Vege a i e mo phological a ia ion
in Chamaedo ea ela io (A ecaceae),
a i s app oach a species delimi a ion
Diego Villa -Mo ales1,2, Alejand a Mo eno-Le elie 2,
Ma celo Rod igo Pace3, Miguel Ángel Pé ez-Fa e a4
1 Posg ado en Ciencias Biológicas, Uni e sidad Nacional Au ónoma de México, Mexico Ci y, Mexico
2 Ins i u o de Biología, Ja dín Bo ánico, Uni e sidad Nacional Au ónoma de México, Mexico Ci y, Mexico
3 Ins i u o de Biología, Depa amen o de Bo ánica, Uni e sidad Nacional Au ónoma de México, Mexico Ci y, Mexico
4 Ins i u o de Ciencias Biológicas, He ba io Eizi Ma uda, Labo a o io de Ecología E olu i a, Uni e sidad de Ciencias y A es de Chiapas, Tux la
Gu ié ez, Mexico
Co esponding au ho : Miguel Ángel Pé ez-Fa e a (pe ez a [email p o ec ed])
Academic edi o : Luiza Teixei a-Cos a♦
Recei ed
30 May 2025♦
Accep ed
26 Augus 2025♦
Published
20 No embe 2025
Abs ac
Backg ound and aims – Chamaedo ea is he la ges genus o neo opical palms dis ibu ed mainly in lowland ain o es s
and mon ane cloud o es s om Mexico o Boli ia. Species delimi a ion in his genus emains p oblema ic due o high
mo phological a ia ion and inconsis en axonomic ea men o such a ia ion. Chamaedo ea ela io , a climbing species
om sou he n Mexico and no he n Cen al Ame ica, exempli ies hese challenges wi h se e al his o ical synonyms and
in o mally ecognized mo pho ypes. This s udy e alua es he mo phological a ia ion in ege a i e cha ac e s be ween
his species’ mos cha ac e is ic wo mo pho ypes, cespi ose and soli a y, o e alua e hei axonomic dis inc i eness.
Ma e ial and me hods – Six popula ions o C. ela io we e sampled in Mexico and Gua emala, ep esen ing bo h soli a y
and cespi ose mo pho ypes. Thi een mo phological lea cha ac e s we e measu ed om ma u e indi iduals. Da a we e
examined using uni a ia e and mul i a ia e analyses (NMDS, k-means clus e ing, PCA, MANOVA, LDA) o e alua e
mo phological a ia ion be ween he mo pho ypes and popula ions. Disc iminan analyses we e also used o assess
classi ica ion in o ei he mo pho ype.
Key esul s – Uni a ia e analyses e ealed signi ican di e ences be ween mo pho ypes in mos measu ed cha ac e s,
mainly hose associa ed wi h lea size. Simila i y analyses eco e ed bo h mo pho ypes as being dis inc om each o he .
PCA showed no able sepa a ion o he mo pho ypes along he i s componen ha summa ized lea and median lea le
size. Bo h MANOVA and LDA con i med signi ican di e ences be ween mo pho ypes and esul ed in high accu acy
classi ica ion.
Conclusion – Lea mo phology clea ly dis inguishes he soli a y and cespi ose mo pho ypes o C. ela io , suppo ing hei
po en ial delimi a ion as sepa a e species. The b anching habi o he cespi ose o m u he ein o ces his dis inc ion.
While cu en esul s suppo species-le el ecogni ion, addi ional e idence om ep oduc i e cha ac e s, niche, and
gene ic di e gence is ecommended o con i m a possible species delimi a ion.
Keywo ds
climbing palm, Gua emala, Mexico, mo phome ics, sys ema ics
INTRODUCTION
Chamaedo ea Willd. is he la ges genus o neo opical
palms, wi h a ound 106 species (POWO 2025). I consis s
o dioecious, mainly small unde s o y palms wi h ei he
pinna e o bi id lea es, and soli a y lowe s in all bu h ee
species (Hodel 1992). They occu mos ly in ain o es s
and mon ane cloud o es s, anging om sea le el o
Plan Ecology and E olu ion 158 (3): 445–456, 2025
h ps://doi.o g/10.5091/plece o.160648
Copy igh Diego Villa -Mo ales, Alejand a Mo eno-Le elie , Ma celo Rod igo Pace, Miguel Ángel Pé ez-Fa e a. This is an open access a icle dis ibu ed unde
he e ms o he C ea i e Commons A ibu ion License (CC BY 4.0), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided
he o iginal au ho and sou ce a e c edi ed.
Plan Ecology and E olu ion is published by Meise Bo anic Ga den and he Royal Bo anical Socie y o Belgium.
RESEARCH ARTICLE
Villa -Mo ales e al.: Mo phological delimi a ion o Chamaedo ea ela io 446
almos 3,000 m a.s.l. (Hodel 1992; D ans ield e al. 2008).
Two cen es o di e si y o he genus ha e been p oposed
based on he as asso men o species in hese a eas: he
moun ainous egions o sou he n Mexico and Gua emala,
and Cos a Rica and Panama (Hodel 1992).
Va ious phylogene ic hypo heses ha e been p oposed
a he species le el in Chamaedo ea, expanding ou
unde s anding o i s in e speci ic ela ionships and b oad
biogeog aphical his o y (Thomas e al. 2006; Cuenca and
Asmussen-Lange 2007; Cuenca e al. 2008; Cano e al.
2022). Howe e , he conside able a ia ion bo h be ween
and wi hin species has no been s udied in de ail, and
esea ch on species limi s wi hin he genus is lacking
(Hodel 1999; D ans ield e al. 2008). This has led o wo
sepa a e scena ios ega ding he same a ia ion p oblem:
a) a high numbe o synonymies a e ound wi hin species
due o names published based on small di e ences in
local o ms and hei in e media es; o b) species which
no able mo phological a ia ion has no been p ope ly
ecognized beyond in o mal o ms o a ie ies.
Species delimi a ion and a ia ion be ween simila
species and species complexes ha e been s udied p e iously
in o he palm gene a. T adi ional mo phome ic analyses
ha e been used wi h posi i e esul s in he delimi a ion o
axa (Hende son 2006; Laubengaye e al. 2012; San os-
He nández e al. 2022). Con e sely, in highly a iable
species complexes, species limi s could no be esol ed
wi h good ideli y using mo phome ic da a (Bo chsenius
1999; A ia e al. 2017). Fewe s udies ha e in eg a ed
mul iple lines o e idence, mainly mo phological, gene ic
and some imes en i onmen al da a, wi h a ied esul s
(Bacon e al. 2012, 2016; A ia e al. 2020). Despi e he
ecognized a ia ion be ween and wi hin Chamaedo ea
species, ew s udies ha e explo ed species limi s in
his genus. Bacon and Bailey (2006) delimi ed wo
simila species, C. epejilo e Liebm. and C. al e nans
H.Wendl., based on mic osa elli e da a complemen ed
by mo phological di e ences. In his unde g adua e
hesis, Ruíz-Cas illejos (2011) delimi ed simila looking
C. glauci olia H.Wendl. and C. plumosa Hodel using
mo phome ic analyses o he lea es. Ano he s udy using
mo phome y on he lea , ye unpublished, seg ega ed C.
enella H.Wendl. and C. geonomi o mis H.Wendl. (Pé ez-
Fa e a unpubl. da a).
A no able example o a high numbe o his o ical
synonyms and a lack o o mal ecogni ion o a ie ies
o o ms is C. ela io Ma . This climbing species occu s
mainly in lowland opical humid o es s along he A lan ic
slope o sou he n Mexico, Gua emala, and Hondu as. I
is less commonly ound a highe ele a ions in mon ane
cloud o es s o he Sie a Mad e O ien al in Mexico
(Fig. 1). A leas i e o he species names and a ie ies
ha e been synonymized unde i (Hodel 1992; Dowe and
Hodel 2021). Mos o hese names we e ini ially published
based on a ia ions in lea size (Ma ius 1849; Wendland
1853; Damme 1905) o habi di e ences (Damme 1905;
Gé ôme 1911).
The a ia ion in his species was p e iously summa ized
in ou in o mally ecognized o ms by Hodel (2013):
1) he mos common o m, a soli a y palm wi h deeply
bi id ju enile lea es and adul pinna e lea es wi h
s ongly e lexed apical lea le s ha help i a ain a obus
climbing habi ; 2) a cespi ose o m wi h ae ial b anching
s ems, pinna e ju enile and ma u e lea es, and a weake
climbing habi ; 3) ano he soli a y o m wi h pinna e
ju enile and ma u e lea es cha ac e ized by hei na ow,
linea lea le s; and 4) a a ian o he i s soli a y o m,
wi h la ge bi id blades e ained a ew yea s in o ma u i y
bu ha e en ually p oduces i s ypical pinna e lea es. The
dis ibu ion and habi a o hese o ms a e also dis inc
Figu e 1. Geog aphic dis ibu ion and sampling si es o Chamaedo ea ela io . Poin s ep esen dis ibu ion o he ba ium specimens
om Villa -Mo ales (2020) and GBIF da a.
Plan Ecology and E olu ion 158 (3): 445–456, 2025 447
be ween hem (Table 1). The hi d o m would e en ually
be seg ega ed as a dis inc species, C. acanensis Pé ez-
Fa ., Villa -Mo . & Hodel, based on mo phological and
gene ic cha ac e is ics (Pé ez-Fa e a e al. 2021).
Dowe and Hodel (2021) p oposed ha he a ious o ms
o C. ela io could be ele a ed o o mal axonomical s a us
gi en p ope esea ch. The delimi a ion and ecogni ion
o C. acanensis p o ed his poin and added o he
ongoing discussion o species limi s wi hin Chamaedo ea,
especially in C. ela io . Wi h now wo climbing species
in he genus ha esemble each o he supe icially, a
eassessmen o he a ia ion wi hin C. ela io is desi able.
Following he publica ion o C. acanensis, Hodel’s (2013)
o ms o C. ela io we e educed o h ee. Howe e , o
his s udy, we compa ed only wo o hese o ms based
on habi di e ences: soli a y and cespi ose (Fig. 2). The
ecogni ion o he emaining o m, cha ac e ized by i s
ea ly lowe ing, will be discussed u he below. This
s udy aims o e alua e he ege a i e mo phological
a ia ion be ween hese wo mos well-known o ms, o
mo pho ypes, o C. ela io , and de e mine whe he he e
is su icien a ia ion o delimi hese as dis inc species.
MATERIAL AND METHODS
Species dis ibu ion
P esence poin s o C. ela io in Mexico we e aken
om he compiled and e ised da abase o he ba ium
ma e ial om Villa -Mo ales (2020). Cen al Ame ican
dis ibu ion in Gua emala and Hondu as was ob ained
om he Global Biodi e si y In o ma ion Facili y (GBIF
2024). The o al dis ibu ion map was isualized in R
.4.3.1 (R Co e Team 2024).
Sampling
We sampled six popula ions o C. ela io in i e locali ies
om Mexico and Gua emala co esponding o bo h
soli a y and cespi ose mo pho ypes (Fig. 1; Suppl.
ma e ial 1). A single locali y, El Mi ado , Ve ac uz,
Table 1. Fo ms o Chamaedo ea ela io acco ding o Hodel (2013) and Pé ez-Fa e a e al. (2021). * P obably based on he a ea
desc ibed in Hodel (2013).
(1) Soli a y o m (2) Cespi ose o m (3) Chamaedo ea
acanensis
(4) Soli a y wi h ea ly
lowe ing
Habi Soli a y Cespi ose, b anching Soli a y Soli a y
Ju enile lea es Bi id Pinna e Pinna e Bi id
Ma u e lea es
(a e lowe ing) Pinna e Pinna e Pinna e Bi id, hen pinna e
Habi a Lowland ain o es Mon ane cloud o es Mon ane cloud o es Lowland ain o es
Ele a ion
(m a.s.l.) 50–1,200 1,000–1,900 1,200–1,900 below 600*
Dis ibu ion
A lan ic slope o
Gua emala, Hondu as,
Mexico
A lan ic slope o Mexico Paci ic slope o
Gua emala and Mexico A lan ic slope o Mexico
Figu e 2. Chamaedo ea ela io mo pho ypes. A. Cespi ose mo pho ype. B. Lea es o cespi ose (le ) and soli a y ( igh ) mo pho ypes.
Scale ba = 10 cm. C. Soli a y mo pho ype. Pho og aphs by Ma ía Fe nanda Ma ínez Vela de (A) and Diego Villa -Mo ales (C).
Villa -Mo ales e al.: Mo phological delimi a ion o Chamaedo ea ela io 448
Mexico had bo h mo pho ypes g owing in syn opy, while
in he o he s only one o hem was ound. Mo pho ypes
we e ecognized and assigned on he ield by hei habi ,
ei he soli a y o cespi ose wi h abo e-g ound b anching.
We collec ed lea es om di e en indi iduals, wi h an
a e age o 10 plan s sampled pe popula ion. Lea es o
C. ela io show signi ican mo phological changes as he
plan g ows and acqui es i s scanden habi as an adul
(S andley and S eye ma k 1958). The ypical, e lexed
apical lea le s a e seen in ma u e indi iduals wi h ully
pinna e lea es, and as such only hese lea es we e sampled
o a oid a ia ion ega ding he age o he indi iduals and
lea ma u i y as much as possible.
To a oid con usion, we ha e used he e m “mo pho ype”
in con as o “ o m” used by p e ious au ho s (Hodel
2013; Dowe and Hodel 2021). We conside his e m o
e lec be e he amewo k o his s udy since sampling
and e alua ions we e based on mo phological, speci ically
habi , di e ences. Ou cespi ose mo pho ype co esponds
o he p e iously men ioned cespi ose o m and ou
soli a y mo pho ype would compa e o he i s soli a y
o m o Hodel (2013). Al hough we we e unable o isi he
locali y o he emaining soli a y o m desc ibed by Hodel
(2013) om he is hmus o Tehuan epec, we conside i
o be a a e, local a ian o ou soli a y mo pho ype o
a numbe o easons ha will be discussed a e wa ds.
Field explo a ion in he a ea whe e his o m inhabi s has
been made p oblema ic due o ecen i es and an up ise
in iolence.
We could no consis en ly ind ep oduc i e s uc u es,
ei he in lo escences o in uc escences, du ing he
sampling ield ips. Mos soli a y plan s had al eady
los hei male in lo escences o we e in ad anced s a e
o decomposi ion o p ope ly measu e hem, while only
a ew indi iduals in some popula ions had imma u e
ui s. Simila ly, we did no obse e measu able lowe ing
and ui ing s uc u es in he cespi ose popula ions,
excep o e y ew ui ing plan s in La Espe anza. As
a esul , due o he small and inconsis en sample size
compa a i e s a is ical analyses could no be pe o med
wi h ep oduc i e cha ac e s.
Mo phome ic analyses
To e alua e he mo phological a ia ion among
popula ions and mo pho ypes, we measu ed 13
mo phological cha ac e s om he lea es, mos o which
ha e been used in p e ious s udies (A ia e al. 2017;
San os-He nández e al. 2022): pe iole leng h (PL), pe iole
wid h (PW), achis leng h (RL), wid h o he achis a i s
middle leng h (mRW), wid h o he achis be ween he
mos dis al lea le s (aRW), lea le numbe on one side o
he achis (LN), basal lea le leng h (bLL), basal lea le
wid h (bLW), basal lea le inse ion leng h (bLI), median
lea le leng h (mLL), median lea le wid h (mLW),
median lea le inse ion leng h (mLI), and dis ance
be ween median lea le s (mLD). All a iables we e log-
10 ans o med be o e he analyses. All s a is ical analyses
we e pe o med in R .4.3.1 (R Co e Team 2024) using
RS udio .2023.12.1 (RS udio Team 2023).
Uni a ia e me hods we e used o s udy he a ia ion
be ween bo h mo pho ypes. S uden ’s - es was pe o med
o all cha ac e s conside ing bo h mo pho ypes using he
s a s base package in R. Fo a mul i a ia e app oach using
Q- ype analyses, we i s explo ed he s uc u e o he da a
and summa ized i s a ia ion by applying a non-me ic
mul idimensional Scaling (NMDS) on a dis ance ma ix
ob ained wi h he Euclidean measu e. Clus e ing o he
da a was also analysed ia a non-hie a chical k-means
clus e analysis which does no o m nes ed clus e s
i e a i ely as in hie a chical clus e ing, ins ead inding
all clus e s simul aneously as pa i ions o he da a (Jain
2010). To assess he op imal numbe o clus e s, we used
he Calinski-Ha abasz index (Caliński and Ha abasz
1974). Bo h NMDS and k-means clus e analyses we e
done using he R package egan .2.6-8 (Oksanen e al.
2024).
Unde a R- ype mul i a ia e app oach, we i s
examined all a iable’s con ibu ion o o al a iance
in a simpli ied manne wi h a P incipal Componen
Analysis (PCA). Componen s wi h eigen alues g ea e
han 1.0 we e ex ac ed and plo ed in a sca e plo . A
p io i g ouping (soli a y s cespi ose mo pho ypes) was
i s e alua ed wi h a MANOVA o assess di e ences in
g oup cen oids. We calcula ed he pa ial e a-squa ed
measu e (η²) o quan i y he e ec o he mo pho ype on
he a iance o he lea cha ac e s, using he R package
e ec size .1.0.0 (Ben-Shacha e al. 2024).
We hen used a Linea Disc iminan Analysis (LDA)
o u he es di e ences in he mo pho ype g ouping
and ind he a iables wi h he highe disc imina ion
powe . Ou a p io i classi ica ion o he indi iduals was
e alua ed using he con usion ma ices p oduced by his
analysis, es ing i s accu acy wi hou c oss- alida ion and
wi h c oss- alida ion ia wo me hods: jack-kni ing and
Mon e Ca lo (1,000 epea s using 40% o he da a as es ).
LDA and jack-kni e c oss- alida ion was done wi h he R
package MASS .7.3-60 (Ripley e al. 2024), while Mon e
Ca lo c oss- alida ion was done wi h he R package ca e
.7.0-1 (Khun e al. 2024). Addi ionally, we epea ed he
disc iminan analysis wi h popula ion g ouping as o
compa e i s esul s wi h he mo pho ype g ouping. The
same e alua ion o he accu acy ia con usion ma ices
was done as in he p e ious g ouping.
RESULTS
Acco ding o he uni a ia e analyses be ween g oups,
he e we e highly signi ican di e ences (p < 0.001) in
almos all a iables among bo h mo pho ypes o C. ela io
(Suppl. ma e ial 2). The only cha ac e s wi h lowe o no
signi icance co espond o he basal lea le s: inse ion
(bLI; p < 0.01) and leng h and wid h (bLL and bLW,
espec i ely; bo h p > 0.1) (Suppl. ma e ial 2). Based on
he signi ican di e ences, he soli a y mo pho ype can
be dis inguished om he cespi ose by ha ing longe
Plan Ecology and E olu ion 158 (3): 445–456, 2025 449
lea es (RL), hicke axes (PW, mRW, aRW), la ge lea le
inse ions (bLI, mLI), and la ge median lea le s wi h
longe dis ance be ween hem (mLL, mLW, mLD) bu
sho e pe ioles (PL).
On he NMDS plo mos specimens ga he ed in wo
mo e o less de ined g oups along he ho izon al axis:
mos cespi ose indi iduals a e ound on he le o he
axis, whe eas almos all soli a y indi iduals a e on he
igh o he axis (Fig. 3A). Abou hal o he specimens
om To on epec (cespi ose) o m a small clus e loca ed
be ween he wo la ge g oups and e en loca ed on he igh
side o he ho izon al axis, sugges ing ce ain simila i y
wi h soli a y indi iduals a he han o o he cespi ose
ones. Indi iduals om El Mi ado g ouped oge he
acco ding o hei mo pho ype, no he locali y. K-means
clus e ing esul ed in an op imal numbe o g oups (k)
o wo (highes C-H index: 43.56). These wo g oups
co espond mos ly o he a p io i assigned mo pho ypes,
he only excep ions being ce ain specimens in he
cespi ose popula ions o La Espe anza (2 indi iduals)
and To on epec (5 indi iduals) ha we e g ouped
wi h he soli a y indi iduals (Fig. 3B). All p e iously
assigned soli a y indi iduals we e g ouped oge he . As
in he NMDS, El Mi ado plan s we e clea ly sepa a ed
acco ding o hei mo pho ype.
The PCA showed ha he i s i e p incipal
componen s explained 91% o he a ia ion in he da a
(PC1: 58.1%, PC2: 18.2%, PC3: 6.4%, PC4: 5.1%, PC%:
3.2%). We show only he componen s wi h an eigen alue
highe han 1.0, being ha he i s (7.56) and he second
(2.37). The a iables wi h he highes con ibu ion o
he PC1 we e hose conce ned wi h he o e all size o
he lea and median lea le s: achis leng h (RL), pe iole
wid h (PW), and median lea le leng h (mLL) and
wid h (mLW). Fo PC2, he a iables wi h he highes
con ibu ion we e lea le numbe (LN) and he ones
measu ed om he basal lea le s: leng h (bLL), wid h
(bLW), and inse ion (bLI) (Suppl. ma e ial 3A). In a
sca e plo o he i s wo p incipal componen s (Fig.
4), indi iduals o each mo pho ype a e dis ibu ed on
di e en sides along he i s componen axis wi h ce ain
o e lap occu ing be ween he cespi ose To on epec
popula ion and soli a y popula ions. Specimens om he
same popula ion g oup oge he wi h a ying deg ees o
o e lap happening wi hin he same mo pho ype. As wi h
he p e ious Q- ype analyses, he indi iduals om El
Mi ado a e clea ly sepa a ed om each o he acco ding
o hei mo pho ype.
Signi ican di e ences be ween mo pho ypes wi h all
a iables measu ed we e ound by he MANOVA analysis
(p < 0.001; Table 2). The p opo ion o o al a iance
explained by he g ouping is high (η² = 0.87), sugges ing
ha mo pho ype di e ences accoun o mos o he
a ia ion in he lea cha ac e s.
Table 2. MANOVA esul s using di e en s a is ics.
Mo pho ype Value App ox F p
Pillai’s T ace 0.872 24.11 < 0.001
Wilks’ Lambda 0.128 24.11 < 0.001
Ho elling-Lawley T ace 6.814 24.11 < 0.001
Roy’s La ges Roo 6.814 24.11 < 0.001
Figu e 3. Q- ype analyses plo s. A. NMDS plo using Euclidean dissimila i y. B. K-means clus e analysis o a ious g oups (k =
2–6). Popula ions a e shown on he ho izon al axis; each ba co esponds o a measu ed indi idual. Abb e ia ions o popula ions
co espond o Fig. 1 and Suppl. ma e ial 1.
Villa -Mo ales e al.: Mo phological delimi a ion o Chamaedo ea ela io 450
Figu e 4. PCA biplo o i s wo p incipal componen s. Abb e ia ions o a iables a e explained in he ex .
Figu e 5. Linea disc iminan analysis plo s. A. Conside ing mo pho ype g ouping. B. Conside ing popula ion g ouping.
Plan Ecology and E olu ion 158 (3): 445–456, 2025 451
Table 3. Con usion ma ix ob ained om disc iminan analyses when conside ing mo pho ype g ouping. Rows indica e assigned
indi iduals and columns p edic ions. Numbe s shown o each assigna ion co espond o non-c oss- alida ed LDA, jack-kni e
alida ed LDA, and Mon e Ca lo alida ed LDA, espec i ely. Mon e Ca lo c oss- alida ion numbe s a e shown in a e age. Highes
numbe s a e shown in bold.
Cespi ose Soli a y
Cespi ose 31 / 29 / 29.4 0 / 2 / 1.9
Soli a y 0 / 1 / 1.4 29 / 28 / 27.2
Table 4. Con usion ma ix ob ained om disc iminan analyses when conside ing popula ion g ouping. Rows indica e assigned
indi iduals and columns p edic ions. Abb e ia ions o popula ions co espond o Fig. 1 and Suppl. ma e ial 1. Numbe s co espond
o non-c oss- alida ed, jack-kni e alida ed, and Mon e Ca lo alida ed LDA, espec i ely. Mon e Ca lo c oss- alida ion numbe s
a e shown in a e age. Highes numbe s a e shown in bold.
MIR-C ESP TOT MIR-S LB YAL
MIR-C 9 / 8 / 6.6 1 / 1 / 2.1 0 / 0 / 0.5 0 / 1 / 0.7 0 / 0 / 0 0 / 0 / 0.1
ESP 0 / 2 / 2 10 / 6 / 5.6 0 / 1 / 1.5 0 / 1 / 0.6 0 / 0 / 0.1 0 / 0 / 0.2
TOT 0 / 0 / 0.1 1 / 3 / 2.6 10 / 8 / 8.1 0 / 0 / 0 0 / 0 / 0.1 0 / 0 / 0.1
MIR-S 0 / 0 / 0.1 0 / 0 / 0.2 0 / 0 / 0.1 8 / 5 / 5.2 0 / 1 / 1.0 1 / 3 / 2.4
LB 0 / 0 / 0 0 / 0 / 0.1 0 / 0 / 0 0 / 1 / 1.3 10 / 8 / 7.7 0 / 1 / 1
YAL 0 / 0 / 0 0 / 0 / 0.3 0 / 1 / 0.6 0 / 1 / 1.9 0 / 1 / 1 10 / 7 / 6.3
Table 5. Bes cha ac e s a disc imina ing be ween mo pho ypes acco ding o s a is ical analysis. Cha ac e abb e ia ions as in
Ma e ial and me hods.
Analysis Bes disc imina ing cha ac e s
-s uden es All excep bLI, bLL, and bLW
LDA (mo pho ypes) RL, PW, mRW
LDA (popula ions) RL, PW, mRW
In he LDA analysis conside ing he wo mo pho ypes,
he highes disc imina ion was gi en by achis leng h
(RL), pe iole wid h (PW), and median achis wid h
(mRW), all cha ac e s associa ed wi h lea size (Suppl.
ma e ial 3B). The specimens o m wo g oups clea ly
sepa a ed om each o he (Fig. 5A). Al hough a single
cespi ose specimen om El Mi ado g aphically appea s
sepa a ed om he es o i s g oup, his was a ibu ed
o i s a he sho and hin achis compa ed o he es .
Rega ding classi ica ion, specimens we e classi ied
co ec ly o hei a p io i assigned mo pho ype in he
disc imina o y analyses wi h high accu acy (LDA: 100%,
jack-kni e LDA: 95%, MC-LDA: 94%) (Table 3).
Conside ing he six popula ions, he LDA’s i s
dimension accoun ed o 59% o he o al a iance, he
second 23%, and he hi d 11%, amoun ing o mo e han
90% o i . The a iables wi h he highes disc imina ion
coe icien along he i s dimension a e he same as o he
LDA be ween mo pho ypes (RL, PW, mRW), while lea le
numbe (LN), pe iole wid h (PW), and median lea le
wid h (mLW) had he highes coe icien s along LD2
(Suppl. ma e ial 3C). The sca e plo o he i s wo linea
disc iminan s shows dis inc clus e s co esponding o
he mo pho ypes along he i s axis, wi h o e lap be ween
co esponding mo pho ype popula ions (Fig. 5B).
Cespi ose popula ions show a highe deg ee o sepa a ion
along he second axis in compa ison wi h he soli a y
popula ions, he coe icien s indica ing ha specimens
om El Mi ado ha e hinne pe ioles and less, na owe
lea le s han hose om La Espe anza and especially
To on epec. Al hough his a ia ion is no ewo hy o
men ion, he sepa a ion be ween mo pho ypes explains
mo e o he a iance, as in e ed by he dimensions (59%
s 23%).
When conside ing popula ions a he han
mo pho ypes, indi iduals we e mos ly classi ied o
hei own popula ion, ollowed by classi ica ion o o he
popula ions o he same mo pho ype (Table 4). In only
e y ew cases in he c oss- alida ed analyses specimens
we e classi ied o a di e en mo pho ype popula ion,
mos ly wo cespi ose indi iduals we e p edic ed o
belong o a soli a y popula ion, and one soli a y
specimen p edic ed in a cespi ose popula ion. No ably,
no specimens om To on epec we e p edic ed o belong
o soli a y popula ions. The accu acy o he classi ica ion
among popula ions a ied be ween me hods (LDA: 95%,
jack-kni e LDA: 70%, MC-LDA: 65.8%).
Villa -Mo ales e al.: Mo phological delimi a ion o Chamaedo ea ela io 452
DISCUSSION
Vege a i e mo phological delimi a ion
All bu wo o he analysed mo phological cha ac e s
o lea es showed signi ican di e ences be ween he
soli a y and cespi ose mo pho ypes o C. ela io , and
bo h mo pho ypes we e signi ican ly dis inc when all
a iables we e compa ed in he MANOVA analysis. In
he simila i y and o dina ion analyses, he mo pho ypes
could be dis inguished sepa a ely wi h only mino
o e lap be ween hem. Specimens we e classi ied o hei
p e iously assigned mo pho ype wi h high accu acy in
he disc iminan analyses. I popula ions we e conside ed
o he disc iminan analyses ins ead o mo pho ypes, he
classi ica ion s ill sepa a ed he la e wi h high accu acy.
The a iables ha mos dis inguish be ween mo pho ypes
a e ela ed o he o e all size and obus ness o he lea es
(Table 5), wi h he soli a y o m’s being la ge wi h bigge
lea le s and longe inse ions o he achis. The cespi ose
mo pho ype, hough smalle in lea es and lea le s, has
longe pe ioles.
Ce ain mo phological o e lap was obse ed,
mainly be ween some cespi ose To on epec and soli a y
specimens, bo h in he simila i y analyses (NMDS and
k-means clus e ing) and in he PCA. Acco ding o ou
aw da a and he PCA coe icien s, lea es o To on epec
specimens we e he mos obus and wi h he la ges
lea le s wi hin he cespi ose popula ions, bu no so as
in he soli a y popula ions. This would explain hei
simila i y wi h soli a y specimens and hei esul ing
posi ion and clus e ing in he Q- ype analyses. The
disc iminan analyses, howe e , g oup his popula ion
clea ly wi h he o he cespi ose ones.
The mos s iking cha ac e used o ecognize
mo pho ypes in he ield, he b anching o he cespi ose
mo pho ype, was no conside ed in he analyses because
o i s lack o a iance and e iden seg ega ion powe .
Ma u e indi iduals o he cespi ose mo pho ype can
p oduce b anches well abo e g ound, one on e e y o he
node, o ming a dense clump abo e he g ound and
nea by smalle ege a ion. No o he Chamaedo ea species
a e known o p oduce p ope ae ial b anches, o he
cespi ose (clus e ing) species o m new s ems (b anches)
om buds on p oximal nodes below o jus abo e he
g ound (Fishe 1974; Hodel 1992). Ne e heless, much
wo k is s ill needed in cha ac e izing he b anching o
many palm species.
The only o he species in he genus known o also
p esen a soli a y and cespi ose habi is C. epejilo e;
howe e , he a ia ion in his species has been associa ed
o i s cul i a ion his o y, he soli a y habi a esul o
domes ica ion (Cas illo Mon e al. 2017). Al hough C.
ela io is used in some a eas in Mexico o handc a ing
o as o namen al, i s ex ac ion is i egula , and no
se ious domes ica ion e o has been ecognized
(Con e as Co és e al. 2018; Rendón-Aguila e al.
2022). Conside ing his, he p esence o wo habi s in C.
ela io does no seem o be a consequence o cul i a ion
no ela ed o human ac i i y. In iew o he uniqueness
o i s b anching habi , and pending p ope de elopmen
s udies, his di e ence should be conside ed an impo an
cha ac e ha could enhance species delimi a ion o bo h
mo pho ypes.
Chamaedo ea ela io ’s mo pho ypes a e also seg ega ed
geog aphically and possibly ecologically (Table 1).
Acco ding o p e ious knowledge and he ba ia eco ds,
he soli a y mo pho ype inhabi s lowland opical humid
o es s usually below 800 m a.s.l., wi h some popula ions
ex ending up o abou 1,200 m. In con as , he cespi ose
mo pho ype is ound in opical mon ane cloud o es s
om 1,000 o 1,900 m a.s.l. (Hodel 2013; Villa -Mo ales
2020). In ce ain locali ies in Oaxaca and Ve ac uz,
Mexico, he soli a y mo pho ype can be ound in opical
o es s on he oo o he moun ains and lowlands
ex ending beyond, while he cespi ose mo pho ype
has been egis e ed a ew hund ed me e s uphill in
mon ane o es s on he slopes o he moun ains. Though
geog aphically e y nea , he di e ence in al i ude and
hus ege a ion is no o ious.
El Mi ado is he only locali y o ou knowledge whe e
bo h mo pho ypes ha e been ound li ing in syn opy. Thei
coexis ence in his si e has been egis e ed his o ically, as
we loca ed specimens collec ed by Liebmann in he 1840s
(a C he ba ium) co esponding o bo h mo pho ypes
in he same a ea. This locali y is geog aphically nea he
species’s no he n limi , and only e y ew popula ions
o bo h mo pho ypes a e known o exis beyond i .
Fu he mo e, El Mi ado is on he uppe bounda ies
o al i udinal ange o he soli a y mo pho ype and on
he lowe ange o he cespi ose one. This could explain
hei p esence in his a ea, p obably being a con ac zone
whe e bo h mo pho ypes can inhabi gi en he p ope
al i udinal ange o each o hem. The ege a ion ype
he e has been egis e ed in di e se collec ed specimens
as mon ane cloud o es (o synonyms sensu Gual-Díaz
and González-Med ano 2014) hough i is a a lowe
al i ude and less dense and humid as he o he s sampled
he e in La Espe anza and To on epec. Al hough mon ane
cloud o es s a e a iable in en i onmen al cha ac e is ics
and composi ion (Gual-Díaz and González-Med ano
2014), a p ope composi ional and ecological s udy o he
si e could lead o a be e unde s anding o i s uniqueness
ega ding his syn opy.
Chamaedo ea species a e known o occu sympa ically
in many a eas (Hodel 1992). Howe e , e y ew esea ch
has been done on how species limi s a e main ained in
hese ins ances. Di e ences in pollina o s and lowe ing
ime ha e been p oposed in o he species in he genus
(Luna e al. 2005; Bacon and Bailey 2006), as well as in
o he unde s o y palms (Knudsen 1999; Bo chsenius
2002; Bo chsenius e al. 2016). Di e en ep oduc i e
iming could also be keeping he mo pho ypes dis inc
in El Mi ado , since imma u e ui s we e only ound in
some soli a y plan s. Howe e , eco ds o he cespi ose
Plan Ecology and E olu ion 158 (3): 445–456, 2025 453
mo pho ype bo h in El Mi ado and o he si es a e e y
limi ed and incomple e o p ope ly assess i s phenology
and compa e i wi h i s soli a y coun e pa . Ecological
s udies pending, ou esul s sugges ha he a ia ion
ound in he lea cha ac e s has li le co ela ion
wi h en i onmen al ac o s and could be e idence o
unde lying gene ic di e gence. This could pe haps be kep
h ough ce ain means o ep oduc i e isola ion ha can
only be seen in his locali y due o hei cohabi a ion.
As s a ed p e iously, we ha e conside ed only wo
mo pho ypes in his s udy based on hei habi di e ences,
a soli a y and a cespi ose one. Hodel (2013) desc ibed a
hi d o m ha esembles he soli a y mo pho ype bu
di e s in size and de elopmen al cha ac e is ics. We
belie e his o be a a e a ian o he soli a y mo pho ype,
a he han a comple ely di e en o m, based on ou
cu en knowledge o C. ela io and he ollowing easons.
Fi s , bi id ju enile lea es o simila size we e seen in o he
locali ies isi ed in Ve ac uz, Mexico. Second, he mo e
obus size is only no ed in i s bi id lea es wi h no e e ence
o a dis inc ion in i s pinna e lea es (Hodel 2013). Thi d,
bo h i s habi a and al i udinal ange all in he ange o
ou soli a y mo pho ype. Fou h, we ha e seen he ba ium
ma e ial o C. ela io om Uxpanapa, Ve ac uz and Los
Chimalapas, Oaxaca ha bo de he epo ed si e and all
a e qui e simila in appea ance. Though we did no include
he ba ium ma e ial in his s udy due o i s incomple eness
and di icul y in assessing homology be ween lea le s, no
ma e ial could be accu a ely e e ed o his o m since all
specimens we e pinna e.
The cen al di e ence o his o m is ha i e ains
i s bi id, o he wise ju enile, lea es o a ce ain pe iod
o ime a e a aining ma u i y (i.e. lowe ing) bu will
e en ually p oduce i s ypical pinna e lea es (Hodel 2013;
Dowe and Hodel 2021). This dis inc ion is hus bes
in e p e ed as de elopmen al a he han mo phological.
The di e ence be ween ju enile and ma u e lea es in
palms has been desc ibed by Tomlinson (1960), bu no
s udies ha e been done o unde s and he de elopmen al
p ocesses esponsible o i . In only one expe imen in he
palm Ca yo a mi is Lou ., ju enile lea es we e induced in
a ma u e indi idual by adminis e ing gibbe ellin (Fishe
1976). I such a change is media ed by gibbe ellin o
simila ho mones in Chamaedo ea is unknown. Al hough
a e, o he species in he genus can also lowe be o e
hei lea es a ain hei maximum numbe o lea le s o
e en de elop pinna e lea es a e yea s o lowe ing and
bi id lea p oduc ion (Hodel 1992).
Scanden habi o Chamaedo ea ela io
No esea ch has been done on he habi o scanden
species o Chamaedo ea. S ill, ield obse a ions on C.
ela io ha e shown he soli a y mo pho ype o ha e a
mo e igo ous climbing habi han he cespi ose one, as i
can each highe heigh s in he canopy and has s onge
suppo s. The la e was no iced du ing his s udy, since
pulling on s ems om below o each he lea es p o ed
ha de in he soli a y o m. The cespi ose mo pho ype,
on he o he hand, has mo e o a sc ambling habi ,
co e ing a wide a ea and en angling wi h mo e plan s
bu no eaching as high. The soli a y mo pho ype’s
la ge lea es migh ha e an adap i e alue and could be
ela ed o i s habi , as shown in o he climbing palms.
When compa ing lea mo phology be ween Desmoncus
o hacan os Ma . (a s ong climbe ) and D. polyacan hos
Ma . (a weake climbe ), Isna d e al. (2005) men ion
ha la ge lea es p o ide D. o hacan os wi h a la ge
ange o encoun e suppo poin s in he ege a ion and
can ensu e i s a achmen o wide suppo poin s (i.e.
b anches). Al hough C. ela io ’s climbing habi is di e en
in a ious ways om ha o species in Desmoncus Ma .,
conside ing he i s lacks mo e specialized s uc u es as
spines o ci i, ce ainly ends in lea size appea o be
simila .
Wi h no spines no p ope acan hophylls in a ci us,
climbing habi in C. ela io seems o be acili a ed by i s
e lexed lea le s, especially dis al ones, and hei ha dened,
callose bases ac ing as suppo (Ace edo-Rod íguez 2020).
Du ing sampling, we no iced ha lea es would lose some
lea le s, a he inse ion o he achis, when pulled down
wi h o ce om unde he canopy. This sugges s ha he
main poin o suppo is indeed he ha dened base o he
lea le s. The leng h o he lea le inse ion o he achis
was seen in si u o be p opo ional o he size o his
callose base, longe inse ions had la ge bases. This could
amoun o he p e ious hypo hesis in ha he soli a y
mo pho ype, as a s onge climbe han he cespi ose one,
would ha e a mo e sizeable suppo sys em based on i s
signi ican ly longe lea le inse ions.
Towa d species delimi a ion
T adi ional mo phome ics ha e been shown o be a
eliable ool in delimi ing axa in palm species (Hende son
2006; Laubengaye e al. 2012; San os-He nández e al.
2022). Ou analyses on he mo phological a ia ion o
he lea show ha he soli a y and cespi ose mo pho ypes
o C. ela io a e signi ican ly dis inc om each o he .
Though no p ope ly included in ou analyses o
easons men ioned p e iously, he b anching habi o he
cespi ose mo pho ype is unique and is o en used as he
main di e ence o seg ega e hem (Hodel 2013).
Though a ia ion in ep oduc i e s uc u es could no
be p ope ly s udied be ween mo pho ypes, we conside
ha he obse ed di e ences in ege a i e cha ac e s
a e indica i e o p obable e olu iona y dis inc i eness
be ween hem, especially conside ing hei cohabi a ion
in he same habi a while main aining hei unique
mo phology (in El Mi ado ). Pa e ns obse ed when
analysing ege a i e cha ac e s we e compa able when
including ep oduc i e cha ac e s in Geonoma Willd.
(Bo chsenius 1999) and Gaussia H.Wendl. (San os-
He nández e al. 2022). Di e ences in lea mo phology
we e likewise con i med by gene ic di e gence be ween
simila species Chamaedo ea glauci olia and C. plumosa
