Journal of Plant Research

https://doi.org/10.1007/s10265-023-01495-z

REGULAR PAPER – TAXONOMY/PHYLOGENETICS/EVOLUTIONARY BIOLOGY

Leaf morphometric analysis and potential distribution modelling

contribute to taxonomic differentiation in the Quercus microphylla
complex
Oscar Ángel De Luna‑Bonilla1,2 · Susana Valencia‑Á3 · Guillermo Ibarra‑Manríquez1 ·
Saddan Morales‑Saldaña1 · Efraín Tovar‑Sánchez4 · Antonio González‑Rodríguez1

Received: 6 June 2023 / Accepted: 31 August 2023

© The Author(s) 2023

Abstract
Mexico is a major center of evolutionary radiation for the genus Quercus, with oak species occurring across different habitat
types and showing a wide variation in morphology and growth form. Despite representing about 20% of Mexican species,
scrub oaks have received little attention and even basic aspects of their taxonomy and geographic distribution remain unre-
solved. In this study, we analyzed the morphological and climatic niche differentiation of scrub oak populations forming
a complex constituted by six named species, Quercus cordifolia, Quercus frutex, Quercus intricata, Quercus microphylla,
Quercus repanda, Quercus striatula and a distinct morphotype of Q. striatula identified during field and herbarium work
(hereafter named Q. striatula II). Samples were obtained from 35 sites covering the geographic distribution of the complex
in northern and central Mexico. Morphological differentiation was analyzed through geometric morphometrics of leaf shape
and quantification of trichome traits. Our results indicated the presence of two main morphological groups with geographic
concordance. The first was formed by Q. frutex, Q. microphylla, Q. repanda and Q. striatula, distributed in the Trans-Mexican
Volcanic Belt, the Sierra Madre Occidental and a little portion of the south of the Mexican Altiplano (MA). The second
group consists of Q. cordifola, Q. intricata and Q. striatula II, found in the Sierra Madre Oriental and the MA. Therefore,
our evidence supports the distinctness of the Q. striatula II morphotype, indicating the need for a taxonomic revision. Within
the two groups, morphological differentiation among taxa varied from very clear to low or inexistent (i.e. Q. microphylla-Q.
striatula and Q. cordifolia-Q. striatula II) but niche comparisons revealed significant niche differentiation in all pairwise
comparisons, highlighting the relevance of integrative approaches for the taxonomic resolution of complicated groups such
as the one studied here.

Keywords Morphometrics · Niche analysis · Population approach · Scrub oaks · Taxonomy

* Antonio González‑Rodríguez Introduction

[email protected]
1 The genus Quercus (Fagaceae) is recognized for its eco-
Instituto de Investigaciones en Ecosistemas y
Sustentabilidad, Universidad Nacional Autónoma de nomic and ecological importance and recently has become
México, 58190 Morelia, México a model clade in different fields of biology (Cavender-Bares
2
Posgrado en Ciencias Biológicas, Unidad de Posgrado, 2019). For these reasons, defining and properly differentiat-
Edificio A, 1° Piso, Circuito de Posgrados, Ciudad ing species in this group is a fundamental task (Valencia-A
Universitaria, Coyoacán, 04510 Ciudad de Mexico, México 2020; Wu et al. 2023). On the basis of phylogenomic anal-
3
Herbario de la Facultad de Ciencias, Departamento de yses (Hipp et al. 2018, 2020), Quercus has recently been
Biología Comparada, Universidad Nacional Autónoma divided (Denk et al. 2017) into two subgenera with eight
de México, Circuito Exterior, s.n, Ciudad Universitaria, sections (subgenus Quercus, including sections Lobatae,
Coyoacán, 04510 México City, México
4
Ponticae, Protobalanus, Quercus and Virentes, and subgenus
Centro de Investigación en Biodiversidad y Cerris, including sections Cerris, Cyclobalanopsis and Ilex).
Conservación, Universidad Autónoma del Estado de
Morelos, Av. Universidad 1001, Col. Chamilpa, CP, Mexico is recognized as a major center of diversification of
62209 Cuernavaca, Morelos, Mexico the oaks, with more than 160 species, included in sections

13
Vol.:(0123456789)
 Journal of Plant Research

Lobatae, Quercus, Protobalanus and Virentes, and at least Quercus, seven in Lobatae and three in Protobalanus);
109 endemics (Galicia et al. 2015; Hipp et al. 2018; McCau- 20 of them are endemic and 11 are shared with the USA.
ley et al. 2019; Nixon 1993; Valencia-A 2004). In California, studies of scrub oak species have revealed
Despite these recent advances in the clarification of the that they frequently form hybrid swarms with adaptations
taxonomy and phylogeny of oaks, considerable confusion to local environmental conditions, which represents an
remains within groups of closely related species, especially interesting scenario for the origin of new species (Ortego
in Mexico (Hipp et al. 2020). Some of these difficulties in et al. 2015; Sork et al. 2016). In the case of Mexico, Sabás-
the delimitation and identification of oak species arise from Rosales et al. (2017) provided a qualitative analysis of the
phenotypic similarity due to recent divergence, the pres- leaf morphology and trichome types of six scrub oak spe-
ence of few taxonomically informative characters in floral cies. These six species are part of the Q. microphylla Née
traits and the great variation in vegetative characters at dif- complex, in which all taxa are strictly shrubby in habit, show
ferent levels (from intraindividual to between populations) elliptic, ovate, oblong or lanceolate leaves, mostly with stel-
(González-Rodríguez et al. 2004; Hardin 1975; Morales- late trichomes on the abaxial side or with some stipitate or
Saldaña et al. 2022; Valencia-A 2004, 2020). Leaves usu- fasciculate trichomes. Intergradation in leaf morphology is
ally display considerable intraspecific variation, since their evident among these species (Sabás-Rosales et al. 2017) and,
morphology depends on an intricate relationship between in addition, their area of distribution is not clearly known,
genetic and environmental factors (Chitwood and Sinha mainly because the original descriptions are not sufficiently
2016; De Heredia et al. 2018). Regarding genetic factors, the precise in terms of diagnostic characters, which possibly has
frequent gene flow among oak species impacts the expres- caused numerous misidentifications and, consequently, the
sion of morphological characters (Cannon and Petit 2020; over- or underestimation of the real distribution of the taxa.
Dumolin-Lapégue et al. 1999; Leroy et al. 2017; Whittemore The complex analyzed in this study is named after Q.
and Schaal 1991). In turn, variation in factors such as light microphylla, a scrub oak species with broad distribution in
exposure, water availability or temperature, and even prun- Mexico (González-Elizondo et al. 1993; Romero Rangel
ing, may result in leaf differences between individuals of et al. 2002; Sabás-Rosales et al. 2017; Valencia-A 2004;
the same population or between branches of the same indi- Valencia-A et al. 2017; Vázquez-Villagrán 2000) and prob-
vidual (Aykut et al. 2017; Borazan and Babaç 2003; Jensen lematic taxonomically. The synonymies reported for Q.
1990). Therefore, the high degree of leaf polymorphism in microphylla involve putative varieties of Q. frutex Trel.
oak trees results from the interrelation between introgres- (Valencia-A 2004) and Q. striatula Trel. (Bartholomew and
sive hybridization, phenotypic plasticity, and adaptation Almeda 2023; González Villarreal 1986; McVaugh 1974),
to environmental gradients (Morales-Saldaña et al. 2022). problems that have produced a “domino” effect since the
In addition to the large intraspecific variation, many of the inconspicuous differences and apparent sympatric distri-
vegetative characters in Quercus also show low interspecific bution between Q. microphylla and Q. striatula involve
differentiation or tend to evolve in a convergent or parallel Q. cordifolia Trel. too (Muller 1944; Villareal et al. 2008).
manner (Jones 1986; Tucker 1974). In the same way, Q. frutex may show confusion with Q.
In Mexico, some oak species complexes have been ana- repanda Bonpl. in distribution and appearance (Romero
lyzed to demarcate taxa, either from a strictly morphological Rangel et al. 2002, 2014). Quercus intricata Trel. adds to
perspective, as for the Acutifoliae group (Romero Rangel the problem by sharing characters such as leaf texture and
2006) or, more recently, using integrative approaches, as for indumentum appearance with Q. frutex and Q. repanda,
the Quercus laeta complex (Morales-Saldaña et al. 2022) in addition to presenting a sympatric distribution with Q.
and the Racemiflorae group (McCauley et al. 2019). In these cordifolia. Finally, the complex also includes Quercus deli-
recent studies, quantitative methods for analyzing vegetative quescens and Quercus mohoriana. However, these are geo-
characters have proven useful for the identification of differ- graphically restricted species in Mexico, with a distinguish-
entiated groups. Such methods include geometric morpho- able phenotype from the rest of the complex (Muller 1944;
metrics (Albarrán-Lara et al. 2010; Morales-Saldaña et al. Sabás-Rosales et al. 2017).
2022; Rodríguez-Gómez et al. 2018; Viscosi 2015; Viscosi In this context, the main objective of this study was to
et al. 2009b), as well as the statistical characterization of contribute to the recognition of six scrub oak species belong-
micromorphological structures that are considered of high ing to the Q. microphylla complex (Q. cordifolia, Q. frutex,
taxonomic value, such as trichomes (Fortini et al. 2015; Q. intricata, Q. microphylla, Q. repanda and Q. striatula)
López-Caamal et al. 2017). by: (i) describing the leaf morphology of each species, from
In comparison to their arboreal congeners, less attention a geometric morphometrics approach, (ii) characterizing the
has been given so far to most of the approximately 70 scrub types of trichomes present in the species, as well as evaluat-
oak species that exist worldwide. At least 31 species with ing if there are differences in the number of trichome arms
this growth form are distributed in Mexico (21 in section among the species, and (iii) defining the potential areas of

13
Journal of Plant Research

distribution and assessing whether there are niche differ- revolute margin and tomentose to slightly glabrescent under-
ences among the species of the Q. microphylla complex. side with stellate trichomes. It is found in xerophytic shrub-
lands and pine forests, on limestone soils or less frequently
on rocky soils, at elevations between 2200 and 2500 m. It is
Materials and methods known from the center and east of the Trans-Mexican Vol-
canic Belt, from the states of Guanajuato, Hidalgo, Puebla,
Study species Tlaxcala and Veracruz (Valencia-A 2004; Vázquez-Villagrán
2000).
The six species in this study belong to the section Quercus Quercus repanda Bonpl. (Fig. 1) also has several erect
(withe oaks). The six species are shrubby in habit, with rhi- stems of 0.2 to 1.5 m in height that emerge from the rhizome,
zomatous growth and with small (0.5 to 2.6 cm) and tomen- making it difficult to delimit individuals. Leaves are obovate
tose to glabrescent leaves. Except for Q. intricata, which is or elliptic-oblong, with a cuneate to rounded base and obtuse
also distributed in the southern USA, the rest of the species to rounded apex, coriaceous texture and revolute-repanding
are endemic to Mexico. margin. The underside is tomentose with sessile fasciculate
Quercus cordifolia Trel. (Fig. 1) grows as clearly differ- trichomes. It grows as small islands in the clearings of pine
entiated individuals with a stem of 0.5 to 1 m in height. The and pine-oak forests and as a scarce element in xerophytic
leaves of this species are papyrose, ovate, ovate-lanceolate shrublands, at elevations of 2,100 to 3,140 m. It is mainly
or ovate-elliptic with a generally cordate base and rounded distributed in the Trans-Mexican Volcanic Belt, in the states
to acute apex, with a flat and entire or sparsely lobed margin. of Guanajuato, Hidalgo, Michoacán, Puebla, Querétaro, San
The underside can be from dense tomentose to glabrescent, Luis Potosí and Veracruz (Valencia-A et al. 2017).
with the tomentum formed by stellate trichomes. The spe- Quercus striatula Trel. (Fig. 1) is also characterized by
cies is found at elevations of 1700 to 3200 m, in pine and erect stems of 0.3 to 1.2 m in height that emerge from the
pine-oak forests in the north of the Sierra Madre Oriental rhizome. It grows as scattered islands and covers large sur-
(Sabás-Rosales et al. 2017; Valencia-A 2004). faces at the edges of forests or in clearings. It has ellip-
Quercus frutex Trel. (Fig. 1) forms extensive patches, tic-lanceolate or oblong-lanceolate leaves with rounded to
with stems measuring 0.6 to 1.5 m. The leaves of the spe- obtuse base and acute to rounded apex. The texture of the
cies are elliptic-oblong or oblong-lanceolate with a rounded leaves is subcoriaceous and the margin is flat to slightly rev-
base and rounded to acute apex. The texture of the leaves olute with tomentose undersides with stellate trichomes. It is
is subcoriaceous, with entire and revolute margins. The generally associated with xerophilous scrub or pine forests,
underside is tomentose with stipitate-fasciculate trichomes. between 2,100 and 3,000 m of elevation. It is distributed on
It occurs in the clearings of oak, pine-oak, pine or fir forests the inner slopes of the Sierra Madre Occidental, in the states
and can be the dominant species in the Quercus scrub of the of Chihuahua and Durango (Sabás-Rosales et al. 2017).
Trans-Mexican Volcanic Belt between 2000 and 3000 m.
(Valencia-A et al. 2017; Vázquez-Villagrán 2000). Collection of botanical material and grouping
Quercus intricata Trel. (Fig. 1) can measure from 0.4 of populations
to 1.5 m in height. This species has ovate, ovate-elliptic or
elliptic-oblong leaves with rounded to slightly cordate base To determine the distribution of the species and to select the
and acute or rounded apex. Leaf texture is coriaceous, mar- collection localities for the different populations, we revised
gins are serrated and notoriously revolute. Leaf undersides the specimens determined under the names Q. cordifolia,
are densely pubescent with stellate trichomes. This species Q. frutex, Q. intricata, Q. microphylla, Q. repanda and Q.
is generally distributed in the periphery of forests at the base striatula deposited in the National Herbarium of Mexico
of mountains or in dry thickets. It is an important part of (MEXU) and the herbarium of the Faculty of Sciences
submontane scrubs, in the transition zones with rosetophy- (FCME), both belonging to the National Autonomous Uni-
lous scrub and stone pine forest, on stony soils. It is found versity of Mexico (UNAM) (Thiers 2016). Specimens that
at elevations from 1600 to 2700 m in the states of Coahuila, had geographic coordinates and concurred with the origi-
Nuevo León and Chihuahua (Sabás-Rosales et al. 2017; nal descriptions of the species and with the type specimens
Valencia-A 2004; Villareal et al. 2008). consulted in the electronic resource JSTOR Global Plants
Quercus microphylla Neé (Fig. 1) has several erect stems (plants.jstor.org) were selected. From this survey, 35 locali-
of 0.2 to 0.7 m in height emerging from the rhizome, mak- ties were chosen and sampled for the six species (Fig. 2,
ing it difficult to delimit individuals. The species grows in Table S1); also taking into account accessibility and safety
dense but not large patches. The leaves are elliptic-lanceolate aspects. During the field work, in some populations we
or oblong-lanceolate with rounded or obtuse base, acute to observed variations in growth form and leaf morphology,
slightly rounded apex, subcoriaceous texture, flat to slightly texture and indumentum, characteristics that deviated from

13
 Journal of Plant Research

Fig. 1  Representative leaves of the six species of the Q. microphylla complex on abaxial view. Leaves of the Q. striatula II morphotype are also included

13
Journal of Plant Research

Fig. 2  Collection sites of the populations of the Q. microphylla complex, accompanied by the population number according to Table 1. Gray
shading indicates elevation (m a.s.l.)

those reported in the original descriptions of Q. striatula, From the “.TPS” file, the matrix of coordinates was
so these populations were designated as a variant “morpho- imported into the Geomorph 4.0.3 (Adams et al. 2016)
type” (hereafter called Q. striatula II) (see Fig. 1). For each package of the R Statistical Software (v4.1.2; R Core Team
of the collected populations, five leaves per individual (from 2021), where the first and last points of each leaf outline
different parts of the shrub) were selected for all sampled were removed to avoid overlap with the apical and basal
individuals at the population (5–20), ensuring that selected landmarks. Subsequently, a Generalized Procrustes Analysis
leaves were mature and without noticeable herbivory marks. (GPA) was implemented to remove size, position, and rota-
Subsequently, the leaves were pressed and dried. tion effects. From the resulting datasets, leaf shape was aver-
aged for each individual using the mshape function of the
Geomorph package. In order to visualize the clustering pat-
Morphological analyses terns of individuals while maximizing the variance between
a priori defined groups (i. e. the six species plus the Q. stria-
Geometric morphometrics. Once dried, leaves were digitized tula II morphotype), a canonical variate analysis (CVA) was
using a flatbed scanner (Hp SJ300) at a resolution of 300 dpi, performed using the CVA function implemented in the Mor-
always accompanied by a size scale. In order to represent the pho 2.11 package (Schlager 2017). Additionally, subsequent
leaf shape, we placed two landmarks at homologous points CVAs were performed to explore differences within the main
of each leaf (the insertion of the petiole with the lamina, and groups of individuals segregated by the previous analysis,
the apex), as well as 35 equidistant points on the leaf outline this time averaging leaf shape for each population. Finally,
(Fig. S1). Then, the tpsDig2.10 (Rohlf 2006) software was linear distances between different pairs of landmarks (i. e.
used to generate a data matrix (x, y coordinates of each of blade length, middle blade width, apical width and basal
the 37 total landmarks) in “.TPS” format. width) were calculated (Fig. S2), to analyze in a traditional

13
 Journal of Plant Research

way the leaf morphological differences between the taxa, to trim the set of bioclimatic layers that were used to elabo-
using the interlmkdist function of the Geomorph package. rate the ENM for each taxon. The selection and editing of the
Boxviolin diagrams were used to represent values of each of ecoregions, as well as the clipping and transformation of the
these linear distances, grouped by taxon. Differences among bioclimatic layers to ASCII format, were performed using
taxa were evaluated using the Kruskall-Wallis test, followed the QGis 3.8.3 program (QGIS Development Team 2019).
by a Dunn test as a post-hoc test (due to the non-normality Because using a large number of bioclimatic layers
of the data), implemented in the R software. can lead to overfitting of the models (Peterson and Naka-
Trichome imaging. Trichomes were observed in two ran- zawa 2008), a pre-selection of bioclimatic variables was
domly chosen individuals from each population collected. performed before making the final ENM. Initially, the 19
Trichomes were scraped from the abaxial side, always from climatic variables were considered to obtain the distribu-
the central part of the leaf blade in the region between sec- tion model for each taxon. From the models obtained and
ondary veins and mounted on a microscope slide using fast- based on the Jackknife test of MaxEnt ver. 3.4.1 k (Phil-
drying nail polish. Subsequently, these preparations were lips et al. 2006), those variables with the highest percentage
observed under an optical microscope (Zeiss Axio Imager. of contribution to the model for each taxon were selected.
A2) with a 10 × objective, and images of the trichomes were Subsequently, by means of a Pearson correlation analysis
digitalized and scaled using an integrated photographic in R software, from each pair of highly correlated variables
camera for further analysis. From each preparation, ten tri- (r > 0.80) one was eliminated, thus obtaining the sets of vari-
chomes were subsequently randomly selected, considering ables for each of the taxa (Table S2).
both zenithal and lateral views, to clearly observe the num- The ENM were performed with the maximum entropy
ber of arms and their pattern of insertion. Trichomes were algorithm in Maxent version 3.4.1 k (Phillips et al. 2006),
characterized according to the different types proposed by since it has been demonstrated with different simulations
Jones (1986). In addition, the number of arms of each tri- that this algorithm generates good predictions, even with
chome was counted and a Kruskall-Wallis test, followed by a reduced number of records (< 10; Pearson et al. 2007;
a Wilcoxon test, was applied to evaluate possible differences Phillips et al. 2006). The parameters for the development of
between taxa, using the R software. the models were those that came by default in the program
(Phillips and Dudík 2008), except for the Extrapolate and
Ecological niche models Do clamping options that were deactivated to avoid artifi-
cial extrapolations in the extreme values of the ecological
Ecological niche models (ENM) were constructed consider- variables (Elith et al. 2011). The model output format was
ing seven entities: i. e. the six species and the Q. striatula logistic, and 75% of the records was used to train the models
II morphotype (Table S2). Presence records for each entity and the remaining 25% to validate them; this partitioning
were obtained from our own field collections and specimen was done randomly. Initially, the area under the curve (AUC)
data from FCME and MEXU herbaria; the final dataset con- value of the receiver operating characteristic (ROC) was
sisted of 15 records for Q. cordifolia, 14 for Q. frutex, 31 for used to evaluate the robustness of the ENM. Those models
Q. intricata, five for Q. microphylla, 13 for Q. repanda, 11 with AUC values between 0.8–0.9 were considered reason-
for Q. striatula, and 27 records for the Q. striatula II mor- ably good and ENM with values above 0.9 were categorized
photype. To avoid overfitting the models, the thinning func- as very good (Peterson et al. 2011). However, considering
tion of the spThin package (Aiello-Lammens et al. 2015) that the usefulness of ROC analysis has been questioned
in R software was used, which randomly selects presence (Lobo et al. 2008; Peterson and Nakazawa 2008), the final
points separated by at least 5 km, using 100 replicates. ENM were also validated by partial ROC analysis, which has
The 19 bioclimatic variables were obtained from the been designed to address the shortcomings of ROC analysis
Worldclim dataset at a spatial resolution of 30 arcseconds (Peterson et al. 2008). This analysis was performed using
(Fick and Hijmans 2017); https://​world​clim.​org/​data/​world​ the online tool NicheToolBox (http://​shiny.​conab​io.​gob.​mx:​
clim21.​html) for the area corresponding to Mexico. These 3838/​niche​toolb2/) with 1000 replicates and an omission
layers were trimmed into smaller areas to be used in the ratio (E) = 0.05.
analyses for each taxon. The criterion for trimming was to Finally, the degree of climate niche differentiation among
reduce the layers according to the accessible area (area M) taxa was measured via Schoener's D (Schoener 1968), using
for each taxon, which was defined based on the ecoregions the niche similarity test implemented in ENM Tools (Warren
file of the World Wildlife Fund (Olson et al. 2001), selecting et al. 2010, 2021). This test is recommended when species
those ecoregions that coincided with the location of the col- to be compared are partially or fully allopatric (Warren et al.
lection sites for each taxon, as well as the areas contiguous to 2010, 2021) and employs randomization to compute a null
them, where they could hypothetically be present (Martinez- distribution and estimate whether the ENMs from different
Meyer 2005). Thus, the resulting areas were used as a “clip” species are more or less like each other than expected by

13
Journal of Plant Research

chance, based on environmental differences in the environ- along the first axis (Fig. 3). These groups differ significantly
ment in which they occur (Warren et al. 2008). The test according to the calculated Mahalanobis distances between
was run using the “background.test” function in R package taxa (Table 1). The first of these two groups (Group I) is
“ENMTools” (Warren et al. 2010, 2021) with 100 replicates, situated at negative values of the first axis and is formed by
sampling a total of 10,000 random points and was analyzed individuals of Q. frutex, Q. microphylla, Q. repanda and Q.
with a two-tailed test. striatula, all of which possess elliptic-oblong leaf morphol-
ogies. Meanwhile, the second group (Group II) is located
towards positive values of the same axis and included indi-
Results viduals of Q. cordifola, Q. intricata and Q. striatula II,
characterized by ovate leaves with a slightly cordate base
Geometric morphometrics (Fig. 3). Interestingly, these groups have a geographic cor-
respondence. Populations composing Group I are distributed
The canonical variate analysis among all individuals indi- mainly on the Trans-Mexican Volcanic Belt (TMVB), the
cated that 75.3% of the total variance in leaf shape was Sierra Madre Occidental (SMOc) and a little portion of the
explained by the first two axes (CV1 = 53.7%, CV2 = 21.6%). south of the Mexican Altiplano (MA), while populations in
The scatter plot of the first two canonical variates revealed Group II are mainly found over the Sierra Madre Oriental
that it is possible to distinguish two main groups separated (SMOr) and the Mexican Altiplano (MA) (Fig. 2).

Fig. 3  Scatterplot of the first two canonical variates (CV1 = 45.9%, CV2 = 18.7%). Each point represents the average leaf shape of an individual.
Representative morphologies of the two main groups along the first canonical axis are shown at the bottom

13
 Journal of Plant Research

Table 1  Mahalanobis distances Quercus species Q. cordifolia Qf Qi Qm Qr Qs

between group means from a
Canonical Variate Analysis Q. frutex (Qf) 6.14
(CVA) of individual mean leaf
Q. intricata (Qi) 5.75 8.02
shape in the Q. microphylla
complex Q. microphylla (Qm) 5.26 3.68 7.73
Q. repanda (Qr) 5.92 4.30 7.89 4.70
Q. striatula (Qs) 5.46 2.84 7.38 3.08 4.22
Q. striatula (QsII) 3.96 7.53 7.28 6.65 7.47 6.70

Values in bold are significant (P < 0.05)

Subsequent canonical variate analysis for the population However, Q. frutex-Q. repanda and Q. striatula-Q. micro-
averages of the samples belonging to Group I explained phylla did not differ significantly according to Mahalanobis
88.7% of the variance on the first two axes (CV1 = 58%, distances (Table 2).
CV2 = 30.7%). The scatter plot suggested two main morpho- Likewise, in the case of Group II, the CVA analysis recov-
logical groupings within this group, one formed by Q. frutex ered 99% of the variation in the first two axes (CV1 = 74.1%,
and Q. repanda (with negative values along the CV1), with CV2 = 25.1%). In this case, it segregated the populations
elliptic-oblong leaves with a slightly attenuated base. On of Q. intricata, characterized by elliptic-ovate forms with
the other hand, Q. microphylla and Q. striatula (with posi- attenuated apex and slightly cordate base, with respect to a
tive values along the CV1) formed another grouping with second morphological group formed by Q. cordifolia and the
predominantly elliptic blades with a rounded base (Fig. 4a). Q. striatula II morphotype, both with ovate leaf blades with

Fig. 4  Scatterplot of population averages for the first two canonical variates corresponding to the morphological Group I (a) and morphological
Group II (b). Representative leaf morphologies of the two main groups defined by the CV1 in each analysis are shown

13
Journal of Plant Research

Table 2  Mahalanobis distances between group means from CVA of population mean leaf shape within the morphological Group I and morpho-
logical Group II in the Q. microphylla complex
Group 1
Q. frutex Q. microphylla Q. repanda

Q. microphylla 6.65
Q. repanda 6.10 7.67
Q. striatula 4.42 3.80 6.95
Group 2
Q. cordifolia Q. intricata

Q. intricata 4.29
Q. striatula II 3.50 5.55

Values in bold are significant (P < 0.05)

cordate base (Fig. 4b). In this case, the comparison between (P < 0.0001) greater than 1 (Table S3), indicating that the
Q. cordifolia and the Q. striatula II morphotype was not models obtained are statistically better than chance.
significant in terms of Mahalanobis distances (Table 2). The potential distribution maps based on the ENM
The distribution of the values obtained for the linear dis- showed three main geographical patterns: (i) taxa with
tances between pairs of landmarks reaffirms the distinct- higher suitability in the TMVB (Q. frutex, Q. microphylla
ness of most taxa in pairwise comparisons for each variable and Q. repanda); (ii) taxa with higher suitability in the
(Fig. 5). Despite the similarities in leaf shape among the SMOc (Q. striatula); and (iii) taxa with higher suitability
taxa within each of the two morphological groups, the com- in the SMOr and MA (Q. cordifolia, Q. intricata and Q.
parisons of their linear dimensions were significant in most striatula II) (Fig. 7). These patterns also coincided with the
cases. In general, the pair of taxa with more discrimination vegetation type where the different taxa were collected: oak
difficulties with these variables is Q. cordifolia-Q. intricata forest margins (TMBV), pine-oak forest clearings (SMOc),
which showed the least differences (Fig. 5). scrub and pine-oak forest (SMOr), and xerophytic scrub with
pine (MA).
Trichome analysis Compatible with what was observed with the morpho-
logical groupings, Q. frutex, Q. microphylla and Q. repanda
Stellate-type trichomes were found in all populations ana- (Fig. 7a, c, g) are codistributed in some parts of the TMBV,
lyzed, with presence of stipitate fasciculate ones in Q. frutex particularly towards the central part. However, the distribu-
and very scarcely in Q. intricata and Q. microphylla (Fig. 6). tion of Q. frutex and Q. repanda is mostly located towards
The number of arms of the trichomes ranged from five to 23 the east, while Q. microphylla shows a fragmented distri-
with a mean of eight trichomes for most taxa. The Kruskall- bution mainly towards the western part of the TMBV. The
Wallis test (χ2 = 81.217, P = 0.0001) followed by a Wilcoxon fourth species of this group, Q. striatula, shows a potential
post-hoc test indicated significant differences in arm number distribution restricted to the SMOc (Fig. 7e). On the other
between Q. cordifolia-Q. striatula, Q. microphylla-Q. stria- hand, the species in the morphological Group II showed sim-
tula and Q. repanda from all the other taxa (Fig. S3). ilar potential distribution areas. Quercus striatula II (Fig. 7f)
showed a discontinuous distribution in the highlands of
Ecological niche models Nuevo León state and southern Coahuila state, similar to
what was observed for Q. cordifolia (Fig. 7b). Quercus intri-
From the Jackknife analysis and Pearson’s test, a total of two cata (Fig. 7d), although sympatric with these two previous
(Q. cordifolia and Q. repanda) to seven (Q. striatula II) non- species in some areas of the SMOr, is also found discontinu-
collinear bioclimatic variables were chosen to perform the ously in central and northern Coahuila and even reaches the
ENM of the different taxa (Table S2). In general, the ENM United States of America, according to records not used in
for the analyzed taxa were categorized as “very good” based this work. The pairwise niche comparison tests were signifi-
on the AUC values, which were greater than 0.9, except for cant in all cases, with Schoener’s D values ranging between
the model of Q. repanda, which had AUC = 0.89 for the 0.069 and 0.688 (Table 3). The lowest D value was obtained
training data and AUC = 0.86 for the validation data and between Q. striatula and Q. striatula II, revealing that the
was thus considered as “reasonably good”. Likewise, par- niche occupied by the two taxa are highly divergent from
tial ROC tests yielded AUC ratios with values significantly each other (D = 0.069; P = 0.009). In turn, the highest niche

13
 Journal of Plant Research

Fig. 5  Box-violin plots of linear distances: a blade length, b mid- ferent letters indicate significant differences among groups according
dle blade width, c basal width, d apical width. Colors correspond to a Kruskal–Wallis test followed by a Dunn test. Q. f = Q. frutex, Q.
to the two main morphological groups in Fig. 3, with red indicating m = Q. microphylla, Q. r = Q. repanda, Q. s = Q. striatula, Q. c = Q.
taxa included in Group I and blue indicating taxa in Group II. Dif- cordifolia, Q. i = Q. intricata, Q. s II = Q. striatula II

similarity was observed between Q. frutex and Q. repanda Leaf analysis from a population perspective,
(D = 0.688: P = 0.029). the first part of the puzzle

Recently, the need to adopt methodologies for taxonomic

Discussion purposes based on data from several individuals and dif-
ferent populations has been highlighted, mainly due to the
Discrimination and circumscription of species at the mor- variation that individuals or populations may present (Di
phological and molecular level, as well as their biogeograph- Pietro et al. 2020). Such variation arises as a result of phe-
ical and ecological characterization, is one of the current notypic plasticity or local adaptation to the different envi-
challenges prevailing for several species groups within the ronments in which populations develop, sometimes leading
genus Quercus (Denk et al. 2017). However, this challenge to the formation of ecomorphotypes, potentially causing
is increased in groups of recently diverged species (Pinheiro a taxonomic overestimation of the number of species (Di
et al. 2018), being particularly complicated when diversifi- Pietro et al. 2020; Proietti et al. 2021). At the other extreme,
cation or lineage formation occurs with few morphological the number of species can be underestimated because of
changes (cryptic or sister species), or when reduced differ- morphological similarities, as reported by Morales-Saldaña
entiation results from non-adaptive radiation, morphologi- et al. (2022) for the Q. laeta complex, with a large number
cal stasis, or phenotypic convergence (Bickford et al. 2007; of morphotypes and synonymies within it, which may cor-
Morales-Saldaña et al. 2022; Nosil and Feder 2012; Valen- respond to different species.
cia-A et al. 2016; Valencia-A 2020). For this reason, among the most powerful tools of cur-
rent morphological analysis, which has been used in differ-
ent taxonomical complexes, geometric morphometrics has
the advantage of providing statistical robustness, while the
analysis of the shape as a whole will also implicitly evalu-
ate different traits quantified by traditional methods (e.g.,
leaf length, width, distance between lobes and/or veins).

13
Journal of Plant Research

Fig. 6  Trichomes found in the different taxa. All trichomes are of the Q. intricata, d Q. microphylla, e Q. repanda, f Q. striatula and g Q.
stellate type with different number of arms, except in Q. frutex which striatula II
shows stipitate-fasciculate trichomes. a Q. cordifolia, b Q. frutex, c

In this way, a detailed analysis of the species hypotheses (Fig. 4a), as reported by Romero et al. (2002) and Valencia-
being tested is possible, including the identification of shape A et al. (2017). However, they differ in all linear dimen-
aspects for which the evaluated groups differ the most, there- sions and also in shape, from elliptic forms with attenuated
fore contributing to the elaboration of an overall diagnosis base (Q. frutex) to obovate forms (Q. repanda) (Figs. 4a,
for each group of samples from a population perspective 5). Finally, in this group there is similarity between Q.
(Pérez-Pedraza et al. 2021; Valencia-Cuevas et al. 2015; microphylla and Q. striatula, both with elliptical forms and
Viscosi et al. 2009a, 2010; Yang et al. 2022). rounded base, which may explain what was reported by
The Q. microphylla complex has been reviewed in few McVaugh (1974) and González (1986).
taxonomic works. Trelease (1924) proposed the Microphyl- On the other hand, it should be noted that the morphotype
lae series, at that time only consisting of the currently rec- Q. striatula II, which has been determined as Q. striatula in
ognized Q. microphylla, Q. frutex and Q. repanda. However, herbaria, clearly differs from the type and specimens con-
the high intraspecific morphological variation of these spe- sidered as Q. striatula in this work, partially explaining the
cies generates a morphological gradation that results in an identity problems between these taxa. In this regard, within
overlap of the forms reported in the original descriptions Group II, the CVA based on population morphological aver-
that, together with the presence of non-glandular trichomes ages clearly separated Q. intricata from Q. striatula II. At
in the indumentum of all of them, has caused diverse inter- the same time, similarity was observed between the latter
pretations on the validity of the taxa (Sabás-Rosales et al. and Q. cordifolia (Fig. 4b), which have had identity prob-
2017). lems according to Muller (1944) and Villareal et al. (2008)
Currently, the Q. microphylla complex as proposed by by presenting ovate leaves with a cordate base. Despite this,
Valencia-A (2020) is composed of Q. cordifolia, Quercus they differ in the four linear dimensions, with all measures
deliquescens, Q. frutex, Q. intricata, Q. microphylla, being smaller in Q. striatula II (Fig. 5). So, apparently the
Quercus mohoriana, Q. repanda and Q. striatula. Our geo- taxonomic complexity of Q. cordifolia has actually occurred
metric morphometrics results subdivided the complex into with Q. striatula II and not with Q. striatula.
two groups (Fig. 3). On the one hand, in Group I, we found Finally, the indumentum in all species presented tri-
Q. frutex, Q. repanda, Q. microphylla and Q. striatula, for chomes of the stellate type, with a mean around eight arms
which the possible synonymy and identity problems between for most of them (Q. cordifolia, Q. frutex, Q. intricata, Q.
the latter and Q. cordifolia, previously discussed by Muller microphylla, Q. striatula and Q. striatula II) clearly sepa-
(1944); Martínez (1977) and Villareal et al. (2008), are mor- rating only Q. repanda (mean of 19 arms) from the rest.
phologically ruled out. On the other hand, morphological In addition, the stipitate stellate trichome type is generally
similarity between Q. frutex and Q. repanda was observed present in Q. frutex (Fig. 6). Therefore, these two taxa can

13
 Journal of Plant Research

13
 Journal of Plant Research

◂Fig. 7  Potential distribution of seven taxa of Mexican shrub oaks. a strong niche divergence. Thus, they could be considered as
Q. frutex, b Q. cordifolia, c Q. microphylla, d Q. intricata, e Q. stria- two different species. In addition, in cases in which there is a
tula, f Q. striatula II, g Q. repanda and h distribution of two main
morphological groups (red = morphological Group I; blue = morpho-
limited morphometric differentiation, as it happens between
logical Group II). Gray shading indicates elevation (m a.s.l) Q. microphylla and Q. striatula, the analysis indicated a high
degree of niche differentiation with an evident allopatric
pattern, so that the low morphological differentiation could
be clearly discriminated from the rest of the complex under be due to phenotypic convergence, as has been reported for
these two traits. Likewise, the number of arms achieved sig- other Mexican oaks (Valencia-A 2020). Even though our
nificant differences between Q. microphylla and Q. striatula, results indicated a tendency toward niche divergence among
species with great similarity in morphometric measurements taxa within the Q. microphylla complex, examples of taxa
(Fig. 5). In this sense, it is still necessary to quantitatively with high degree of niche similarity were also observed,
evaluate the density of trichomes in the leaf lamina, as well which were accompanied by morphological similarity (e.
as the epidermis type to find out if there are differences g., Q. cordifolia-Q. striatula II, Q. frutex-Q. repanda and
among the different taxa (Fortini et al. 2015; López-Caamal Q. microphylla-Q. repanda). However, the different types
et al. 2017; Scareli-Santos et al. 2013; Tschan and Denk of evidence combined allows discrimination among taxa in
2012). almost all cases.
Finally, paradoxically Q. microphylla is the species of the
complex with the widest previously reported distribution.
Ecological and geographical patterns within the Q. In spite of this, during our field work it was very difficult
microphylla complex to find populations of this species, so apparently, the high
taxonomic complexity of the group has resulted in an over-
The geographic distribution of the taxa in the Q. microphylla estimation of the distribution of this species, mainly towards
complex is one aspect of notorious disagreement among dif- the north of the country. In addition, one of the problems
ferent authors (González-Elizondo et al. 1993; Romero Ran- detected during the herbarium revision is that several speci-
gel et al. 2002; Sabás-Rosales et al. 2017; Valencia-A 2004; mens were indicated to have a tree growth form in the col-
Valencia-A et al. 2017; Vázquez-Villagrán 2000). From a lection data. This alternation between tree and shrub habit
total of 21 pairwise comparisons of potential distribution can occur in species such as Quercus opaca or Quercus gri-
models, in six comparisons (28.6%) we found some degree sea. However, the species that comprise the Q. microphylla
of overlap in the geographic ranges of the taxa, while in 15 complex are strictly shrubby. But, apparently the shape, and
comparisons (71.4%) they showed non overlapping distribu- particularly the size of the leaves, generates this confusion at
tions (Table 3). This suggests that, in most cases, allopatry the moment of identifying specimens, contributing to over-
is the predominant distribution pattern among species of the estimate the distribution of the species.
Q. microphylla complex. In addition, when we contrasted the
potential distributions between the two main groups detected Q. microphylla, Mexico’s largest shrub oak complex
through geometric morphometry, it was also possible to
observe that these two groups are largely allopatric. Species The Q. microphylla complex represents the shrub oak com-
of Group I are distributed towards the center and west of the plex with more species in Mexico (eight of at least 32 spe-
country, through the TMVB, SMOc and the southern portion cies present in the country). In addition to the species stud-
of the MA, while taxa in Group II are found mainly towards ied in the present work, Q. deliquescens and Q. mohriana
the center and northeast of Mexico, in the south and north are part of the complex, since they share the characteristics
of the MA and the SMOr (Fig. 7). Therefore, this zoning, of shrubby habit, stellate trichomes and elliptic-oblong to
together with morphometric analyses, allows an approxima- ovate leaves. In the case of Q. deliquescens, it was not pos-
tion of the identity of species from their distribution, making sible to locate any population corresponding to the speci-
it possible to rule out the presence of some taxa in certain mens deposited in the herbaria, besides being presumably
regions. restricted to the state of Chihuahua. In turn, Q. mohriana is
Complementarily, the background test allowed us to mainly distributed in the United States of America and in
quantify the degree of niche similarity among taxa and to Mexico it is only found in the north of the state of Coahuila.
incorporate ecological evidence to clarify taxonomic contro- Therefore, both species were not considered for the present
versies within the Q. microphylla complex. Firstly, the cli- work. However, all these species together result in a set of
matic niche comparisons allowed us to support the hypothe- taxa with high morphological/taxonomic complexity within
sis that Q. striatula and Q. striatula II actually represent two the genus Quercus.
distinct biological entities, with a remarkable morphometric Moreover, our results provide valuable information that
differentiation, a clear allopatric distribution pattern and a can complement the poorly detailed descriptions of most

13
 Journal of Plant Research

Table 3  Results of pairwise Quercus species pairwise comparison Range overlap Schoener’s D P-value
niche comparisons tests based Background test
on Schoener’s D for the seven
shrub oak taxa in the Quercus Q. cordifolia-Q. frutex Allopatric 0.191 0.009
microphylla complex based on
Q. cordifolia-Q. intricata Parapatric 0.472 0.009
100 replicates
Q. cordifolia-Q. microphylla Allopatric 0.680 0.039
Q. cordifolia-Q. striatula Allopatric 0.168 0.009
Q. cordifolia- Q. striatula II Parapatric 0.556 0.009
Q. cordifolia- Q. repanda Allopatric 0.471 0.009
Q. frutex-Q. intricata Allopatric 0.077 0.009
Q. frutex-Q. microphylla Parapatric 0.355 0.009
Q. frutex-Q. striatula Allopatric 0.109 0.009
Q. frutex-Q. striatula II Allopatric 0.129 0.009
Q. frutex- Q. repanda Parapatric 0.688 0.029
Q. intricata-Q. microphylla Allopatric 0.388 0.029
Q. intricata- Q. striatula Allopatric 0.098 0.009
Q. intricata-Q. striatula II Parapatric 0.461 0.009
Q. intricata-Q. repanda Allopatric 0.276 0.009
Q. microphylla-Q. striatula Allopatric 0.296 0.009
Q. microphylla-Q. striatula II Allopatric 0.284 0.009
Q. microphylla-Q. repanda Parapatric 0.568 0.019
Q. striatula-Q. striatula II Allopatric 0.069 0.009
Q. striatula-Q. repanda Allopatric 0.275 0.009
Q. striatula II-Q. repanda Allopatric 0.209 0.009

species. And, consequently, they can serve as a basis for within the analyzed populations. Therefore, evaluating the
different types of studies in the genus. In this sense, it has different species hypotheses with molecular data can greatly
been suggested to carefully analyze different evidence simul- contribute to clarify the identity of these variants.
taneously (e. g. morphometric, ecological niche modeling,
Supplementary Information The online version contains supplemen-
molecular and chemical data, etc.), with a wide represen- tary material available at https://​doi.​org/​10.​1007/​s10265-​023-​01495-z.
tation of individuals and populations; and in this way, to
assign a set of morphological characters covering the known Acknowledgements We thank Gonzalo Contreras, Goretty Mendoza,
distribution of the species to also assign its locus classicus Ricardo Gaytán, Javier Piña and Douglas Castillejo for their collabora-
tion in the field work. Gonzalo Contreras helped us with the statistical
(name linked to a type specimen, also linked to a geographi- analysis. We also thank to MSci Violeta Patiño Conde for her assis-
cal locality), as a way of adding additional ecological infor- tance in obtaining microscope images. This work constitutes a partial
mation enabling further identity of the species (Di Pietro fulfillment of the Posgrado en Ciencias Biológicas of the National
et al. 2020). Autonomous University of Mexico (UNAM). Oscar De Luna, thanks
Posgrado en Ciencias Biológicas-UNAM and the Consejo Nacional
Finally, the need for molecular analysis of species bound- de Ciencia y Tecnología (CONACYT) for the PhD scholarship and
aries was highlighted. On the one hand, the strong separa- financial support to develop graduate studies at UNAM (scholarship/
tion between the morphotypes of Q. striatula suggests the CVU:717866/742699). The authors are grateful for funding from the
presence of two different species (they are morphologically, Dirección General de Asuntos del Personal Académico (DGAPA)-
Programa de Apoyo a Proyectos de Investigación e Innovación Tec-
geographically and ecologically distant). Therefore, the mor- nológica (PAPIIT IN210020). We are also grateful for additional
photype Q. striatula II may remain under some synonymy or funding received from the Association de l'Arboretum des Pouyouleix
be undescribed. On the other hand, Q. microphylla and Q. through Béatrice Chassé.
striatula are morphologically grouped, and show intraspe-
Data availability Original data used for this study are available upon
cific variation that deviates from the description of each spe- request to the corresponding author.
cies, mainly in aspects such as leaf consistency, appearance
of the indumentum and the way in which the shrub develops. Declarations
Also, the similarity between Q. microphylla and Q. frutex,
mainly in geographically close populations, can still gen- Conflict of interest On behalf of all authors, the corresponding author
states that there is no conflict of interest.
erate taxonomic confusion. All these variations generate
uncertainty about the existence of other unrecognized taxa

13
Journal of Plant Research

Open Access This article is licensed under a Creative Commons Attri- Elith J, Phillips SJ, Hastie T et al (2011) A statistical explanation of
bution 4.0 International License, which permits use, sharing, adapta- MaxEnt for ecologists. Divers Distrib 17:43–57. https://​doi.​org/​
tion, distribution and reproduction in any medium or format, as long 10.​1111/j.​1472-​4642.​2010.​00725.x
as you give appropriate credit to the original author(s) and the source, Fick SE, Hijmans RJ (2017) WorldClim 2: new 1-km spatial res-
provide a link to the Creative Commons licence, and indicate if changes olution climate surfaces for global land areas. Int J Climatol
were made. The images or other third party material in this article are 37:4302–4315. https://​doi.​org/​10.​1002/​joc.​5086
included in the article’s Creative Commons licence, unless indicated Fortini P, Antonecchia G, Di Marzio P et al (2015) Role of micro-
otherwise in a credit line to the material. If material is not included in morphological leaf traits and molecular data in taxonomy of
the article’s Creative Commons licence and your intended use is not three sympatric white oak species and their hybrids (Quercus
permitted by statutory regulation or exceeds the permitted use, you will L.). Plant Biosyst 149:546–558
need to obtain permission directly from the copyright holder. To view a Galicia L, Potvin C, Messier C (2015) Maintaining the high diver-
copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. sity of pine and oak species in mexican temperate forests: a
new management approach combining functional zoning and
ecosystem adaptability. Can J for Res 45:1358–1368. https://​
doi.​org/​10.​1139/​cjfr-​2014-​0561
References González Villarreal LM (1986) Contribución al conocimiento del
género Quercus (Fagaceae) en el estado de Jalisco. Instituto de
Adams DC, Collyer M, Kaliontzopoulou A, Sherratt E (2016) Geo- Botánica. Universidad de Guadalajara, Guadalajara
morph: software for geometric morphometric analyses González-Elizondo SM, González-Elizondo M, Cortés-Ortiz A
Aiello-Lammens ME, Boria RA, Radosavljevic A et al (2015) (1993) Vegetación de la reserva de la biosfera “La Michilía”,
spThin: an R package for spatial thinning of species occur- Durango, México. Acta Bot Mex. https://​d oi.​o rg/​1 0.​2 1829/​
rence records for use in ecological niche models. Ecography abm22.​1993.​668
38:541–545. https://​doi.​org/​10.​1111/​ecog.​01132 González-Rodríguez A, Bain JF, Golden JL, Oyama K (2004) Chloro-
Albarrán-Lara AL, Mendoza-Cuenca L, Valencia-Avalos S et al plast DNA variation in the Quercus affinis-Q. laurina complex in
(2010) Leaf fluctuating asymmetry increases with hybridization Mexico: geographical structure and associations with nuclear and
and introgression between Quercus magnoliifolia and Quercus morphological variation. Mol Ecol 13:3467–3476. https://d​ oi.o​ rg/​
resinosa (Fagaceae) through an altitudinal gradient in Mexico. 10.​1111/j.​1365-​294X.​2004.​02344.x
Int J Plant Sci 171:310–322. https://​doi.​org/​10.​1086/​650317 González VLM (1986) Contribución al conocimiento del género
Aykut Y, Emel U, Tekin BM (2017) Morphological variability of Quercus (Fagaceae) en el estado de Jalisco. Universidad de Gua-
evergreen oaks (Quercus) in Turkey. Bangladesh J Plant Taxon dalajara, México
24(1):39–47. https://​doi.​org/​10.​3329/​bjpt.​v24i1.​33004 Hardin JW (1975) Hybridization and introgression in Quercus alba. J
Bartholomew B, Almeda F (2023) Nomenclator botanicus of Arnold Arboretum 56:336–363
Fagaceae in Latin America. Phytotaxa 580:1–97 Hipp AL, Manos PS, González-Rodríguez A et al (2018) Sympatric
Bickford D, Lohman DJ, Sodhi NS et al (2007) Cryptic species as parallel diversification of major oak clades in the Americas and
a window on diversity and conservation. Trends Ecol Evol the origins of Mexican species diversity. New Phytol 217:439–
22:148–155. https://​doi.​org/​10.​1016/j.​tree.​2006.​11.​004 452. https://​doi.​org/​10.​1111/​nph.​14773
Borazan A, Babaç MT (2003) Morphometric leaf variation in oaks Hipp AL, Manos PS, Hahn M et al (2020) Genomic landscape of the
(Quercus) of Bolu, Turkey. Ann Bot Fenn 40:233–242 global oak phylogeny. New Phytol 226:1198–1212. https://​doi.​
Cannon CH, Petit RJ (2020) The oak syngameon: more than the org/​10.​1111/​nph.​16162
sum of its parts. New Phytol 226:978–983. https://​doi.​org/​10.​ Jensen RJ (1990) Detecting shape variation in oak leaf morphology: a
1111/​nph.​16091 comparison of rotational-fit methods. Am J Bot 77:1279–1293.
Cavender-Bares J (2019) Diversification, adaptation, and community https://​doi.​org/​10.​1002/j.​1537-​2197.​1990.​tb113​80.x
assembly of the American oaks (Quercus), a model clade for Jones JH (1986) Evolution of the Fagaceae: the implications of foliar
integrating ecology and evolution. New Phytol 221:669–692. features. Ann Missouri Bot Garden 73:228–275
https://​doi.​org/​10.​1111/​nph.​15450 Leroy T, Roux C, Villate L et al (2017) Extensive recent secondary
Chitwood DH, Sinha NR (2016) Evolutionary and environmental contacts between four European white oak species. New Phytol
forces sculpting leaf development. Curr Biol 26:R297–R306. 214:865–878. https://​doi.​org/​10.​1111/​nph.​14413
https://​doi.​org/​10.​1016/j.​cub.​2016.​02.​033 Lobo JM, Jiménez-valverde A, Real R (2008) AUC: a misleading meas-
De Heredia UL, Duro-García MJ, Soto Á (2018) Leaf morphology ure of the performance of predictive distribution models. Global
of progenies in Q. suber, Q. ilex, and their hybrids using multi- Ecol Biogeogr 17:145–151
variate and geometric morphometric analysis. Iforest 11:90–98. López-Caamal A, Ruiz-Amaro LDC, Zepeda-Rodríguez A et al (2017)
https://​doi.​org/​10.​3832/​ifor2​577-​010 Micromorphological character expression of the hybrid Quercus
Denk T, Grimm GW, Manos PS et al (2017) An updated infrage- × dysophylla and its parental species (Q. crassifolia and Q. cras-
neric classification of the oaks: review of previous taxonomic sipes). Bot Sci 95:375–389. https://​doi.​org/​10.​17129/​botsci.​875
schemes and synthesis of evolutionary patterns. In: Gil-Pelegrín Martínez M (1977) Los encinos de México. Anales del Instituto de
E, Peguero-Pina JJ, Sancho-Knapik D (eds) Oaks physiological Biología: México Comisión Forestal. pp 230
ecology. Exploring the functional diversity of genus Quercus L. Martinez-Meyer E (2005) Climate change and biodiversity: some con-
Springer International Publishing, Cham, pp 13–38 siderations in forecasting shifts in species’ potential distributions.
Di Pietro R, Conte AL, Di Marzio P et al (2020) A multivariate Biodiv Inf. https://​doi.​org/​10.​17161/​bi.​v2i0.8
morphometric analysis of diagnostic traits in southern Italy and McCauley RA, Cortés-Palomec AC, Oyama K (2019) Species diversi-
Sicily pubescent oaks. Folia Geobot 55:163–183. https://​doi.​ fication in a lineage of Mexican red oak (Quercus section Lobatae
org/​10.​1007/​s12224-​020-​09378-0 subsection Racemiflorae) the interplay between distance, habitat,
Dumolin-Lapégue S, Kremer A, Petit RJ (1999) Are chloroplast and and hybridization. Tree Genet Genomes. https://​doi.​org/​10.​1007/​
mitochondrial DNA variation species independent in oaks? s11295-​019-​1333-x
Evolution (N Y) 53:1406–1413. https://​doi.​org/​10.​1111/j.​1558-​ McVaugh R (1974) Flora Novo-Galiciana: Fagaceae. University Her-
5646.​1999.​tb054​05.x barium, University of Michigan, Ann Arbor

13
 Journal of Plant Research

Morales-Saldaña S, Valencia-Ávalos S, Oyama K et al (2022) Even Trans-Mexican Volcanic Belt. Botany 96:121–131. https://d​ oi.o​ rg/​
more oak species in Mexico? Genetic structure and morphological 10.​1139/​cjb-​2017-​0116
differentiation support the presence of at least two specific enti- Rohlf FJ (2006) tpsDig 2.10. Department of Ecology and Evolution,
ties within Quercus laeta. J Syst Evol. https://​doi.​org/​10.​1111/​ State University of New York, Stony Brook, NY
jse.​12818 Romero RS, Rojas ZEC, Aguilar EML (2002) El Género Quercus
Muller CH (1944) Fagaceae. J Arnold Arboretum 25:437–450 (Fagaceae) en el Estado de México. Ann Mo Bo Gard 89(4):551–
Nixon KC (1993) Infrageneric classification of Quercus (Fagaceae) and 593. https://​doi.​org/​10.​2307/​32985​95
typification of sectional names. Ann Sci For 50:25–34 Romero Rangel S (2006) Revisión taxonómica del complejo Acutifo-
Nosil P, Feder JL (2012) Genomic divergence during speciation: causes liae de Quercus (Fagaceae) con énfasis en su representación en
and consequences. Philos Trans Royal Soc B: Biol Sci 367:332– México. Acta Bot Mex 76:1–45
342. https://​doi.​org/​10.​1098/​rstb.​2011.​0263 Romero Rangel S, Rojas Zenteno EC, Enríquez A, de Lourdes M
Olson DM, Dinerstein E, Wikramanayake ED et al (2001) Terrestrial (2002) El Género Quercus ( Fagaceae ) en el Estado de México.
ecoregions of the world: a new map of life on earth: a new global Ann Missouri Bot Garden 89:551–593
map of terrestrial ecoregions provides an innovative tool for con- Romero Rangel S, Rojas Zenteno EC, Rubio Licona LE (2014)
serving biodiversity. Bioscience 51:933–938. https://​doi.​org/​10.​ Fagaceae. Flora Del bajío y de Regiones Adyacentes 42:
1641/​0006-​3568(2001)​051[0933:​TEOTWA]​2.0.​CO;2 87–110. https://​doi.​org/​10.​1016/​0034-​6667(84)​90063-0
Ortego J, Noguerales V, Gugger PF, Sork VL (2015) Evolutionary Sabás-Rosales JL, Siqueiros-Delgado ME, Valencia-Ávalos S,
and demographic history of the Californian scrub white oak spe- Enríquez-Enríquez ED (2017) Reconocimiento taxonómico de
cies complex: an integrative approach. Mol Ecol 24:6188–6208. seis especies arbustivas de encinos (Quercus secc. Quercus:
https://​doi.​org/​10.​1111/​mec.​13457 Fagaceae). Polibotanica. https://​d oi.​o rg/​1 0.​1 8387/​p olib​otani​
Pearson RG, Raxworthy CJ, Nakamura M, Townsend Peterson A ca.​44.2
(2007) Predicting species distributions from small numbers of Scareli-Santos C, Sánchez-Mondragón ML, González-Rodríguez A,
occurrence records: a test case using cryptic geckos in Madagas- Oyama K (2013) Micromorphology of Mexican Oaks (Quercus:
car. J Biogeogr 34:102–117. https://​doi.​org/​10.​1111/j.​1365-​2699.​ Fagaceae). Acta Bot Mex 104:31–52
2006.​01594.x Schlager S (2017) Morpho and Rvcg-Shape Analysis in R: R-Pack-
Pérez-Pedraza A, Rodríguez-Correa H, Valencia-Ávalos S et al ages for geometric morphometrics, shape analysis and surface
(2021) Effect of hybridization on the morphological differen- manipulations. Statistical shape and deformation analysis. Else-
tiation of the red oaks Quercus acutifolia and Quercus grahamii vier, Amsterdam, pp 217–256
(Fagaceae). Plant Syst Evol 307:1–14. https://​doi.​org/​10.​1007/​ Sork VL, Riordan E, Gugger PF et al (2016) Phylogeny and intro-
s00606-​021-​01757-0 gression of California scrub white oaks (Quercus section
Peterson AT, Nakazawa Y (2008) Environmental data sets matter in Quercus). Int Oaks 27:1–14
ecological niche modelling: An example with Solenopsis invicta Trelease W (1924) The american oaks. Mem Natl Acad Sci 20:1–255
and Solenopsis richteri. Global Ecol Biogeogr 17:135–144. Tschan GF, Denk T (2012) Trichome types, foliar indumentum and
https://​doi.​org/​10.​1111/j.​1466-​8238.​2007.​00347.x epicuticular wax in the Mediterranean gall oaks, Quercus sub-
Peterson AT, Papeş M, Soberón J (2008) Rethinking receiver operating section Galliferae (Fagaceae): Implications for taxonomy, ecol-
characteristic analysis applications in ecological niche modeling. ogy and evolution. Bot J Linnean Soc 169:611–644. https://​doi.​
Ecol Modell 213:63–72. https://​doi.​org/​10.​1016/j.​ecolm​odel.​ org/​10.​1111/j.​1095-​8339.​2012.​01233.x
2007.​11.​008 Tucker JM (1974) Patterns of parallel evolution of leaf form in new
Peterson AT, Soberón J, Pearson RG et al (2011) Ecological niches world oaks. Taxon 23:129–154. https://​doi.​org/​10.​2307/​12180​
and geographic distributions: monographs in population biology. 95
Princeton University Press, Princeton Valencia-A S (2004) Diversidad del género Quercus (Fagaceae) en
Phillips SJ, Dudík M (2008) Modeling of species distributions with México. Bot Sci 75:33–53. https://​doi.​org/​10.​17129/​botsci.​1692
Maxent: new extensions and a comprehensive evaluation. Ecog- Valencia-A S (2020) Species delimitation in the genus Quercus
raphy 31:161–175. https://​doi.​org/​10.​1111/j.​0906-​7590.​2008.​ (Fagaceae). Bot Sci 99:1–12. https://d​ oi.o​ rg/1​ 0.1​ 7129/b​ otsci.2​ 658
5203.x Valencia-A S, Sabas Rosales JL, Soto Arellano OJ (2016) A new
Phillips SB, Aneja VP, Kang D, Arya SP (2006) Modelling and analy- species of Quercus, section Lobatae (Fagaceae) from the Sierra
sis of the atmospheric nitrogen deposition in North Carolina. Int Madre oriental, Mexico. Phytotaxa 269:120–126. https://​doi.​org/​
J Global Environ Issues 6:231–252 10.​11646/​phyto​taxa.​269.2.5
Pinheiro F, Dantas-Queiroz MV, Palma-Silva C (2018) Plant species Valencia-A S, Flores-Franco G, Jiménez-Ramírez J, Mora-Jarvio M
complexes as models to understand speciation and evolution: a (2017) Distribution and diversity of Fagaceae in Hidalgo, Mexico.
review of South American studies. CRC Crit Rev Plant Sci 37:54– Bot Sci 95:660–721. https://​doi.​org/​10.​17129/​botsci.​1020
80. https://​doi.​org/​10.​1080/​07352​689.​2018.​14715​65 Valencia-Cuevas L, Mussali-Galante P, Piñero D et al (2015) Hybridi-
Proietti E, Filesi L, Di Marzio P et al (2021) Morphology, geometric zation of Quercus castanea (Fagaceae) across a red oak species
morphometrics, and taxonomy in relict deciduous oaks woods in gradient in Mexico. Plant Syst Evol 301:1085–1097. https://​doi.​
northern Italy. Rendiconti Lincei 32:549–564. https://​doi.​org/​10.​ org/​10.​1007/​s00606-​014-​1151-4
1007/​s12210-​021-​01001-4 Vázquez-Villagrán L (2000) Flora del valle de Tehuacán-Cuicatlán.
QGIS Development Team. 2019. QGIS Geographic information sys- Fagaceae. Universidad Nacional Autónoma De México Instituto
tem. Open Source Geospatial Foundation Project. http://​qgis.​ De Biología 28:1–39
osgeo.​org Villareal JA, Encina JA, Carranza MA (2008) Los encinos (Quercus:
R Core TeamR Core Team (2021). R: A language and environment Fagaceae) de Coahuila, México. J Bot Res Inst Tex 2:1235–1278.
for statistical computing. R Foundation for Statistical Computing, https://​doi.​org/​10.​1111/​conl.​12288.​This
Vienna, Austria. URL https://​www.R-​proje​ct.​org/ Viscosi V (2015) Geometric morphometrics and leaf phenotypic plas-
Rodríguez-Gómez F, Oyama K, Ochoa-Orozco M et al (2018) Phy- ticity: assessing fluctuating asymmetry and allometry in European
logeography and climate-associated morphological variation in white oaks (Quercus). Bot J Linnean Soc 179:335–348. https://​
the endemic white oak Quercus deserticola (Fagaceae) along the doi.​org/​10.​1111/​boj.​12323

13
Journal of Plant Research

Viscosi V, Fortini P, Slice DE et al (2009a) Geometric morphomet- Warren DL, Matzke NJ, Cardillo M et al (2021) ENMTools 1.0: an
ric analyses of leaf variation in four oak species of the subgenus R package for comparative ecological biogeography. Ecography
Quercus (Fagaceae). Plant Biosyst 143:575–587. https://​doi.​org/​ 44:504–511. https://​doi.​org/​10.​1111/​ecog.​05485
10.​1080/​11263​50090​27752​77 Whittemore AT, Schaal BA (1991) Interspecific gene flow in sympatric
Viscosi V, Lepais O, Gerber S, Fortini P (2009b) Leaf morphologi- oaks. Proc Natl Acad Sci USA 88:2540–2544. https://​doi.​org/​10.​
cal analyses in four European oak species (Quercus) and their 1073/​pnas.​88.6.​2540
hybrids: a comparison of traditional and geometric morphomet- Wu Y, Hipp AL, Fargo G et al (2023) Improving species delimitation
ric methods. Plant Biosyst 143:564–574. https://​doi.​org/​10.​1080/​ for effective conservation: a case study in the endemic maple-leaf
11263​50090​27231​29 oak (Quercus acerifolia). New Phytol 238:1278–1293. https://d​ oi.​
Viscosi V, Loy A, Fortini P (2010) Geometric morphometric analysis org/​10.​1111/​nph.​18777
as a tool to explore covariation between shape and other quan- Yang K, Wu J, Li X et al (2022) Intraspecific leaf morphological vari-
titative leaf traits in European white oaks. Tools for identifying ation in Quercus dentata Thunb.: a comparison of traditional and
Biodiversity: Progress and Problems 257–261 geometric morphometric methods, a pilot study. J For Res (Har-
Warren DL, Glor RE, Turelli M (2008) Environmental niche equiva- bin). https://​doi.​org/​10.​1007/​s11676-​022-​01452-x
lency versus conservatism: quantitative approaches to niche evo-
lution. Evolution (N Y) 62:2868–2883. https://​doi.​org/​10.​1111/j.​ Publisher's Note Springer Nature remains neutral with regard to
1558-​5646.​2008.​00482.x jurisdictional claims in published maps and institutional affiliations.
Warren DL, Glor RE, Turelli M (2010) ENMTools: a toolbox for
comparative studies of environmental niche models. Ecography
33:607–611. https://​doi.​org/​10.​1111/j.​1600-​0587.​2009.​06142.x

13