long. 1-2 mm dian 
glabrous above, rufous lurluraceous-lepidole helir 
lnjli>iesi<-mes elect, bipimialcly paniculate. 7-31 
3-15 cm. pyramidal, usually longer than the leave 
the racliis. branehlels. abaxial bract surfaces, ar 
pedicels denselv and minutely appressed rufoi 
iiullliaceoils-lepidole. the branches loose I v col 
gesled into 5- to 12-flowered con nibs: pediinch 
nearly obsolete to 3.2 cm long, the lower branch, 
subtended by leaves; inflorescence unkiioui 

blades. I)i.l elliptic |«> oblanoeolate. 1.8-5,1 X 0.5 
2.4 cm, the secondary veins 13 to 25 pairs; inflo- 
rescence lirancli brad petioles >ittii hn In the leal' 
Made petioles. nearly obsolete to 12 nun long; 
Moral brads caducous, membranous, ovate. 1-1.4 
X 1-1.3 nun. apically acute. prominently punctate 
and piuiclale-lineale. glabrous ada\ially. the mar- 
gins irregular, minutely erose. hyaline, sparsely 
glandular eiholale; pedicels slender. 1.5-3.5 nun 
long, inconspicuously punctate and punetate-liii- 
cale. Flan crs 5-meroiis. while or light pink: calvx 
lobes chartaceous. orbicular, 2.2-3 X 2.2-3 mm. 
apically rounded, prominently pellucid to orange 
punctate and punelale-lineate. scattered fuifura- 
ceons-lepidole abavially. the margins mm 
nutely erose. hyaline, sparsely glandular ciliolate; 
corolla nieml.ranoiis. 5.1-7.1 mm long, the tube 
1.5-2.2 mm long, sparsely (iirluraeeous-lepidote 
abaxiallv. glabrous ada\ially. the lobes narrowly 
ovale. 4.2-1.7 X 2-2.5 mm. apically acute, prom- 
inently orange punclale and pimclate-lineate. com- 
pletely glabrous, the margins entire, hyaline; sta- 
mens 4.2-5 mm long, the lilaments 2.5-3.4 mm 
long, the slaminal tube 1-1.2 mm long, the apically 
free portions 1.5-2.2 mm long, the anthers ovoid to 
oblongoi.l. 2.2-2.7 X 1.1-2.1 mm. apically apicu- 
late. basally dee|)ly cordate, the connective con- 
spicuously punctate; pistil 5.2-5.,'! mm long, gla- 
brous, the ovary ovoid. 1.1-1.2 mm long, the style 
4.4-^1.7 mm long, inconspicuously punctate, the 
ovules 31 to 30. Fruits globose. 8-11 mm diam.. 
prominently punctate and punctale-lineate. 

Dulrihiilton. \idi\ia ilunlapiuua is endemic to 
the Osa reiu'iisula. runlarenas. Costa liu.i. growing 
from sea level to 900 m in elevation. 

Ecology and conservation status. \rdisia dun- 
lapiana occurs along water courses in lowland wet 
forests. Hecause il thriyes in medium lighl exposure 
conditions, il is resilient enough to stand some in- 
tervention, and is locally common in some areas. 

ElMiiology. This species was named in honor 

K. Xfiitihn r>t,:> Ml!. I\II. \1( h: Cu.l.-n ,l«- < .olht.M ( ..,.,-i.,l 

'■!>■ I2\la\ I '('M ill). /,'. \fiinlur & F. Ouesada ,i265 (CM. 
FTC. IMS. MO): |{eseiv:i I oteslal Coll.. Ol.lee. O-.i IVn- 

■n but differs by its leaf blades nit 
lyx chartaceous, corolla membram 
sj>arsely furfuraceous-lepidote outsi 

M X 


Forest. II July | 

Osa Peninsula. I'lava Ca.n- 
■.20 June 1991 (fl), P. Har- 

icmanimUM. FTC. MO. 

I'S): Caique \aeional ( Wcny ado Sirena. I Talus Forest. 
o.InK !•«!«) ,f|,. C. Kenan l2l(i[C.\\. III,. MO); Canton 
• I-' Oolhlo. liain Ii.ieslry Gallon. I lov I ill., n. 
Farm. « Aug I *><> 1 (IV). ./. Mann 1(17 (CK. MO): near 
airport area. I mi. \\ ol limeonde Osa. ,'! \iijj. |-»()7(ll). 
/' Karri, L'lt'i:: |0. Mlli: I'au|iie Yi< lonal ( loreoy ado. S. 

i I K-. I\|i. Mlli. Canlun ile (..illilu l'.iM|lir \arional Cor- I si|ninas 2.1 \,, P>'i, ( ||, 1/ S , \ / /;,„ ,„/„ 
I Hi (FTC. I\H. MO): Canton de Colli).. I'arque \aeional 

Ithi. . 

(Don Oue.ho. \\ ..I li(N I e- 
1 1. A. Ihainsm .»"^|C. FTC), 
i 1056 (C. MO). 

. Ardisia dwyeri Fundell. Wrighlia 4: 145. 
1970. Auriculardisiu duyeri (Fundell) Fundell. 
Phytologia 49: 344. 1981. TYI'F: Panama. 
Panama: Cerro Jefe. roadside thicket. 29(H) It. 
|884 rn], 20 Aug. 1967 (fl). ./. Ihcyer & S. Hoy- 
den 8082 (holotype. Fl.!. F neg. 55051!. l.\, 
neg. 71-167!; isotypes, GH!. MO!. US!). Figure 

: .siu loniilal,,,!,, Fundell. Wrighlia 6: 70. 1070 

I' 10:. ■>,!:$. I0JJI . 'H I'F: Panama. I*a 
Fl l.lano-Cartf Road. 0.» km from Inter-Am. 
Highway. I 100-1200 II. |3.T, .<(><> ni|. 21! Dec 
' " '••uiki&H. Hansen 1 1 72 (liol. 
: isotype, MO!). 

. Man I I 

Shrubs or small trees 1.5-8 in tall 
di. mi. Iliancldets slender, terele. 4-5 
densely furfuraceous-lepidole. Irares > 
tic. 4-15.4 X 1.1-5.7 < 

(, 10 

basally cuneale. deeurrenl on llie petiole, incon- 
spicuously punctate and punctale-lineate. glabrous 
above, densely to sparsely furfuraeeous-lepidole 
below, the midrib impressed above, prominently 
raised below, the secondary veins 24 to 32 pairs, 
indistinct to proimnulous above and below, the 
margins entire. Hal; petioles slender, marginal.'. 5- 
10 mm long, glabrous aboye. densely to sparsely 
furfuraeeous-lepidole below. Inflorescences creel, 
pinnately to bipinnateb paniculale. 4-9 X 2.5-5 

Volume 90, Number 2 

he- leaves, densely cupuliforni and furluraceous- manae. Ardisia dwyeri is most closely related to 1. 
epidote. the branches loosely congested into 5- to resca because of its short caly \ lohes. Ion- |>edieels. 
I 1-llovvered corymbs; peduncle 1-1.8 cm long, the thin braiichlels, wide corolla lohes. and long all- 
ow rr branches snhteii.led |,\ leaves: inflorescence thers. However. Ardisia dwyeri can he separated 
iracts unknown: inflorescence branch bracts ca- from A. resca by its coriaceous leaf blades, shorter 

1 foliaceous. oblong. 1.2- pedicels to ;> mm long. 

5.7 X 4.9-10.4 mm. apically acute, ineonspicu- lyx lobes to 3.2 mm wide with rounded apices. 

ously punctate and punctale-lineate. glabrous chartaceous. longer and wider corolla lobes to 6 X 

above, densely furfuraceous-lepidole below, the 3.3 mm. longer and w idci • anthers to 4.2 X 1 .4 mm. 

midvein inconspicuous or impressed above, prom- longer styles to 5 nun long, and larger fruits to 9.8 

inenllv raised below, the secondary veins indistinct, mm m diameter. 

the margins entire, flat: floral bracts caducous. The type of Ardisia conglobata is unique only for 

membranous, ovate. 2-5.7 X 2.4-3.1 mm. apically its small axillary inflorescence of congested (lower 

acute, prominently punctate and punctate-lineale. heads, hut because it is in bud. this is not signifi- 

glabrous above, lurluraceoiis-lepidote below, the cant. In all other respects it matches the type of A. 

margins irregular, minutely erose. hyaline, sparsely dwyeri. 

glandular ciholate: pedicels stout. 4-5 mm long. 

inconspicuously punctate and punclale-lmeale. fur- >/"'<"'"'"> examined. I'WWIV Colnii: Sanla Kila 

furaceous-lenidole Flowers Vnicous ,,„<■„ e ""'^ '"' mL '' "' IV: "" iis " ,m,ls " H >" 2 " J " K '" u ' ! ,IIl 

li i.i. ou ,»< ol< . / lours .> „ ml. lis, , /Vw7 r/ <;/ „„„„ ( ,, ,. (f (L>| _ m ^ ( Vnnam{i: , ,., ,„ 

u '• h - hl l"" k '" |»i"-|>l'': calyx !"!'«■> coriaceous. J,|,.. .|,, |ir> |„. N ,„ I( | ,,„|, u ,,,. ;; N , n p,..- (|L ,,, , ; 

oblate. 1.1-1.5 X 2.8-3.2 mm. apically rounded. Mcl'lw,s,m 712., (I I. YIO): Kl Uano <; ; „n |{„ a .l. 1(1 km 

proniinenlK pun. late and piuiclatedineale. -la- '""" lnl '' \""'n.a,i I lw>.. I ( >,-i. I«>7 I ill. 1. 1. \ Mo,, X 

brousadaxia,, V -'......-... ■ I ..., j^^ii^ 

margins irregular, m.m.lely erose. hyaline, sparsely n] V( |() ||lK |()7(( „,,,_ (; S;/// . y/ „ .,„, (V , ()| 

glandular ciholate: corolla . hailaeenus. 6.3-7.5 

mm long, the tube 1.3-1.5 mm long, the lobes ,,*,.. . n ,,,.„.,„ „• 

. r , v o •) •» n -li •''• »'dl-ia eiieuneaia ihi.l.'lll I'ipoly A l(|c- 

ovate, 5-6 X 2.3-3.3 mm. apically acute, promi- , , ... , 1() rio .,,.,„ . . , • .. . 

,, , , . , '. " ,, , kelson. Sida 18: 512. 1998. Aunculardisia eu- 

nenlly punctate and puiietale-hneale. glabrous ada- , , , M m . . — 

• ii r c , . , . . . cuneata Lund. II. Invtoloma .... II 1 '. 

x,ally. lurlura.eou.-lepulo,.. on the ,ubc and me- , , u . , , , . 

■ i . I ■ i - 

Ardisia eaeuneata (Lundell) l.undell. Phylo 

gia 61: 63. I f »:',7. „,„„. imal. I i I'l I'm. una. 
stamens 5-4 mm long he filament, 2.5-3 mm Sjm B|as; ^ ^ fmm ( . a| ^ 

long, the stamrnal tube 0.8-.., mm long, the api- a ,„ „, , „„ ,„ y „_,, ^ .^ 

i'mlLioui ' v> ,,,, , , -> .■' , i »' Vi ", ;""' n m j,,k |,,:; ' '"'- '■■ ,y " w * c <i " h " h > :j59H 

dn < M)01( ;j- ." . ' ■ ■ mill. .i|»i< ai > cipit - (hololyp,-. II.!: isolvp.'. V|()!l. liouiv 13. 
ulale-euspidate. basal ly cordate, the connective 

conspicuously punctate: pistil 5.8-6 mm long, gla- S/in/fts 2-2 S in tall. Hranehlels slender, terete, 

brous, the ovary globose to oblong. 0.9-1 mm long. 2-3 mm diam.. densely rufous furfuraceous-lepi- 

prominently punctate, the style 4.9-5 mm long. dote, glabresccnt. I saves with blades membranous. 

conspicuously punctate, the ovules 21 to 27. Fruits oblong to narrowly oblong arrow l> oblanceolale. 

globose. 8-9.8 mm diam.. prominently punctate. 4.1-12.4 X 0.9-3.1 cm, apically long acuminate to 
caudate, with an acumen 10-18 mm long, basally 
cuneate. decurrent on the petiole, prominently 
punctate and puii.late-lineale above and below. 
Panama, but also has a disjunct population along mos ,| v <r| a |>rous above, densely rufous furfura- 
the Santa Rita Ridge in Colon. Panama, growing ceous-lepidote. the midrib impressed above, prom- 
Iron. 335 to 1007 m in elevation. iiuntly raised below, the secondary veins 15 to 21 
Ecology and conservation status. Ardisia dwyeri pairs, prominulous above and below, the margins 
occurs in roadside thickets and premontane and entire, flat: petioles slender, canaliculate. 4-7 mm 
cloud forests. Because of its restricted distribution long, glabrous above, furfuraceous-lepidotc below, 
it should be considered threatened. Inflorescences erect, bipinnately paniculate. 3-9 X 
Etymology. 'I Ins species was named in honor 2-0 cm. pyramidal, shorter than the leaves, the pr- 
of John Dvvycr. curator at the Missouri Botanical duncle. branches, and pedicels densely furfura- 
Garden and author of the Kubiaceae in the Flora ceous-lepidote. the branches loosely congested into 
of Panama. W ,„ 7-Howe.ed corymbs; peduncle obsolete to 8 

bract- cadueou-. membranous, oblong. 1.5-2.1 X 
0.4-0.6 mm. apieallv acute, glabrous adaxially. fur- 
furaeeous-lepidole. the margins irregular. minutely 
erose. hyaline, sparsely glandular eiliolate; floral 
bract- similai In the inlloie-eenee hraiieh hraels. 
hut 0.4-0.7 X 0.2-0.3 mm: pedicels slender. 3.6- 
4.8 mm long, prominently |)imetale and pmielale- 
lineate. mixed eupulilorin and furfurai -cous-lopi- 
dote. Flowers . r >-merous. light pink oi purple: calvx 
lobes memhranous. ovate to narrowly ovate. 1.4- 
1.6 X 0.8-1 mm, apicalK acute. prominently pune- 
late and |)iin. tali -I ineule. glahroiis adaxially. Inr- 
liiraeeous-lepidole. the margins irregular, minuleh 
erose. hyaline, sparsely glandular eiliolale: corolla 
uicmhraiiou-.. :!. 5-3.0 mm long, the luhe 0.5-1 mm 
long, the lohes narrow h ovate. 2.1-3.1 X 1.7-1.8 
mm, apieallv acute, prominenth punctate and 
puiictale-lineale. glahroiis adaxially. furluraeeou-- 
lepidole abaxially. the margins ■ -tit i r-« -. hvaline: -la 
mens 2.8-2. ( > mm long, the filaments 1.3-1.6 mm 
long, the stamina! luhe 0.4-0.0 mm long, the api- 
calK free portions 0.0-1.2 mm long, the anthers 
narrowly ovoid. 1.6-1.7 X 0.6-0.7 mm. apieallv 

conspicuously punctale: pistil 1.2-4.3 mm long, 
glahroiis. the ovary ovoid. I-I.l mm long, the slvle 
3.1-3.3 mm long, epunctate. the ovules 7 to 0. 
Fruits unknown. 

Distribution. Ardisia 
the Vi-agandi area in San Ilia-. I', mania, growing 
at 300 to 400 m in elevation. 

Ecology and conservation status. Ardisia eucu- 
nculu occurs in river flood plains, because of its 
restricted distribution, it should he considered 

Etymology. The specific epithet was derived 
from the (deck "eu" meaning well. good, thorough- 
ly, completely, or truly and "-cuneate" referring to 


i Ardisia si 

jbg. Auriculardisia sect 

. Pal- 

manae. Ardisia eucu 

■neata is most easily coi 


with A. ■ 

:..,•/..;.;.'. ;>,■.•-•. 

i (see under that species for 

similarities). However. Ardisia eucuncata is 


distinguished bv its 

branchlets to 3 mm in c 


ter, pedi. 

■els to 1.8 

lo 3.1 

X 1.8 m. 

n. anthers t 

1.7 X 0.7 mm, and the 


to 3.3 mi 

n long. 


ns examined. 

PANAMA. San Bias: Ni 


di,Wedar Trail, 19 Jul? 

, 1986(A), J. McDonaghel 

al. Itiil 

180. 1971. Auriculardisia Jimhrillijera (Lun- 
dell) Lundell. Phytologia 49: 344. 1981. 
TYPE: Costa Rica. Heredia: Tirimbina, 700 ft. 
[213 m), 2 June 1971 (fl, fr). G. Proctor 32238 
(holotvpe, LI.!. F neg. 55663!; isotypes. F!, 
Lb!). Figure 44. 

Trees 3-25 m tall, 4.2-30 cm diam. Branchlets 
slender, terete. 4.5-7.5 mm diam.. denseh lurfu- 

ol.long elliptic to oblaneeolale. 17.2-31.8 X 4-0.3 
cm. apieallv acuminate, with an acumen 6-21 mm 
long, basalh acute, decurrent on the petiole, in- 

, on-pi. i slv pin let ale anil punctale- 1 ineale aho\e. 

prominently punctate and p lale-lineate helow. 

sparselv liirluraeeoiis-lepidote above. densely fur- 
luraceous-lepidote below, the midrib impressed 
above, prominenth raised helow. the secondary 
veins 33 lo 15 pairs, inconspicuously raised above 
and below, the margins entire, inrolled: petiole- 
-lender, marginate. (>-l5 mm long. 2-3 mm diam.. 
-pai'selv furluraceoii-depidote above, densely fur- 
lur.ieeou-lepidole below. Inflorescences creel, tri- 
piiinately paniculate. 11-37 X 7-27 cm. pyrami- 
dal, usually longer than the leaves, the rachis. 
branchlets. abaxial bract surfaces, and pedicels 
fiu-furaeeous-lepidole: the branches congested to 
loosely congested into 5- to 11-flowered corymbs: 
peduncles 1.6-3.0 em long, the lower branches 
-ublended bv leave-: inflorescence brails unknown: 
inllorescence branch hraels unknown: Moral hiael- 
caducous, membranous, ovate. 0.9-1.4 X 0.9-1.5 
mm. apieallv acute, promineiillv punctale and 
punetale-lineate. glabrous above, lurfuraceous-lep- 
idote below, the margin- irregular, minutely erose. 
hvaline. sparsely glandular eiliolate: pedicel- sloul. 
2-3.5 mm long, inconspicuously punctale anil 
punetale-lineate. densely furfuraceousdepidote. 
Floicers 5-inerous. w hile or light yellow: calv \ lobe- 
coriaceous, suborbicular. 2-2.8 X 2.2-3 mm, api- 
eallv rounded, prominently punctate and punetale- 
lineate. glabrous adaxially. spar-eh furluiaeeou- 
lepidote abaxially. the margins irregular, iniinilely 
erose. hyaline, sparsely glandular eiliolate: corolla 
eharlaeeous. 6.2-0.*' mm long, ihe lube 1.4-1.9 
mm long, the lobes elliptic lo lanceolate. 1.5-5.2 
X 2.3-3 mm. apieallv acute, prominently punctate 
and punelate-liiieale. glabrous throughout, the inui- 



finibrillifrra I 1 II W r l t 

filaments 2.0-2.8 nun long. ihe -lamuial tub- 0.8- 
1.2 mm long, the apieallv free portions 1.4-2 mm 
long. Ihe anthers ovoid. 2.6-2.9 X 1.3-1.5 mm. 

apieallv apieiilale. basallv deeply cordate, ihe con- 
nective eonspieiiou-lv piinelate: pistil 6.1-7.3 mm 

long, ulalio.u-. | niiieiillv punetale and punctate- 

1,1 -- 

style 5-5.7 nun I Miii.-.. prominently punctate ami 
punctate-lineale. (lit- ovules 46 to 51. Fruits glo- 
bose, 7-9.5 mm dium.. prominently punctate and 

a is distributed 
from R San Juan in \ »,_ tl I I i 

Costa Rica (Majuela. Heivdia. Liindn). and Pana- 
ma (Panama. Yeraguas. San lilas) to Chocd. Colom- 
bia, growing from 10 to 800(-l200) in in elevation. 

Ecology and conservation status. Ardisia fim- 
hrillifcra is localb inlretpienl in lowland tropical 
wet forests or. rarely, in premontane forests. It lias 
been collected in locations dial are fairb remote 
and maiiv ol them are protected, so we see no im- 
mediate threat to this species. 

Etymology. The specific epithet was derived 
Irom the Latin "limhri" or "hinhrill" and "ifera" 
referring lo the c.ilw lolies hearing a fringe of glan- 
dular cilia. 

Within Ardisia suhg. \uriculardisia sect. I'al- 
manac. Ardisia Jimhrillifera is easily distinguished 
from A. hagenii h> its narrower calyx lobes to A 
mm wide, shorter corolla lobes to 5.2 nun long. 

i long. 


to 5.7 . 

[0 :," 

Sprrnn.-ns ,-xamincd. \l< \I!\CI \. Km San Juan 
Buena Vista a I k \\ of the Km. San Juan delta. 13 Sep. 
|<>,".-, ll,,. /. M,n,,.,r \ A />'. /,'-/;.■,. _'//*'■: Ml M . MO) 
,„,,, (an,. Chonlaleno 20 km NK of Kl Castillo. 1P.-2I 
Apr. 1978 ill). I), \,-ill & I'. \incclli .Hn<)2 (H\MY I.I.. 
MO. \ i t: Km San Juan. 2 km K ol Buena \ ista. I I Sep. 

P.:;2 ,i, i; /;,.,.„ ;iu dlWlY \lo. \,. I n: Mpi... 

,1, I Castillo. Helena I ud io-\lai/. along Canoel Bason. 3 
km Irom the mouth of I ! I o Barlola. 3 I Dee. I 09( , (1,1. /,'. 
R„rd„ rl al. .11.1.; ill! II,. MO|: Mpin. Id Castillo, lie 

; of Rio 
97 (fr), R. Rueda , 

, liainmi. (.nrdillei 

null i .,,l|„ 

!')')() (In. 
tinea l.a !• 

I Km I 

I dr.. 

ilar &■ II. Sclav, dl 940 (I Mi. MO); Refugio Na< 
Fauna Silvestre Barra del Colorado. I.lanura de 
> Lindo. 24 July 1995 (fr), F. Araya 792 
(CR. K. KTC. INR, MO); Hacienda Tape/eodlda. la Sue, 
le. 20 air km W 10 Mar. I078i.ll,. C. I),, 

I rfa. Sender,. 
I 01,8 (II). ./. 


/..I'l'O (MO): tinea Indio. slopes ol 
1971 (fr). 1. Cain 217,0 ( \|( )): ,>„ II 
l.lano-Carti road, near Nusagandi. along Wedei Trail. .'5 1 
(),(. 1 092 lii). C. Mrl'luTsun & U. Richardson 159HH 
(ITO, MKXU. MO. PMA). San I 


, MO); 

PM u 

\ of Kseuela \grfcola Alio de I'iedra. I7<)el. l«»7l d.i. 
,S. Mori & J. kallunl., 230/ iMOl: \ W of Sanla he. I .«". 
km \ of Kseuela \m„ola Mlodc hedra. 23 Keh. 1975 
il.-i. \ Mori X ./. kolhada 177.1 1 1. 1.. MO); 7 kin W of 
Santa Ke on new road past agricultural school. 12 \pr. 
1071(11). M. W///9/(I.L.MO. 1 S..C(ll.O\llll\.t:ii..- 
,-,V M| le Kmsiicio. /„na de I raha. Cerrodel (aiehillo. 

I , w. 108, 0,)./;. Cmdmus ::::.; (.1 \i M. mo>: k.,,s ;i1i 

Juan Basin, .(car IWordd. 2" Ma,. 1070 (fr). K. Form, el 
al. /.!>/ (COL, M0). 

13. Ardisia furfiiraeea Standi.. J. Wash. Acad. 
Sei. 17: 525. 1927. \uri<ulard,s„, fiirjaracca 
(Standi.) Lundell, Phytologia 49: 344. 1981. 
TYPK: Costa Rica. Heredia: Cerro de Las La- 
jas, IN of San Isidro, 2000-2300 m, 7 Mar. 
1926 (fl), P. Standley & J. Valeria 51556 (ho- 
lotype, US!, LL neg. 1971-34!, US neg. 2372!). 
Kigure 45. 

\rdisio duriii.scida l.imdell. Wrighlia 7: 2 1. I OP, I. Swi. 

1600-1800 m, 6-7 Keh. 1077 (II). 11. Hun,;. (,. \,, 
conti & J. Gentry 10288 (holotype, F!, F neg. 68149!; 

diriculardisiti trichomata Lundell. Phytologia 63: 73. 
I9H7. Svn. nov. \r,lis,a Inchomato (luml. Ill lun- 
dell. I'hMologia 63: It..}. 1987. TYPK: Cosla Rica. 
San Jose: roadside leading from Alto U Palma to 
Rajo I,a Hondura, C 

2-6 m tall, to 3.7 cm t 

mixture of densely ciipiiliform a 
lepidote scales. Leaves with blade 

long to elliptic. 16.2-17.8 X 3.8-13.2 cm. apicallv 

acuminate, with an acumen 0.3-1.0 cm long In.- I^rilmhon. \rd,sui jurfnracea is erulemie I, 

sally acute to cuncatc. decurrenl on the petiole. ,,,e "»>»'"<"»"»' "'"^ Costa Rica in Carlago. 1 le- 

inconspicuously punctate and punctate-lineate, gla- n>1,ia - a,Ml San J(,s ^ 8 n,win * al ,2()0 t( ' 2(MM> '" "' 

ril. impressed al,!'',''. ,„".., ii,!-i',llx ,.,!•..' ,1 ]lu""L ""™ '™™ "' l-'non.ane to montane we. lore., 

secondary veins 17 lo 01 pair., pron.iuulou. al>ove a " ° "' """ u "' I'amueni- ,.( lorests found in 

marginale. 0.7 2.2 nil Ion-, dabn,..- above, below '* k "" M " lmm ;l "' l '""'^ *">"H "«'...Imt of eolloc- 

vvilh a mixture of densely cupuliform and lurfura- " ,ms a,ul s,,ou,<l l,e '-""s^red threatened at this 

ceous-lepidote scales. Inflorescences erect, hi- or '" ne " 

H| I I | l It 17.4-46.7 X 15.2-2') 6 Etymology. The specific epithet means scurfy, 

em. pyramidal, as long as or longer than the leaves. ' m, '"' <l vvil1 ' l"'"'"''^ ^"- 'H'-mug lo the ves- 

wilh a N„M,„v ol ,I,,h.K cupulilorin ami lurfura- " Um ' ,l,m "- ll< "" ''"• l ,l: '"'- 

ceous-lepidote scales, the branches terminally con- Vh,,1,n Ardlsm sul ^ \""<ulardisia seel. Pal- 

gesled into 5- to 8-flovvered corymbs; peduncle """""'' ■^'■"'"Mfunm-a is most closely related lo 

nearlv sessile lo 2.7 cm long, the lower branches 1 " '"'V'"™""' '>> ^"'"' <>Hhe flexuous hrancl.lels 

sulilended l>v leaves; inflorescence hracls unknown: :m<l ,,I,I<M1 ^ '" na,louK •-Hi|>«ic- leaf blades. Ilou- 

inflorescenee branch bracts earlv caducous, mem- rVe ' ; *«*''»«> M/urna-a can be distinguished from 

branous to chartaeeous. ovate to oblong. 1.3-1.8 X l ""V 1 "'™"'" by its narrower calyx lobes to 1.7 

0,4-0.5 cm. apicallv acute to rounded, or nearb so. ,I,m wi,,t ' a '" 1 wi,,,T <omll;i l,,l)es lo 2 - r > """ »'<•<'- II ' - ' ' ' f " ' ^ M ml i I 

linealions. glabrous above, below with a mixture of 2 ' 2 X ' mm - a1 " 1 lon - , ' r sl > l,>s «" S - 2 »»» l( "'^ 

densely cupuliform and furfuraceous-lepid<,|«. " a ' , l ,rars ,hal Lim,,f>11 < IWI <- 19 » 7 ) «•'«' '»»• 

scales, the midrib ineoiispiciious. the secondary ,,ott ' tl,< ' similarities between \rdisia jurjumcea and 

wins obscure, ihc margins entire, flat: floral bracts ("'I'l'lalions represented b\ the types of \nlisio dn- 

early caducous, membranous, ovale 8-13 X riiiscnla i\iu\ \iiri,-nhirdisio iri< -homala. The types 

0.8-1.3 mm. apicallv acule. proniiiienllv punctate °' ^ r ' lls '" limits, idn and \un< ulardisui trichomata 

and punctate-lineale! glabrous above. scattered fur- £ " v i,lr " ti(al *° ,lu " ,)( ' -^disia furfurucva in all re- 

furaceous-lepidole below, ihe margins irregular, mi- s l H ' ( ' ts - 

nutely (.valine, sparsely gland. dar-ciliolale: Spenmens examined. COSTA KICA. Cartago: Tapan- 

pedicels sloul. 0.8-2.1 mm long, inconspicuously If Ueserw. 7 Dee. I <»i!2 111). /. < :„mc: l'r_>;;j 1 1 I |l'|. MOi: 

punetate-lineate. densely furfuraceous-lepidote. Canton ile I'arais... I'ar.|.ie Yieimial Tapani.'. \alle del 

Flowers r>- or 6-n.erous. while: calvx lobes ,{, ;"''''« z «'''- s «' 1 " 1 ""> \"', S os. N,^ n , Trail. 20 

,,,„,. ,„!,„ ,,| ,, I,,,,,. I - , ' | , , - ■''''> 1W1 «rl. C Mora e, al. 2<>l (l\B. M0). Heredia 

.■'"" " ," ,'" ,Na "- '™ X 1| -'' mm - I'ar'l...- Nan Itn.uli.. « a, nil,, Iran.,., Trail. I„ ,,,(,, 

ap.cally rounded. often prominently an.l ineo.ispic- ,,f irail. 20 .lane 1002 uier.i. /,'. liable & C. (;„dl 9.T> 

"ouslv pil.lclale and pun<-lal,-li..eale. U labrou> H" I "C- IMS. \K »•: l'.m,ue \at,<mal liianl,,, ( .an.llo. I,an- 

Sep. 1992 (fr), H. Boyle & \. .S«„„- 11)52 < r TC I 
I'): aloni; Hi.. Mi, lialj, I. \ I I.mi. i« ^l,,|„- ,,| \..|.-.,,, I 
. 12 Apr. I0n(» |fl). XI. Cnnum 7060 (C|{. F\T,. 

ally, densely furfuraceous-lepidote abaxiallv. 
the margins irregular, minutely erose. hyalii 
sparsely glandular-ciliolate; corolla membranoi 
5-5.3 mm long, the tube 0.7-1.4 mm long. I 
lobes ovate. 3.9-4.3 X 2.3-2.5 mm. apicallv acu 
sparsely prominently punctate and punetate-li 
eate. glabrous throughout, die margin entire, hva- 

line: stamens 5.8-6 mm long, the fila.nenls 3.0- 1. 1 li ' Ardida generalensis KickelsnnK I'ipolv 
mm long, the staminal lube 0.6-0.7 nun Ion- the ,lov - ' ^ l>K: (: ° s,;l l{ir;l - Siin -h»se: basin o 

ap.callv free portions 3.2-3.5 mm long, the anthers (^em\. 67.VXM) „,. Mar. 1012 (II). I. SL 

ovoid. 2-2.2 X 0.8-1 mm. apuallv apiculate. ba- S()25 < h « ,,( » t > I »«■- US!: isotype. MICH!). Kij 

late: pislil 5.6-6.5 . 

. Auriculardisia 

such as Cybionthus / 

. (I.uudcll) \gnslini. because 

old col 


(nee einai<;malis) perfaeile ivmgiioscilur. 

Trees to 6 iti tall. Branchlets slender, terete, 3-4 
mm iliam.. densely and minuteh appresscd rufous 
furfuraceous-lepidotc Leans with lilades membra- 
notis. elliptic lo narrow l> ol>o\ate. 4.8-1 1.5 X 1.7- 
4.2 cm, apical ly acute, with an acumen 4-7 mm 
long, basallv acute to cuneate. decurrent on the 
petiole. inconspieiiouslv punctate and pundate-li- 
neate above and below, glabrous above. denseh 
and miiiutcU appresscd rufous lurfuraccous-lcpi- 
dote below, the midrib impressed above, promi- 
nently raised below, the secondary veins 21 to 25 

entire, flat; petioles slender, canaliculate, 4-8 mm 
long. 0.5-1 mm diam.. glabrous above. Inflons 
cences erect. bipinnalelv paniculate. 1.8-7.2 X 3- 
4.5 cm. columnar to narrowly pvramidal. longer 
than the leaves, the rachis. brancldcls. and pedicels 
densely furfuraceoiis-lepidole. the brandies rarclv 
subtended bv leaves, the branches terminating in 
3- to 7-flowered corvmbs: inflorescence bracts and 
branch brads unknown: floral brads caducous, 
membranous, ovate. 1.2-1.8 X 0.6-1.1 mm. api- 
callv acute. inconspicuous!) pundale and punc- 
laledineate. glabrous throughout, the margins irreg- 

be considered threatened. 

Etymology. The specific epithet refers to the lo- 
cality of El General where the type was collected. 

Within Ardisia subg. Auriculardisia sect. Pal- 
inaiiar. \nlisia generalensis is most similar lo -1. 
daricnensis (see undci that species for similarities). 
However, it is easily separated bv the shorter and 
thinner, canaliculate petioles to 8X1 mm, mem- 
branous leaves with less numerous secondarv veins 
to 25 pairs, and the apically acute calyx lobes. 

15. Ardisia gifiantea bickdson <.K Pipoh. sp. 
nov. TYPE: Panama. Veraguas: N\Y of Santa 
Fe. 8.8 km from Escuela Agrfcola Alto de Pie- 
dra. Pacific slope, 21 Dec. 1974 (fr), S. Mori, 
./. Kallunki, T. Cochrane. H. Cochrane. H. Han- 
sen, R. Kowal & M. Nee 4003 (holotype, MO!: 
isotype, LL!). Figure 47. 

I',, ,,,|CI 


5.1 cm long, nearly epunclate. Fionas 5-merous. 
white or pink: calyx lobes coriaceous, ovate. 3.4- 
3.7 X 3.2-3.1 mm. apicallv acute, niostlv epunc- 
late. glabrous adaxiallv. lui liuaeeousdepidote aba- 
xiallv. the margins entire, hvaline; corolla 
chartaceous to coriaceous. 7.9-8.1 mm long, the 
lube 3.3-3.0 nun lonu. the lobe- narrowly ovate to 
lanceolate. 4.4—4.7 X 2.2-2.5 mm. apically acute, 
mostly epunclate, glabrous throughout, the margins 
entire, hvaline; stamens 6.1-6.2 mm long; the fil- 
ament 3.3-3.4 mm long, the staminal lube 2.1-2.3 
mm long, the apically free portions 1-1.3 mm long, 
llie anthers narrowly ovoid lo laneeoloid, 3-3.1 X 
1.2-1.3 mm. apically apiculale. basally lobalc. the 
connective iiienn-picuoiislv pundale: pistil 8-i!.3 
mm long, glabrous, the ovarv oblong. 2.6-2.7 mm 
long, the stvle 5.4-5.6 mm long, epundate. the 
ovules 25 to 29. Fruits unknown. 

Distribution. Ardisia generalensis is endemic to 
the basin of El General in San Jose. Gosta Rica, 
growing from 675 to 900 m in elevation. 

Ecology and conservation status. Ardisia gene- 
ralensis occurs in wet forests. The area contains a 
number of Daridi and < .hoco Holistic (dements. 

.;..', in ii,, 2 1 1U ■' 1.2-1.1) em ad apices rotundatis 
fnec aeulis) perfacile cojmoscitur. 

Trees to 8 m tall, to 30 cm diam. Branchlets stout, 
terete. 8-10 mm diam.. denseh appresscd rufous 
furfuraceoiis-lepidolc. Lean's will) blade- membra- 
nous, elliptic. 45.8-46.6 X 15.7-21.5 cm. apically 

I •' 2 


acule. decurrent on the petiole, prominently punc- 
tate and punclate-liueale. glabrous above, furfur; 
ceous-lepidole. the midrib impressed above, pron 
inenllv raised below, the secondarv veins 48 lo 55 
pairs, promiuentlv raised above and below, the mar- 
gins entire, inrolled: petioles slender, margin; 
1.2-1.7 cm long. 4-5 mm diam., glabrous abo 
furfuraceous-lepidote below. Inflorescences en 
hi- to tripinnately paniculate, to 32 X 25-28 c 
pvramidal. shorter than the leaves, the racl 
branches, and pedicels furluraceous-lepidole. the 
branches loosely congested into 3- to 7-flovvered 
corvmbs: peduncle obsolete: inflorescence and 
branch brads unknown: floral brads unknown; 
pedicels slender. 9-13 mm long, inconspicuousb 
punctate and punclale-lineale. furfuraceous-lepi- 
dote. Flowers 5-merous. color unknown: calyx lobes 
chartaccous. oblate. 1.1-1.6 X 3.3-3.6 mm, api- 

cally rounded, prominently punctate and punctate- riaceous, elliptic. 1,2-21.2 X 2.6-8.3 cm, apically 

lineate, glabrous adaxially. sparsely to densely fur- acuminate, with an acumen 0.8-2.1 cm long, ba- 

I'uraceous-lepidote ahaxialb. llie margins irregular. sally obtuse or acute, decurrent on the petiole, 

minutely erose. hyaline, sparsely glandular cilio- prominently punctate and punclate-lineate, with a 

hUe: corollii. stamens, and pistil unknown. Fruits lew scattered scales above, densely and minutely 

globose. 9.5-13.5 cm diam.. prominently punctate. apposed l-'mminniib liiiluraceous-lepidote be- 

A ,. . . . . low. the midrib impressed above, prominently 

Distnbunon. g.ganl.a i> no, b. ^ ^ _ ^ ^ ;}? to ^ ^ 

from the type , oil, ,,■„„. Iron, „. a, .he s, „. la ^^ ^ ^ M ^ ^ ^ 

Agrfcola Alto de Ptedra in \eraguas. I'anama. '.,,,,, • , . ant - ,„ 

• , rm , im n - n I v ,i,n protmnulous below, the margins neatly entire to 

growing at 500 to 1000 m in elevation^ ^^ ^.^ 

glabrous above. den-el\ lernigineous furfuraceous- 
l.-pidole below. Inflorescences erect, hi- to tripin- 
nately paniculate. 1-20 X 1 22.5 cm. pyramidal. 
n>uallv longer than the leaves, the rachis, branch- 
lets, abaxial surfaces of all bracts, and pedicels 

densely ferrugim - In, luraeeous-lepidole. the 

branches loosely congested into 5- to I 1 -flowered 
corymbs; peduncle nearlv obsolete to 2.1 cm long, 
the lower branches often subtended by leaves; in- 
florescence bracts absent: inflorescence branch 
bracts very early caducous, membranous, ovale. 
1.0-3.2 X 0.5-1.5 mm, apically acuminate, prom- 
inently punctate and punclate-lineate. glabrous 
adaxially. ferrugineoiis lurliiiueeous-lepidote aba- 
densk. Meddel. Dansk Naturhist. p0ren xially, the margins irregular, minutely erose hya- 
Kj0benhayn 1861: 128. 1862. Auriculardisia line, sparsely glandular ciliolate; Moral bracts sim- 
glundidosnmarginala (Oerst.) Lundell. Pin to- ilar to the inflorescence branch bracts, but 1.2-1.8 
logia 51: 285. 1083. TYPK: Costa Rica. Car- X 0.4-0.6 mm; pedicels slender. 7.1-8.7 mm long, 
tago: in motile basil. 8,000 0000 ft. |2438 prominenlly punctate and punclale-liiieale, ferru- 

2713 m|. Jan. 1847 (fl). A. Oersted 25 gineous furfuraceous-lepidole. Flowers 5-merous. 
(holotvpe. C!. V neg. 22018,!: isolvpes. C!. VI light given, white or cream to light pink: calyx 
Lb!. M!|. figure 48. lobes membranous to chartaceous, widely ovate to 

orbicular, 2.5 2.5 X 2.5 2.5 mm, apically acute lo 

population biology 

and < 


status are unknown 

Etymology. The 

epithet is based 

on the i 


tionallv lame leaf blades. 

Within Ardisia s 


[uriculardisia sect 

. I'al- 

manae, Ardisia niuunlru i> nin-l -dm 

,1a, lo 1. 


rreana, but is distil 


■d b\ il- larger leaves lo 

46.6 cm long, and 


dive oblat 

e calyx 


Ardisia gigantea is 1 


only from i 

, Iruitin-. 


imen: the holotype 

is ol 1 

j single h 

iaf with 

an at- 

:ii-hed uillorescence. The 

LL isotyp 

Lundell, Phylologia 57: 150. 
f985. Syn. nov. Ardisia leptopoda (Lundell) Lundell. 

. pmminciilly piuielale and punctal<-l 

Phytologia 61: 65. 1986. i 

l>,»d,i 1 1 uimI.1I ) l'i|)(,l\ A. IlickeiMHi. Sida lii: 5I.J. ,| n | ( . nbaxiallv. the margins irregular, minutely 

'j" ( ' ' [ ^ ; -;;- ^ " ■" " ;;; ^ "'; | ■"';_ ,,«... I.v i. . ^.i.m-K glandular c.l.olate; corolla 

""; ' " s '" l ( "'" |l " | , " '", , , '" (i . ' i membranous. 6.2-6.4 nun long, the tube 1.2-1.5 

' i " , '. i ", ,,, ' / leer, &'k mm long, the lobes ovale. 4.5-5 X 3.2-3.8 mm. 

Foster Fill I (l,olm M >e. 15 I neg. (>['.:\2M: i>.,i\|,es. M |, acute, prominently punclate and punctate- 

DUKK!. LL!, MO!, US!). liueale. glabrous througlioul. the margins entire, 

Shrubs or small trees 1.5-12 m tall, 7.6- 18 cm In aline; stamens 4.5-4.7 mm long, the filaments 

iam. Branchlets slender, terete. 3-6.5 mm diam.. 2.2-2.4 mm long, the staminal tube 0.7-1 mm long. 

ensely appressed lernigineous luiiuiaceoiis-lepi- the apically fn 

ate. leaves with blades membranous to nearly co- thers narrowly 

/ ol. WO.i (MO); It I 

H H 

Volume 90, Number 2 

. Auriculardisia 

l\ cordate, the connective i-on-| h<-ik «n-l\ punctate; 
pistil 6.2-6.4 mm long, prominently punctate, gla- 
brous, the ovary oblong. 1.1-1.5 mm long, the style 
1,9-5 mm long. prominently punctate and pune- 
tate-lineale. the ovules 29 to 33. Fruits globose. 
6.8-9.2 mm diam.. prominently punctate. 

Distribution. Ardisia glandulosomarginata is 
found in all provinces of Costa Rica except Gua- 
nacaste. and is common from the Costa Rican bor- 
der to central Panama, growing from 1675 to 3056 
m in elevation. 

Fculog\ and conservation status. Ardisia glan- 
dulosomarginata normally occurs in very wet cloud 
to elfin forests with occasional populations at the 
upper limit ol prcnionlaiH' and montane forests. It 
has been collected in areas with remnant forests, 
and apparently lias -Mine tolerance lor disturbance. 
Therefore, we do not believe it is threatened at this 

Ft\molog\. The specific epithet refers to the 
prominently raised punctations along the abaxial 

Within \rdisia subg. \uriculardisia sect. Fal- 
manae. \rdisia glandulosomarginata can be sepa- 
rated from I. croalii by the ovate calyx lobes to 2.5 

rule 8881 (CK. I TO. INK. MO): Canton di- Coin linis. 
Zona Protectnra I .as labia-. Ciienea Terraba-Sieipe. !i kin 
\K,,I l',ogre>o. I'll,.!,. I«)«)7llli. /■'. <;„mi,oa& I. I'icado 
1052 (INB, MO). Sail Jose (anion de l'<t,/ /, l< sdon, 

d«e \b ofCerro 
W. 1980 (II. IV). 
i Me, 10-13 mi. 
r), J. Duke 8028 

_>!>.-,.: (MICH. MO): ra. 3.7 k 
l>„„(a nn load through Kajol 
II. Stevens 18212 (M<>1. Pan; 
beyond Ooofv bake. 12 Feb. 

47. Ardisia hagenii Lundell. Wrightia 1: 59. 
1968. Auriculardisia hagenii (Lundell) Lun- 
dell. Phytologia 49: 344. 1981. TYPK: Pana- 
ma. Chiri(|iif: Boquete region, Horqueta. |6500 
ft.] 1981 ni, 17 Apr. 1940 (fl). C. von Hugen 
& W. von Hagen 2008 (holotype, MO!, LL neg. 
71-166!: isotype, NY!). Figure 49. 

I'lolologia '.I 
(jiiiima (bumf 
I'm;; \\ p|.; : 
\1aoho. 0!!.-.( 

I '«;:>, ifh. /;. : 

, 7: Jf.7. P'KI. 

ehiriquiana (I 

|o;:o. H..IH. in- 

|».l> X UlckoN 

Speeimens examined. COSTA RICA. 
nas de San Pedro <l<- Sail Ramon. 25 Oct. 1025 lib. I. 
/Irenes I /'>.', (K \ , ); \olcan Poas. S slope of crater. 20 
\,,r. 1030 (II). (, Cufndnnli 88') lO. \\ ): Canton .1.- \llaro 
Km/. Zareern. 17 Jan. 1938 (fl). \. Smith ll-l In (fl. Car- 
la-o: along -lieam \allevs on the S\\ slope of \olean 
Ira/... \\ "I Sabanilla. 7 Mar. I0!>,7 Mil. II. Hurgrr el nl. 
I2().i8 lb. M. TK\). Il«>r«*ilia: K10 \uellas. upper Km 
Pallia. K slope of \olean Karba near the Continental Di- 
• i.le. 2 Job 1083 (fr). K. Harringer & /.. Cbrislms.m 
:;:!,",-> \ il ): Ilaneho (lores. 22 lei, 1890 III). 1. Tonduz 

J (BR); 

- Field 

Rio Puerto Viejo just K of its junction with the Kin Sa- 
rapiquf. Sndero Tre, liios at 2500 m line. 7 Jan. 109.5 
(fl), R. Wilbur 63199 (DUKE. MO). Limi'm: Cordille.a I 1 
lanianca. headwaters of the uunaniecl \\ branch oi llieKio 
Teribe. between the Rio Sini and the Conlinenlal Divide 
at Cern, Kekom. along river. 21 27 Mar. PM'.I llll. (.. U« 
ridse el ul. 25719 1 1 j .. V10): Refugio Kami del Colorado 

, 11 1 \\ 10 \,,i I0O0 Hh 1/ (Nl 

yum 9761 (CK. K'fC. MO). Pi.ntaieiias: Canton de I'ere/ 

-ulardisia loraanu Lundell. Wrightia 7: 273. 19Kb 
Smi. nov. Ardisia toroana (I.undell) Lundell. Phyto- 
logia 01: 07. I9K6. num. inval. Ardisia toroana (Lun- 
dell) Pipoly & Ricketson, Sida IB: 511, 1998. 
TV PL: Panama. Kocas del Torn: 15 km up the Chan- 

ubs or small trees to 11 m tall, to 12.5 cm 
Brancldets stout, terete, 3.5-5.5 mm diam.. 
ly and minutely appressed rufous furfura- 
lepidote. U'aves with blades chartaceous to 
■jptic to narrowly elliptic, 13.5-24.6 
X 3.6-7.7 cm, apically acuminate, with an acumen 
0.6-1.6 cm long, basally acute, decurrent on the 

petiole, inconspicuously | etale and punetate-lin- 

eale above and below, mostly glabrous above, 
densely and minutely appressed rufous lurluia- 
ceous-lepidote below, more so along the midrib and 

i... b 

minently raised below, the secondary veins 42 
19 pairs, slightly impressed above, prominently 
>ed below, margins entire, flat: petioles slender, 
>8 mm long, 1-3 mm diam., glabrous above, fur- 

luraeeous-lep s e.ect, hi- to wider. n»n,i, -eon-, orhicula, calw lobes to 3.7 mm 

Iripinnately paniculate. 9.1-20.5 X 4.2-20.5 cm. wide, longer corolla lobes to 5.0 nun long, longer 

pyramidal, longer than the leaves, the rachis. anthers to 5.8 mm long, longer shies to 7.1 mm 

branehlets, abaxial bract surfaces, and pedicels long, fewer secondary veins of the leaf blades, and 

densely furfuraceous-lepidote, the branches loosely thicker branehlets. 

congested into 4- to 9-flowered corymbs; peduncles The von Hagen collections are alvpical of the 
0.8-3.2 mm long, the lower branches subtended by species. The inflorescence branches appear fasei- 
leaves; inflorescence bracts unknown; inflorescence ated. causing some branches to he swollen and re- 
branch bracts usually persistent, chartaceous. duced. giving the inflorescence a "columnar" ap- 

2.5-4.5 X 1.5-2.2 i 

conspicuously punctate and pimclalc-hneale, fur- I nfortunately. I.undell failed to understand the 

furaceous-lepidote abaxially. glabrous adaxially, the atypical nature of his type and never annotated any 

margins megulai, minutely erose. hyaline, sparseb oilier specimens as Anlisia hagenii. Instead, lun- 

glandular ciliolate; floral bracts similar to the inflo- dell described Auriculanlisiu chiriquiana Luridell. 

■nitrations to char- which has normal branches and an inflorescence in 

1.4-2.3 X 1.6-2.8 n 

young bud. The type o 

late-hneale, lurhiraccoiis-lepidole. Flowers 5-me- gin of its calyx lobes. However, the types (one- 
rous, pink to red-violet; calyx lobes coriaceous, or- sponding to both Auriculardisia chiriquiana and 1. 
bicular, 2i>-2.8 X 3.5-3.7 mm, apically rounded. t annum match Anlisia hagemi in all other respects. 
coii>picu..n-.|\ hi.,, -k puiicUte ami punclalc-liiieale. 

f " rf ~-'" ■•">- -•' ■> -i»">. . ;;;::;;;:, ; i;::::::l,:;i:i- ;,,,':•::;;:: .,;^;:;t 

Hie lll.llgll.,,, o,,,s, hvall.X ^.ai-eb glai.dl. (,-,.,. Ii,.„ I, , I! |.„, I •>,-.() ,th \l <„,,,„ 1,1 I'U 

lar ciliolate; corolla ehnrlaceous. 8.<>-8.9 mm Ion-. (MO. I'M \l: SL and \K <>l Changuinola I km from IHUK. 

the lube 1.2-1.3 mm long, the lubes ovate. 5.3-5 N •'""■ l<,1!0 "D- ]/ '"'"" •'' "'■ '"^ (MO. I'M A): along 

punctate and pinu-tate-bneal,-. glabrous. I,n„,gl I. „.„„„. ,,„„„ |p||(( .| mr „.„-.,. ,,„,,_ ? ,,,.,. ,„.,- (h) ,. 

the margins entire, slightly hyaline; stamens 7.2- SUI'hason 71127 (K ITO. II. \l(). I'M \). Chiriqui: along 

7.3 mm long, the filaments 4.3-4.4 mm long, the ma<l between Koiluna Lake and Chiii<|ui Oiande. I.fvf) 

stuminul tube 1.2-1.3 mm long, the ap.calb In, ^ »« d .... ow , I „,,,,„ , I ,k. « Ma, |«JH> (.1) / , ,«,„ 

portions 3.0-3.2 mm long. ,he aulhc lanecoloul. , ,„, „, „,.„ ^ ,,„ ,,,,„ v|| „ ,„ Nm ,„.,, „'„ ,, 

3.6-3.8 X 1. 1-1.4 mm. apically apiculale. basalK //„„„„,■/ :>7U, ,M(),. S slopes ofC-rm I'ate Macho along 

lobate, the connective conspicuoiislv punetale; pis- I{| " ''"I" x ll". II Nov. I'»<':i (IV). S. Knapp el nl. 20H(i 

tils 8.6-8.9 mm long, glabrous, the ovarv oblong (ll ~ V1(> " N>l l: M, ' ar h """ ,a |);,,n - "'""J? <><''><>'da de 

I 8-2 mm Ion- the stvle (,') " I ,„.,, I „, ■ ,T W "'''- ^ "' ' "' "'• ll l,ixi,l «"- r> l)vr - N85 <">' <'' W '" 

l.o _ mm Ion,. II.. slyl, (,.-,. „„,, long. , on- /% , rsi>n -,,, (| ,,. (; ,, m)y ^ u> ^ ^ 

spieuously punctate, the ovules 50 to 55. Fruits \la,l„, „ I ,|, !<>;;„ ( ||.. ,, \h I'L-,,,,,, lV ' M. \lerello H2<M, 

globose, 5.2-6.7 mm diani.. inconspicuously punc- (KTC. LL. MO): I'oiiuiia Mam region, above IN edge of 

tate and punctat Ii .1 i.conspieuouslv costate. lak <-^~ V- I"-" (Hi <-■ l/.VV/mw! «MV |K fTC. 

\IL\l. MO. \V I'M \): SK .lopes and suumiil ..I Com, 

Distribution. Anlisia ha genii is endemic to l>a "' Mi ! l ' h ": ,ri , lils , '"»" Ul " |,;,i " Xl «'- > kl " ™ "< |{ <- 

w ^„.,. n |» ln „ i, ,,,. ..... .|..elr. L'oMav 1'ij; k. .s vl,ma et ,,l. 48.V (I ,L. MO): 

western I anama. .,, lUas ,lcl Ion. and Chm,,,,,. ,,,„.„„„ | )am ,,.„,„„ ,,,„,„ (hl ,. |)ra(|a A| ,. ||a s „, ( ,_ 

growing from KMM) to 2 KM) ,„ in elevation. „.„„! Divi.le. IT, Jan. I«>i!<) (II). <;. \l ( Vhers,m l.iMi I 

Ecology and conservation status. Anlisia hage- (''''<- MKXl . MO. I'M A). 
nii occurs in montane wet. cloud, and elfin forests. 

While it is certainly not common, at this time there 48 ' Ar « ,i ^' 1 ' lii«> (I.undell) Pipoly & Ric- 

are no data to suggest the species is threatened. ketson. Sida 18: 513. 1998. Auriculanlisiu hu- 

Ftymology. This species was named in honor gonensis I.undell. Wrighlia 7: 268. 1984. Ar- 

of Christine and \\n||»a.i» von lluueii. collector of disiu hugonensis (I.undell) I.undell. I'hvlologia 

the 6 ': ^- '^6. num. inval. TVPK: Colombia. 

Within Anlisia subg. Auricu/anlisia sect, f'al- Chnco: Mpio. dc (,)uibd6. Corregimienlo de 

manae, Anlisia hagenii may be most easily con- Uiayabal. liio Hugo,., ca. 80 m. 12 Sep. 1976 

fused with A. pseudoracemiflora because of 'its el- (,r) - K ' Fon ' r " lK: ,{ - J<"""iillo 2812 (holotype, 

liplie to narrowly elliptic leaves and inllorescence M !: isot >l ,< '- [V10! >- Fi P urt * 5() - 

much longer than wide. However, \nlisia hagenii Trees 10 m tall. Branehlets 4.5-6 mm diam.. 

is easily separated from 4. pseudoracemiflora bv its smooth, densely appressed rufous furluraccous-lcp- 

Volume 90, Number 2 

umen I ) mm I i r III acute. I invtil on 
:' petiole, inconspicuously punctate and pundate- 
leate, glabrous above, lurfuraceous-lepidote be- 
#. tin- midrib impressed above, pronunentlv 



Inflorescences erect. Iripinnalely paniculate. 11.8- 
13.1 X 0. 5-10. 5 cm. pyramidal. mosll\ long.i than 
the leaves, denselv cupuliiorin and liirfiiraceous- 
lepidote. the blanches looseb congested into 3- to 
7-Mowered corymbs; peduncles obsolete to 5 mm 
long, the lower branches subtended by leaves; bi- 
ll bracts unknown; II branch 
bracts unknown; Moral bracts unknown; pedicels 
stout, terete. 1-1.8 nun long, inconspicuously 
punctate and punclate-lineale. lurluraceous-lepi- 
dole. Flowers 5-merous. color unknown: calyx lobes 
..VI. {5 nun. apical 

ute to rounded, 
tate-lineate. glabrous adaxially. glabrous to 
ely lurfuraceous-lepidote abaxiallv, the mar- 
irregular, minutely erose. hyaline, sparsely 

Distribution. Ardisia hugonensis is known only 
•om the type and is endemic along the Ufo llugon 
i (aioco. (ioloinbia. growing al about .">0 in in el- 

Ecology and conservation status. Ardisia hugo- 
ensis occurs in the wettest of Neotropical forests, 

true pluvial forest that may receive up to II m 
f rain per year. Willi increasing expansion ol hu- 
iau populations in the area, the species should be 
)nsi«lered threatened. 

Etymology. The specilic epithet refers to the 
fo Hugon from where the type was collected. 

Within Ardisia subg. Auriculardisia sect. Pal- 
anae. \rdisia hugonensis is most closely related 

A. smurjilana because of its short calyx lobes 
al are as long as or much longer than wide, short 
•dicels. and very long and wide coriaceous leal 
ades. However. A. hugonensis differs from ,4. 
mrjituna by its smooth branchlets. smaller leaf 
ades to 1 1,9 X 4,1 cm. slender, canaliculate, 
oiler petioles to 6 mm long, and chartaceous. 
ate. larger calyx lobes to 1.6 X 1.8 mm. 

>. Ardisia knappii (Lundell) I'ipoly & Kicket- 
son. Sida 18: 513. 1908. Auriculardisia knap- 

knappii (Lundell) Lundell. I mvlologia 01: 65. 
1086. nom. inval. TYPE: Panama. San Bias: 
23-29 km from Pan-American Highway on El 
Llano-Carti Road. 0O°22'N. 078°60'W, 300- 
400 in. 28 Oct. 1081 (II). S. Knapp 1843 (ho- 
lotype. Lb!; isolype. MO!). Figure 51. 

Shrubs or treelets 2.5-5 n. tall. Branch/els stout, 
terete. 8-14 mm diam.. densely and iniiiulcly ap- 
pressed ruloiis lurfuraceoiis-lepidole. glahreseent. 
h-aves with blades membranous and chartaceous. 
elliptic or oblong. 24.5-51 X 0.2-15.1 cm. apically 
acuminate, with an acumen 5-18 mm Ion-, hasalh 
cuneate. decurrent on the |)etiole. prominently 
punctate and |iunclale-lineate. sparsely lurfura- 
ceoiis- 1. 'pidole ab.i\c. detach I in f ill aceousdepi- 
dote below, the midrib impressed above, promi- 
nently raised below, the secondare veins 60 to 70 
pairs. | i | above, prominently raised be- 

low, the margins entire, inrolled; petioles stout, 
marginale. 3-8 cm long, sparsely furfuraceoiis-lep- 
idote above, densely furfuraceoiis-lepidole below. 
Inlloresiciiccs ci.< I. Inpimialeb paniculate. 9-20.5 
X 5.5-17 cm. pyramidal, shorter than the leaves, 
densely mixed < upiililoriii and hufiiraceons-lepi- 
dote. the branches congested, nearly glomerulate. 
into 15- to 25-Mowered corymbs; peduii(4es obso- 
lete to I cm long, the lower branches subtended by 
leaves: inflorescence brads unknown: inllorcscence 
branch brads . .m I u. on-, menibraiious. ovale to ob- 
long, l.L-2.0 X 1,4-1.0 mm. apically acule. the 
veins unknown, prominently punctate and punc- 
tale-lineate. glabrous above, furluracenus-lepidole 
below, the margins irregular, minutely erose. hva- 
line. sparsely glamlulai cili..lal<-: Moral brails sim- 
ilar to the inflorescence branch brads, but ovate. 
1.2-1.5 X 0.7-1 mm; pedicels slender. 1-1.7 mm. 
prominently punctate and punctale-lineate. lurfu- 
raceous-lepidoie. Flowers 5-merous. purple to ma- 
genta; calyx lobes 
1.3—1.5 mm. apically ; 
and punctate-lineate. glabrous adaxially. furfu 
ceousdepidote, the margins irregular, imiiulc 
erose. hyaline, sparsely glandular ciliolate; coro 
membranous. 5.1-5.2 mm long, the lube 1.2-1 
mm long, the lobes ovate, 3.8-4 X 1.8-2 mm. a| 

'dc. glabrous adaxially. mixed lepulole abaxiallv. 
the margins entire, hyaline; stamens 3.0-4.1 mm 
long, the filaments 1.8-1.0 mm long, the staminal 
lube 0.8-0.0 mm long, the apically free portions 
0.0-1.1 mm long, the anthers narrowly ovoid. 2.3- 
2.4 X 1-1.1 „„„, apically apiculale. basally sub- 

prominently punctate, the ovules 12 to 15. Fruits 

globose. 7.5 iihii diam.. prominently punctate. 

Distribution, Ardisia knappii is found on Cerro 
Canta (iallo in llie Indio-Maiz. in liio San Juan. 

Nicaragua, and along llie I'll Llann-Carli h'nad in 
Panama. \eragua>. and San Was in Panama. It is 
not currently known Irom Costa llica. It grows from 
to 450 m in elevation. 

Ecology and conservation status. Ardisia knap- 
pii occurs in tropical wet Ion-Ms along ndgos and 
sleep slopes. Not enough is known of its population 
biology to accun 

Etymology. This species was named in honor 
of Sandra knapp (P>\1). preeminent authority on the 
systematics of Neotropical Solanum. 

Within \rdisia stibg. \uriculardisia sect. Fal- 
manac. Ardisia knappii is most closely related to A. 
pulrerulenla because ol its short calyx lobes that 
are as long as or longer than wide, short pedicels, 
and large leal blades. However, ■\rdisia knappii dif- 
fers from A. pulverulenta by its thicker branchlets 
to 14 mm in diameter, longer and wider l.-al blades 
to 51 X 15.4 cm. wider inflorescence to 17 cm 
wide, longer and wider calyx lobes to 1.6 X 1.5 
mm louver and wider corolla lobes to 4 X 2 nun. 
wider anthers to 1.1 mm wide, and longer styles hi 

3.9 r 

Mpio. <le San Juan del Norte. Reserve lndio-Mai/. 10 
|,„|i„. <;,.,-,•„ Caiila Oallo. 1 I Sep. \>>'K\ ill). /,'. Ii,„;l„ , 

,,/. i:.->i (ill i r.i. ir. Sep. i«)«j!', ill). /.'. nurd,, rl ,,!. Hi, I 

(IIU.K. MO). 16 Sep. I'Wi! (II). R. Kurd,, rl at. Hh.v 
tllll.K. MO). l'\N\M\. Boras del lor., I -end,, ,1 
\e,agiias. SK sale ol island. !{() Mar. I9<)() ((I). I'. M. Pi 
trrson X522 (MO. I S|. Panama: 1 1.2 mi. (on new ma, 

flufl 442.1 (\.\.. MO); 10 mi. from the P 
way on the El Llano-Carl i Road. 21 
Knapp el al. 4738 (LI.. MO); Kl Uano 

1974 (fl. Ir). S. Mori & J. Kallunki 
Llano-Cart f road, near Nusagandi. alo 
1 Nov. 1992 (fr). G. MvPherson & M 
(1TC. MO. PMA). San Bias: trail ah 
vide. 25 July 1986 (fl). ./. McDonagh e 

station. I Mas 1 992 (fl), R. Paredes < 

083°42'W. 0-200 m. 13 July 1977 (fr). R. Lies- 
ner3266 (holotype. MO!: isotype. CR!). Figure 

Trees 2.5-8 m tall, 4-10 cm diam. Branchlets 

slender, terete. 2-7 nun diam.. densely cupulilorm 
lepidole and slipilale slellale-loineutellous. Leaves 
with blades chartaceous. elliptic 6.6-17.4 X 2.6- 
5.8 cm. apically acuminate, with an acumen 7-17 
mm long, basally acute, deeurrent on the petiole, 
inconspicuously punctate and punctated itiealc. -fi- 
brous above, cupuliform l.-pidote below, with ad- 
ditional stipilate-slellate Irichomes along the mid- 
rib, the midrib impressed above, prominently 
raised below, the secondary veins 21 to 41 pairs, 
prommuloiis above ami below, the margins entire, 
flat; petioles slender, canaliculate. 8-15 mm long. 
1-3 nun diam.. glabrous above, veslilure below as 
in branchlets. Inflorescences erect, bipinnately pa- 
niculate. 4-8 X 3-7 cm. pyramidal, usually shorter 
than the leaves, cupulilorm lepidole and stipilale- 
stellate tomenlose. the branches loosely congested 
into 5- to 12-flowered corymbs; peduncles 0.5-1.2 
cm long; inflorescence bracts unknown; inflores- 
cence branch bracts membranous, oblong. 0-18 X 
3.2-4.2 mm. apically acute, prominently punctate 
and punctate-hneate. glabrous above, vestiture as 
in branchlets below, the margins irregular, minutely 
erose. hyaline, sparsely glandular ciliolate: floral 
bracts similar lo the inflorescence branch bracts, 
but spatliulale. enclosing the bud (similar lo 6W.v- 
santhus). 2.5-4.2 X 1.5-2.5 mm: pedicels slender. 
4.5-6.2 mm long, inconspicuously punctate and 
punctate-lineal.-, sparsely furfuraceous-lepidole 
and siipitale-slellale. Elnwrs 5-inerous. pale or- 
ange to light pink or pink-tan; calyx lobes charta- 
ceous. orbicular. 1.7-1.9 X 1.7-1.9 mm, apically 
acute, prominently punctate and punctate-lmeate. 
glabrous adaxiallv. sparsely furfuraceous-lepidole 
abaxially. the margins irregular, minutely erose, hy- 
aline, sparsely glandular ciliolate: corolla m.-iubra- 
nous. 8.5-8.7 mm long, the tube 2.1-2.3 mm long, 
the lobes lanceolate. 6.2-6.4 X 2.5-2.7 nun. api- 

-. .,.(, .,.;; 

. WKWI. I'M\). long, the 

eb , 

I lube 1.2 


Ardisia liesneri Lundell, Wrightia 6: 100. 
1980. Auriculardisia liesneri (Lundell) Lun- 
dell, Phytologia 49: 344. 1981. TYPE: Costa 
Rica. Puntarenas: Osa Peninsula, Coreovado 
National Park, slopes above Llorona, 08°36'N. 

rally free porth 
linear-lanceoloid. 4.2-4.3 X 1-1.1 mm. apically 
apiciilale. basally sagittate, the connective conspic- 
uously punctate: pistil 7.1-7.3 mm long, glabrous. 
the ovary oblong. 1.2-1.1 mm long, the style 5.7- 
5.9 mm long, mostly epuiictate. the ovules 33 to 

!>-h^ e t '.-fe^- 

n;li-.|\ J 'i.tic ; : puncla i ill. ale 

Distribution. Ardisia liesneri is endemic to the 

Osa Peninsula of Piuitarotias. Costa IJk a. growing 
from sea level to 400 m in elevation. 

Ecology and conservation status. Ardisia lies- 
neri occurs as an understory clcinenl in primary 
lowland vvel forest, lis extremely limited distribu- 
tion makes it \ u lot Tabic to llireal. c\ .-u llioiu li nu-| 
of its known range occurs in protected lands. 

Etymology. This species was named in honor 
of Ronald L. I.iesner. senior curatorial assistant at 
the Missouri Botanical (mrden. 

Within Ardisia subg. Aurieulardisia sect. Pal- 
manae. \idis . i - that arc most 

-mill, ii to tho-c ill L i nissiprdi, I'llala (sec under 
that species for similarities). However. 1. liesneri is 
easily separated from A. crassipedieellata by the 
mixture of cupulilorm lepidole and glandiilai -!< I 

orbicular calyx lobes 1 .9 X 1 .9 mm. smaller corolla 
lobes 6.1 mm long, longer and narrower anthers to 
4.5 X 1.1 mm long, and longer styles to 5.9 mm 

Specimens examined. COST\ l!IC\. Puulurena- 
lirsmu Lorestal Oolfo Dulee. luo ranges nl ( !,-, rn <|e ( >,o. 
eonflueiiee of Kin Puson anil I i 10 liuienn. o Aug. 1001 
(II). «. lgv„/„/^0(CK. I'l'O. IMS. V1()»: ( ianldn .le ( )-,. 
Kesena Lorestal (.nil,. Dulce. Peninsula .le Osa. I.«.s Vln- 
n.». Palna Chal. .10 . lulv 199.1 (Ir). K. Aguilur & W. V«/,o, 

„W>/ (CK. LIO. IM5. MOl: C.vmadn Nairn, \„k. 

along l.lnmna trail to San Pednllo. 21 Julv P>77 (Ir). C. 
Hartshorn 1HU2 il. ITO. MOl: Kc.-ria I'me-lal Coll., 
Dulce. Osa Peninsula. Iroeha <!.■ La larde road. 10 km 
SW of l.a Palma, S of Kincon .le Osa. along ri.lge L of 

the Urn Mincoii sail, x. 28 \|.r pkm; Mb. /;. //,„„„/-/ ,v 

/.'. Ilnhles IhT.lT (CK. IML MO): Kesena Inre-lal Oollo 
Dulee. Osa Peninsula. Kanelio Ouemado. ea. 15 km \\ of 
Kmeon on ridge at MY end of vallev. near Ida Canadn. 
.10 May 1988 (II). H. Hummel el id. 16HW (CK. WW.. I.I,. 
M J : XL. MO. \ I" ): Canton de ()-a. Kesena I'me-lal Colin 
Dulee. ILinrhn Ouemado \alle\. ea. 15 km \\ of Klllenn. 
S side of \allex ale,,, <,>, e „ d <> l.iadoii.i and liio IL 
vilo. 11 Sep. 1000 (IV). H. Hummel el al. I 7.". ! I lIML 
MO); Canton de O-a. Kesersa I n,e-lal (..die, Dulee. 1,'au- 
, l„. nue.nado. ea. 15 km \\ ..( Kmeon. in range l.elme 

Lni,. Ill - Oi i.i I. V .III II ■ • I""2 llrl /-'. Hummel \ 

li \f!uil,a H1572WH,. IML MOl: Parque NaeinnalCor- 
emado. Sirena. Los I'atos Lore-I. 20 Mas |08<> (111. C. 
Kernan & P. Phillips 1126 (CK. LTC. MO): Canto,, de 
Osa. Keserva Forestal Colin Dulee. Peninsula de Osa. 
Kanelio Ouemado range. 21 Sep. 1005 (Ir). 1. Marin & 
H. Gutifrrez 66 (IML MO): Canl.m de Osa. Cerro Kanelio 
Ouemado. Kmeon. 20 Vug. P><>! llrl. ./. Marin Hit (CK. 
1 .!, MO \<; » . i , -• ,, KiiK-iin. permanent 
plot I kniol l$OSCOS\ statin... <> Sep. l""2d,I.A. linen- 
sen 95 (WW,). 

Chiricjuf: vicinity Fortuna Dam. forested slopes 
along ridge at S boundary of watershed. 
08°45'N, 082°15'W, 1250 m, 28 Apr. 1986 (fl, 
fr), G. McPherson 9107 (holotype. MO! [uni- 
or illustration, see Pipoly (1991b: 524, fig. 1). 

villi fine Ion- 

Tree to 4 m tall. Hranchlets s 
gilii.lnial ridges. 5-7 mm (ham., with a dense mix- 
lnre ol 'c iipuliform and furfuraoeous-lopidole s. ale-. 
traces with blades coriaceous, elliptic. 7-18 X 4- 
7.7 cm. apically a. ule to shot l-aeiiminale. with an 
acumen 5-6 nun long. Iia-all\ obtuse to rounded, 
deeurrent on the petiole, inconspicuously 
and punetate-lineate. glabrous above, with a dense 
mixture of cupulilorm and fin luiaceoiisdepidolo 
scales below, the midrib impressed above, promi- 
neiitb raised below, the secondary wins 59 to 18 
pairs, proininulous above and below, the margins 
entire, rexolule: petioles slender, marginate. 1.1-2 
.in long, glabrous oi >par>el\ I'm luraceous-lepnlole 
above, with a dense mixture of cupulilorm and lur- 
liiraeeoiis lepidotc -ealcs below. I n/lnresi fin e.s pen- 
dent, pinnate to bipmnalc paniculate. fV-15 X 8-9 
em. globose, shorter than the leaves, vestiture of 
the raehis. abaxial bract surfaces, branchlets and 
pedicels similar to the branchlets. the branehe- 
loosely congested into 4- to 9-flowered corymbs; 
peduncles 2.5-2.7 mm long; inflorescence bracts 
unknown: inflorescence branch bracts membra- 
nous, lanceolate, 1.5-5.5 X 0.6-1.6 mm, apically 
acute, the veins inconspicuous: floral bracts similar 
to the inflorescence branch bracts except acute. 
1.8-2 X 0.8-1 mm; pedicels slender, recurved. 6- 
I I mm long, inconspicuously punctate and pun. 
tate-lineate. Fionas 5-merous. pah- pink; calyx 

apically acute. promiuenlK punctate and punetate- 
lineate. glabrous adaxially. with a dense mixture of 
cupulilorm and furfuraceous-lepidote scales aba- 
xiallv. margins irregular, minutely crose, hyaline, 
sparsely glandular ciliolale: corolla membranous to 
chartaeeous. 9.9-10.1 mm long, the tube 2.7-2.9 
nun long, the lobes narrowly ovate. 7-7.8 X 5.4- 
5.6 mm. apically acute, prominently punctate and 
punctate -lineate. glabrous throughout exce|)t the 
tube sparsely lurliuaccousdopidole basally aba- 
xially. the margin entire, hyaline: stamens 6.9-7.1 
mm lonu. the filaments 5.5 5.0 mm long, the -ta- 
minal tube 1-1.2 mm long, the a})ieally free por- 
tions 2.4-2.5 mm long, the anthers lanoeoloid, 4- 
4.1 X 1.2-1.3 mm, apically apieulate. basally 
lobate, the conneeli\<' conspicuously punctate: pis- 
til 14.4-15.1 mm long, glabrous, the ovary ovoid, 
5.5-5.7 mm long, the style 9.1-9.4 mm long, prom- 

^lobosr, 5-7 mm diam.. 
icoiispieiiou-h punelale. 

Distribution. Ardisia lundelliana is known only 
(mi the hololype collection, growing around the 
ortuna Dam in Cliiriqui. Panama, al ;>r I I 250 

Ecology and conservation status. Ardisia lun- 

delliana is a ridge-top species in montane forests. 

Because of ils restricted distribution, il should be 

considered threatened. 

Etvmolo<x\. This species was named in honor 

of ihe lute Cyrus Longworth l.undell. specialist in 

Myrsinaeeae for over 60 years. 

Within \rdisia siibg. Auruu/ardisia sect. Pal- 

nianar. \rdisia lundelliana is one ol a number of 

species that have branchlels covered with a mixture 

of dense ciipuliiorm and 1 in 1 in 

scales. In >lcrilc condition, \rdisia lundelliana max 

be most easib confused with 1. con glomerula (see 

under that species lor similarities). However. \r- 

disia lundelliana is easib se])arated from A. con- 

glomeiola because ol Hs pendent inlloies. enre mill 

much louder, recurved pedicel- to 1 1 nun long and 

much larger anthers to 1,1 mm. 

52. Wdi-iu m. phr.-somi Pipoly. Novon 4: 38. 
1994. TYPK: Colombia. Antioquia: Mpio. 
Front ino, area culled Mum', in W-Central part 
of Antioquia. ea. L5 km from Nutibarra, 
06°40'N, 076°2()'W, 1875 m. 3 Nov. 1988 (II 
buds), G. McPherson. J. Zarucchi. F. Rolddn 
& 0. Escobar 12954 (holotype, HUA!; iso- 
types. MO!, US!). 
For illustration, see Pipoly (1994: 38, fig. I). 
Trees 4r-6 ni tall, to 7 cm diam. Branchlets flex- 
uous. slender to stout, subterete. 5-7 mm diam.. 
with a mixture of deiiselv ciipuliiorm and lurlura- 
ceous-lepidole scales, Ia'uics with blades charla 
ceous. oblong to narrowly elliptic. 18-33 X 5.8- 
8.5 cm. apicalh ahrupth acuminate, with an 
acumen 1.2-1.5 cm long, basallv acute, decurrenl 
on the petiole, prominently punctate above and be- 
low, essential!) glabrous above, with a mixture of 
densely ciipuliiorm and furluiaeeous-lepidote 
scales, more deiiselv so along the midrib, the mid- 
rib impressed above, prominent I v raised below, the 
seeondarv veins 3(1 to 38 pairs, proniinulous above 
and below, the margins entire, revolule; petioles 
stout, canaliculate. 1.2-1.5 cm long, essentially 

cupulilonii .nnl lurluraceou-lepidnle scales. Inflo- 
rescences erect. In- to qiiulripimiulelv paniculate. 
21-30 ■ 20 .".(I cm. pviainidal. nearlv as long as 

form and furfuracenii — lepidole scales, the hi. inches 
loosely congested into 5- to 84lowered corymbs: 
peduncle 2.(>-3.l (in long, the lower branches sub- 
tended bv have-: inlloresceiice bracts and branch 
brads unknown; Moral brads earlv caducous, eliar- 
taceous. lanceolate, 1-1.4 X 0.3-0.4 mm. apically 
attenuate, proniiiientlv punctate and puiidale-liii- 
eale. glabrous above, below with a mixture of 
densely cupuliloiin and I'm iiiraceous-lcpidote 
scales, the margins entire, deiiselv lepidde: pedi 

eels slender. 1.2 1.8 mm long. incoiis| nislv 

punctate and punclate-liueate. wilh a mixture of 
densely cupuliloiin ami I in 1 1 1 1 . i < ■■■■■us-lepidote 
scales. Flowers 5-nieroiis. (ream; calyx lobes ehar- 
taeeous. suborbicular to oblate, 1.4-1.8 X 1.7-1.9 
nun. apicalh rounded lo obluse. deu-elv and prom 
inenth punctate and punctaledincale. glabroii- 
adaxiallv. furfuraeeoiis-lepidole abaxiallv. the mar- 
gins irregular, nuiiiilelv erose. hvaliiie. sparselv 
glandula.-ciliolate; corolla coriaceous. 1.2-1.1 mm 
long, the tube 0.3-0.5 mm long, the lobes oblong. 

iiently punctate and punctate-lineate. glabrous 
ihrouul I. tin r i km in 1 1 entire, erose. hvaline: sta- 
mens 3.2-3.6 nun long, the filaments 1-1.2 nun 
long, the stami.ial tube 0.3-0.5 mm long, the api- 
cally free portions 0.5-0.9 mm long, the anthers 
narrowly ovoid to lanceoloid. 2.4-2.6 X 1-1.2 mm. 
apicalh apiculate. basallv sagittate, the connective 
• ■punctate: pistil 3.5 3.'> nun long, glabrous, the 
ovary ovoid, eostate. 5-angled. 1.4-1.0 nun long. 
1-1.2 nun diam.. the style 2. 1-2.3 mm long, epunc- 
tate. the ovules 24 to 35. Fruits unknown. 

Distiibuiion. Ardisia mephersonii is known onlv 
from the \l i j in area of Mpio. de Iroutiiio in Anli- 
ing from 1700 to 1990 in in 


because o| i|- leslrii led di- 

Etymology. This species was dedicated to ( 
don McPherson of the Missouri Botanical Card 
colleague and specialist in Madagascan Kuphorl 

Within Ardisia subg. \uriculurdisiu sect. / 
manae, Ardisia mephersonii is most closely re I a 
to 1. Jurfuruceu (see under that species for simi 
ities). However, Ardisia mephersonii can be disl 

, eoimincl. (.('I <l\' \ \N|i,u|.li.. '■ | : 

I. valine, sparsely glandular eiliolale; floral brads 
similar to the inflorescence branch hracls. hut 0.6- 

km \Y (,l \ulihara. A \1..r. 1002 M.-r.l. 1. Cmtn r, „l. « • ' X 0.5-0.8 mm: pedicels stout. 0.5-3 mm long, 

Thiol (\IO|: \lpi... ilc Ftuntino. km \.\ of mad \utihara- prominenllv punctate and punctate-lineate, furfu- 

La |{lan(|iiita. ivjiiim <>l Mum. Ml- <!<■ C.evas. (» \..\. raceous-lepidote. Flowers 5- or 6-merous. white. 

.088 (fl l,u,l). ./- /«„<„•/,/ ,/ «/. 72.?2 (MO). Ught pink pink _ b | U( , ( , red: ( . a]yx lobes charta . 

eeous. ovate. 2.6-2.8 X 1.8-2 mm. apieallv aeute. 

53. Ardisia mejjistophylla Wright, a 1: |)mmmen|K punc . talt , a „d punctate-lineate, gla- 

147. 1970. Aunculurdisia megistoplnila (Ia.ii- |)mus a( | axia || v . f UI .f.,raceous-lep.dote ahaxially, the 

dell) Lundell. Phvlolog.a 10: 3 1-1. 1081. maii , ms im ,, lt | ar , miiuile | v erosr . | na linr. sparsely 

TYPE: Colomhia. Choc,',: Costa del Pacific... .| aM( | u | a r ciliolate: corolla memhranous, 7-7.2 mm 

ensenada de I Iria. 5 June 1050 (lr|. A. Fer- , |)M ^ „ u , |(l|)e , ; -_, p, mm |()ng _ ,, le , () , )es ()Vate to 

mimic: 211 (hololvpe. US!. IT neg. 1071-09!). mimm \ y ovale< 5.4_5.5 X 2.2-2.3 mm. apieallv 

h -""' ^- acute, prominently punctate and punctate-lineate, 

\nlisia atrahi L.mdell. Wrigl.tia (v. 60. 1 070. Svn. n..i. glabrous throughout, the margins entire, hyaline: 

\iirir„lunlisi<, ulruia (l.undrlli lundell. IM.x i..l..gia stamens 5.5-5.7 mm long, the filaments 2.7-2.8 

49:312. 1081. TV PL: Panama. (.•Ion: bio U.unche mm | (H1 jr, ||,e stamiriul tuhe 1-1. 1 mm long, the 

apieallv free portions 1 .0 I .!'. nun l.m.u. the anther- 

tier &'l lanceoloid. 3.1-3.3 X 0.0-1 mm. apieallv apicu- 

rl)„,le W>L> Vl,nlui M ».'. Vl()!."V"neg. 55670!. IT late, hasally cordate, the connective conspi. uoush 

•g. 1070-,'V}!: isntype, LL!). punctate; pistil 6.4-6.8 mm long, glahrous. the ova- 

macroslachva Lundell. Wrighlia 6: 81. 1970. Svn. ,. y moi( | , () (> blong. 1-1.2 mm long, the style 5.4- 

5.6 mm long, epunctate. the o\ules 1 1 to 13. Fruits 

Labryellow>ever"r7se-.'rcl.^''a"n.'p; globose. 0-0 mm dian... prominently punctate. 

300-700 ft. |91-213 

Nov. Xurimlanlisia mmrostuchxa (l.undcll) Lundell 
Phvloluiua 10: .ill. 1081. Til'K: Par 

Distribution. Ardisia megistophyllu ranges from 

inlial Panama to Choco, Colombia, growing from 

55666!). 91 to kk)() „i in elevation. 

Trees 2-8 n, tall. Branches stout, terete. 0-15 Ecology and conservation status. Ardisia megis- 

mm diam.. denselv rufous furluraceous-lcpidote. "'I' 1 " 11 " l > (, " ,s '" l "» la,1(l P U " Ial " vermt * f, ^ sts - 

litres with blades chartaceous. elliptic to ohlong Apparently, it .s a la.rh rare and should be 

or narrowly ohlong. 23.2-61.1 X 7.1-21.3 cm. api- <-»"s.dere<l threatened. 

eallv aeute. with an acumen 6-17 mm long. basalK Ftvmologv. The specific epithet was derived 

aeute. decurrent on the petiole, prominently punc- l '"" 1 "»' { '^' k ""'cgislo," very big or very large, 

late and punetate-lincale above and below, nearb ; "" 1 "P"* 1 "- leaves ' 

glabrous above-, furluraceous-lcpidote below, the Within W '*'" *"'*• ^'u'durdisia sect. Pal- 

midrib impressed above, prominently raised below. »""""■ W "'" >neg,stoplnlla .s most closely relat- 

the secondary veins 70 to 85 pairs. pron.inentK «' (l "> A - "A"""""""- «- "'^'"- »■ <r«ssip<>s. and 

largins entire, inro- '" ""/"A'"""" because <>l its long <alv\ lobes and 

_, 1.4-5.5 cm long. 4- llli(k Proles. Ardisia megistophyllu can easily be 

. glabrous above, furfuraceous- separated Iron, all dice bv its longer calyx lobes 

lepidote below. Inflorescences erect, hi- to l «> 2 " 8 mm Um ^ l «'»^' r (,m,lla l(,1,f * s to 55 mm 

tripinnately paniculate. 12-37 X 9-19 cm. pvra- '""*• '""^ " llll<1 ^ »» 3 ' 3 mm lon ^ a, » l 1,,n ^ r 

midal. shorter than the leaves, the branches eon- slvl, ' s l " '>■<> mm lo "^ 

gested into 9- to 13-flovvered corymbs: peduncles ''^ 1 >P^ '»< Ar(ii >»« " tr "'" ls lim( l ue onl > ,or lts 

obsolete to 2 cm long, the lower branches subtend- "M ,ll > ll,n --' 1 - l^l"'^ '""1 -l^htly larger calyx 

ed bv leaves: inflorescence brad, unknown; inflo- '«» 1 ^- T,w * 'M"' of 1 " "'"™'^/'>« is unique only 

reseence branch bracts caducous, membranous. '"' -■ — l-r-il.l l^ M.b— I.- p.l.nle- and sbghtlv larger 

to oblong. 3.2-8.3 X 1.2-1.8 mm, apically 
, prominently punctate and punctate-lineate. 
his above, furfuraceous-lepidote below of 

calvx lobes. 

leading li|) . 

|-\\\V1\. <:«,■!•■: 

le flat scale- with llx- inai-iiis entire |n,m l'.a . Inula near 12 Oct. 1077 ifl. fr). ./. J 
teeth, the midrib impressed above. :>,<m (VIO). Colon: !■: >anta liita Kidge. luml.e, i. 

3952 (MO); Santa Kita 
nsisthnius Hwy., 25 Sep. 
. MO). San Bias: Cerro 
e Nevers el al 4046 (MO). 
icion, 4 Dec. 1967 (fr), W. 
)Z). COLOMBIA. Choco: 
)os Boras del Rib Mulata. 
of El Valle, 8 Aug. 1976 
'8 (LL, MO). 

. Ardisia nigropunctata Oerst., Vidensk. 
Meddel. Dansk Naturhisl. F0ren KJ0benhavn 
1861: 127. 1862. Aurieidardisia n^ropunclata 
(Oerst.) Eundell, I 'hytologia 54: 285. 1983. 
TYPK: Cosla Rica. Cartago: Monte Irasii [Ira- 
zu|. 8(MK)-9()(K) ft. 12438-2743 m|. Jan. 1847 
(fr). A. Oersted 2HI) (lec-Uity|M>. designated 
here. C!, F neg. 22953!). Figure 54. 

siu elumlalensis Mr/, in Engl., Pllan/rnr. I\. 236 ( 
9): 90. 1902. Svn. nov. Aurirulanlisia 
(Me/) Eundell, Phytologia 54: 285. 1983. TYPE: 
Panama. Boras del Torn: Eaguna de Chiri<]uf and its 
nrighborhood. Nov.-Der. 1885 (fl). J. Hart 136 (Irr- 

nuimmost, I. Irll. Wrightia I: 60. 196}{. S\n. ,„,\. 

inruhinlisia manimosa (Eundell) Eundell, Plivto- 
ini 51: 285. 1985. TYPK: Nicaragua. Granada: 
mmit of Ml. Moinbacho, near Granada. 1160 in. 
Dec. P>10 (II). 1. Cram H6<> (l.ololvpr. A!. EE 
%. 1971-65!). 

Trees 1.5-10 in tall. 2.6-25 e.n diani. Hram 
slender, terete, 2-7 mm diam., densely and 
nutely appressed ferrugi neons fulfil rareous- 
dote. Iraves with blades membranous to eorian 
oblong or elliptic to oblanceolate. 6-^15 X 2-K 
apically acute to acuminate or rounded, wit 
acumen ;>-2.-> mm long, basallv obtuse to am 

punctate and punetate-lineate, mostly glal 


ddlcra lalamaiica. around Ties Colinas. ICOO-I,'!. 
m 20 Mar. 1984 (fl. fr), G. Daridse. 6'. Ilerrera (. 
& R. Warner 23645 (holotvpe. 1,1.!: isotypes. INB n 

w Eundell. Pliylnlugia (,:J: .1. 

» K.5 1087 IM'I ( osla lii< i 
■ leading from Alto la Palma 

liirac.-nii^-lcpidole. the midrib impressed abo\e. 
piomiiieullv raised below, the seeondan \eins 30 
to :-)0 pairs, hullate aho\e. prominently raised be- 

petioles slender, marginate. 4-18 mm long. 1-3 
mm diam., glabrous above. lurfiiraeeous I. -pidote 

nieiilale. 7-52 X 5-28 em. pyramidal, usually lon- 
ger than the leaves, the rachis straight to genicu- 
late, the peduncle. lachis. bunches, and pedicels 
furfuraceous-lepidote. the blanches loosely con- 
gested into 3- to 9-flovvered corymbs: peduncle 
nearly obsolete to 1.2 cm long, the lower branches 
subtended by leaves; inlloresceiicc bracts unknown: 
iiilloreseenee branch bracts caducous, membra- 
nous, lanceolate to oblong. 1 .3-1.3 X 0.5-1.2 cm. 
apically acute, prominently punctate and pimclale- 
line.lte. glabrous above, lurluraceoiisdepidole be 
low. the margins irregular, minutely erose. hyaline, 
sparsely glandiil.u . iliolale: Moral biaels similar to 
the inllorescence branch bracts, but 0.5-2.1 X 
0.2-0.5 nun; pedieeb -lender, lurluraccoiisdepi- 
dole. I'lnirers 5- or 6-meroiis. while |o pink oi light 
purple: calw lobes membranous to chartaceous. 
ovate. 1.7-2.4 X 1.2-2 mm, apically acute, prom- 
inently punctate and pimctate-lineate. glabrous 
ada\ially. Iiirluraeeou — lepidolc. I h< margins erose. 
hyaline, sparsely glandular ciliolate; corolla iiiein- 
branous. 1.9-5.8 mm long, the tube 0.8-1.1 mm 
long, the lobes narrowly ovale to lanceolate. 3.9-5 
X 1.5-2.5 mm. apically acute, prominently punc- 
tate and punctate lineale. glabrous ada\iallv. lur- 
luraeeoiis-lepidole abaxiallv. the iiiargiiis entire, 
hyaline: stamens 3.3-1,5 mm long, the filaments 

,;: i "- i( 


. PL iitl.i. A'. 11,//,,/ 

long, the apically free j 

lube 0.50.8 
I. 1-2.8, mm 

Missouri Botanical Garc 

date, the connective conspicuous!) punctate: pi-til 
H.5-9..'} nun long, ylnhrous. prominently punclalc. 
the ovary oblong. I -I. I mm Ion-, the style 7..~>-ii.:i 
mm long, proiiuncnllv punclalc and punclalc-lin- 
eale. the ovules 12 to 20. Fruits globose. 5-9 mm 
• ham., promuicnllv punctate. 

Distribution. Ardisia nipropum lain i- 

3200 i 

Ecology and conservation status. Ardisia nigro- 
puncluta occurs in moist forests, from prcniontane 
to cloud forests, ami is always found along tlic Ici- 
est margins. Its tolerance to disturbance and broad 
habitat iaii;je combine to give the species great re- 
sdiencv. and it is not considered threatened. 

Flymolo^. The specific epithet refers to the 
black punctations throughout the plant. 

Common Names. "High Ridge blossom berrv" 
(P. Gentle :i<M2); "blossom berry" \l>. Gentle 0106); 
"Asib" (P. Gentle 7187): "Blossom berry grape" (P. 
Untie :WI): "Manchador" |.S. Record 25); "Uva 
Montanera" (R. Rueda et al. 2674). 

Pipoly (unpublished data) has observed that this 

Within Ard 
uinae. Ardisia nigrapunctata i 

s unique only for its 

but matches 

The type of 1 

wilh slronglv 

which l,undcll(|0(>b) characterized as "mamn 

or nipple-shaped. The type of Auriculardisia 

drata is unique for its slightly larger and mor. 

brous leaf blades with conspicuously | 

. nigropunvtata 
. The type ol \ Auriculardisia 
1 unique only for its large, glabi 

kvhich are similar I, 

the fruits match those 

of Ardisia niirropunctata e\- 


Specimens examined. 

BELIZE. Cayo: Chiquibul, 

Raspaculo Camp, 28 Mai 

. I«>.)( H |,). 1. Moun, Pi 17 (WW. 

MO). Staim Creek: Slat 

in Creek \allev. 17 mi.. 7 Keh. 

MHO (II). l\ Gentle ,1205 

(A. k. 1,1. |2|. MICH. M. 1 S): 

Slanii Creek \alle\. Mountain Cm Hidge. M) Mar. I'MO 

(fr). P. Gentle .1294 (A. 

1.1. MICH. MO. \H Toledo: 

southern Maya Mountain* 

-. liladen Yiline Heserve. river- 

laden Hrailcli. i! Mav l<><> ( , (fV|_ 

G. Daridse & M. Meadow 

s:;.»r.w iuich. itc. \to. ski.c 

southern Maya Mountain. 

-. niaden Nature- Reserve. West 

17 iIIl D. Hollands. H. Kid 105 

(RUM, MO. SKI.). Gl m 

CubilgUitz and Yakapur, 

44888(F). Izabal: l.os Ai 

ideslo Lntre Kins. 1 Mar. I<>2(, 

(1,1. |2|). HOM)l HAS. Cones: Cienaga tract .war \» 
A/ul, Lake Yojoa, 31 Dec. 1952 (fl), L Williams & 
Williams 18793 (IS). Grariaa a Dios: Camp Tiro. 2 i 


al C. The collection from barba is labeled as 1. 
Oersted 28G. .mil lli. Ii.i-.ii eollcelion is labeled as 
I. Oersted 2111). We hereby designate the I. Oersted 
281) collection at C as the lectotype because it is 
by far the more complete collection. 

In his original description. Me/ (1902) 
three collections for Ardisia chontalensis. R. Tale 
228 from Nicaragua (in young fruit), li. Seemann 
5<) from Chonlales. Nicaragua (in bud), and ./. Hart 
186 from diiriqiu. Panama (in flower), all from k. 
All three sheets are fragmented with no attached 
leaves. We here select the original k specimen of 
J. Hart PAG as the lectotype because it is in flower. 

Ardisia ni^ropundata. as with many species of 

over a broad range of leaf and floral part sizes, and 
corresponding to the type of .4. chontalensis is 

■ n IViia lilaiira. Cataeamas. 2!i \|,r. IT, 7 (1,1. 
,1. Malmu li. II. Uii it. N , ). Simlu Iturhur.i. < „„,,, 
trad, area K ol Lake V,,oa. <> Lei,. I <>Sl (ll 1. /! U/,„o//,i 
(L). Yoro: Cerm l.elueen Kio Ouan Ouan and H.o Texa- 
guat. L ol Cerro Ouan Ouan. S ol San Jose in ||,e |{,„ 
i Nombre de Dios. (, 

I i-tt-t I Wi\G.[ \ Oli.mtalcs: Can, ( 

/ el „/. : 

■ ■a. ■'! I 


will, buffer v. 

\\l\ \| ■ 

<) Apr. IWH (fr). A'. Rueda & I. Goronado till I (III LL. 
Mil,. Malawi,, a: Maei/o- de Lena- Blaneas. SL side 
di. linage of Mu. la.,.!., II Ouelaad.iu. -lope- \ A \\ ,,| 

H<la. Nm Martin, on I .order uilh l)e|.t. Ju -a. lii-20 

Jan. I<«!2 III). II. >l,,rns n „/. 20>>»<> ||)l KL. II\M\. 
MO. \U Noeva Srpmia: along Ouel.rada Taslash. 
hiancl. ..I IlioSnlonl.. :i kin s„| .|ala|,a. (, \,,r. 1<>77 (It). 
I). Xedl /o/i'XMO,. Kf„ San Juan: .alley of |{,o Indio. 

2 1 lei, |<>77, 

i. H. ' 

1 ■ ■ •- ' i .' !•' 

is S of Hacienda 

21 Jan. 1 98.3 no. 

/' 1/„,yv/o /«/r.W (MO. \H Zelava: NF Nicaragua, re- 
„o„ol Iha-manV Bluff. ...-:.. Nalinn 151. 16 Dec. 1027 
(III. /■: l-jwlrsitif! '<> IK K): Monkev Point. Cam. I.I Palo. 
1.5 km from raiiM.ii. 25 Ocl. 1981 (In. I\ More,,,, 12.177 
I MO. Ml: near Itil Tingnia. km N\\ of Bonanza. Pi 
Max 1978 (111. />. W/.W77 |||\\1\. MO): Bonanza, on 
grounds of Neptune Mining Co.. 26 Pol, 1979 (|,i. ./. /'/- 
i„,h .*..-'_' (IIW1\. MO. Ml: P-larmn I'Apen ni.-nlal II 
Beoreo. 7 Pol,. 1985 |i, I. I). AVo, .W.7 (MO. N1 ). Without 
loealitv: s.,1. (II. frl. E. Encdriehslhal (K|. 1867 1868 
111)./,'. 7,,/e JJ.S'iki. CO>| \ Pl« \. \laj..ela: E slopes of 
\olean \1i.avall«->. \\ ol Bijauua. neai the Bio Zapole. I I 
12 Feb, 1982 (III. II llnrii,; el ol. I l<> 70 |CB. P. I T.N \ I: 
Reserve Forestal. San Ramon. 18 Pol). 1983 (hi. I. (.,,, 
ml,,,! 355 ,1.1. MO. Ml. Curliuso: n.-a. bridge nve, Bio 
Orande do Orosi at Tapanli. 2 Dec. PC'llll. IV I. T. \nloim, 
,aV2(F. FF);Turrialba.Tauitis. between I laeienda Moiav la 
and Calaveras. ■". \ug. I<»05 111), 6. //<■,, e„, ,'i_T ,' H.|{. I , 
Ciiaiia<a-I«- (.anion do Tilaran, San Cerardo Abajo. Rio 
Cairn \r K n,. linoas Ouesada and Avcc. 5 Doc 1991 (II). 
A'. ««•//» 6'. <fc E. Cruz 4298 (PTC IMP MO). Ile.edia 
Canton de Savapii|iu. Pari|no Naeional Braulio Carrillo. 
I'uoslo PI Ceiho. on ridge ere-l 250 in F ol'Transool trail. 
5 Mar. 199 I (slcr.). />'. Hovle el al. 2920 (PTC. IMS. MO): 
M,.nle Baiha |ISaiv ,|. Max 1847(A),. ' 

I S of Ouebra- 

\K> I'a, 

345. 1981. TYPE: 
slopes SK of Cerro Punla, 65(K)-7000 ft. [1981- 
2134 m], 22 May 1971 <fr). C. Proctor 32020 (ho- 
lotype, LL!, F neg. 55644!; isotypes, F!, F neg 

;w Pimdell. Wrighlia 6: «,. PC'), >\u. tmi. 
\urienlardisia gentry! (Pundoll) Pnndell. Phvlologia 
49: 344. 1981. TYPE: Colombia. Chocd: N ridge of 
\lt„ do Buev. PSW of PI \alle. 500-1 150 in. 8 Aug. 
1976 (fl), A. Gentry & M. Fallen 17318 (hololvpe. 

\,,l,,;„ ',',me,,^ns,s l.nndoll. Wrighlia 6: 85. PC". M,. 

Phvlologia l'»: 3. 15. 1081. TYPE: Nicaragua. Rivas: 
Isla Ometepe. Pago d< 

Clean Madera-. 
summit. !2(M)m. 
■ncelli 3298 (ho- 

Limon: liraulio < arrillo National Park, 
da Con/ales. II \ug. 1992 (slcr.). \l. Bbhlke 100 (P); 
Pinion Kivor. Nov. I "0 I (III. //. I'illie, s.„. (M. PS). I'un- 
larona-: ( anion de O-a. Pila Co-tena. fila Cruce-. head- 
waters of Bio Piedras Blaneas. Corn, \iiguciana. W -lope. 
9 Dee. 1993 (IV), H. Homme! el al. 1927,8 (CIC FTC. IMP 
MO). San Joso: Canton t \e Pore/ Zeledon. Cn. n, a IV-rra- 
l,a Sierpc. Estacion Santa Plena. 13 Sop. 1997 (slcr.). E. 
Mfaro & M. Set;, in, 13117 (I NIP MO): Canton do Perez 
Zeledon. Clurripo. Cordillera Talamanea. Camin.. a Corn, 
Cliirrip.'.. 31 JnP I'^H, hlrr.l. /.'. Co mho,, AJ I. Ilojas 733 
(I Mi. MO). P\N\M\. Boeas del Toro: 1.5 mi. W of 

Almiranle. 15 Oct. I«N,5 ill). K. Ilium I 173 l\KD. , 

headwaters of liio Culebra ca. 5 km PNP of Cerro Pate 
Macho. II Feb. 1979 (f r |. /,'. Hummel (>1 I0\ MOi: u.imh 
,,l Chili, |in Lagoon. Water \alle\. 1 Dee. 1910(11). //. ,-„/, 
Wcle! 1323 (Li. <:i,iri M „.: Porluna Dam area. N fork of 

i (Pundell) Pun- 

IJ,,, Pnnidenria .'. k.u_SE ol \clii..l.- near W border of 

,11. Ml';. V<ra»ua>: from Santa fc past \g School 
to base of Cem. Tut i- along InM -tream. flows from Cerro 

Tuli. 4 Feb. 1977 (ster.), ./. Eolsom 1593 (MOi: PI - «le 

Calivo. 1811 (IV). A. Ine,h,,l,sih„l 013 (W [2|. P neg. 

. Ardisia palmana Donn. Sm.. Rot. Guz. 27: 
434. 1899. Aurkulardisia palmana (Donn. 
Sm.) Lundell, Phytologia 49: 345. 1981. 
TYPE: Cosla Kica. San Jose: in sylvis prope 
La Palma. 1460 m, Sep. 1898 (fl), A. Tonduz 
12632 (7460) (holotype, US! [2]; isotypes, 
BM!. CR not seen. F!, F neg. 68244!, G! [3], 

2 1 Pel,. 1978 (IV). II \eill \ I 

lotype, M0!). 
culanlisid Pundell. Phvlolngia 
1981. Syn. nov. Ardisi 
dell. Phvlologia 61: 63. 1986. nom. iuval. \rdisia 
euruhiginosa (Lundell) J. P. Morales. Phvlologia 83: 
111. 1997. TYPE: Panama. Bocas de Ton,: Cordil- 
lera Talamanea. 2-5 airline km N\\ of the peak of 

border. 09°03-0PN. 082 50-5PW. 
i, 1 & 9 Mar. 1984 (fr), G. Daridse, L 
*rrera, C. Chacon & I. Chacdn 25486 

.!: i-„upes. MO!, NY!). 

I: 1(). 1968. Ardisia azaharemis Pundell. Phytologia 
61: 62. 1986, nom. inval. Ardisia azaharensis Pipob 
& Kicketson. Sida 18: 511. 1998. TYPE: Costa Rica. 
Alajuela: 15 km NW of San Ramon by air. Cerro 
Azahar. headwaters of Bio San Pedro, by road. 9 km 

left again on jeep road 15 km to lop of ridge. 
I0 09'30"N. 081 3l-:i5'\\. 1400-1500 m, 14 May 
1983 (fr). R. Eiesner. E. Judziewicz. J. Gdmez-hturito. 
B. Pfrez G. & A. Carvajal 15575 (holotvpe. PL!: iso- 
types. P!. MO!). 

Small to large trees 2-22 m tall. 10-35 em diam. 

Hranelileis slender, terete. 3-10 mm diam.. densely 
liiiiurneeiiiis-lepidole. Emins with l.lado membra- 
nous or ehartaceous. elliptic or oblong. 8.3-32.7 X 

2-19 mm long, basallv acute, decuirent on the pet- 
iole, the midrib impressed abme. prmniiientb 
raised below, the secondary veins 52 to 65 pairs. 
pmimimhnis above in. I belou. pininnieiil l\ [ >u 1 1< 
late and punetale-liiieale. deuselv fiiifuraeeons-lep- 

Missouri Botanical Garden 

idote, mostly will, a-,- above, tin- mar- 
canaliculate. 5-12 mtii long, sparsely to densely 
furfuraceous-lepidole. Injlores, enees erect, tripin- 
nately paniculate, 14-27.5 X 14-38 em, pyrami- 
dal, longer than the leaves, mixed eiipiiliform ami 
llllluraceous lepidole. the branches loosely eon- 
bested into .'i- to 12-llowered eorvmbs: peduncles 

tended hv I. ;i\ <■-: uilloresceiice brads unknown: in- 
llorescence lii'anch bracts unknown: Moral brads 
caducous, membranous, ovate. 1.6-1.9 X 1-1.2 
mm. apically acute, glabrous adaxiallv. lurfura- 
ceous-lepidote ahaxiallv. the margin minutely 
erose. hyaline. sparsely glandului ciliolale: pedicels 
stout, 1-2.5 mm long, inconspicuously punctate 

and | elate lineal.-, etipuliform le|iidole. Flowers 

5-nierous. while to light pink: ealw lobe- .1km I . i 
ceous to coriaceous, orbiculai to oblate. 1.1-1.4 X 
1.4-1.9 mm. apically acute to rounded, prominent- 

xiallv. sparsely lurluraoeous-lepidote abaxiallv. the 



. hyaline, sparsely 

, 4-4.8 i 

long. lli. lube 1.1-1.4 mm long, the lobes narrow Iv 
ovate to ovate. 2.7-3.7 X 2.2-2.5 mm. apically 

glabrous throughout, the margins entire, hyaline; 
stamens 4.1-4.4 mm long, the filaments 2.3-2.5 

apically bee portions I .(>- I .9 mm long, llie anthers 
narrowly ovoid to lanceoloid. 1.9-2.5 X 0.8-1.2 
nun. apically apiculale. basally subeordate. the 
connective conspicuously punctate; pistil 4.9-5 
mm long, glabrous, the ovary oblongoid. 1.3-1.4 
mm long, the style 3.5-3.9 mm long, epunctate to 
inconspicuously punctate, the ovules 30 to 33. 
/■'/nils globose. 5-*J mm diam.. prominently piuie- 

Distribution. Ardisia palmana is distributed 

•om Rivas. \i. -aia-Mia. lluouidioiil Costa liica to 
ocas del Tom. and Cluriqui. ranarna. un-wiug I'rom 
00 to 2850 ... in elevation. 

Ecology and conservation status. Ardisia pal- 
>ana occur- in primary and -ec. Miliary lorc-l-and 

However, \rdisia palmana may be separated bom 
■1. anchicayana by the shorter calyx lobes to 1.4 
mm long, shorter corolla lobes to 3.7 mm long, 
shorter anthers to 2..i mm long, and shorter styles 
to 3.6 mm long. 

"Forets de l.a Palma. 1159 ,,,. 25 l\ I8<>8. I,/. 
Tonduz I2M2." However, two different handwritten 
label- ,-,\\\ be found, one from the Herb. Instil. 
I'hvsieo-Ceogr. Nat. ( iostaricensis. the other from 
the Herbier Roissier. In John Donnell Smiths orig- 
inal dcs( nption he Latinized the tv})e collection 
stating. "In sylvis propre La Palma. all. I 100 .... 
Sept. 1898. Tonduz n. I2M2 herb. nat. Cost." John 
Donnell smith mad.- mass di-l ribut Ions of this col- 
lection with labels titled "Fx Hands Cualemaleii- 
sibus Neennn Sab adoreu-ibus. I londiireiisibus. \i- 

I'oMiiell Siniili."" |h,. remaining label lor this spe- 
cies slates. "7160. Ardisia Palmana. Don... Sin. in 
Hot. Caz. xxvii. 434. La Palma. Prov. San Jose. Cos- 
ta Rica. alt. 1460 m. M.[ month | Sept. 1898. Leg. 
[collector] Tonduz. (n. I2.M2 herb. nat. Cost.)." 
Thus the number 7I()0 has often been associated 
with this collection, which is the Donnell Smith 
Herbarium number rather than Tondu/s collection 
nuniber. It is important to note that the hololype at 
IS consists of two sheets comprising a sinule gath- 
ering. one with the inflorescence and the other with 
the branehlel apex. Therefore, no leelotv pificalion 

corresponding to the type of Ardisia rufu is un 
only for its slightly more rufous indument. The 
of A. Ixx/uclensis is unique only for its thi. 
leaves and apparently few flowers; however, it i 
fruit and appears to have lost most of its flow 
The type of A. gentryi is unique for its smaller f 
ers. which are primarily in bud. I he type o 
omctcpensis is in young fruit and only unique 
slightly larger and more numerous ovules 
slightly thicker pedicels. The type of Auricularo 


Etymology. The specific epithet was i 
<.m the nam*- of the type locality. La Pain 
ase, Costa Rica. 

Within Ardisia subg. Auriculardisia seci 

nd thicker calvx lobes. The type of 1. 
is in fruit and unique for its slightly sin 
anil thicker, shorter pedi. <>\~. I loweyei. 

mined. NICARAGUA. Rivas: near 

- slopes of Volenti Maderas above Ral- 
, II Sep. l9K.'i(H). M. Nee & W. Rahlvlo 

T. 28089 (MO). COST\ l!IC\. Vhiju.-la: xirii.ii> ,,l Da,,,- 

/areero. :> Dec. I <>Ti2 (IV). It. Urn son 200 \ (MO. WIS): 
Canton (If Sari ICimoii. Cordillera filarali. Monleunle. 
Sari (.cranio liiol.^ical Siiili«.ii. 3 I Mai. I 0'C, ,|| : , />. /■„, 
»n.< I'/.Vi (CIL FTC. IMl. MINI . MO): licsorva Foreslal 
San liamon. ca. Colonia I'almarena. 20 Julv 1981 (II). ./. 
/'//w/v "'^ < M() - M - TKX > : I'almi'-". Ht--i..n ,,l /arm,,. 
12 0,1. 10.57 (III. I. Smith .1(11 (I. MICH. MO). Carlago: 
\,,|,,„,Ti,,,,all,a. I7N-,,. l'X..Hlll. /.. /.Vr/wrr// /6662(C): 
ucmh ,,1'O.iebrada Casa lilaiica. lapanli. 20 Sep. 100) 
(II). 1/. OVr/urn, -W// (I.I.. MO): Panamcrican Niglmav S 
of Carta-io. between T.-ja - and Fmpalme. I.a Trinidad. 
ahoiil .!<> II km from Nm Ice. II Sep. |0<>| (III. A. /,■/», 

.7/56 (K. NY. IS 

,1 l.a Ca 

rimeja al>, 

»nl 10 

km S 

of Kl Tejar. Co, 

Williams el ul. - 

do. i.m 

in de 

n Cacao, 17 

Dec. 1000 (lr). C 


INB. mo; 





Tierras Morenn- 

.. Km 

San Lorenzo. 20 

j (in. ■;. 

R Irifi 

: 226 

MO). Ileredia: 

,..,. Par<|i 

Mranlio Carrilo. 1 

I>I|,I1(IM \l 

1000 (IV). 1). 

21 Apr. 1 

Hire,,, 22,1 (FTC 

slope of Central 

3296 (A, MO. N 

3 Mar. 1020 (lr 


Her.Mlia & Sai 

,' Jo,,.; ( 

Cerro Ilondura. : 

(DUkK. F); Los 

IO:57 (Irl 

/570(F): Canton 

ipe. U.Drr. 

1003 (IV). (',. \ar 

mis <-i ul. , 

'628 (CK. K). Lin 

ion: C 


lillera Talamanca. 

(F): Monies,- 

rde. npp, 

■1 comma 

nitv. 2 

1 Feb. I0«5 (lr). IF 

//,//«■/■ /.;/ri 

San Jo> 

>e:U 1 

ncnial Hi 

vide, NFof San Jeronimo, 

10 17 Nov. 

1069 (fl 

IV). II. Hinder 

1 Dl kl'. F. M 

Zona Prot 

de Eseazii. N slope 

of Cerro Kah 

• ill.- Mir, 

,a- l!a- 

in, 9 Oct. 1991 (II). 

/. »/„„//,.>■ /. 

tains F of San Jose. 

2;; on. loci 

): forels de Kaneho 

Flores, 22 F, 

?/.?/ (UK); Lablazo. 

23 Jan. 1935 (fr), M. 

10 ni) 

. PANAMA. Boras 

" "„ 1 1 ii, . ' _ il I Mo, < hi. Monle A/ill. I.I mi. \ «»l Fiilre Kros on F slopes 

16093 (LL. M0. US). 

36. Anlisiu panamensis Luiidcll. \\ right ia 3: 
198. 1966. Ardisia pallidiflora Standi., J. 
Wash. Acad. Sci. 17: 523. 1927. nom. superb 1 ., 

11011 Ardisia pallidiflora Kidl.. J. Straits Branch 
Hoy. Asiat. Soc. 61: 27. 1912. Auriculardisia 
panamensis (Uundell) Lundell. Phytologia 49: 
345. 1981. TYPE: Panama. Chiriquf: between 
Alto de las Palmas and top of Cerro de la Hor- 
queta, 2100-2268 m. 18 Mar. 1911 (fl). //. 
Plttier 3255 (holotype, US!. UL neg. 1971-79!. 
LIS neg. 2380!). Figure 56. 

Shrubs 1-2 m tall, to 3 em diam. Hranrhlets slen- 
der, terete, 1-3.5 mm diam.. sparsely to densely 
and mmule|\ minus lurluia, ■cniis-lepidole. Leans 
with blades membranous, elliptic. 2^8.4 X 0.8-3.2 

mm li> rig. basally aculc. de-current on the peliole. 
prominently punctate and punclatc-liiicale above 
and below, glabrous above. spars<'ly and mi iiulcly 
lurluraceous-lepidote below, the midrib impressed 

above, prominently raised below, llie sec lary 

veins 28 to 36 pairs, prominulous above and below, 
the margins entire. Hat: petioles slender, canalicu- 
late. 5-16 mm long, glabrous above. Im Im.i, , -on — 
lepidole below. Inflorescences erect, pinnately. rare- 
ly bipinnalely. paniculate. 3-6.5 X 2.5-8 cm, 
pv ramidal. usually as long as the leaves, sometimes 
shorter or longer than the leaves, densely to sparsc- 
Iv hirluraccoii>-|epidole. the brain lie- loosely con- 
gested into 6- to I l-flowercd corymbs: |)eduncles 
0.9-1.8 cm long, the lower branches subtended bv 
leaves: inflorescence bracts unknown: inflorescence 
biarieh bracts caducous, membranous, oblong. 5. 1 
4.2 X 0.8-1.1 mm. apicalb acute, prominently 
punctate and punctate-lineal,', glabrous above. I'ur- 
luraccou-.-l<-pidiile below, the margins irregular, mi- 
nutely hyaline, sparsely glandular ciliolate; 
floral brads similar to the inflorescence branch 
bracts, but 1.4-2.6 X 0.5-0.9 mm; pedicels slen- 
der. 12-15 mm long, cpunctate to inconspicuously 
punctate and punctale-liiicate. usually glabrous, 
rarely with a few scattered furfuraccoiis-lepidotc 
scales. Flowers 5- or 6-mcrous. white, pale pink or 
light purple: calyx lobes membranous, orbicular to 
ovate, 0.9-1.2 X 1-1.2 mm. apically acute to 
rounded, with lew prominently punctate and punc- 
late-lineate. glabrous throughout, the margins irreg- 
ular, minutely crose. hyaline, sparsely glandular 
ciliolate: corolla membranous. 4.3-4.5 mm long. 
the tube 1.1-1.5 mm long, the lobes ovate, 3-3.2 
X 2.2-2.4 mm. apically acute, with few promi- 
nently punctate and punclate-lineale. glabrous 
throughout except spaiselv lurluraeeoiis-lepidote 
on tube abaxially. the margins entire, hyaline: sta- 
mens 3.1-3.2 mm long, the filaments 1-1.1 mm 
long, the- staminal lube 0.5 ().<> mm long, the api- 
cally free portions 0.5-0.6 mm long, the anthers 

<f-U I 

lale-api. ill. lie. basally deeply cordate, tile connec- 
tive conspicuously punctate; pistil 4-4.1 mm long, 

glabrous, llie ovary oblong. 1 .2-1. .'5 mm long, api- 
..illv prominently punciate. (lie style 3.2-3.4 mm 
long. I'pinielate. llie ovules 16 to 18. Fruits globose. 

(> Ml linn ili.iin.. prominently punelale. 

Distribution. Ardisia panamensis is endemic to 
( rrni llon|iieta in Cluriqiu. I I 

1750 to 2268 m in elevation. 

Ecology and conservation status. Fieldwork by 

Pipolv has re\calcd thai \idi\ia panamensis occurs 
along llie forest edge at lie junction of the montane 
and cloud forests. It is locally common, lint should 
he considered threatened because it borders the lile 

/olie prelel-l'ed for coljcc | il.llll.ll 1< HI-. 

Etymology. The specific epithet was derived 
from the type location. 

Within Ardisia subg. Auriculardisia sect. Pal- 
manae. \rdisiu panamensis belongs to a group of 
species including 4. duxeri and 4. vescu because 
of its short calyx lobes, ong pedicels, thin bran, h- 
lets. wide corolla lobes, and long anthers. Hovvc\cr. 
\rdisia panamensis isea-ilv separated fr llie oili- 
er species by its extremely long pedicels to 15 mm 
long, and on mosl specimens die highly geniculale 
branches of the inlloresceiice and persistent sec- 
ondary branch brads arc highly distinctive 

l<> <> I pans. pi n i h! hi I Hi ii 

gins entire, inrolled; petioles slender, eanali. ulal. . 
5-11 mm long. 2 ''> mm diam.. Iiirluraeeoiis-lepi- 
dole above and hclow at first, then glabreseenl 
above. Inflorescences a pseudoraeemose pinnate 
panicle. 17-22.5 X 3-8 cm usually shortei than 
the leaves, the radii-, branches, abaxial floral brad 
surfaces, and pedicels fiufuraceoiis-lepidote. the 
branches loosely congested into 3- to 5-llowered 
corymbs, the lower branches often subtended b\ 
leaves: inflorescence bracts u 
branch bracts caducous, r 
4.1 X 1-1.5 mm, apieally acule. prominently punc- 
tate and piinctale-lineate. glabrous above, furliira- 
ccous-lepidole below, tl margins irregular, mi- 
nutely erose. hyaline, sparsely glandular rib. .late: 
Moral brads similar to the inlloresceiice branch 
bracts, but 1.6-2.5 X 0.5-0.9 mm: pedicels stout. 
1.8-3.3 mm Ion", inconspicuously punctate ,md 
piuiclali' lincale. lurluraceous-lepidote. Fionas 5- 
merous. white to light pink: calyx lobes charta- 
ceous. ovate. 2.6-2.8 X 1.8-2 mm. apieally round- 
ed, prominently punctate and punctate dm. ale 
glabrous adaxiallv. luiltua. e,,u-lcpidote abaxiallv. 
the margin entire, minutely erose. hyaline, sparsely 
glandular ciliolate; corolla charlaceous. 6.4-6.6 
mm long, the tube 1.2-1.3 mm long, the lobes 
ovate. 4.7^1.8 X 2.4-2.5 mm. apieally acute, 
proinineiitlv punelale and puiiclale-liiieale. gla- 

Speeintens t 

med. PANAMA. Cliiriqui: Cordillera brous thn 

..I 1 ill'. I Inn | in |,i (Inn,;, li It'll I rail In sllllllllil. 2 .1,111. I ' » , ."> 

(III. T. Cnrhrane el ,//. 6277 |K. MO): Om. Il.»r<|uela. I! 
km N\\ of H.»(|uete. at lop. 20 Oct. 1980 ill). I'. Unas * 
/,'. Drcssle, Pi.VJ \\\. I I: I'll... Mo.|.iele. (.run I l..ri|iiela. 
28 June 1981 (lb../. I'ipnh 7065(1)1 KK. ITC. MO. W 
IIA. I C): trail tn llonuieta. 1.1 Max 1971 (II. III. 
<;. I'roeh,, :il<><)7 (I.I. |2|l: S slopes nl Cain ||„rqiirta. N 
nl li.x|llclr. 21 Jan. 1071 |!l. IV). It II ilhui <>l ,//. I.SI.y', 


57. Ardisia pseiidui a. cinilloi a Pipolv. ( ialda-i.i 
16(78): 270. 1991. TYPE: Colombia. Narifio: 
Mpio. de Barbacoas. on the road from Barba- 
coas to 1050 tn, 7 Aug. 1962 (fl). /.. 
Mora 2292 (holotvpe. COP!: isotvpes. PSO!. 
US!). Figure 57. 
Trees 30-40 m tall. Branchlets stout, terete, 6-7 
mm diam., densely and minutely rufous lurliira- 
ceous-lepidote. h'aves with blades charlaceous. el- 
liptic, 12.6-29.6 X 5.1-9.7 cm, apieally acute. 
with an acumen 2 7 inn long, basally obtuse, de- 
current on the peli, ile. iioiniiieiillv punelale and 
puiielate-lineale. densely In rf u raceous-lepii lote. of- 
ten glabreseenl above, lb - midrib impressed above, 
prominently raised below, the secondary veins 55 

mens 0-0.2 nun b > 1 1 - . llie filaments 2. ( >-3 mm long. 
the staminal lube 1-1.1 mm long, llie apieally free 
portions 1.8-2 mm long, the anthers laneeoloid. 
3.1-3.2 X 1.1-1.2 mm, apieally euspidale-apieu- 
late basally cordate, the connective conspicuously 
punelale: pistil 7.0- 7.8 nun long, glabrous, the ova- 
ry globose. 2.3-2.5 mm long, the style 5.1-5.5 mm 
long, slender, erect, inconspicuously punctate, the 
ovules 28 to 35. Fruits unknown. 

Distribution. Ardisia pseiidoraeeiii/jlnra is 

ilein cordillera. near Barbacoas. in Narino. Colom- 
bia, growing at 1050 m in elevation. 

Ecology and conservation status. Ardisia pscu- 
doracemiflora oceans in one of the lloristicallv rich- 
est areas of llie ( ordill.-ia < Icidental ol < iolombia. 

well over 800(1 mm of rain per year. The area is 
also known to house many endemics, including 
(llusia garciabarrigac. C. niambiensis. ('.. telvognna. 
\rdi\ia niambiensis. four more undescribed Mvrsi- 
naceae. and many new species from other families. 
especially \raceae and Ce: 

Within Ardisia subg. Auriculardisia sett. Pal- 
uinae, irdisia pseudoracemiflora ma\ be nm-i eas- 
y confused with 1. Ihiiirmi he. aiise ol its elliptic 
■a\es and psendora. rmusc inflorescence. Howev- 

I. I. ji-n -ihliHii, ■eim/lui<i i- < a-ib s.-parnlrd limn I. 
uiifiiii In lis narrower, rliarla. ron- rnlys l< >bes In 
nun wide, shorter corolla lobes lo (>.(> mill long. 
loiter anthers lo 3.2 mm Ion-, shorter styles to 5.5 
mi long, hit U'i uuinbei ol -r.-ondnry \ eins of the 
■af blades, and thinner hranchlets. 

margins i ri < -nhir. minutely erosc. hyaline, sparsely 

mm long, the tnhe 0.6-0.9 mm long, the lohes nar- 
rowly ovate. 1.7-1.8 X 1-1.2 mm. apieally acute, 
prominently punelale and puiirlale-lineale. via 
hmns ihrnughoiil. the margin- .nine. hyaline; sla- 
inens 1.7-1. J? mm long, the hlainenls 0.7-0.8 mm 
long, die slannnal tnhe 0.3-0.1 mm long, die api- 
eally free portions 0.3-0.5 mm long, the anthers 
ovoid. 1.1-1.2 X 0.6-0.7 mm. apicallv apicnlate. 

lale: pislil l.5-l.() mm long, glahrons. I In ■ 
globose, 0.2-0.3 mm long, the style 1.2-1. 
long, prominently punctate, the ovules 12 I 

58. \nli*i:i |...hrn,l,,,l;, Mr/, in I „.J.. I'llau 
zenr. IV. 236 (Heft 9): 88. 1002. Aurivuhmlisia 
pulvertdenta (Me/.) Lundell. Phytologia 54: 
285. 1083. TYPF: Panama. Veraguas: Cap 
Corrientes. Keh. 1818 (II). R. Seemann Km 
(leetotype, designated hy Lundell (1968). K!. 
LL neg. 1071-88!; isoleetotype. BM!). Figure 

Shrubs or tiers. I!,, in, hh-is -Inidrr. Icicle, exfo- 
liating. 3 3.5 mm diam.. Inrlinarenii- lepidole. 
h-ares with blades niembraiious. elliptic. 18.5-20 
X 6.7-7.0 cm. apieally acute, with an acumen 5- 
7 mm long, basally obtuse, decnrrenl on the petiole. 

mron-pi. uoiisb pum -late and j tale line. lie. via 

brous alio\e. tiiiliiiae.on-. lepidole below, the mill- 
rib impressed abme. prominently raised below, the 
secondary veins 3(> to 12 pair-, piominiilnn- above 
and below, the margins minutely erenulale. rc\o- 
I ut«-; petioles slender, marginal*-. 3 5 mm long, gla- 
brous above, densely lurhuarrous-lrpidnlr h.low. 
hi(lore.<.eeiiee< .-reel, pinnale oi bipi nnalrly panic- 
ulate, 11-13.5 X 2.5-1,2 cm. pyramidal, shorter 
than the Irnves. I u it in a. eon-- lepidole. the bran* lie- 
loosely congested into 5- lo 7-flowered corymbs; 
peduncles obsolete; inflorescence bracts unknown; 
inflorescence branch bracts caducous, membra- 
nous, ovale. 1.5-2.2 X 1.2-1.3 mm. apieally acute, 
prominently puuelale and'. -la 
brous a<la\iall\. I in I ma. eon- l.-pidolr abaxialh. the 
margins irregular. minutely erosc. hyaline, sparsely 
glaiKlular ciliolale; floral brails similar lo the inflo- 
rescence branch bracts, but 1-1.3 X 0.5-0.7 mm; 
pedicels slender. 1.5-2.5 mm long, prominently 
punctate and punelale-lineate. furliiraeeoiis-lepi- 
<lote. Flowers 5-merous. a|)pearing light pink or 
red; calyx lobes membranous, ovale. 1.2-1.3 X 

Etymology. The specific epithet 

le/- description ol the "Mores pnlve 
i)l i" meaning the lepidote scales of 
overed with a fine bloom or powdery 
Ardisia pulverulenta is most closely 

rower inflorescence lo 1.2 em wide, shorter and 

narrower calyx lobes lo 1.3 X 0.8 mm. shorter and 

narrower corolla lobes to 1.8 X 1.2 mm. i i 

anthers to 0.7 mm wide, and shorter styles lo I.I 

mm 1 lg 

59. Ardisia rueclae Kicketson & Pipoly, sp. nov. 

TYPK: Nicaragua. Rio San Juan: Mpio. de San 

Juan del Norte, Keserva Indio-Mafz, down riv- 

er 10 km from Cerro Canta Gallo. La Chiripa 

hunting trail, 1 1°07'N. 083°54'W, 100 m, 18 

Sep. 1998 (fl). R. Rueda. 1. Coronado. W. Ve- 

lasquez & y; Ruin 8765 (holotype, M0!; iso- 

type. HULK nol seen). Figure 59. la.i.iiiaiii loliarem <-lliptieain. prilirellos i.s.,iie 

Maceis)iierui.sM-.-.iiHlariis:«5ad53(iH-c2:{ i Hl3l|-j.igis. 

lol.ulis ralv.mi- .nal.s (,„.„ mlu.ula, alms) .lenique ant- 

rete. 5.5-5.5 nun diam.. den-el v 1 1 hi i n i 1 1 1 1 1 . • I \ ap- /elavu. mowing ,il .>() to 112 in in elev at ion. 
pre— ed i iiIimi- lurfuraceoii--lepidnle. I. cores with l\colog\ and consecration stains. \rdisia rnedti 

blades coriaceous, elliptic. 5.5-15.6 X 2.7-7.1 cm. occurs in tall, pluvial forests on lateritic soils, am 

apically acuminate, with an acumen 5-15 mm long. label data indicate that it is locallv common. Itain 

basallv acute, decurrenl on the petiole. prominently Tail in this area of Central America is comparabl. 

glabrous above, densely and minutely appressed ince of Colombia. Because of the remoteness of tb 

lurfuraceous-lepidole below, the midrib impressed populations, and the (act that the land it occurs oi 

above, prominently raised below, the secondary is protected, it is probable that this species doe 

veins 58 to 55 pairs, inn.n-pi< •uuiisly raised above ll(l1 hice immediate threat. 

and below, the margins entire. Hal: petioles slender. Etymology. It is our pleasure to name this spe 

marginale. 6-1 I mm long. 2-5 mm diam., glabrous <''<'* '■' ll( ""»'- «>< Knardo Hueda. dean of arts am 

above, densely and minutely appressed rufous fur- st-it>M,es at ••»«* l uiversidad Nacional Autonoma d. 

luraceous-lepidole below. Inflorescences erect, hi- Nicaragua-Leon, who is a specialist in Verbena 

pi.uialeK paniculate. 4.8-10.5 X 4-7.5 cm, pyra- reat '- 

midal. usually shorter to slightly longer than the Within \rdisia subg. \nricnlardisia sect. Pal 

leaves, the rachis. branchlets. abaxial bract surlac- """""'• »"//*"/ rncdae is similar to 1. dunlapituu 

es. and pedicels densely and n.inuIeK rufous fur- he< ' ;1 "^ ,,f '•* Hliptic leaves, pedicels up to 1.5 mn 

furaceous-lepidote. the branches congested into 5- ,on «- a,,<l s,vl< ' s "P "> *-» »"» '""^ l»»l "<"> b( 

to 0-flovvered corymbs; peduncles 1.1-2.8 em long. s, 'P'»ated from it by lb,- semiterete branchlets. co 

the lower branches subtended bv leaves; inflores- ,ia <'< ,()lls U '* 1 ' l>latl«-M with more secondary veins 

rence bracts unknown: inflorescence branch bracts ova,r ,;llvx '"'^ ;1,ld sh " rt «' r ;mt,,, ' ,s '" 2<) '"» 
unknown: floral bracts caducous, membranous. g ' 

DVate, 0.8-1.2 X 0.8-1.2 mm, apically acute. h.ralypes. MCAIi \<4 \. Km San Juan: \1,,i... ,1, 

promineilllv punctate and pilllclale-lineale. gla- ^an Juan «lrl N.,i1r. lir.-.M-\a,-\1ai/. Ceim Kl Ci^inlr 

hrous above, denselv and minutelv appressed ru- 2I ■''''> l "" ,, <*»■'•■>• '<■ AW " <' "'■ "■>" OILI.K. \1(» 

<r"-" ; '••'■"'■"'- i * - ;; J, i-;:;.r^^;,;;',^^:^ 

alar, minutely erose. hyaline, sparsely glandular | a>li; \| |M „ ,| t . Wvil ,„,„„,,. i;,..,,,,, 1,,,1,,,-M,,,,. | {l( 

iliolate: pe.hcels stout. 1.8-1.5 mm long, incon- I'ijihave <-nli<- .1 ran.. l!i|a-na \ Kl < .em, Cliiripa. I-Clan 

.picuously punctate and ,,.,.. elale-lii.eate densely 1W< ' < IV >- A ' /l '"'< / " '"' "'■ '""-'-' < IU IK - MO). <">. H. 

, , . , , , .' / liuedn el at. KHIJT 1 1 1 1 | \\. MO). | .", | a „. IWO (fr) R 

.ml nunu.elv appressed rufous furfuraceous-lepi- ^ ,., „,_ „„.„ m , , . M()) 

lote. blowers :>-merous. yellow to purple: calyx 

obes coriaceous, ovate. 2.1-2.6 X 2..5-5 mm. api- f>() AnliHia siniirfit ana Rjekelson & Pipolv. sp. 

•ally roundel, prominently punctate and punctate- Mm , T yp, <; . ( ; | om l )ia . V alle del Cauea:Baj< 

.neale. glabrous adaxiallv. densely lurluraeeous- C a lin,a. Concesion Pulpapel-Buenav entura. 

epnlole abaxiallv. the margin* irregular, carrelera Nacional km 28. ca. I IK) ,„. 0.C55' N. 

■rose, hyaline, sparsely glandular ciliolate: corolla (,77°W. 2(. July 1080 (II. fr). M. Monsalre li. 

■oriaceous. 6.1-6.7 mm long, the lube 2-2.1 mm .?/// (hololypr. CI \ C!: isolypes. FIT,!. MO!). 

ong. the lobes lanceolate. 4.4-4.6 X 1.8-2 mm. Figure 60. 

ipically acute, proniineiillv pun. tale and punch 

UK-ale. idabrous lluoiiuhoiil. the marmiis ent 

stamens 5-5.5 mm long, the filaments 1.8-2.2 . 

ong. the staminal tube 0.8,-1 mm long, the apically l.ranarris). inll<.rescenliis 25-32 ( 

ree portions 2.8-5 mm long, the anther ovate. 2.(>- 

5.0 X 0.8-1 nmi. apically apinilale. ha-alb d.-eplv 

■ordate. the connective conspicuously punctate; Unknown habit or height. Bnmcldels stout. I< 

>istil 4.5-5.1 mm long, glabrous, the ovary oblong. rete. horizontally checking and exfoliating. 8-1 

-1.6 mm long, the style A~A..) nun long, pronu- mm diam.. densely lurluraeeoiis-lepidote. Iron 

scalvcinis 0.8-1.0 X 0.6-1.1 Im.n 1.2- 

with blades eoriac.'oiis. elliptic lo oblong. 25-25. 
)ose. 7.5-8.2 nun diam.. incon- X 5-5.6 em. apically acute, with an acumei 
te. nun long, basallv acute to cuneale. decurrent . 

petiole, incou.-picuously piuiclate and punclat 
\rdisio niedae is endemic lo the eale. glabrou- above. Iiiiluraceous-lepidote I 

inarginate. I 1.5 mm Ion-, l: 1 . 1 1 1 m i j - .1! .. >\ <-. I111I11 
nic-cii- lepidole below. Inflorescences (Meet. bipin- 
nate I.) Iripinnalely paniculate. 2 1-25 X 28-32 cm. 
pyramidal, ii-uallv lender llian ihe leaves, iur ("in n- 
eenn- lepidole. lli« Iiraiic hr-, loo-olv congested into 

I. mi-. den-elv IiiiIiii.iic.ii- I. pi, lole: inllnrescence 
ami l.r:ni. Ii hi. ids unknown; Moral brads unknown: 
pedicels slender. I I..", mm Ion-. iiieoii-pieu..ii-l\ 
punctate and punctate-lineate. densely to sparsely 
lurfiuaeeous-lepidote. linnets 5-merous; ealw 
lobes membranous, orbicular to oblate, 0.8-1 X 
0.6-1.1 mm. apicallv acute to rounded, prominent- 
ly punctate ami | el;il< -I iuealc sparsely furfura- 

niemhranous, 3.9-1.1 mm long, the tube 0.9-1.2 
1 1 long III II 11 I, lanceolate. 2.6-2.9 X 

1.1-1. 6 mm. apicallv acute, prominently punctale 

gins entire, hyaline: stamens 2.8-3.7 mm long, the 
filaments 1.9-2.1 mm long, the slanuiial luhe 0.2- 
0.1 mm long, the apicallv Tree portions 1.5-1.7 mm 
long, the anthers ovoid. 1.6-1.7 X 1.6-1.7 mm, 
apicallv . 1 1 >i< ulatc hasally loliale. the connective 
conspicuously punctale: pistil unknown. Fruits glo- 
bose. I. '1—1.6 mm diam.. conspicuously punctate 
and punctate-lineate. 

01. W.lisia la.ariae l.undell. I'hvlnlogia Ol: 67 
1986. Auriculardisia tarariae (l.undell) l.un- 
dell. l'hytologia 63: 75. 1987. TYI'K: Cost; 
Rica. Union: Cordillera Talan.anca. Atlantic 
-lope ( < in. Taiaiia. locally known as Ire- I'i 
cos. 09°09'N. 082°58'W. 2400-2600 m, 1( 
Sep. 1981 (fl). G. Davulse. G. Herrera Ch. S. 
M. Gravum 28882 (holotype. LI,!; isotype 
MO!). Figure 61. 

Treelets to 1 m tall. Branehlets slender, the in- ridge- In lining up to 5 angles. 3-5 mill 
diam.. densely furfiiraceous-lepidolc. I raves with 
lilade- inenihraiious lo eh.n laee.ui-. elliptic to oh- 
laneeolate: 2.2-16.6 X 1-5.2 cm. apicallv acumi- 
nate, with an acumen 1-11 mm long, hasally cu- 
neate. decurrenl on the petiole to the stem, 
prominently punctale and punclaledineate above 
and below, glabrous above, densely furfuraeenus- 
lepidole below, the miilril. impie— ..I abo\e. prom- 
inently raised below, tin secondary veins 23 to 26 
pairs, promiiiulous above and below, the margins 

niarginale. <> |6 mm long, glabrous abo\e. f'u rl"u-' below, Inflorescences erect, bipin- 
nalely paniculate. 9-22 X 1-19 cm. pyramidal. 
lougci than the leave-, die laeln- and branches mi- 
nutely rufous eiipulifbrm lepidote. the branches 

Distribution. Ardisia smurfitana is known only 

peduncle obsolete to 0.5 cm long, the lower bra 

Iron, the type collection in Hajo Calima in Yalle del 

es subtended by leaves: inflorescence bracts 

Cauca. Colombia, growing at about 100 in in ele-*. oblon» ■> 3-3 2 X 01-0 8 mil/ 

^fToIo^v and conservation status. Ardisia smur- 

sparsely glandular ciliolate; floral bracts simil 
the inflorescence branch bracts, but 0.9-2 X 

Jitana occurs in lowland pluvial loresl. because il 

is apparently eiidemn 111 the in.i-l valuable limbci 
stand on the continent, il should be considered 

Etymology, The specific epithet does not refer 

0.1 mm; pedicels slender. 1.5-6 mm long, reeu 
upward, inconspicuously punctate and punc 

lo blue, elf-like creatures, but rather, is named for 
the Sinurlil Carton Colombia Company, whose gen- 

lineale. sparsely furfuracoous-lepidote. Flouei 

erous subsidies provided a group of investigators 

orbicular. 1.2-1,1 X 1-1.2 mm. apicallv a< 

vvilh opportunities to better study and understand 

prominently punctale and punclale-lmeate. 

the flora of Bajo Calima. 

brous throughout, ihe margins irregular, mini 

Within Ardisia subg. Auriculardisia sect. I'al- 

erose, hyaline, sparsely glandular ciliolate: co 

manae. \rdisia smurfitana is most closely relaled 

membranous. 3.6-3.8 mm long, the lube 0.6 

lo A. hugonensis (see under that species for simi- 

mm long, the lobes ovate lo narrowly ovate. 2 

larities). However. 1. sniiujitana differs from 1. I111- 

X 1.6-1.7 mm. apicallv acute, prominently pi 

gonensis by its bian.lil.-i- I101 i/onlallv checking 

tale and punctate-lineate. glabrous throughout 

and exfoliating: larger leaf blades to 23.5 X 5.6 

margins entire, hvaline; stamens 3.8-4 mm 1 

cm; shorter, stonier, niarginale. petioles!,) 1.3 mm 

Ihe (darnels 2.5-2.6 mm long, the staminal 

mm long, the anlhers ovoid. 1.5-1.7 X 0.7-0.9 
linn. apicallv -ill ill lale-a| bas;illy deeply 
cordate, the eoimecti\e conspicuously punctate: 
pistil 3.7-3.0 mm long, glabrous. I lie ovary globose, 
0.7-0.8 mm diam.. the style 3-3.1 mm long, epimc- 
tate. the ovules 18 to 19. Fruits globose. 3-5 mm 
diam.. prominently punctate. 

Distribution. Ardisia larariae is known from the 

( lord i Hera Ta lama ilea n-\ir the ( .oMu Pica-- Panama 

border in 1 .iiiion and Pi nl 

arenas. Costa Pica, grovv- 

ingat 2100 to 2750 m in 


Ecology ond conserrati 

'on status. An!;: > ■ :.:<<: 

rioc occurs al tin inlerficeol upper cloud and elfin 

forest. In this geographic 

area, llie forests are dom- 

inated bv oaks, but also 1 

lave sizeable Podocarpus 

population^ as well. Because it is known I'rom so 

lew specimens, its curre 

nt conservation status is 

Etymology. The specific epithet was derived 
from the type locality on Cerro Tararia. 

Within Ardisia subg. Auriculardisia sect. Pal- 
ma/i'ic. \rdisia larariin is mo-l closely related to 
\. tenuicaulis and 1. tenuis because of its very 
small calyx lobes, less than 1.1 mm long. Ion- ped- 
icels, narrow corolla ln'ies. and short anthers. Ar- 
disia larai/ai • an easil\ he dislin; in-lied from both 
related ta\a by its larger calyx lobes to 1.4 X 1.2 
mm. longer and w ider corolla lobes to 3 X 1.7 mm. 
wider anthers to 0.9 mt l wide, longer styles to 3.1 
mm long, and smaller fruits lo 5 mm in diam. lei. 

raise el „/. _Vio.»6 ( 

RioTerbi an. I I lie 16.. Si.. 
etal. L>W6o(U„ MO). Punt; 
Parque Indigena Pa A mist ... 
l.icimi Mtai.ii.a. path I'mm ll 
del Silencio, 

5 (fr), L Angulo /W(l\li. \l<>, 

(►2. Vnli^i:i feuui.a.d - I u !, , V\ , ighli.i (,: I |0. 

1980. Icaeorea tenuicaulis (Lundell) I.undell. 

Phytologia 49: 352. 1981. TYPK: Panama. 

Chiriquf: Fortuna Dam site. 1400-1600 m, 15 

Sep. 1977 (fr). J. Folsom, R. Dressier & K. 

Dressier 556 1 (holoiype, MO!). Figure 62. 
Shrubs to 1 m tall. Pranchlels slender, angled, 
1.5-3 mm diam.. -parse I v and minutely ferruguu-- 
ou- liiiluiaceousdepidote. Isares with blades mem- 
branous, elliptic. 3.7-8.8 X 1.6-3.9 cm, apically 

long, basalb acute lo oblnse. de. urreul on the pet- 

eeoii--lepidote. the midrib impressed above, prom- 
inently raised below, the secondary veins 20 to 25 

pail-,, proi Ions above and below, the tertiary 

\cins above and below, the margins entire. Hal: pet- 
ioles slender, canaliculate. 5-7 mm long, glabrous 
above, sparsely l.u Inraceoiis lepidote below, the 
margins entire or with small teeth. /;;/<V„, .- 
terminal pendent, bipinnately paniculate. 1-8 X 1- 
7 cm. pyramidal, longer than the leaves, sparsely 
furluraeeous-lepidole. the branches loosely con- 
gested into 5- to 7-llouered corymbs: peduncles | 
6 mm long, the lower branches subtended by 
leaves: inlloresceuce brads unknown: inllore-eeiii-c 
branch brads caducous, membranous, lorate. 1.9- 
3.7 X 0.5-0.9 mm, apically acute, prominently 
punctate and punctate-lineate. glabrous adaxialb. 
usuallv glabrous abaxially or with a few scattered 
scales, the margins irregular, minutely erose. hya- 
line, rarelv glandular ciliolate; floral brads similar 
I.. the uillorescenee branch brads, but 0.8-1.1 X 
0.2-0.3 mm: pedicels slender, recurved, 4-5.2 mm 
long, epunetate lo inconspicuously punctate and 
pniK lale lineate. nearly glabrous or sparsely fin 
furaccous-lepidolo. Flowers 5-merous. light pink: 
. buds] calyx lobes membra- 

apically acute, prominently pun. tale and punctate 
lineale. glabrous llrmughoul or sparsely liirliua 
ceoiis-lepidole abaxially. the margins irregular, 
minutely erose. hyaline, sparsely glandular cilio- 
late: corolla membranous. 2.7 3 mm long, the tube 
0.3-0.4 mm long, the lobes ovale lo narrowly ovate. 
2. 1-2. () X 1 .2-1 .4 mm. apically acute, prominently 
punctate and pinrdaie lineale. glabrous throughout, 
the margins entire, hyaline: stamens 2-2.2 nun 

lube 0.1-0.2 mm long. 1 1 1 e - apically free portions 
0.1-0.3 nun long, the anthers ovoid to narrowly 
ovoid. 1.8-1.9 X 0.5-0.7 mm. apically apiculate, 

hasallv lobate. die . oiiueeliye conspicuously punc- 
tate: pistil 2.2-2.4 mm long, glabrous, the ovary 
ovoid, 0.8-0.9 mm long, the style 1.3-1.5 mm long, 
epunelale. the ovules 5 lo 7. Fruits globose. 7-8 
mm diam.. inconspicuously punclate. 

Distribution. Ardisia tenuicaulis is known only 
from the Fortuna Dam area of Chiriquf. Panama, 
growing from 1200 to 1600 m in elevation. 

Ecology and conservation status. Ardisia lenui- 

caulis occurs in premonlaiie wel forests. It is known 
only from one locality, and given llie intensity willl 
which the area has been surveyed, we should as- 
sume the species is rare and thus threatened. 

Etymology. The specific epithet was derived 
om the Latin "teuui" neaning slender or lliin and 
-eaulis" meaning stem. 

Within Ardisia subg. Auriculardisia sect. Pal- 

.. , , . - .,<-, ■•• :- ! Mi , ■ .1 i ..,il. 

However, the extremely small calyx lobe- il 
more closely with \. tenuis, from which il is sepa- 
rated by the longer' ea w to 1.2 mm long, longer 
corolla lohes to 2.() inn long, longer anthers to 1.9 
mm loim. shorter style to I .5 mm long, ami imieli 

lis is most closelv related to 1. tarariae and 1. ten- 
uis because of its very small calyx lohes to 1.2 X 
1.3 mm. longer pedicels 5.2 mm long, narrower co- 
rolla lobes I,. |. 1 mm v ide. and shorter anther- to 

Speenwn examined. P\\ \\1 \. Clnriqui: vi< inilv of 
Kortuna Dam. m vallev S of lake. 2.i Dee. 19o0 (II). (,. 
Mvl'hvrsnn & J. \mndu 10114 (MO). 

63. Ardisia tenuis Lundell. Wrightia 4: 149. 
1970. Icacorea tenuis (Lundell) Lundell. 
Phvtologia 49: 352. 1981. Aurlculardisia ten- 
uis (Lundell) Lundell. Wrightia 7: 273. 1984. 
TYPE: Panama. Darien: Cerro Pine. 2500- 
4500 ft. 1762-1372 m], 9-10 Aug. 1967 (fl). 
./. Duke & T. El ins 13762 (holotype, LL!, LL 
neg. 1979-4!; isotvpes, GH!, MO!, US!). Figure 



earn, Lundell. Phvtologia 48: 131. I9[!l. S> n. 

\uri( uhutlisia pinenna (Lundell) Lundell. 
logiu 49: 345. 9KI. TMMv Panama. Dane: 

Pirn-, riilpe top near lianrl.o I'la-tic. 1 1>(H) m. 
I July 1977 (fl). ,/. holsam. li. Ilarlmnn cv R. 
rr (hololvpe. LL!: isotvpes. \1K\l!. 

Shrubs 3 m tall. Branchlets slender, terete. 1-3.5 
mm diain.. densely cupuliforin and furfuraceous- 

lepiilole. Lares with blade- membranous, elliptic 
to narrowly ovate. 2.1-11 X 0.6-3.7 cm. apically 
caudate-at uiniiiale. with an acumen (v-19 mm long, 
ba.-ally ohluse to rounded, decurrenl on the petiole, 
prominently piiiiclale and piuiclale-lineale. gla- 
brous alio\e. sparsely f irfuraceous-lepidotc below 
across ihe blade, more densely mixed eupuliform 
and furluracenus-lepidete along the midrib, the 
cm, lull impre-scd above prominently raised below, 
the secondary vein- 25 to 3 1 pairs, prominulous 

eieiiululc Hal: petiole- -lender, marginate. 4-9 nun 
long, glabrous above, densely eupuliform and fur- 
luraceoiis-lepidole below. Inflorescences erect, bi- 
pinnate to tripinnalely paniculate. 5-16 X 5-11 
cm. pyramidal, longer than the leaves, indument of 

pulale with the margins entire or deeplv toothed. 
the branches loosely congested into 5- to 9-flowered 
corvmbs; peduncle nearly obsolete to 0.'> cm long. 
the lower branches subtended by leaves; inflores- 
cence brads unknown: inflorescence branch bract- 
caducous, membranous, oblong or ovate. 1 .3-(>.2 X 
0.3-2.5 mm. apically acute, or short petiolate. the 
midrib impressed above, prominently rai-d below, 
prominently punctate and pimclate-lineale. gla- 
brous above, furfuraeeous-lepidole as in leaves be- 
low, the margins irregular, minutely erose. hyaline. 
sparsely glandular eiliolale: Moral brads similar to 
the iullorescencc branch brads, but 0.4-0.7 X 
0.1-0.2 mm: pedicels slender. 4.5-6.2 mm long, 
inconspicuously punctate and punctate-lineate. 
densely eupuliform and hu luru. .■oii--lepido|e. 
hloicers 5-meroiis. (ream or while: calyx lobes 
charlaceous. ovate. 0.9-1.2 X 0.7-0.8 mm. apical- 
ly acute, prominently punctate and pum tate-liu- 
eale. glabrous adaxially. mixed lepidole ahaxiallv. 
the margins irregular, minulelv erose. hyaline. 
spaisdv eiliolale: coroll.i inembraiious. 
2.9-3 mm long, the tube 0.5-0.6 mm long, the 
lobes ovale to narrowly ovate. 2.3-2.4 X 1.4-1.5 
nun. api-allv acule. prominently punctate and 
punctate-lineate. glabrous throughout, the margins 
entire, hyaline: stamens 2.1-2.3 mm long, the fil- 
aments 1.1-1.3 mm long, the slaminal tube 0.3- 
0.4 mm long, the apically live portions 0.7-1 mm 
long, the anthers ovoid. 1-1.1 X 0.6-0.7 mm, api- 
cally cuspidate-apieulate. basally siibcordate. the 
connective conspicuously punctate: pislil 3.1-3.6 
mm long, glabrous, the ovarv ovoid. 0.5-0.7 mm 
long, the style 2.6-2.9 mm long, prominently punc- 
tate and punclale liueate. the ovules 4 to 6. Fruits 
globose. 1,3-4.7 mm diain.. prominently punctate. 

Distribution. Ardisia tenuis is known from Cerro 
Pine in Darien. Panama, and on Mturas de Nique 
in Choco, Colombia, on the Panama-Colombia bin- 
der. The collection bearing the label Lturance 114 
is probably mislabeled, because the flora of Cerro 
Chapon in lioyae.i i> totally unrelated to that of the 
Choco. and we know of no non-weedy species of 
angiosperms that are shared between the lu... It ba- 
been collected at 750 to 1500 m in elevation. 

Mlhough locally abundant, because of its restricted 
distribution it should be considered threatened. 

Etymology. The specific epithet was derived 
from the Latin meaning thir 

linn branchlets and iiillore-eencc 

Inselv related I 

alyx lobes, less than I.I linn long with I,,,,- |„ ,|- 
ct'ls. narrow (.nulla lubes. IMI | ^horl anthers. \r- 
'isa tenuis can easily he distinguished from A. la- 
armr In ils smaller calw lobes to 1.2 X 0.8 mm. 
Iiorler and narrower corolla lobes to 2.4 X 1 .5 mm, 
laimuei anthers to 0.7 Mini wide, shorter sivles to 
.0 mm long, and larger fruits to 4.7 mm in di- 

be sep 

calyx lobes to 0.8 i 

orolla lobes to 2.4 ) 

The type of Ardusia pirrenna is unique or 
slightly larger inflorescence and leaf b 
wevcr. these fall well within die range of 

Speeimen* ,-u,mnn;l. I'W \ \l \. Ih.rien: Ccrro Pino. 
II \pr. PH>7 (II. IV). Y. Hnsian Mi (MO); Scrranfa <lc 
Pirre. along ascent of Serraina do Pine ah<i\e Cana Cold 
Mine between UioCaiia ..... I I!,,, I'm uelia l.'lli.lo. 27.lub 
l«)7(. (II). '/J C.roal :171UH> [\\\)\; Par.|l,c \aconal .lei Da- 
rieri. slopes olOeirn Mali. Ii.-a.lnatcrs of S branch of liio, ca. 22 km K of Pucuro. 21 Oct. 1087 (II). //. 
Cuadros el at. :!»:i" (MO): ,,.!■■. •|. r .n, a \ of Ccrro Pine, 
between I .-in. I'm,- lop ml h'.mrlio Plastieo, II \o\. 
1077 (II. Ir). ./. l-olsnm el ,il. o;"/r.\IO): summit of Com, 
Cine. 20 Dec. 1 072 (II). I. Centn & 1. Cleieell 70l<> if. 
MO): Par.|iio Yi.ional .lei Marion. Panama Colombia bor- 
der, near gold mine at headwaters of \ branch of Ido 
Purlin, slopes of Ccrro Taoaro.nia. ca. O km N of Ccrro 
Mali. 27 0.1. I«»J7 (Ir). 6. ,le \erers el „l. ll.Vi 1 1 .1 .. \l(»: 
Ohrpigana District, crest of Cana-Oiiasi trail. 15 Mar. 
POO (Ir). M. Tern X A'. Terrs I Mi (I). COI.OMHIA. 
IDoiibllnl. |.iobabl\ represents a label mi\-ii|>- Itovacu: 
Cliap.ii. region. 100 ,„i. N\\ ,,| li.^ola. 27> Max \'>A2 (Ir). 
t. hneranee III a. II. \1< I | < hoc. V S\\ ,-i.ijr.- leading 

. Ar.lisia tvsnuii bundell. Wrighlia I: 165. 
1971. Auriculardism tysonii (bundell) l.nndell. 
I'bytologia 49: 345. 1081. TYPK: Panama. 
Panama: Ccrro Jefe. in C/usia forest. 2700- 
3000 ft. |82.i-0l4 m|. 27 Jan. L966 (Ir). E. 
Tyson, J. Dwyer & K. Blum 3279 (holotvpe. 
MO!, I.I. neg. 71-117!: isolvpes, I.I.!. V neg. 
55646!, MO!). Figure 64. 

'sin fulonensis l.iindell. Wrighlia (r. HO. 1070. Syrr. rrov. 
\urieitl„r,lisiti ptlonensis (l.nndell) bundell, Phvto- 
logia 10: .iir>. 1 08 1. TV PK: Panama. Code: ha Mesa 
above Kl Yalle de Anion, ca. 2 km W of Ccrro Pilon. 
on slopes of sleep knife-like ridge, 900-030 m, 22 

Shrubs or trees 3-7 in tall. 4-12 cm diarr 
Hrunelitels stout, terete. 5-8 nun diam.. densely fur ens depidole. I,, ne-. with blades coriaceous 
elliptic, 3.7-12.2 X 2.1-6.2 nun, apically acuh 

i long. 

ii. I pnnclalc-liiieale abo\e and below. glabrous 
bo\e. densely furfiiraeeoii-.-|epiilolc below, the 
in hib impressed abo\e. promiiieiilly raised below. 
le secondary \eins 27 to 38 pairs, prominiiloiis 
bo\e and below, the margins entire, flat: petioles 
hauler, marginal.'. !! 17 nun lung, glabrous abo\e. 
eous-lepidole below. In/loiesceiiees erect. 
bipinnali'ly paniculate. I 1 X 1-12 can. pyrami- 
dal, usually longer than the leaves, sparsely liirlu- 

inlo 5- In 7-llow. red cor\ nibs: peduncles nearb ob- 
solete to 2 .an long, ihe lower branches subtended 
by leaves: iulloreseenee bracts unknown: iriflores- 

to oblong. 0.8-1.5 X 1.2-1.8 mm. apicalb acute, 
pronimeiilly punctate and puiietale-bneale. gla- 
brous above, furluraeeous-lepidote below, the mar- 
gins irregular, minutely erose. hyaline, sparsely eihulaie; Mural biacls similar to the inflo- 
rescence branch bracts, but 1-1.4 X 0.5-0.8 mm: 
pedicels stout. 0.7- 1. P. mm long, inconspicuously 
punctate and piuiclale-lineale. lurfuraceoiis-lepi- 
dole to glabrescent. Flowers 5-merous. while to 
light pink: calw lubes membranous to ehaitaceous. 
orbicular to ovate, 1.2-1.6 X 1 .4-1 .6 nun. apically 
acute to rounded, prominently punctate and punc- 
lale-lineale. densely lurfuraceous-lcpidole. glabres- 

sparseb glandular ciliolale: .orolla membranous. 
1.1-1,2 mm long, the tub.' 1.2-1.3 mm long, the 
lobes narrowly ovate. 2.0-3 X 1.6-1.0 mm. api- 
cally acute, prominently punctate ami punclale-liii- 
eate. glabrous throughout, the margins entire, 
hyaline: stamens 3.2-3.3 mm long, the filaments 
1.3-1.1 mm long, the staimnal lube 0.3-0.4 mm 
long, the apically free portions 1-1. 1 mm long, the 
anthers ovoid. 2.2-2.4 X 0.7-0.0 mm. apically 

ousb punctate: pistil .'', 3.2 nun lung, glabrous, the 
ovary oblongoid. 1.2-1.3 mm long, lire style 1.0-2 
mm long, prominently punctate, the ovules 23 to 
25. bruits globose. 5.5-7 mm diam.. prominently 

Distribution. Ardisiu tysonii is known hum t 
rea around Ccrro Jefe and Ccrro A/.ul in Panan 
u- area around Ccrro Tute in Veraguas. and I 

Etymology. This species was named in he 
of Edwin I.. Tyson, the collector of tlie type. 

Within Ardisia subg. \uriculardisia sect, i 
manae. Ardisia tysonii is most closely related t 

ever. A. tysonii can easily he separated In 

epala hv its thicker branchlets to 8 mm 

er. wider calyx lobes to 1.6 mm in dial 

er and narrower corolla lohes to 3 X 1 .<j 

tr anthers to 2.4 nun long, and shorter 

mm long. 

indell (1979) failed to compare the type , 

pilonensis. which is in fruit, with that 

ii. Examination of hoi h revealed that the 

S/.,v,„„.„s examined. P\\ \M \. Panama: ..........1 of 

Cerro Jefe. on road below lower. 20 Jan. lOHl (111. //. 
Chun-hill I2HU (I. I.. MO. I'M \): Cerro Jefe. along trail in 
forest al c„«l of roa.l. ri-l.l hand fork. \\ past radio lour,. 
30 Sep. 107!! ill. In. /,' //,„„„„./ /.'.'// (V10): Dllo. Pana- 
ma. Cerro Jefe. I.elou anlenna. IT, June lOiil (ster.l. /. 
/'//W. nun (\|(). \U alonu hud I, on, <.«.„«> A/ul lo 
Cerro Jefe. 10 Jan. 1000 (frl. /•.. Tyson .I.VW (MOl: Cerro 

al.o\e lo, mnen \iri.nrl. P, Jan. I'»77, (fr). A'. Wilbur & J. 
LutcM, /"/o.mIH KKl:C.-m, lower. 2 I Jan. IW7 
(1,1. /. \,ddvsf,ino X ./. wv,m/,/ -.'..'/ (MO. \V I'M \). V,- 
rajruas: \\\ of Santa I'e. 2 km Iron. Kseuela \grfeola 
Mto de I'iedra. ridue lop helow summit of Cerro Tute. 28 
Mar. I<J7.-, (IV). .S. Mori £ ./. Kallunk, .V.Vl ill . MO); 
Cerro Tute. ea. 10 km \\\ of Santa IV. on ridge top. 10 
Mav 107.-, („■). N. Mon (U7II (I.I.. MO). 

I.undell. WrightiaO: 112. 
1080. Auriculardisia ungiiiensis (I.undell) I.un- 
dell, Phytologia 49: 345. 1981. TYPE: Colom- 
I'i.i. Choen: S rn . una del Darien. \\ of I iigiu'u 
near Panama border. 550-1000 m. 25 July 
1070 (fl). A. Gentry. II. h-dn & /.. Form, 17032 
(holotv|)e. I.I.!: isotvpes. ETC!. MO!). Figure 


MO!. I'M A!). 

Small shrubs or trees 4-25 m tall, to 20 cm 
Branch/els slender, terete. 5-5 mm diam., d 
ami minutely rufous lurfuraceous-lepidote. 
glahrescenl. leaves with hlades coriaceous, e 
7.8-17.1 X 2.9-6.6 cm, apically acuminate 
an acumen 4-9 mm long, hasally acute, dec 

tatc-lineate. glahrescenl ahove. densely liirliua- 
eeousdcpidole helow. the midrib impressed ahove. 
prominently raised below, the secondary veins 21 
to 29 pairs, inconspicuous ahove and helow. the 
margins entire, flat; petiole slender, canaliculate. 
7-19 mm long. 1-2 mm diam.. glahrous ahove. fur- 
furaceous-lepidole helow. Inflorescences erect, hi- lo 
tripinnately paniculate. 6.7-11.9 X 8.1-12.6 cm. 
pyramidal, longer or shorter than the leaves, the 
racliis. hranchli-K. ,1 «.i\ ,., I l.rai I -mlaces. ami ped- 
icels densely furfimiceous-lepidote. the hranches 
loosely congested into 3- to 7-flowered corynths: 
peduncles 0.5-2.3 cm long, the lower hranches 

subtended hv Ir.urs; inllni esccn. e bl.lrl- ahscllt: 

inllorescence hraneh hracts unknown: lloral hraets 
caducous, memhranous. ovate. 0.9-1.1 X 0.5-0.7 
mm. apicallv acute, proininenlly punctate and 
pimctate-lineale. glahrous ahove. the margins irreg- 
ular, minutely erose. hyaline, sparsely glandular 

spicuously punctate and | >u 1 1< lale -lim-.ili-. indu- 

branchlets. Flow* 

calyx lohes charlaceoiis 


apicallv hro 
ly puiiclate 
all\. -par-ely 

,"',.-.,■:■■ . 

lell. Wrifrhtia 7: 270. I«>8I. 
r>yn. nov. Ardisia nelmlosa (I.undell) I.undell. Phyto- 
logia 01: 07). I OHO. until, nival, \rdisia nelmlosa 
ll.iiiid.-lll Pipolv ,\ 10, kelson. Sida 01: 7,13. 1008. 
X\ YY-. Panama. Panama: Cerro Jefe. 850-000 m. 20 
Oel. 1080 (fr). k. SMsma IWiO (holotvpe. I.I.!: iso- 
types, KM!, MO!). 
icidardisia parriflora I .iiudell. Wrighlia 7: 271. 1034. 
Syn. nov. INoii Ardisia parriflora Talhot. Syst. List 


miniitels erose. hyaline, sparsely glandular cilio- 
late: corolla coriaceous, 4.8-5.1 mm long, the tube 
1-1.8 mm long, the lohes ovale to lanceolate. 3.1 
4.1 X 1.6-2.2 mm. apically acute, prominently 
punctate and piuirtale-liiieale. glahrous throughout. 

filaments 1.5-1.8 mm long, the staminal tube 0.7- 
0.9 mm long, the apically free portions 0.6-1.1 mm 
long, the anthers ovoid to lanceoloid. 1.8-3 X I- 
1.1 mm. apically apiculate. hasally lohate. the con- 
nective conspicuously pun. lale: pistil 4.7-5.4 mm 
long, glabrous, the ovary ovoid to ohlong. 1.2-1.1 
mm long, the style 3.5-1 mm long, prominently 
pi ni elate and pilllelutc-lincatc. the ovules 15 lo 13. 

Pirre in Darien, Pana.r 


and conservation status. Ardisia 

guiensis oc 
|...rc-l -. ;n '.! 

curs in premontane wet forests, cl 
elfin forests. It has been found in ai 

Ite lairlv resilient, doing well in full-light >i I u;il ion-. 
Because of its ability to li\ *■ in lull sunlight, we do 
not believe it is lineal ned at this t i i ti« '. 

_ t. Name< for the town of I'ngufa, 
Chocd. Colomhia. vvhe e the type was collected. 

Within \rdisia suhg. Aitrieulardisia sect. Pal- 
manae. \rdisia un^iio nsis may he separated from 
both 1. frlandiilosomarfiinata and 1. croatii by its 
inconspicuous abaxial leaf punctations. flat <pialer- 
nan venation, smallei calyx lobes to 2 X 3.1 nun. 
dense long glandular-eiliolate along the entire mar- 
gin, and smaller corolla lobes to 4.1 X 2.2 mm. 

The type of \rdisia iiniiuiensis is in very young 
fruit, while tin- l\pe of \urieulardisia nebulosa is in 
mature fruit, and the tvpe of 1. pun [flora is in bud. 
The types of both 1. nebulosa and 1. pun [flora are 
notable onl\ lor the more densely black punelatious 
of the dotal parts, will no other significant feature 


■imens examined. 

PANAMA. Darien: Cerro Fine, 

9-10 / 

W. 1967 (fl hud) 

. /. i)ukv& T. Klin* h.i:;f>f>L>\\.i.. 

MO. MOCZ); forest trail 

\ from Knsriiada del Cuavalio. 

SE Jaque, 13 Jan 

. |9i',:! Ifrl. \. („,i,r,Hul rl al. 2 In 

(BM } 

AO): Parque Nacioi al del Darien, W slope of Cerro 

Mali. ■ 

\ \ S branches ol Win Pucuro. 

ca. 18 

km K of Pucuro. 

■2 Oct. 1987 (IV). />'. Ihmunel el 

al. 16432 (CAS, F, FTG 

. IT. MFXF. MO. l>M\):Sof Fl 


region called Altl 

old Cam i no Real 

. . 1 ..„!,.... 22 Aug. 1987 

MePherson (FIT,. I.E. MO). Panama: 6.5 

an b) 

road \ of I .ago ( 

:,rro \zul. 13 Jan. 1971 (fr). M. 

>«7(EE, MO, NY) 

; Cerro Jefe, 4.7 mi. above Ooolv 


27 Dec. 1980 (fr), 

K. Sytsmaet al. 2!>,2!i (IT. MO I. 

6. Vnlisia y.sra I .i ndell. \\ «-«■ 

Auriculardisia resea (l.undell) l.undell, Phyto- 
logia 49: 345. 1981. TYPE: Panama. Code: 
near Continental Divide along lumbering road 
8.4 km above El Cope, 1 km beyond sawmill, 
900 m, 19 Jan. 1978 (fl, fr), B. Hammel 952 
(holotype, MO!. F neg. 55676!). Figure 66. 

i! ,,, ! II I i i I I I I, i n I" I 

TYPE: Panama. Code: I. a Mesa, above El Valle de 
Anton, ca. 2 km tt el' Cerro Pilon on slopes of sleep 
knife-like ridge. 'MM-030 m. 22 Jills l«>7r, III bed). 
T. Croat 37161 (hololvpe. MO!. V neg. 55601!: iso- 
type, EL!). 
Shrubs or hers to I in tall. Branehleis slender, 
■rele. longitudinally ndged. exlolialing. 2 1.5 mm 

am., densely appressed rufous In till I areoiis- lepi 
te. glabrescent. I.euies with blades membranous. 
iplie. 3.1-9.6 X 0.8-3.1 em. apically acuminate, 
th an acumen 5-12 mm long, hasally obtuse to 
unded. decurreut on the petiole, prominently 
in. Ial« and piuiclalediii.-ule above and below, ler- 
prominulous above and below, gla- 
brous above, sparsely furfuraceous-lepidote below, 
the midrib impressed above, prominently raised be- 
low, the secondary veins 27 to 36 pairs, prominu- 
lous above and below, the margins entire, flat; pet- 
ioles slender, canaliculate. 5-7 mm long, glabrous 
above, sparsely below. Inflo 
reseenees .'reel, pinnate to bipiiiiialelv paniculate. 
5-7 X 3-4.5 cm. pyramidal, shorter than the 
leaves, appressed furfuraceous-lepidote, the 
branches loosely congested into 5- to 7-llovvered 
corymbs: peduncles nearly obsolete to 0.8 em long, 
the lower branches subtended by leaves: inflores- 
cence bracts unknown: inflorescence bianeh bracts 
caducous, membranous, oblong, 2.7-3.8 X 0.8-1.3 
mm. apicallv acute, prominently punctate and 
punctate-lineate. glabrous throughout, the margins 
irregular, minutely erose. hyaline, sparsely glan- 
dular eiliolate: floral bracts similar to the inflores- 
cence branch bracts, but 1-1.6 X 0.5-0.7 mm; 
pedicels slender. 0-9 mm long, inconspicuously 
punctate and punctate-lineate. furfuraceous-lepi- 
dote or often glabrous. Flowers 5- or 6-merous. 
deep pink to pale purple: calyx lobes membranous 
to chartaceous. ovate. 1.4-1.5 X 1.1-1.2 mm. api- 
callv acute, prominently punctate and punetale-lin- 
eale. sparsely hull u aceoiis-lepidote. the margins ir- 
regular, minutely hyaline, -p.iiselv glaudulai 
eiliolate: corolla membranous. 1.5-4.7 mm long, 
the tube 1.2-1.3 mm long, the lobes ovate. 3.3-3.4 
X 1.9-2 mm, apically acute, prominently punctate 
and punctate-lineate. glabrous throughout except 
sparsely furfuraceous-lepidote abaxially, the mar- 
gins entire, hyaline; stamens 4-1.2 mm long, the 
filaments 1.9-2 nun long, the staminal tube 0.8- 
0.9 mm long, the apically free portions 1-1.2 mm 
long, the anthers lanceoloid. 2.3-2.4 X 0.8-0.9 
nun. apicallv apieiilate. hasally cordate, the con- 
nective conspicuously punctate: pistil 4.4-4.5 mm 
long, glabrous, the ovary ovoid. 1-1.1 mm long. 
epunctate. the style 3.4-3.5 mm long, epunctate. 
the ovules 9 to 11. Fruits globose, 4-5 mm diam., 
prominently punctate and punctate-lineate. 

'■disia resea is known lioni only 

vide. It is apparenlb tare and should llms be eon- 25: 1 l<">. 189".. \uri< iilunlisiu i>/eim>holr\a 

>idered threatened. (Donn. Sm.) Lundell, Phytologiu 49: 345. 

Etymology. The specific epithet refers to the 198) . TYPK: Costa Rica. Alajuela: Potrero del 

bladder-like shape of the flower buds. Alto, Volcan Poas, 2450 m, 20 July 1888 (fl). 

Within Ardisia sub-. Aurirulardisia sect. Pal- //. Pittier 389 (leetotype, designated here, US!, 

manae, Ardisia rrsra is most closely related to A. US nog. 2.582'. I I ■■«■■■. I«>7 1-85!; isolectoty- 

dwyeri (see under that species for similarities). pes, BM, BR!, CR!, G! [2|, W!). Figure 67. 
However, A. vesca differs from A. dwyeri bv its 

, i r i i i i r l » o l'- i i ' " "'" '"■ : I Phytologia 

membranous leaf blade,, longer pedueb to 9 nun , ( , , ^^ ^ ^ 

long, membranous or chartaceous much narrower Tom . l)[||w . r Rf() < :o | u , m .. 2 5(K^(MK) in. Aug. 19«.5 

calyx lobes to 1.2 mm wide with acute apices. ({], f r ), /,. ,;„„„■;. R_ (,h„eun. I. Charon X I,. Ilenvra 

shorter and narrower membranous corolla lobes to 21919 (holntype, 1,1.!; isotypes, MO!, I'MA!). 

3.4 X 2 mm, shorter and narrower anthers to 2. 
X 0.9 mm. shorter stvles to 3.5 mm long, an 
smaller fruits to 5 mm in diameter. 

The type of Ardisia minima is in very young bu 
ami i- notable only lor its slightly larger leaves, an 
although both the types of 4. vest 
were described at the same tin* 

. . " conspu 

lineate, glabrous above, densely furfuraceoiis-lepi- 

Speeunen examined. I'AN AM A. Co<-l«'-: vicinity of La dote below, the midrib impressed above. 

Mesa, II El Yalle. on \ slope ..I' Cerro (-ailal. 12 , ,, , ,it> i n< ■ ii 1 1 v raised below, the secondary veins 6 to 

July 1987 (fl), G. MePhe,on 11230 (IX, MO). ^ ^ ^.^ ;(| _, ^ M< ^ , he margjns ep _ 

tire, revolule; petioles slender, canaliculate, 0.4- 

1 .5 mm long, glabrous above, furfuraceous-lepidole 

pinnately [.anieulale, 3.4-13.0 X 2.4-8.2 cm, py- 

spicuously [)unctale and punctate-lineate. 
luifnraceous-lepidote like the branchlets, the 
branches terminating in 3- to 8-llowered corymbs; 
peduncle 2.1 <».2 <iii long: inflorescence bracts 

<;. \l lilCI lARDIsn SKCT. Pleurobotryae 
iisia subg. Auri 

bracts caducous, i 

0.5-2.1 mm. apically acute to round* 
ly punctate and punctate-lineate. gl 
furfuraeeous-lepidote below like the I 
isia subg. Aurirulardisia. section margins erose. hyaline; lloial bracts 

- sifirnoiileis inside! 

and is defined by its florescence branch bracts, but 2.4-3.9 X 0.5- 

ateral inflorescences, long naked pedun- mm; pedicels slender, terete. 6.2-10.5 mm lc 
f).2 cm long, dense i 1 1 aiilhesis, inconspicuous!) pu 

i most plant parts, and pendent flowers on late and punctate lineate. densely appressed 

iall\ sigmoid pedicels. furaceous-lepidote. i'lourrs 5-merous. white to Ii 

pink or yellow; calyx loh 

Iisia pleurobotrya Dorm. Sm., Rot. Gaz. ular to ovate, 1.4^2.8 

\nl\ia plenmholna. \. I'lnu.aniu l.r.meli. -11 Detail of aba\ial leaf surf; 

puiietatf-liuenle. densely lilt i 1 1 1 .1. ooils-l,-pii|,,|e me- 
dially, the margins irregular, croso-fimbrinle. tlie 
fimbriae composed oi simple or stellate trie-homes; 

1.3-1.8 mm long. the lobes ovale. 3.9-4.9 X 2.8- 
3.1 mm, apically acute, proinineiilly pimclate and 
pimclate-lineate. glabrous, tlie margins entire; sta- 
mens 3.8-4.1 mm long, tl filaments 1.4-1.6 mm 
long, the slaminal lulie 0.7-0.9 mm long, incon- 
-pi' - ! n >i i — I \ putiel.ile Hid pun, -late lincalc. the api- 
cally tree portions 0.3-0.0 mm long, the anthers 
nin wlv v I 2.4-2.0 X 0.7-1.1 mm. apically 
subulate, basalb sagittate, dehiscent In suhapical 
pores, the connective pimclate dorsalK; pistil 3.3 
4.5 mm long, (he ovary obturbinate, 0.9-1.1 mm 
long, the slvle 2.3-3.5 mm long, oiispi, isb 

Fruit glohosc. 7.7-9.1 mm diam.. proimncntlv 
punctate and piiiictaledineate. glabrous. 

Histrihuliuii. \idi\ni i>l,-iin>l>i,lr\d is found 
throughout the high mountains of Costa Rica and 
in liocas del Tom and < :hii i«|in. Panama, from I KM) 
to 33(H) m in elevation. 

h-o/o{;\ and loiisrrriilmii status. IVimarv and 
disturbed or remnant, oak. montane, cloud, and elf- 
in forests. It appears to he quite common in pro- 
served it thriving in public, open areas on Volcan 
it is not threatened at this time, 
fie epithet was derived 
meaning lateral, in a 
nd "botrya, meaning hunch, 
which refers to the elongate lateral inflorescence 
branches with clustered corymbs. 

Common Name. "Tucuico" (Standley 35090). 

\rdisia i>li-im>hotr\<i ha- asymmetric i.ilw I. .he- 
wilh auriciilate bases and Imiuiaccoiis-lcpidote in- 
dument. which clearly place it in subgenus \uri- 
cu/ardisia. However, the strictly lateral inflores- 
cences with long naked peduncles with pendent 
flowers on long pedicels clearly mark this specie-. 
\o other taxa within subgenus \uriculardisn, lia\e 

Although Donnell Smith's types are generally as- 
sumed to be at IS. his original description of \r- 
ilisin I'lriuoholtMi listed lour different collections 
without designating ;i type, necessitating the nerd 
to select a lectotvpe. We hereby designate- the //. 
Piliivr :m collection at IS as the lectotvpe. Tins 
collection has generally I •• -<-n regarded as the type 
and lias been photographed by both 1.1. (neg. 1971- 
85) and US (neg. 2382). Lundell annotated this 

Lundell's concept of Aitriculardisia pleurobotrya 
var. parva came solely from his holotype, now at 
Lb, which has eight inflorescences in very young 
bud and one inflorescence slightly older, with 
smaller leaves. However, the isotype at MO has two 
branches, one similar to the holotype. but vm|Ii larg- 
er leaves more typical of Ardisia pleunthoii \a. The 
other branch is in fruit with mature calyx and 
matches the specie's in all respects. Lundell 
1 198. lb: 230) stated that variety par, a "appears to 
be a diminutive form of Auriculardisia plcumhuii \a 
(Donn. Sm.) Lundell. and all parts of the variety 
arc smaller than the species." which was true given 
the nature of the material he had to work with. 

Stamens examined. COSTA IOC A. AlajuHa: Yolean 
I'oas. near crater, s.d. III). 1. Undress.,,. I \\ ): near \oloan 
I'oas. .> May I'K.o (II. fr). I. Jimenez I/. :<>!<>, I I: Pa ,,,,„■ 
Naoional \ol,an I' I 'J July |<){!l (III. ./. /ty.,/, 7 1 J 1 
NY. TKX, DS); slopes of Volcan 

,M Kl . I. ... 
I'oas In,,., rim 

1971 (fl. fr). R. Wilbw 

Teeri I.U,77 (DIKK. IT). 

\olca ( alba, yi.-inilx ol I inc.. I >,„ mad 

(II. fr)./-! \lmeda.-,70i\n k t.|: \ol, -an Ira/. 
Cll mule 10 from Carlago lo I he « rater. 21 
til. k. l;„,,„,ee, e, al. .1212 (CAS. I. NV 
hclou cater ol loi.i.ll,;, \. .!,,„„,. 2(, |„K 
Irnl 07<> (K II.