Evolutionary Stasis in Euphorbiaceae Pollen

doi: 10.1111/j.1420-9101.2012.02494.
Evolutionary stasis in Euphorbiaceae pollen: selection

and constraints
A. MATAMORO-VIDAL*, C. A. FURNESS, P.-H. GOUYON*, K. J. WURDACK & B. ALBERT
*Departement Systematique et Evolution, Museum National dHistoire Naturelle, UMR CNRS 7205 OSEB, Paris, France
Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, UK
Department of Botany, National Museum of Natural History, Washington, DC, USA
Universite Paris-Sud 11, UMR CNRS-AgroParisTech 8079, Ecologie Systematique et Evolution, Orsay, France
Keywords:
Abstract
comparative analysis;
evo-devo;
microsporogenesis;
morphospace;
pollen aperture pattern.
Although much attention has been paid to the role of stabilizing selection,
empirical analyses testing the role of developmental constraints in evolutionary stasis remain rare, particularly for plants. This topic is studied here with a
focus on the evolution of a pollen ontogenetic feature, the last points of callose
deposition (LPCD) pattern, involved in the determination of an adaptive
morphological pollen character (aperture pattern). The LPCD pattern exhibits
a low level of evolution in eudicots, as compared to the evolution observed in
monocots. Stasis in this pattern might be explained by developmental
constraints expressed during male meiosis (microsporogenesis) or by selective
pressures expressed through the adaptive role of the aperture pattern. Here,
we demonstrate that the LPCD pattern is conserved in Euphorbiaceae s.s. and
that this conservatism is primarily due to selective pressures. A phylogenetic
association was found between the putative removal of selective pressures on
pollen morphology after the origin of inaperturate pollen, and the appearance
of variation in microsporogenesis and in the resulting LPCD pattern,
suggesting that stasis was due to these selective pressures. However, even in
a neutral context, variation in microsporogenesis was biased. This should
therefore favour the appearance of some developmental and morphological
phenotypes rather than others.
Introduction
Given the Darwinian process of descent with modification, evolutionary change is to be expected as a common
occurrence. Although such evolutionary change occurs in
many cases, several examples of lineages exhibiting little
morphological change for an individual character, for
character complexes or for the whole organism are known
(Futuyma, 2010). It may be inferred that a character
exhibits little morphological change in a lineage when the
rate of change of the character is slower than it theoretically could be (see Bradshaw, 1991; Williams, 1992;
Futuyma, 2010 for examples). The amount of morphological variation of a character can be visualized by
determining the theoretical morphospace of all the posCorrespondence: Alexis Matamoro-Vidal, Department of Biological Science,
Florida State University, Tallahassee, FL 32310, USA.
Tel.: +1 850 645 8521; fax: +1 850 645 8447; e-mail: amv@bio.fsu.edu
sible states that can exist for the character and then
looking at how the morphospace has been filled during the
course of evolution (Raup, 1966; McGhee, 2007). Morphospaces are delineated by the fact that all the members
of a lineage have inherited a limited number of developmental modules (Breuker et al., 2006; Muller, 2007;
Wagner et al., 2007) and that the morphological variation
allowed by these modules is finite (Lauder, 1981; Wake &
Larson, 1987; Raff, 1996; Salazar-Ciudad, 2006; Schlosser,
2007). Therefore, the evolution of a character may only
occur within the limits allowed by these modules unless
new modules are acquired and added to the previous ones.
The utilization of morphospace allows testing for the
occurrence of stasis or phylogenetic conservatism between
related lineages that share common developmental modules. If it is found that, for a given character, a lineage
explored a smaller subset of the morphospace than
another lineage, it may be stated that this character
exhibits stasis in the former relative to the latter.
2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

JOURNAL OF EVOLUTIONARY BIOLOGY 2012 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1077
1078
A. MATAMORO-VIDAL ET AL.
Neo-Darwinians consider two classes of mechanisms

supposedly involved in the lack of change: failure in the
generation of variation caused by developmental constraints or failure in the fixation of new variation
caused by stabilizing selection (Charlesworth et al., 1982;
Maynard Smith, 1983; Williams, 1992; Gould, 2002;
Hansen & Houle, 2004; Eldredge et al., 2005). Evolutionary stasis may thus be explained by the fact that
phenotypes corresponding to the unfilled parts of
the morphospace are always counter-selected. An alternative is that developmental constraints (i.e. biases in the
production of variant phenotypes or limitations on
phenotypic variability caused by the structure, character,
composition or dynamics of the developmental system
Maynard Smith et al., 1985) appeared during the history
of the lineage, preventing switches from the filled parts of
the morphospace to the unfilled ones. Such constraints
may, for example, be expressed as the nonequiprobability
of mutations, resulting in nonequiprobability of developmental changes (Stoltzfus & Yampolsky, 2009); the
impossibility of some reversals to an ancestral
state (Teotonio & Rose, 2001) or the easier derivation
of some character states from a particular character state
than others, for example, because of the structure of the
genetic network or the nonlinearity of the genotype
phenotype map (Oster & Alberch, 1982; Bradshaw, 1991;
Weill et al., 2004; Wagner, 2011). Alberch (1982) suggested a thought experiment to identify the roles of
natural selection and developmental constraints to
account for the empty spaces of the morphospace. It
consists in removing the selective pressures acting on the
character and studying variation of the character in this
context. If the character remains unchanged, then
constraints are likely to be acting. If it changes, then it
is probable that evolutionary stasis was due to stabilizing
selection (see Amundson, 1994; Raff, 1996: 298 for
comments on this experiment). We used this approach to
test for the action of developmental constraints and
natural selection on the evolution of pollen grains.
Pollen grains are formed through microsporogenesis
and microgametogenesis. Microsporogenesis results in a
tetrad composed of four haploid microspores (future
pollen grains), obtained through the division of a diploid
mother cell by meiosis and the synthesis of intersporal
walls composed of callose. Callose begins to be deposited
on the borders and or at the middle of the plane between
two adjacent microspores and progressively separates
their cytoplasms. As a result, at the late microsporogenesis
stage, some areas of the plane remain unfilled by callose.
We suggest naming these areas last points of callose
deposition (LPCD). The LPCD are variable in number and
position, as three features of microsporogenesis can
change the LPCD pattern (Ressayre et al., 2002a). These
features are the cytokinesis type (two states successive
and simultaneous), the mode of callose deposition (four
states centripetal according to the tetrad, centripetal
according to the cleavage plane, centrifugal according to
the tetrad and centrifugal according to the cleavage plane)

and the form of the tetrad (three states tetrahedral,
rhomboidal and tetragonal). By determining all possible
combinations of variants of these features, a theoretical
morphospace for the LPCD pattern can be elaborated. This
morphospace contains sixteen possible states (tetrahedral
and rhomboidal tetrads cannot be produced through
successive cytokinesis). A subset of this morphospace is
illustrated in Fig. 1 (only LPCD patterns observed in this
study are shown).
All theoretically possible LPCD patterns do not appear
to be observed equally in nature: some have never been
observed, and others dominate large clades. For example,
in early-divergent angiosperms and monocots, microsporogenesis features determining LPCD pattern are quite
variable (Furness & Rudall, 1999a,b; Furness et al., 2002;
Penet et al., 2005; Nadot et al., 2006, 2008; Sannier et al.,
2006). In the eudicot clade, the huge majority of species
have the same LPCD pattern (Wodehouse, 1935;
Blackmore & Crane, 1998; Furness & Rudall, 2004;
Nadot et al., 2008). This LPCD pattern is shown in
Fig. 1c.1. Evolution of the LPCD pattern thus seems to
represent a case of evolutionary stasis in eudicots relative
to its evolution in early-divergent angiosperms and in the
monocot clade. The stasis observed in eudicots could be
caused by developmental constraints arising from the loss
of the capacity to generate variation in the cytokinesis
type, the mode of callose deposition and the tetrad form.
Alternatively, selective pressure could also be involved as
the LPCD pattern may determine the number and
position of the apertures in the pollen wall (Wodehouse,
1935; Heslop-Harrison, 1968; Dover, 1972; Sheldon &
Dickinson, 1983, 1986; Barnes & Blackmore, 1986;
Blackmore & Barnes, 1990; Blackmore & Crane, 1998;
Ressayre et al., 1998, 2002a, 2005; Ressayre, 2001;
Blackmore et al., 2007; Albert et al., 2010). Apertures
are thin areas of the outer pollen wall (the exine),
through which the pollen tube will grow at germination.
They are variable in number and position (e.g. Erdtman,
1952; Kesseler & Harley, 2004; Hesse et al., 2009). The
aperture pattern (i.e. number and position of the apertures) has been found to influence the fitness of pollen
grains (Dajoz et al., 1991, 1993; Till-Bottraud et al.,
1999). Separate mechanisms are involved in the induction and the positioning of apertures. Induction is
achieved through specific cellular mechanisms (reviewed
in Ressayre, 2001) that prevent deposition of the exine
matrix in the areas destined to become apertures (Scott
et al., 2004). In eudicots with equatorial aperture patterns, positioning of the apertures in the pollen wall
depends on the LPCD pattern, as apertures are located in
the areas of the pollen wall adjoining LPCD (Blackmore
et al., 2007). Figure 1 shows the aperture patterns
expected for six different LPCD patterns, according to
Ressayre et al. (2002a). The typical LPCD pattern of
eudicots may result in pollen grains with three equatorial
apertures (Fig. 1c-1), and it may also lead to pollen

Stasis in Euphorbiaceae pollen
1079
Mode of callose deposition

Tetrad form
(a) Tetragonal
4 cleavage planes
View
1
Centripetal according to the
cleavage plane = CpP
2
Centripetal according to the
tetrad = CpT
Whole tetrad
1
4
Section
(b) Rhomboidal
5 cleavage planes
Whole tetrad
1
2
5
4
Section
(c) Tetrahedral
6 cleavage planes
Whole tetrad
5
Section
(planes 4, 5, 6 only)
Fig. 1 Tetrad forms and modes of callose deposition accounting for six different patterns for the number and position of the last points of
callose deposition (LPCD). Only centripetal callose deposits via simultaneous cytokinesis are represented, for tetragonal (a), rhomboidal (b)
and tetrahedral (c) tetrads. These three tetrad configurations differ in the number of cleavage planes of callose that are synthesized to separate
the microspores during microsporogenesis (46 planes, respectively, for tetragonal, rhomboidal and tetrahedral tetrads). Numbering of planes
is arbitrary for tetragonal and rhomboidal tetrads. For tetrahedral tetrads, planes numbered 13 are those that are roughly parallel to the image
plane, and planes numbered 46 are those that are orthogonal to the image plane. For each tetrad form, the formation of the cleavage planes is
illustrated according to the mode of callose deposition. 1 Callose deposition centripetal according to the cleavage plane (abbreviated CpP):
callose begins to be deposited at the border of each plane and progresses towards the planes centre. 2 Callose deposition centripetal according
to the tetrad (abbreviated CpT): callose begins to be deposited at the border of the tetrad and progresses towards the tetrads centre. For
each tetrad form mode of callose deposition combination, the formation of the cleavage planes is illustrated with a view of the whole tetrad
and with a view of a section of the tetrad, to help with the interpretation of photographs in Figs 46. The section represented is in the plane
delimited by the red points (seen as dark grey points in print edition) that figure on the view of the whole tetrad. Callose is represented in blue
(seen as grey surfaces in print edition). Arrows on the view of the section indicate the direction of callose progression. White areas on whole
tetrads are the places where callose was not yet deposited at the stage represented in this scheme. For each LPCD pattern, a pollen grain (grey
sphere) with the corresponding aperture pattern is illustrated. According to Ressayre et al. (2002a), an aperture (coloured black) is expected to
occur in each area of the pollen adjoining the areas where callose is deposited last.
morphs with higher aperture numbers obtained through

rearrangements of microtubule arrays (Ressayre et al.,
2002b). This flexibility in the pollen aperture number
associated with this LPCD pattern has been proposed as a
possible key innovation underlying eudicot success
(Furness & Rudall, 2004). Variation in aperture number
allows pollen to adapt to different environments:
increased aperture number increases the rapidity of
pollen germination but decreases pollen longevity (Dajoz
et al., 1991, 1993; Till-Bottraud et al., 1999). This putative
adaptive flexibility could thus possibly explain the
evolutionary stasis in the LPCD pattern in eudicots.
As pointed out by Mayr (1961), evolutionary changes

are determined by two levels of causation: a proximate
causation (expressed as a set of physiological changes)
and an ultimate one (expressed as changes orchestrated
by natural selection). These two types of causation may
be identified for the LPCD pattern, in which changes are
determined by bottom-up ontogenetic processes (microsporogenesis the proximal cause) and by downstream
selective pressures (aperture pattern the ultimate
cause). As suggested in Alberchs hypothetical experiment (Alberch, 1982), studying variation of the LPCD
pattern in cases where the selective pressures acting on it

1080
have been removed should allow testing for the action of

developmental constraints. In species producing pollen
grains lacking apertures (inaperturate pollen), none of
the cellular mechanisms known to induce apertures are
present (Furness, 2007). Thus, we assume that at least
some of the selective pressures acting on the mechanism
positioning the apertures (i.e. the LPCD) are removed. If
the observed stability of the LPCD pattern among
eudicots is due to developmental constraints, it may be
expected that the LPCD pattern will remain unchanged
despite the loss of selective pressures. Conversely, if the
LPCD pattern exhibits unusual variation in inaperturate
species, this will strongly indicate that its stability was not
due to developmental constraints but to stabilizing
selection.
Inaperturate pollen evolved independently numerous
times throughout the eudicots, although generally in
groups containing a relatively small number of species
(Furness, 2007). In Euphorbiaceae s.s. [sensu stricto is
delimited here as Euphorbiaceae minus the recently
segregated Peraceae (Wurdack & Davis, 2009)], inaperturate taxa fall into two separate groups of differing size:
(i) c. 11 species of tribe Plukenetieae in subfamily
Acalyphoideae (Gillespie, 1994) and (ii) c. 1500 species
in subfamily Crotonoideae [i.e. the clade of inaperturate
crotonoids sensu Wurdack et al. (2005)]. The latter
group is the focus here given its exceptional species
richness. These two groups are widely separated in
phylogenies of Euphorbiaceae (i.e. Wurdack et al.,
2005; Tokuoka, 2007) and represent at least two losses
of apertures. We have made new observations on the
LPCD pattern in Euphorbiaceae s.s. and compared in a
phylogenetic context the variation of the LPCD pattern
among inaperturate and aperturate species. This comparison allows testing for the presence of relative developmental constraints. Under the hypothesis that
developmental constraints are not acting on the LPCD
pattern, we would expect to find a phylogenetic correlation between the loss of the apertures and the appearance of variation in the LPCD pattern. Such a correlation
has been tested here using phylogenetic comparative
methods (Pagel, 1994). This is, to our knowledge, the first
time that the thought experiment of Alberch (1982) has
been conducted empirically.
Material and methods

Assessment of aperture pattern and
microsporogenesis features
Data on aperture pattern were scored from an extensive
survey of the palynological literature for the Euphorbiaceae s.s. (Erdtman, 1952; Punt, 1962; Diaz Zavaleta &
Palacios Chavez, 1980; Thanikaimoni et al., 1984;
Nowicke, 1994; Lobreau-Callen & Suarez-Cervera,
1997; Nowicke et al., 1998; Takahashi et al., 2000;
Suarez-Cervera et al., 2001; Nowicke & Takahashi,
2002; Sagun et al., 2006). For the study of microsporogenesis, a data set of microsporogenesis features was
obtained using information from the literature for 16
species and our own observations for 17 additional
species (Table 1). Data on microsporogenesis were
obtained for aperturate species representing all the major
clades of the family. For inaperturate species, data were
obtained only for species of the inaperturate crotonoids.
Data were not obtained for species of tribe Plukenetieae
(clade A8) because these were not available from the
living collections.
For the description of microsporogenesis, fresh material was obtained from the living collections of botanical
gardens or from plants cultivated in a greenhouse at the
Laboratoire Ecologie, Systematique, Evolution, Orsay,
France (Table 1). Anthers were dissected from unopened
buds of various sizes to find the ontogenetic stages
corresponding to callose deposition and pollen assembled
in mature tetrads. Anthers were squashed on a slide and
mounted with aniline blue solution (aniline blue, 1 g L)1;
K3PO4, 21.2 g L)1; glycerol 187.4 g L)1), according to a
protocol modified from the study of Arens (1949). Aniline
blue staining causes fluorescence of the intersporal
callose walls when illuminated with UV light. Slides
were observed under a Zeiss Axiophot fluorescence
microscope (DAPI filter, excitation at 345 nm, emission
at 425-nm-long pass). For each species, pollen morphology, tetrad form, cytokinesis type and the mode of callose
deposition during intersporal wall formation were recorded. When more than one tetrad configuration was
observed for a given species, a survey of the proportion of
each tetrad form was carried out by counting a number of
tetrads ranging from c. 200 to c. 300 per plant. Chi-square
tests were performed using R (R Development Core Team,
2008) to test (i) whether each tetrad form was produced in
similar proportions within plants and (ii) whether the
distribution of the proportions of tetrad forms was equivalent among the heteromorphic species (i.e. species for
which individuals produce more than one form of tetrad).
Then, a phylogenetic framework was used to study the
evolution of microsporogenesis in relation to aperture
pattern.
Phylogenetic analysis
Phylogenetic relationships of the Euphorbiaceae s.s.
species for which we acquired data on pollen morphology
and microsporogenesis were obtained using DNA
sequences of rbcL and trnL-F genes (Wurdack et al.,
2005; Tokuoka, 2007; Sagun, 2008) (Table S1). Although
molecular data were not available for some of our
investigated species, the resulting species sampling is
sufficient to yield statistically significant results with the
test of Pagel (1994). We used the alignment matrix of the
combined data for rbcL and trnL-F in Wurdack et al.
(2005). Taxon sampling was modified by removing taxa
that lacked morphological data and by adding new taxa

Table 1 Studied specimens, with origin of species for which new

data are provided, and references for data obtained from the
literature.
Species for which new data are provided
Species
Origin (accession number)
Baloghia inophylla (G. Forst.)

P. Green
Bocquillonia spicata Baill.
Museum national dHistoire naturelle,

Paris (# 52146)
Jardin Botanique de la Ville de Paris
(no AN)
Cultivated, seeds from Royal Botanic
Gardens Seed Bank, Kew
(# 351517)
Cultivated, seeds from Rancho
Santa Ana Botanic Garden
(# 20565)
Jardin Botanique de la Ville de
Paris
Caperonia serrata (Turcz.)

C. Presl
Croton californicus Mull. Arg.
Dalechampia roezliana Mull. Arg.

(synonym of Dalechampia
spathulata (Scheidw.) Baill.)
Eremocarpus setigerus (Hook.)
Benth.
Jatropha gossypiifolia L.
Jatropha integerrima Jacq.
Jatropha multifida L.
Jatropha podagrica L.
Hura crepitans L.
Macaranga tanarius (L.) Mull. Arg.
Manihot esculenta Crantz
Mercurialis annua L.
Mercurialis perennis L.
Pedilanthus tithymaloides (L.)
Poit. subsp. tithymaloides
Ricinus communis L.
Cultivated, seeds from Hedgerow

Farms (# CSYOLF1007)
Parc Botanique de Launay (no AN)
Paris (# 18839)
Paris (# 8400)
Royal Botanic Gardens, Kew
(2008-952)
Paris (# 55963)
Paris (no AN)
Paris (no AN)
Royal Botanic Gardens, Kew
(# 1984-2163)
Species for which data were obtained from the literature

Species
Reference
Acalypha alnifolia Klein ex Willd.

Acalypha brachystachya Hornem.
Acalypha ciliata Forssk.
Acalypha indica L. (synonym
of Acalypha ciliata Wall.)
Acalypha lanceolata Willd.
Acalypha malabarica Mull. Arg.
Codiaeum variegatum var. pictum
(Lodd.) Mull. Arg.
Euphorbia corollata L.
Euphorbia dracunculoides Lamk.
Euphorbia dulcis L.
Euphorbia microphylla Heyne.
Euphorbia peltata Roxb.
Euphorbia vermiculata Raf.
Hevea brasiliensis Muell.
Jatropha curcas L.
Micrococca mercurialis Benth.
Kapil (1960)
Kapil (1960)
Mukherjee (1964)
Johri & Kapil (1953)
Thathachar (1952)
Mukherjee (1964)
Albert et al. (2009)
Lyon (1898)
Mukherjee (1960)
Kapil (1961) and Murgia et al. (1986)
Mukherjee (1960)
Mukherjee (1965)
Rao & Devi (1974)
Rao (1964)
Liu et al. (2007)
Rao (1962)
1081
that were available in GenBank (i.e. from Tokuoka,

2007; Sagun, 2008) for which we acquired data on pollen
and microsporogenesis. For the additional sequences,
alignments were easily adjusted by eye using BI O ED I T
version 7.0.5.3 (Hall, 1999), except for some ambiguous
indels in the trnL-F sequences. These were removed
unless informative substitutions were present. Outgroups
were Humiriaceae (Vantanea guianensis Aubl.) and Peraceae (Pera bicolor [Klotzsch] Mull. and Clutia pulchella L.).
They were chosen according to previous studies on the
systematics of Euphorbiaceae s.s. and Malpighiales
(Wurdack et al., 2005; Tokuoka, 2007; Wurdack & Davis,
2009). Rafflesiaceae have inaperturate pollen and have a
phylogenetic placement near the root of Euphorbiaceae
(Wurdack & Davis, 2009), but they lack plastid data and
were not included here (see Discussion).
The phylogenetic relationships of species were inferred
under the assumption of maximum likelihood (ML)
using PhyML 3.0 (Guindon & Gascuel, 2003; Guindon
et al., 2010). The most appropriate model of nucleotide
substitution was determined with MO D E L T E S T version
3.7 (Posada & Crandall, 1998) to be the general timereversible model (GTR + I + G). Parameter values were
estimated during running, and node support was estimated with a ML bootstrap analysis with 100 replicates.
The substitution rate matrix was 1.37049, 2.51808,
0.43019, 0.82579 and 3.43486. Nucleotide frequencies
were (A) = 0.30530; (C) = 0.18853; (G) = 0.22164; and
(T) = 0.28453. The estimated proportion of invariable
sites was 0.469, and a gamma distribution was assumed
for rates at invariable sites. The presented tree was
obtained from a single run. A BioNJ starting tree was
used, and tree topologies were estimated using the
nearest neighbour interchange (NNI) method.
Analysis of character evolution
Pollen and microsporogenesis characters obtained from
species sampled by us and from published information
(summarized in Table 2) were optimized onto the most
likely tree with relative branch lengths obtained from our
ML analyses, using Mesquite (Maddison & Maddison,
2008). Characters were optimized using maximum parsimony (MP), and maximum likelihood using the Mk1
model which assumes that any particular change (from
state 0 to 1 or state 3 to 2, for example) is equally
probable. Maximum likelihood complements the qualitative result assessed with MP optimization, as it takes
into account branch lengths and provides an estimate of
the probability of the character state at each node. In
addition, ML reconstruction is more conservative than
MP in reconstructing ancestral states at nodes for which
few data are available.
Three character states were described for the aperture
pattern: apertures equatorial, global or absent. Microsporogenesis features were coded as binary characters as
follows: (i) cytokinesis: simultaneous or heteromorphic

1082
Table 2 Records of tetrad form, mode of callose deposition and cytokinesis type in species of Euphorbiaceae s.s. investigated, according to their
aperture pattern.
Species
Aperture
pattern
Tetrad form (percentage of each)
Acalypha alnifolia
Acalypha brachystachya
Acalypha ciliata
Acalypha indica
Acalypha lanceolata
Acalypha malabaria
Bocquillonia spicata
Caperonia serrata
Dalechampia roezliana
Euphorbia corollata
Euphorbia dracunculoides
Euphorbia dulcis
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Td
Td
Td
Td
Td
Td
Td
Td
Td
Td
Td
Td
Euphorbia microphylla
Euphorbia peltata
Euphorbia vermiculata
Hevea brasiliensis
Hura crepitans
Macaranga tanarius
Mercurialis annua
Mercurialis perennis
Micrococca mercurialis
Pedilanthus tithymaloides
subsp. tithymaloides
Ricinus communis
Baloghia inophylla
Codiaeum variegatum
Croton californicus
Eremocarpus setigerus
Jatropha curcas
Jatropha gossypiifolia
Jatropha integerrima
Jatropha multifida
Jatropha podagrica
Manihot esculenta
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Equatorial
Td ) Tg*
Td
Td
Td
Td
Td
Td
Td
Td ) Dc*
Td
Equatorial
Inaperturate
Inaperturate
Inaperturate
Inaperturate
Inaperturate
Inaperturate
Inaperturate
Inaperturate
Inaperturate
Global
Td
Td
Td
Td
Td
Td
Td
Td
Td
Td
Td
) Tg ) Dc*
) Tg ) Dc*
) Tg ) Dc*
) Tg*
) Tg*
) Tg ) Dc
(70 %) ) Rh (25 %)
) Rh ) Tg
(77 %) ) Rh (19 %)
(51 %) ) Rh (29 %)
) Rh
(89 %) ) Rh (10 %)
(61 %) ) Rh (32 %)
(72 %) ) Rh (24 %)
(83 %) ) Rh (15 %)
(80%) ) Rh (20%)
) Tg (5 %)
) Tg (4 %)
) Tg (19 %)
)
)
)
)
Tg
Tg
Tg
Tg
(1
(7
(4
(2
%)
%)
%)
%)
Callose
deposition
Cytokinesis
Reference Figure
?
?
CpP
Centripetal
Centripetal
Centripetal
CpP
CpP
CpP
?
Centripetal
?
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
?
Centripetal
?
?
Centripetal
CpP
CpP
CpP
CpP
?
CpP
Simultaneous
?
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Kapil (1960)
Kapil (1960)
Mukherjee (1964)
Johri & Kapil (1953)
Thathachar (1952)
Mukherjee (1964)
Fig. 4
Fig. 4
Fig. 4
Lyon (1898)
Mukherjee (1960)
Kapil (1961); Murgia
et al. (1986)
Mukherjee (1960)
Mukherjee (1965)
Rao & Devi (1974)
Rao (1964)
Fig. 4
Fig. 4
Fig. 4
Fig. 4
Rao (1962)
Fig. 4
CpP
CpP CpT
CpP CpT CfT
CpP
CpP CpT
Centripetal
CpP CpT
CpP CpT
CpP
CpP CpT
CpP CpT
Simultaneous
Simultaneous
Heteromorphic
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Simultaneous
Fig. 4
Fig. 6
Albert et al. (2009)
Fig. 6
Fig. 6
Liu et al. (2007)
Fig. 5
Fig. 5
Fig. 5
Fig. 5
Fig. 6
Td, tetrahedral; Tg, tetragonal; Rh, rhomboidal; Dc, decussate; CpP, centripetal according to the cleavage plane; CpT, centripetal according
to the tetrad; CfT, centrifugal according to the tetrad. * denotes the literature reports on tetrads that have been reinterpreted (see Results).
? denotes missing data.
(simultaneous and successive no taxa were monomorphic for successive); (ii) mode of callose deposition:
centripetal according to the cleavage plane or heteromorphic (centripetal according to the cleavage plane and
centripetal according to the tetrad); and (iii) tetrad form:
tetrahedral or heteromorphic (tetrahedral, rhomboidal
and tetragonal). The resulting data matrix is shown in
Table S1.
Correlation between aperture pattern and
microsporogenesis
To test for correlated evolution between aperture pattern
and microsporogenesis, the aperture pattern was binarycoded (apertures equatorial vs. global absent), as correlation tests can only be performed on binary traits. The
coding of microsporogenesis was the same as that used

for character optimization. This coding is suitable for our
aim of testing whether there is variation in microsporogenesis once aperture pattern is not equatorial.
The test for correlated evolution was carried out using
the BA Y E S DI S C R E T E method of the BA Y E S TR A I T S computer package (Pagel, 1994; Pagel & Meade, 2006).
BA Y E S DI S C R E T E analyses correlated evolution among
pairs of binary traits. The result of the test does not rely
upon any reconstruction of the ancestral character states
(Pagel, 1994). The method consists in comparing two
models of evolution: one in which the traits evolve
independently and one that assumes correlated evolution. Evidence of correlated evolution is found when the
model assuming correlated evolution fits better than the
independent model. The two methods of analysis

implemented in BA Y E S DI S C R E T E [ML and Markov chain

Monte Carlo (MCMC)] were used. In the ML method, the
program estimates the likelihood of the two models. The
difference between them is tested using a likelihood ratio
statistic [LRS = 2(log-likelihood(dependent model) )
log-likelihood(independent model))], which is nominally
distributed as a v2 (4 d.f.). The MCMC method estimates
the Bayesian posterior distributions of the likelihoods
of the data given the model (dependent or independent).
The log-harmonic mean of the likelihoods is compared
between two runs (one assuming correlated evolution and
the other independent evolution). Relative fit of the two
models is tested using the Bayes factor [BF = 2(log
[harmonic mean(better model)] ) log[harmonic mean
(worse model)])]. Any positive value of the BF favours
the model assuming correlated evolution, but conventionally a ratio > 2 is taken as positive evidence, > 5 is
strong and > 10 is very strong evidence. Analyses were
run with the default values of the software.
To account for phylogenetic uncertainty and branch
length heterogeneity in the analyses, LRS and BF were
estimated from a sample of 100 tree topologies obtained
from a Bayesian analysis. This sample was obtained by
analysing our alignment matrix with Bayesian inference
using MR BA Y E S version 3.1.2 (Huelsenbeck & Ronquist,
2001; Ronquist & Huelsenbeck, 2003). The evolutionary
model was set to the GTR + I + G model with gammadistributed rate variation across sites and an estimated
proportion of invariable sites. Analyses were performed
as two independent runs of two million generations, with
sampling every 100 generations. Likelihood values
reached a plateau within 500 000 generations. These
were deleted as burn-in. The tree files of the two runs
were combined in a single file from which 100 trees were
randomly sampled.
Results
Phylogenetic analysis
Maximum-likelihood analyses resulted in a single tree
with a score of log-likelihood = )19082.20 (Fig. 2). The
phylogeny obtained with our analysis recognized all
the subclades (Adenoclineae s.l.; Articulated crotonoids;
C12; Pseudanthial; H12; Alchorneoids; A18) found in
the two previous studies on Euphorbiaceae s.s. (Wurdack
et al., 2005; Tokuoka, 2007). The three major clades in
the previous studies (Crotonoideae s.l.; Euphorbioideae;
and Acalyphoideae s.s.) also coincided with those of our
analysis except that our analysis suggests the monophyly
of the Crotonoideae s.l., whereas the other two exclude
the Adenoclineae from this group. Crotonogynopsis groups
with the Alchorneoids instead of expected clade A3 (i.e.
Wurdack et al., 2005; Tokuoka, 2007), but this difference
does not affect our conclusions. A substantial difference is
in the relationships between the major clades. The
Wurdack et al. (2005) analysis grouped Euphorbioideae
1083
with Crotonoideae s.l., whereas both Tokuokas study

and the current study group Euphorbioideae with
Acalyphoideae s.s. This difference may be due to the
species sampling, which is higher in the Wurdack et al.s
(2005) analysis relative to our own and to that of
Tokuoka (2007). Regardless, relationships between these
clades are poorly supported in all the analyses. Changes
in the relationships between these clades do not affect
the results of our comparative analysis.
Optimization of the pollen aperture pattern
Our survey of the palynological literature provided data
on aperture pattern for 779 Euphorbiaceae species
representing 256 genera. Of these, only species for which
phylogenetic data were available were utilized [101
species, representing 91 of the 152 genera included in
the phylogenetic analysis of Euphorbiaceae s.s. by
Wurdack et al. (2005)]. This sorting considerably reduced
the sampling, but it allowed us to work in a phylogenetic
framework. Although fewer species were included in this
study, the sampling is an adequate representation of the
full phylogenetic tree of Wurdack et al.s (2005) study, as
it contains taxa representing all the terminal clades in
their analysis, except for the small Erismantheae clade.
Of the 101 species in the analysis, the majority (73)
produce pollen with equatorial apertures (usually three
apertures, but sometimes more). To determine whether
the predominance of this character state is due to
phylogenetic conservatism or to homoplasy, the aperture
pattern was optimized onto our phylogeny (Fig. 3a). The
two methods used for the optimization of the aperture
pattern (MP and ML) were congruent: all the character
states inferred at each node in MP were statistically
supported in ML using the likelihood decision threshold
of two [i.e. a state was preferred when its log-likelihood
differed by two units or more from the one of the other
states, as suggested by Pagel (1999)], except for two
nodes at the base of Crotonoideae s.l. where there was
disagreement between the two methods. Figure 3a
shows that aperture pattern is plesiomorphically equatorial and underwent two (maybe three) independent
transitions across the tree. In the clade formed by
inaperturate crotonoids + articulated crotonoids + Gelonieae, parsimony optimization indicates that the aperture
pattern changed to a global pattern and subsequently to
an inaperturate one in the inaperturate crotonoids. In
clade A8, it was transformed directly to an inaperturate
pattern. One reversal to the equatorial pattern was
inferred, in the articulated crotonoids.
Description of microsporogenesis in Euphorbiaceae
Our observations (summarized in Table 2) reveal that the
species of Euphorbiaceae studied that produce pollen
with an equatorial aperture pattern differ in aspects of
microsporogenesis from those species that have an

77
56
100
73
87
67
100
96
80
98
100
99
98
50
66
76
99
98
100
100
72
52
100
Strophioblachia fimbricalyx
Codiaeum variegatum
Sagotia racemosa
Sandwithia guyanensis
Joannesia princeps
Jatropha curcas
Jatropha gossypiifolia
Jatropha podagrica
Beyeria leschenaultii
Baloghia inophylla
Tannodia cordifolia
Domohinea perrieri
Ostodes paniculata
Dodecastigma amazonicum
Ricinodendron heudelotii
Baliosperm motanum
Plagiostyles africana
Nealchornea yapurensis
Dichostemma glaucescens
Neoguillauminia cleopatra
100
Aleurites moluccana
Garcia nutans
Cavacoa aurea
53
99
Outgroups
Cheilosoideae
Endospermum moluccanum
Omphalea diandra
Ditta myricoides
Suregada africana
Elateriospermum tapos
Glycydendron amazonicum
Hevea brasiliensis
97
Manihot esculenta
Manihot grahami
100
Clutia pulchella
0.03
C2
C1
84
Acalypha lanceolata
Acalypha ciliata
Tragia fallax
A3
A2
A4
A8
A7
A6
A5
A1
Pseudanthial
Inaperturate crotonoids
Articulated crotonoids
Gelonieae
Adenoclineae s.l.
Euphorbia pulcherrima
Croton californicus
Maprounea guianensis
Homalanthus populneus
100
H1
Gymnanthes lucida
64
Dalembertia poplifolia
94
Pachystroma longifolium
100
Hura crepitans
Colliguaja integerrima
66
H2
Hippomane mancinella
100
99
Grimmeodendron eglandulosum
Bonania cubana
Pseudagrostistachys ugandensis
Cyttaranthus congolensis
84
100
Discoglypremma caloneura
Amyrea sambiranensis
Mareyopsis longifolia
81 70
Alchorneopsis floribunda
Alchorneoids
100 Gavarretia terminalis
Crotonogynopsis usambarica
100
Aparisthmium cordatum
85
100
Alchornea laxiflora
Alchornea cordifolia
100
Macaranga tanarius
98
Trewia nudiflora
Mallotus japonicus
100
Seidelia triandra
Leidesia procumbens
68
Chiropetalum tricoccum
99
Caperonia palustris
Philyra brasiliensis
100
100
Lasiocroton bahamensis
Adelia ricinella
98
Caryodendron orinocense
Adenophaedra grandifolia
83
Plukenetia volubilis
73
82
Dalechampia spathulata
Astrococcus cornutus
Cnesmone javanica
70
97
Platygyna hexandra
Monotaxis grandiflora
Adriana tomentosa
Ricinus communis
Discocleidion rufescens
94
Koilodepas bantamense
Cephalocroton mollis
Chrozophora tinctoria
99
Sumbaviopsis albicans
Melanolepis multiglandulosa
100
Lobanilia bakeriana
100
69
Micrococca capensis
Claoxylon australe
Homonoia riparia
100
Pycnocoma macrophylla
Argomuellera macrophylla
Mareya micrantha
82
100
91
63
86
96
63
55
76
58
87
84
71
100
Pera bicolor
Neoscortechinia kingii
Klaineanthus gaboniae
Acalyphoideae s.s.
Euphorbioideae
Crotonoideae s.l.
Fig. 2 Relationships of the species of Euphorbiaceae investigated in this study. The tree was obtained from a combined analysis of the markers rbcL and trnL-F using a maximum-likelihood
reconstruction method (log-likelihood = )19 082.20). ML-estimated branch lengths are shown. Numbers next to branches are the maximum-likelihood bootstrap values (only values
> 50 are shown). The names of major clades follow the study of Wurdack et al. (2005).
Euphorbiaceae s.s.
88
Vantanea guianensis
1084

(a)
(b)
Gelonieae
Crotonoideae s.l.
1085
Articulated crotonoids
Euphorbiaceae s.s.
Inaperturate
crotonoids
Euphorbioideae
Vantanea guianensis
Pera bicolor
Clutia pulchella
Neoscortechinia kingii
Klaineanthus gaboniae
Endospermum moluccanum
Omphalea diandra
Ditta myricoides
Suregada africana
Elateriospermum tapos
Glycydendron amazonicum
Hevea brasiliensis
Manihot esculenta
Manihot grahami
Croton californicus
Sagotia racemosa
Sandwithia guyanensis
Joannesia princeps
Jatropha curcas
Jatropha gossypifolia
Jatropha podagrica
Garcia nutans
Cavacoa aurea
Strophioblachia fimbricalyx
Codiaeum variegatum
Aleurites moluccana
Beyeria leschenaultii
Baloghia inophylla
Tannodia cordifolia
Domohinea perrieri
Ostodes paniculata
Dodecastigma amazonicum
Ricinodendron heudelotii
Baliospermum motanum
Plagiostyles africana
Nealchornea yapurensis
Dichostemma glaucescens
Neoguillauminia cleopatra
Euphorbia pulcherrima
Maprounea guianensis
Homalanthus populneus
Gymnanthes lucida
Dalembertia populifolia
Pachystroma longifolium
Hura crepitans
Colliguaja integerrima
Hippomane mancinella
Grimmeodendron eglandulosum
Bonania cubana
Pseudagrostistachys ugandensis
Acalyphoideae s.s.
A8
Aperture pattern
Max. parsimony
Max. likelihood
Cyttaranthus congolensis
Discoglypremma caloneura
Amyrea sambiranensis
Mareyopsis longifolia
Alchorneopsis floribunda
Gavarretia terminalis
Crotonogynopsis usambarica
Aparisthmium cordatum
Alchornea laxiflora
Alchornea cordifolia
Macaranga tanarius
Trewia nudiflora
Mallotus japonicus
Seidelia triandra
Leidesia procumbens
Chiropetalum tricoccum
Caperonia palustris
Philyra brasiliensis
Lasiocroton bahamensis
Adelia ricinella
Caryodendron orinocense
Adenophaedra grandifolia
Plukenetia volubilis
Dalechampia spathulata
Astrococcus cornutus
Cnesmone javanica
Tragia fallax
Platygyna hexandra
Monotaxis grandiflora
Adriana tomentosa
Ricinus communis
Discocleidion rufescens
Koilodepas bantamense
Cephalocroton mollis
Chrozophora tinctoria
Sumbaviopsis albicans
Melanolepis multiglandulosa
Lobanilia bakeriana
Micrococca capensis
Claoxylon australe
Homonoia riparia
Pycnocoma macrophylla
Argomuellera macrophylla
Mareya micrantha
Acalypha lanceolata
Acalypha ciliata
Tetrad form
Max. parsimony
Max. likelihood
Equatorial
Tetrahedral
Global
Heteromorphic
Inaperturate
Fig. 3 Mirror tree of Euphorbiaceae s.s. comparing optimizations of the aperture pattern (a) and tetrad form (b). Colour of the boxes near
species names indicates the state of the corresponding character (box absence indicates data absence). Branch colours indicate the ancestral
states obtained using MP optimization. Pie charts on the nodes indicate the ancestral states obtained using ML. Pie chart colours represent the
probability that this node is at a particular state. Analyses were conducted on the phylogram shown in Fig. 2 but are shown here on a
cladogram for clarity. The names of major clades follow the study of Wurdack et al. (2005).

1086
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
(l)
(m)
(n)
(o)
inaperturate or a global pattern: all the species with

equatorial apertures share the same microsporogenesis
character states (tetrahedral tetrads, CpP mode of callose
deposition and simultaneous cytokinesis). All the species
with inaperturate pollen or with global apertures show
intraindividual variation (i.e. heteromorphism) in one or
two microsporogenesis features: tetrad form and mode of
Fig. 4 Pollen morphology and microsporogenesis for nine Euphorbiaceae species producing pollen with equatorial apertures. (a)
(c) Dalechampia roezliana. (a) Triaperturate
pollen (equatorial section). (b) Tetrahedral
tetrad with apertures arranged in Fischers
configuration (i.e. apertures are grouped in
pairs in the tetrad arrows). (c) Planes 13 of
a tetrahedral tetrad during callose wall formation. Callose deposition follows the CpP
mode (Fig. 1c-1, whole tetrad). Callose (ca)
begins to be deposited at the border of each
plane and progresses towards the centre of
the plane (asterisks). At the stage represented here, the centre of each plane remains
empty of callose. (d)(f) Ricinus communis. (d)
Triaperturate pollen (polar view). (e) Tetrahedral tetrad (same configuration as in b). (f)
Planes 46 of a tetrahedral tetrad during callose wall formation. Callose deposition follows the CpP mode (Fig. 1c-1, section).
Callose begins to be deposited at the border
of each plane (arrows the central border is
common to the three planes) and progresses
towards the centre of the plane. (g)(i) Pedilanthus tithymaloides subsp. tithymaloides. (g)
Triaperturate pollen (longitudinal section).
(h) Tetrahedral tetrad (same configuration as
in b). (i) Planes 13 of a tetrahedral tetrad
during callose wall formation. The callose
deposition follows the CpP mode (same
configuration as in c). (j)(l) Mercurialis
annua. (j) Triaperturate pollen (equatorial
section). (k) Tetrahedral tetrad (same configuration as in b). (l) Planes 46 of a
tetrahedral tetrad during callose wall formation (same configuration as in f). (m)(o)
Mercurialis perennis. (m) Triaperturate grain
(polar view). (n) Tetrahedral tetrad. Apertures are grouped in pairs (arrows) in accordance with Fischers configuration.
(o) Planes 46 of a tetrahedral tetrad during
callose wall formation (same configuration as
in f). (p)(r) Bocquillonia spicata. (p) Triaperturate pollen (equatorial section). (q) Tetrahedral tetrad with apertures (asterisk) in
Fishers configuration. (r) Planes 46 of a
tetrahedral tetrad during callose wall formation (same configuration as in f). (s)(u)
Hura crepitans. (s) Triaperturate pollen
(equatorial section). (t) Tetrahedral tetrad.
Apertures are grouped in pairs (arrows) in
callose deposition. Detailed microsporgenesis sequences

for each species are given below.
Species with equatorial apertures

All the species investigated with an equatorial aperture
pattern produce pollen grains with three equatorial
apertures (Fig. 4a, d, g, j, m, p, s, v), except Caperonia

Fig. 4 (Continued)
accordance with Fischers configuration.
(u) Planes 46 of a tetrahedral tetrad during
callose wall formation (same configuration
as in f). (v)(x) Macaranga tanarius. (v)
Triaperturate pollen (polar view). (w) Tetrahedral tetrad. (x) Planes 46 of a tetrahedral
tetrad during callose wall formation (same
configuration as in f). (y)(a) Caperonia serrata. (y) Hexa-aperturate grain (polar view).
(z) Tetrahedral tetrad. Apertures are grouped
in pairs (arrows) in accordance with Fischers
configuration. (a) Planes 46 of a tetrahedral
tetrad during callose wall formation (same
configuration as in f). All photographs are
taken using fluorescence microscopy except
(e) and (j), which are with light microscopy.
Scale bars 10 lm unless mentioned otherwise.
(p)
(q)
(r)
5 m
5 m
5 m
(s)
(t)
(u)
(v)
(w)
(x)
5 m
(y)
serrata, which produces pollen with six apertures

(Fig. 4y). All these species produce pollen grains that
result from tetrahedral tetrads in which apertures are
arranged in Fischers configuration (i.e. apertures of
neighbour grains of the tetrad are grouped in pairs
(Fischer, 1889); Fig. 4b, e, h, k, n, q, t and z). Arrangement of the apertures in the tetrad could not be
determined in Macaranga tanarius because in this species
the microspores are released from the tetrad very early.
Tetrads are formed by the simultaneous synthesis of six
intersporal cleavage planes (Fig. 4c, f, i, l, o, r, u, x and a).
Cytokinesis is thus of the simultaneous type. During
formation of the intersporal cleavage planes, callose
deposition begins at the borders of each of the six
cleavage planes, and deposition progresses centripetally
towards the centre of each plane (Fig. 4c, f, i, l, o, r, u, x
and a). Callose is thus always deposited according to the
CpP mode (Fig. 1c-1).
1087
5 m
5 m
(z)
( )
Species with inaperturate or global aperture patterns

The four species of Jatropha examined (Jatropha podagrica (Fig. 5ad), Jatropha integerrima (Fig. 5eh), Jatropha
gossypiifolia (Fig. 5il) and Jatropha multifida Fig. 5mp)
produce inaperturate pollen (Fig. 5a, e, i, m). Several
types of tetrad are involved in pollen production in this
genus: these four species produce, within the same
stamen, tetragonal, rhomboidal and tetrahedral tetrads,
as shown in Fig. 5 (b, f, j, n: tetragonal tetrads; c, g, k, o:
rhomboidal tetrads; d, h, l, p: tetrahedral tetrads).
Assessment of the proportion of each tetrad type
(Table 2) revealed that the four species produce a
majority of tetrahedral tetrads (6189%), a relatively
high number of rhomboidal tetrads (1032%) and a
few tetragonal ones (17%). Variation in the mode of
callose deposition was observed for J. podagrica, J. integerrima and J. gossypiifolia. In these species, both CpP
(Fig. 5c, f, j) and CpT (Fig. 5d, h, l) modes of callose

1088
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
(l)
(m)
(n)
(o)
(p)
Fig. 5 Pollen morphology and microsporogenesis of four inaperturate Jatropha species. (a)(d) Jatropha podagrica. (a) Inaperturate pollen.
(b) Mature tetragonal tetrad. (c) Rhomboidal tetrad during callose wall formation. Callose deposition follows the CpP mode (Fig. 1b-1, whole
tetrad). Callose begins to be deposited at the border of each plane and progresses towards the centre of the plane. (d) Planes 13 of a tetrahedral
tetrad during callose wall formation. Callose deposition follows the CpT mode (Fig. 1c-2, whole tetrad). Callose begins to be deposited at the
border of the tetrad and progresses towards the centre of the tetrad. Note the absence of callose at the centre of the tetrad (arrow), allowing us
to distinguish this configuration from the one where callose deposition is centripetal according to the cleavage planes. (e)(h) Jatropha
integerrima. (e) Inaperturate pollen. (f) Tetragonal tetrad during callose wall formation. Callose deposition follows the CpP mode (Fig. 1a-1,
section). Callose begins to be deposited at the border of each plane (arrows the central border is common to the four planes) and progresses
towards the centre of the plane. At this stage, the centre of each plane (asterisks) remains empty of callose. (g) Mature rhomboidal tetrad. (h)
Planes 46 of a tetrahedral tetrad during callose wall formation. Callose deposition follows the CpT mode (Fig. 1c-2, section). Callose begins to
be deposited at the border of the tetrad and progresses towards the centre of the tetrad (arrow). (i)(l) Jatropha gossypiifolia. (i) Inaperturate
pollen. (j) Tetragonal tetrad during callose wall formation. Callose deposition follows the CpP mode (same configuration as in f). (k) Mature
rhomboidal tetrad. (l) Planes 46 of a tetrahedral tetrad during callose wall formation. Callose deposition follows the CpT mode (same
configuration as in h). The arrow indicates the centre of the tetrad, which remains empty of callose at this stage. (m)(p) Jatropha multifida.
(m) Inaperturate pollen. (n) Mature tetragonal tetrad. (o) Rhomboidal tetrad during callose wall formation. (p) Tetrahedral tetrad during CpP
callose wall formation (Fig. 1c-1, section). All photographs are taken using fluorescence microscopy. Scale bars 10 lm.
deposition were found. This variation was observed

within the same stamen for J. podagrica and J. integerrima, whereas it was observed at the scale of the plant
for J. gossypiifolia. Callose deposition in J. multifida was
found to follow the CpP mode (Fig. 5p). As in the case

of the triaperturate species, cytokinesis type is simultaneous in all the Jatropha species investigated (Fig. 5c, d,
f, h, j, l, o, p).

(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
(l)
(m)
(n)
(o)
(p)
1089
Fig. 6 Pollen morphology and microsporogenesis for three inaperturate and one panto-aperturate species of Euphorbiaceae. (a)(d) Croton
californicus. (a) Inaperturate pollen. (b) Mature tetragonal tetrad. (c) Rhomboidal tetrad during callose wall formation. Callose deposition
follows the CpP mode (Fig. 1b-1, section). Callose begins to be deposited at the border of each plane (arrows) and progresses towards the centre
of each plane. (d) Planes 46 of a tetrahedral tetrad during callose wall formation. Callose deposition follows the CpP mode (Fig. 1c-1, section).
(e)(h) Baloghia inophylla. (e) Inaperturate pollen. (f) Mature tetragonal tetrad. (g) Rhomboidal tetrad during callose wall formation. Callose
deposition follows the CpT mode (Fig. 1b-2, whole tetrad). Callose begins to be deposited at the borders of the tetrad (arrows) and progresses
towards the centre of the tetrad (asterisk). At the stage represented here, the centre of the tetrad remains empty of callose. (h) Planes 13 of a
tetrahedral tetrad during callose wall formation. Callose deposition follows the CpP mode (Fig. 1c-1, whole tetrad). Callose (c) begins to be
deposited at the border of each plane and progresses towards the centre of the plane (asterisks). At the stage represented here, the centre
of each plane remains empty of callose. (i)(l) Eremocarpus setigerus. (i) Inaperturate pollen. (j) Tetragonal tetrad during callose wall formation.
Callose deposition follows the CpP mode (Fig. 1a-1, whole tetrad). Only two of the four cleavage plane formations are visible on this view
(the two other planes are orthogonal to the picture plane and thus callose deposition cannot be observed). Callose (c) begins to be deposited at
border of the cleavage plane and progresses towards the centre of each plane (asterisk). (k) Rhomboidal tetrad during callose wall formation
stage. Callose deposition follows the CpT mode (Fig. 1b-2, whole tetrad). (l) Planes 46 of a tetrahedral tetrad during callose wall formation.
Callose deposition follows the CpP mode (Fig. 1c-1, section). (m)(p) Manihot esculenta. (m) Panto-aperturate pollen. (n) Rhomboidal tetrad
during callose wall formation (same configuration as in c). (o) Planes 1 and 2 of a tetrahedral tetrad during callose wall formation (same
configuration as in h). (p) Planes 46 of a tetrahedral tetrad during callose wall formation. Callose deposition follows the CpT mode (Fig. 1c-2,
section). Callose begins to be deposited at the border of the tetrad and progresses towards the centre of the tetrad (arrow). All photographs
are taken using fluorescence microscopy except (a), which is with light microscopy. Scale bars 10 lm.

1090
Microsporogenesis in Croton californicus (Fig. 6ad),

Baloghia inophylla (Fig. 6eh) and Eremocarpus setigerus
(Fig. 6il) is similar to that in Jatropha. These species also
produce inaperturate pollen (Fig. 6a, e, i). Again, these
species produce, within the same stamen, three tetrad
configurations, as shown in Fig. 6 (b, f, j: tetragonal
tetrads; c, g, k: rhomboidal tetrads; d, h, l: tetrahedral
tetrads). As in Jatropha, the different forms of tetrad were
produced in unequal proportions [tetrahedral tetrads
(5177%), rhomboidal tetrads (1925%) and tetragonal
ones (419%)]. Callose deposition in C. californicus
follows the CpP mode exclusively (Fig. 6cd). In B.
inophylla and E. setigerus, both CpT (Fig. 6g, k) and CpP
(Fig. 6h, l) modes of callose deposition were observed.
This variation was found within the tetrads produced by
a single flower. Cytokinesis is simultaneous (Fig. 6c, d, g,
h, k, l).
Manihot esculenta (Fig. 6mp) produces pollen with
panto-aperturate pollen (Fig. 6m). Variation was found
in tetrad form, as this species produces tetrahedral
(Fig. 6op) and rhomboidal (Fig. 6n) tetrads, in unequal
proportions (80% and 20%, respectively). Tetragonal
tetrads were not observed. Callose cleavage planes were
synthetized through CpP (Fig. 6no) and CpT (Fig. 6p)
modes. Cytokinesis is simultaneous (Fig. 6np).
A regular pattern was found for variation in tetrad
form: all species with an inaperturate or global aperture
pattern produced a majority of tetrahedral tetrads.
Within each species, the distribution of the proportion
of each type of tetrad presents a significant departure
from 1 3 (36 < v22 < 296; P < 10)7) (Table S2). This
variation was not homogeneous at the interspecific level
as the species studied differ significantly from each other
in the distribution of their respective tetrad-form proportions (v214 = 198.8; P < 2.10)16).
Optimization of microsporogenesis
Our observations coupled with data recorded in the
literature were used to optimize microsporogenesis
features on our tree. Prior to optimization, some descriptions recorded from literature were reinterpreted in the
light of current knowledge about relations between
aperture pattern and microsporogenesis. In particular,
some studies report the presence in species with an
equatorial aperture pattern of tetragonal and decussate
tetrads in addition to tetrahedral ones (Table 2). We
suspect that the description of tetragonal and decussate
tetrads probably results from observational artefacts
and or terminological confusion. A decussate tetrad is a
tetragonal tetrad for which one cleavage plane underwent
a 90 rotation (Fig. S1). Once they are mature, decussate
tetrads may be indistinguishable from tetrahedral ones.
Indeed, in these two types of tetrad, microspores are
arranged in a very similar way (Fig. S1). The only way to
distinguish between them is to look at cytokinesis and
callose deposition. Tetrahedral tetrads are formed by
simultaneous cytokinesis through the synthesis of six

cleavage planes separating the microspores (Figs 1 and S1).
Decussate tetrads can only be obtained from tetragonal
tetrads produced by successive cytokinesis through the
synthesis of three cleavage planes (Fig. S1; Blackmore
et al., 2007; Nadot et al., 2008). Tetragonal tetrads can also
be produced by simultaneously synthesizing four cleavage planes, but this configuration cannot result in a
decussate tetrad. In the literature, cytokinesis and or the
callose deposition stage was described for all the species
supposed to produce decussate and or tetragonal tetrads
(except for Euphorbia dulcis). Cytokinesis was simultaneous and occurred through the formation of six cleavage
planes in all the species for which data were available
(Table 2). Consequently, the occurrence of decussate
tetrads in these species is not plausible because the
authors who described decussate tetrads never reported
the dyad stage that characterizes successive cytokinesis
(Table 2). The occurrence of tetragonal tetrads is also
questionable because callose deposition stages with
three or four cleavage planes were never described.
Based on the fact that cytokinesis was described as
simultaneous in all species, we conclude that the
reports of decussate and tetragonal tetrads in these
species, in addition to tetrahedral, are incorrect.
We consequently coded the character as tetrahedral
in these species. Callose deposition was always described as centripetal, without a precise description of
the mode (except for Acalypha ciliata, which is CpP).
The resulting data matrix is shown in Table S1.
Figure 3b shows the optimization of tetrad form on
our tree using ML and MP. The optimization conducted
by ML recognizes that variation in tetrad configuration
may have appeared at the base of the clade formed by
inaperturate and articulated crotonoids. MP reconstruction fails to identify the point at which variation may
have appeared, due to the lack of data for the more
basal clades. However, optimization of the tetrad form
on the Wurdack et al.s (2005) topology (in which the
Acalyphoideae s.s. clade is external relative to the
Crotonoideae s.l. and Euphorbioideae) using ML and
MP methods indicates unambiguously that the ancestral
state is tetrahedral and that variation may have
appeared at the base of the clade formed by inaperturate
and articulated crotonoids (Fig. S4). In the articulated crotonoid clade, the panto-aperturate M. esculenta
exhibits variation in tetrad form but, as the related
triaperturate Hevea brasiliensis is tetrahedral, the
ancestral state for the clade could not be resolved
unambiguously. Both MP and ML analyses recognize
the equatorial clade (Acalyphoideae s.s. + Euphorbioideae) as being plesiomorphically tetrahedral, but
this result needs to be treated with caution due to the
large amount of missing data.
Results are similar for the mode of callose deposition
(Figs S2 and S5). Both reconstruction methods suggest
that the CpP mode of callose deposition is fixed for the

equatorial clade and that variation may have appeared

at the base of the clade formed by inaperturate and
articulated crotonoids. Again, inferences regarding the
most basal nodes of the tree must be treated with
caution due to the large amount of missing data.
Finally, cytokinesis is exclusively simultaneous for all
taxa, except for the inaperturate Codiaeum variegatum,
which has simultaneous and successive cytokinesis
(Fig. S3).
Test for correlated evolution
The tests for correlated evolution between traits indicate
correlated changes in two of the three pairs of traits
tested. LRS were relatively high for the pairs of characters
aperture pattern tetrad form (LRS4 = 17.16; P = 0.0018)
and aperture pattern callose deposition (LRS4 = 11.5;
P = 0.021) relative to that for the pair aperture
pattern cytokinesis (LRS4 = 2.01; P = 0.73). LRS were
significant for the first two pairs, indicating that in these,
the model for which the two characters are allowed to
evolve independently is a worse fit for the data than the
model in which the characters evolve in a correlated
fashion. Similar results were obtained with the MCMC
method. Bayes factor values were 15.3, 1.14 and )5.2,
respectively, for the pairs of characters aperture
pattern tetrad form; aperture pattern callose deposition;
and aperture pattern cytokinesis. According to Pagel &
Meade (2006), such values suggest very strong evidence
of correlated evolution for the pair of character aperture
pattern tetrad form, weak evidence for the pair aperture
pattern callose deposition and no evidence for the pair
aperture pattern cytokinesis.
Discussion
Increasingly, in the literature, an important role is given
to mutational and developmental systems in limiting the
direction that can be taken by a lineage during its
evolution (e.g. Bradshaw, 1991; Lynch, 2007; Stoltzfus &
Yampolsky, 2009; Braendle et al., 2010; Futuyma, 2010).
Other studies are inclined to argue that limitations on
morphological diversity are caused by the action of
natural selection rather than mutational biases or developmental constraints (e.g. Beldade et al., 2002; Eldredge
et al., 2005; Wiens et al., 2006). In this study, we show
that limitations in the evolution of pollen morphology
and in the evolution of its developmental system are
likely to be caused by selective pressures acting on pollen
morphology. However, even in the absence of these
selective pressures, a bias in the variability of the
developmental system remains.
Stasis in pollen morphology and development
The aperture pattern is relatively conserved in the
Euphorbiaceae s.s.: most species in this group (for which
1091
palynological data are available) produce pollen grains

with (generally three) equatorial apertures. Our optimization of pollen aperture pattern indicates that this is the
plesiomorphic condition and thus that the predominance
of this pattern is not due to homoplasy. Within Malpighiales, inaperturate pollen also occurs in Linaceae,
Malpighiaceae, Phyllanthaceae, Rafflesiaceae and Salicaceae (Furness, 2007). These occurrences, with the
exception of Rafflesiaceae, map to well-nested clades
within the most robust phylogenies for each family (i.e.
Chase et al., 2002; Kathriarachchi et al., 2005; Davis &
Anderson, 2010; McDill & Simpson, 2011) and therefore
represent losses that do not effect optimizations presented here. Rafflesiaceae are highly modified parasites
that have been placed either sister to Euphorbiaceae s.s.
or sister to Peraceae + Euphorbiaceae s.s. (Davis et al.,
2007; Wurdack & Davis, 2009). Rafflesiaceae pollen is
reported to be inaperturate, and its microsporogenesis
needs further investigation (see Furness, 2007, 2008).
Although not included in this study, it appears to
represent an independent loss of apertures and would
not impact our conclusions. Only two (possibly three)
departures from the equatorial aperture pattern were
observed. Euphorbiaceae s.s. comprise c. 6300 species
(Wurdack et al., 2005). This group is quite diverse with
respect to other morphological, phytochemical and
palynological features (Punt, 1962; Webster, 1994;
Wurdack et al., 2005) and is ancient with an estimated
divergence of 11069 Ma (Davis et al., 2005; Van Ee
et al., 2008; Forest & Chase, 2009). Despite such potentialities for change (high number of species, observed
change in other features and ancient divergence time),
aperture pattern remains relatively conserved as shown
by our optimization. This result is in agreement with
general palynological data, suggesting that aperture
pattern is quite conserved at the eudicot scale. Although
important diversity is known for aperture pattern, most
genera of eudicots have pollen grains with (generally
three) equatorial apertures (Wodehouse, 1935; Erdtman,
1952). Such conservatism may be considered as an
example of evolutionary stasis (Muller, 1984), as this
pollen morphology is a synapomorphy for the eudicot
clade (Donoghue & Doyle, 1989; Doyle & Hotton, 1991),
and its first occurrence in the fossil record corresponds
approximately to 124 Myr (Hughes & McDougall, 1990).
Here, the pattern suggested at the eudicot scale is
confirmed for Euphorbiaceae s.s. using representative
taxon sampling and a phylogenetic framework.
Evolutionary stasis in pollen aperture pattern was
found to be associated with stasis in its developmental
system (i.e. microsporogenesis and LPCD). All the species
producing pollen with equatorial apertures investigated
here were found to undergo microsporogenesis via the
same LPCD pattern, which is the LPCD accounting for
pollen with three or more equatorial apertures (illustrated in Fig. 1c-1). Our optimizations suggest that the
species of the Euphorbioideae and Acalyphoideae s.s.

1092
clades producing pollen with equatorial apertures are

associated with this LPCD pattern. These results need to
be treated with caution due to the limited data obtained
for microsporogenesis, but are nevertheless consistent
with the results of other workers indicating that equatorial apertures develop from this LPCD pattern (Ressayre
et al., 2002a; Blackmore et al., 2007). The ancestral state
for microsporogenesis is not resolved for the basal nodes
of our topology. However, an optimization of microsporogenesis characters on the topology of Wurdack et al.
(2005) (which is based on a larger taxon sample and in
which the Acalyphoideae s.s. clade diverged before
Crotonoideae s.l. and Euphorbioideae) indicates unambiguously that the ancestral state for microsporogenesis
and LPCD is the one accounting for pollen with three
or more equatorial apertures. Whereas this single
LPCD predominates in eudicots, several LPCD have
been described in the monocots (Wodehouse, 1935;
Blackmore & Crane, 1998; Furness & Rudall, 2004; Nadot
et al., 2008). The results shown here support this general
view.
Bias on the fixation of variant phenotypes caused by
selective pressures
It might be hypothesized that developmental constraints
are acting on the LPCD pattern and selective pressures
are acting on the aperture pattern. So both processes
might account for the evolutionary stasis of the entire
system, as the LPCD pattern is part of the ontogenetic
determination of the aperture pattern (Furness & Rudall,
2004). Discrimination between these two hypotheses has
been achieved here by realizing the thought experiment
of Alberch (1982). For this, studying microsporogenesis
in species with an inaperturate or a global aperture
pattern is appropriate because in these, a partial or total
disconnection of aperture position from microsporogenesis occurs (Wodehouse, 1935; Blackmore & Crane,
1998; Furness, 2007). Putative selective pressures due
to aperture positioning may not have any effect on
microsporogenesis and on the resulting LPCD pattern in
these species.
Species of Euphorbiaceae s.s. investigated with inaperturate or global aperture patterns were found to be
variable, at the level of the plant, for the three microsporogenesis features determining the LPCD pattern. For
tetrad form, intraindividual variation was found in all
nine inaperturate species examined and also in a species
with a global pattern. This suggests that variation in
tetrad form may occur easily and that there are no
constraints on this feature, as suggested previously for
some monocots (Nadot et al., 2006; Pereira Nunes et al.,
2009). Variation in the mode of callose deposition was
found in six of the nine species for which data for this
character were available. This indicates that variation in
this character may appear relatively easily in the context
of a loss of the equatorial aperture pattern. The cytoki-
nesis type seems to be less labile, as variation in this

feature was found in only one inaperturate species. The
association between the occurrence of the inaperturate
or global pattern and the occurrence of variation in the
LPCD pattern is not due to chance, as shown by the
comparative analysis. We found a significant phylogenetic association between loss of the equatorial aperture
pattern and appearance of variation in two of the three
microsporogenesis features studied here (tetrad form and
mode of callose deposition). This test is informative
despite the fact that it is based on a single evolutionary
transition from triaperturate to inaperturate pollen. The
robustness of such a test depends on the number of
phylogenetically independent occurrences of the two
binary traits for which a correlation is being tested and
also on the number of nodes of the phylogenetic tree
used: the higher the number of nodes of the tree, the
lower the probability that changes in the two characters
occur at the same node by chance if the two traits evolve
independently. The appearance of variation in tetrad
form and mode of callose deposition is thus not independent of the release of selective pressures on the LPCD
due to loss of the equatorial pattern.
In the context where the LPCD pattern is disconnected
from aperture patterning, six different LPCD patterns are
observed here (Fig. 1) and many others were described in
Codiaeum variegatum with inaperturate pollen (Albert
et al., 2009). Where the LPCD pattern determines the
aperture pattern (i.e. in species possessing equatorial
apertures), a single LPCD pattern is observed (Fig. 1c-1).
A possible explanation for this is that in the context
where the aperture pattern and LPCD pattern are
connected, selective pressure acting on the system
prevents variation. A putative selective pressure is that
acting on aperture number. Thus, according to Ressayre
et al. (2002a), the five alternative LPCD patterns result in
pollen grains with one, two or (rarely) three apertures
(Fig. 1), whereas the predominant LPCD pattern results
in pollen with three or more apertures (Ressayre et al.,
2002b). Variation in aperture number has been shown to
be linked to pollen grain fitness. It may be that the
predominant LPCD pattern has been positively selected
because of the flexibility in aperture number it provides
(Furness & Rudall, 2004).
Bias in the production of variant phenotypes caused
by developmental constraints
Variation in LPCD pattern occurs where it may be
assumed that some selective pressure has been released.
However, this variation appears to be partially biased
because only six LPCD patterns (of the sixteen that are
possible) are observed. Two biases remain. Firstly, LPCD
patterns resulting from a switch in cytokinesis were rarer
than those resulting from a switch in tetrad form or
callose deposition (bias 1). Secondly, LPCD patterns
resulting from a change in two or three microsporogen-

esis features were rare or absent, respectively, whereas

changes resulting from a change in a single microsporogenesis feature were observed in all the inaperturate
species (bias 2). Additional selective pressures and or
developmental constraints are thus likely to be acting on
the system. It may be assumed that where microsporogenesis is connected to aperture pattern, developmental
phenotypes (LPCD patterns) and corresponding pollen
phenotypes (aperture patterns) resulting from a switch in
cytokinesis should be rarer than those resulting from a
switch in tetrad form or mode of callose deposition (bias
1). In addition, phenotypes resulting from a switch in
two or three microsporogenesis features should be rarer
than those resulting from a switch in only one feature
(biais 2). These biases may constitute an example of a
developmental constraint sensu Maynard Smith et al.
(1985), that is, a bias in the generation of phenotypic
variation due to the properties of the developmental
system.
It would be interesting to assess microsporogenesis and
LPCD in other inaperturate lineages (e.g. in Apocynaceae). This would provide additional phylogenetically
independent samples to conduct the thought experiment
of Alberch and would allow to test whether the biases
observed here are produced repeatedly. Assessing the
molecular basis of variation in microsporogenesis would
allow to investigate whether this variation is constrained
by the topology of the gene network (Wagner, 2011), as
well as to test for the presence of molecular signatures of
selection on genes functionally linked to microsporogenesis. To date, knowledge about the molecular pathways
involved in the control of microsporogenesis does not
allow to carry out such studies (but see Magnard et al.,
2001).
Conclusion
The pattern observed here indicates that evolutionary
stasis in pollen aperture morphology and development
may be due, in part, to stabilizing selection because
removing the presumed target of selection permits the
appearance of variation. However, not all possible variants occur, indicating that constraints also contribute for
stability. In the debate concerning selection and constraints, it appears that neither of these two factors can
totally account for evolutionary stasis. It is the interaction
between them that may determine the direction of
evolution.
Acknowledgments
The authors thank R. Bateman, J. Becerra, P. Chesselet,
J. Doyle, D. Hinsinger, S. Magallon-Puebla, S. Nadot,
A. Ressayre, H. Sauquet, M. Lopez-Villavicencio and
three anonymous reviewers for helpful comments and
discussion; V. Sagun for sharing his Acalypha sequence
alignments; A. Dubois, M. Geze, C. Raquin, L. Saunois
1093
and S. Siljak-Yakovlev for technical help; and staff of

botanic gardens (Parc Botanique de Launay; Museum
national dHistoire naturelle; Royal Botanic Gardens,
Kew; Rancho Santa Ana; and Hedgerow Farms) for
facilitating access to living material. A.M.V., B.A. and
P.H.G. received financial support from the Action
Transversale du Museum: Formes possibles, Formes
realisees (Museum national dHistoire naturelle).
A.M.V. thanks the Bentham-Moxon Trust for funding.
References
Alberch, P. 1982. Developmental constraints in evolutionary
processes. In: Evolution and Development (J.T. Bonner, ed.), pp.
313332. Springer-Verlag, New York.
Albert, B., Gouyon, P.-H. & Ressayre, A. 2009. Microsporogenesis variation in Codiaeum producing inaperturate pollen grain.
C. R. Biologies 332: 507516.
Albert, B., Nadot, S., Dreyer, L. & Ressayre, A. 2010. The
influence of tetrad shape and intersporal callose wall formation on pollen aperture pattern ontogeny in two eudicot
species. Ann. Bot.- London 106: 557564.
Amundson, R. 1994. Two concepts of constraint: adaptationism
and the challenge from developmental biology. Philos. Sci. 61:
556578.
Arens, K. 1949. Prova de Calose por meio da microscopia a luz
fluorescente e aplicacoes do metodo. Lilloa 18: 7175.
Barnes, S.H. & Blackmore, S. 1986. Some functional features in
pollen development. In: Pollen and Spores: Form and Function (S.
Blackmore & I.K. Ferguson, eds), pp. 7180, Linnean Society
Symposium Series. Academic Press, London and New York.
Beldade, P., Koops, K. & Brakefield, P.M. 2002. Developmental
constraints versus flexibility in morphological evolution.
Nature 416: 844847.
Blackmore, S. & Barnes, S.H. 1990. Pollen wall development
in angiosperms. In: Microspores: Evolution and Ontogeny (S.
Blackmore & R.B. Knox, eds), pp. 173192. Academic Press
Limited, London.
Blackmore, S. & Crane, P.R. 1998. The evolution of apertures in
the spores and pollen grains of embryophytes. In: Reproductive
Biology (S.J. Owens & P.J. Rudall, eds), pp. 159182. Royal
Botanic Gardens, Kew.
Blackmore, S., Wortley, A.H., Skvarla, J.J. & Rowley, J.R. 2007.
Pollen wall development in flowering plants. New Phytol. 174:
483498.
Bradshaw, A.D. 1991. The Croonian Lecture, 1991 genostasis
and the limits to evolution. Philos. T. Roy. Soc. B 333: 289305.
Braendle, C., Baer, C.F. & Felix, M.A. 2010. Bias and evolution
of the mutationally accessible phenotypic space in a developmental system. PLoS Genet. 6: 113.
Breuker, C.J., Debat, V. & Klingenberg, C.P. 2006. Functional
evo-devo. Trends Ecol. Evol. 21: 488492.
Charlesworth, B., Lande, R. & Slatkin, M. 1982. A neodarwinian commentary on macroevolution. Evolution 36:
474498.
Chase, M.W., Zmarzty, S., Lledo, M.D., Wurdack, K.J., Swensen,
S.M. & Fay, M.F. 2002. When in doubt, put it in Flacourtiaceae:
a molecular phylogenetic analysis based on plastid rbcL DNA
sequences. Kew. Bull. 57: 141181.
Dajoz, I., Till-Bottraud, I. & Gouyon, P.-H. 1991. Evolution of
pollen morphology. Science 253: 6668.

1094
Dajoz, I., Till-Bottraud, I. & Gouyon, P.-H. 1993. Pollen aperture

polymorphism and gametophyte performance in Viola diversifolia. Evolution 47: 10801093.
Davis, C.C. & Anderson, W.R. 2010. A complete generic
phylogeny of Malpighiaceae inferred from nucleotide
sequence data and morphology. Am. J. Bot. 97: 20312048.
Davis, C.C., Webb, O.W., Wurdack, J.W., Jaramillo, C.A. &
Donoghue, M.J. 2005. Explosive radiation of Malpighiales
supports a mid-Cretaceous origin of modern tropical rain
forests. Am. Nat. 165: E36E65.
Davis, C.C., Latvis, M., Nickrent, D.L., Wurdack, K.J. & Baum,
D.A. 2007. Floral gigantism in Rafflesiaceae. Science 315: 1812.
Diaz Zavaleta, G. & Palacios Chavez, R. 1980. Contribucion al
conocimiento de la morfologa de los granos de polen de los
generos mas comunes de la familia Euphorbiaceae de Mexico.
B. Soc. Bot. Mex. 39: 2562.
Donoghue, M.J. & Doyle, J.A. 1989. Phylogenetic studies of seed
plants and angiosperms based on morphological characters. In:
The Hierarchy of Life (B. Fernholm, K. Bremer & H. Jornvall,
eds), pp. 181193. Elsevier, Amsterdam, Netherlands.
Dover, G.A. 1972. The organisation and polarity of pollen
mother cells of Triticum aestivum. J. Cell Sci. 11: 699711.
Doyle, J.A. & Hotton, C.L. 1991. Diversification of early
Angiosperms pollen in a cladistic context. In: Pollen and Spores:
Patterns of Diversification (S. Blackmore & S.H. Barnes, eds), pp.
169195. Clarendon Press, Oxford, UK.
Eldredge, N., Thompson, J.N., Brakefield, P.M., Gavrilets, S.,
Jablonski, D., Jackson, J.B.C. et al. 2005. The dynamics of
evolutionary stasis. Paleobiology 31(Suppl. 2): 133145.
Erdtman, G. 1952. Pollen Morphology and Plant Taxonomy. Angiosperms. Almqvist and Wiksell, Stockholm, Sweden.
Fischer, H. 1889. Beitrage zur vergleichenden Morphologie der
Pollenkorner. J.U. Kerns Verlag, Berlin.
Forest, F. & Chase, M.W. 2009. Eurosid I. In: The Time Tree of Life
(S.B. Hedges & S. Kumar, eds), pp. 188196. Oxford University Press, New York.
Furness, C.A. 2007. Why does some pollen lack apertures? A
review of inaperturate pollen in eudicots. Bot. J. Linn. Soc. 155:
2948.
Furness, C.A. 2008. Successive microsporogenesis in eudicots,
with particular reference to Berberidaceae (Ranunculales).
Plant Syst. Evol. 273: 211223.
Furness, C.A. & Rudall, P.J. 1999a. Inaperturate pollen in
monocotyledons. Int. J. Plant Sci. 160: 395414.
Furness, C.A. & Rudall, P.J. 1999b. Microsporogenesis in
monocotyledons. Ann. Bot. - London 84: 475499.
Furness, C.A. & Rudall, P.J. 2004. Pollen aperture evolution a
crucial factor for eudicot success? Trends Plant Sci. 9: 154158.
Furness, C.A., Rudall, P.J. & Sampson, F.B. 2002. Evolution of
microsporogenesis in angiosperms. Int. J. Plant Sci. 163: 235
260.
Futuyma, D.J. 2010. Evolutionary constraint and ecological
consequences. Evolution 64: 18651884.
Gillespie, J. 1994. Pollen morphology and phylogeny of the tribe
Plukenetieae (Euphorbiaceae). Ann. Mo. Bot. Gard. 81: 317
348.
Gould, S.J. 2002. The Structure of Evolutionary Theory. Harvard
University Press, Cambridge and London.
Guindon, S. & Gascuel, O. 2003. A simple, fast, and accurate
algorithm to estimate large phylogenies by maximum likelihood. Syst. Biol. 52: 696704 (software available from http://
www.atgc-montpellier.fr).
Guindon, S., Dufayard, J.F., Lefort, V., Anisimova, M., Hordijk,

W. & Gascuel, O. 2010. New algorithms and methods to
estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst. Biol. 59: 307321.
Hall, T.A. 1999. BioEdit: a user-friendly biological sequence
alignment editor and analysis program for Windows
95 98 NT. Nucl. Acids Symp. Ser. 41: 9598.
Hansen, T.F. & Houle, D. 2004. Evolvability, stabilizing selection
and the problem of stasis. In: The Evolutionary Biology of
Complex Phenotypes (M. Pigliucci & K. Preston, eds), pp. 130
150. Oxford University Press, Oxford.
Heslop-Harrison, J. 1968. The pollen grain wall. Science 161: 230
237.
Hesse, M., Halbritter, H., Zetter, H., Weber, M., Buchner, R.,
Frosch-Radivo, A. et al. 2009. Pollen Terminology An Illustrated
Handbook. Springer Wien, New York.
Huelsenbeck, J.P. & Ronquist, F. 2001. MRBAYES: Bayesian
inference of phylogenetic trees. Bioinformatics 17: 754755.
Hughes, N.F. & McDougall, A.B. 1990. Barremian-Aptian
angiospermid pollen records from southern England. Rev.
Palaeobot. Palyno. 65: 145151.
Johri, B.M. & Kapil, R.N. 1953. Contribution to the morphology
and life history of Acalypha indica L. Phytomorphology 3: 137
151.
Kapil, R.N. 1960. Embryology of Acalypha Linn. Phytomorphology
3: 137151.
Kapil, R.N. 1961. Some embryological aspects of Euphorbia dulcis
L. Phytomorphology 11: 2436.
Kathriarachchi, H., Hoffman, P., Samuel, R., Wurdack, K.J. &
Chase, M.W. 2005. Molecular phylogenetics of Phyllanthaceae inferred from five genes (plastid atpB, matK, 3 ndhF,
rbcL, and nuclear PHYC). Mol. Phylogenet. Evol. 36: 112
134.
Kesseler, R. & Harley, M. 2004. Pollen: The Hidden Sexuality of
Flowers. Papadakis, London.
Lauder, G.V. 1981. Form and function: structural analysis in
evolutionary morphology. Paleobiology 7: 430442.
Liu, H.F., Kirchoff, B.K., Wu, G.J. & Liao, J. 2007. Microsporogenesis and male gametogenesis in Jatropha curcas L. (Euphorbiaceae). J. Torrey Bot. Soc. 134: 335343.
Lobreau-Callen, D. & Suarez-Cervera, M. 1997. Le pollen des
Crotonoideae Apetales (Euphorbiaceae): ultrastructure
de lexine. Rev. Palaeobot. Palyno. 98: 257291.
Lynch, M. 2007. The frailty of adaptive hypotheses for the
origins of organismal complexity. Proc. Natl Acad. Sci. USA 104:
85978604.
Lyon, F.M. 1898. A contribution to the life history of Euphorbia
corollata. Bot. G. 25: 418426.
Maddison, W.P. & Maddison, D.R. 2008. Mesquite: a modular
system for evolutionary analysis, version 2.5. See http://
mesquiteproject.org.
Magnard, J.-L., Yang, M.C., Learly, M. & McCornick, S. 2001.
The Arabidopsis gene tardy asynchronous meiosis is required
for the normal pace and synchrony of cell division during
male meiosis. Plant Physiol. 127: 11571166.
Maynard Smith, J. 1983. The genetics of stasis and punctuation.
Annu. Rev. Genet. 17: 1125.
Maynard Smith, J., Burian, R., Kauffman, S., Alberch, P.,
Campbell, J., Goodwin, B. et al. 1985. Developmental
constraints and evolution: a perspective from the Mountain
Lake conference on development and evolution. Q. Rev. Biol.
60: 265287.

Mayr, E. 1961. Cause and effect in biology. Science 134: 1501

1506.
McDill, J. & Simpson, B.B. 2011. Molecular phylogenetics of
Linaceae with complete generic sampling and data from two
plastid genes. Bot. J. Linn. Soc. 165: 6483.
McGhee, G.R. 2007. The Geometry of Evolution. Cambridge
University Press, Cambridge.
Mukherjee, P.K. 1960. Embryology of two Euphorbiaceae. Proc.
Indian Acad. Sci. B 53: 217229.
Mukherjee, P.K. 1964. Further contribution to the embryology
of the genus Acalypha Linn. (Euphorbiaceae). Proc. Natl A. Sci.
India B 34: 129141.
Mukherjee, P.K. 1965. Contribution to the embryology of
Euphorbia peltata Roxb. Proc. Natl A. Sci. India B 35: 327
337.
Muller, J. 1984. Significance of fossil pollen for angiosperm
history. Ann. Mo. Bot. Gard. 71: 419443.
Muller, G.B. 2007. Evo-devo: extending the evolutionary
synthesis. Nat. Rev. Genet. 8: 943949.
Murgia, M., Wilms, H.J., Cresti, M. & Cesca, G. 1986. Ultrastructure of pollen development in Euphorbia dulcis L. 1.
Diploid plants. Acta Bot. Neerl. 35: 405424.
Nadot, S., Forchioni, A., Penet, L., Sannier, J. & Ressayre, A.
2006. Links between early pollen development and aperture
pattern in monocots. Protoplasma 228: 5564.
Nadot, S., Furness, C.A., Sannier, J., Penet, L., Triki-Teutroy, S.,
Albert, B. et al. 2008. Phylogenetic comparative analysis of
microsporogenesis in angiosperms with a focus on monocots.
Am. J. Bot. 95: 14261436.
Nowicke, J.W. 1994. A palynological study of Crotonoideae
(Euphorbiaceae). Ann. Mo. Bot. Gard. 81: 245269.
Nowicke, J.W. & Takahashi, M. 2002. Pollen morphology, exine
structure and systematics of Acalyphoideae (Euphorbiaceae),
Part 4. Rev. Palaeobot. Palyno. 121: 231336.
Nowicke, J.W., Takahashi, M. & Webster, G.L. 1998. Pollen
morphology, exine structure and systematics of Acalyphoideae (Euphorbiaceae) Part 1. Rev. Palaeobot. Palyno. 102: 115
152.
Oster, G. & Alberch, P. 1982. Evolution and bifurcation of
developmental programs. Evolution 36: 444459.
Pagel, M. 1994. Detecting correlated evolution on phylogenies: a
general method for the comparative analysis of discrete
characters. Proc. Roy. Soc. Lond. B Bio. 255: 3745 (software
available from http://www.evolution.rdg.ac.uk).
Pagel, M. 1999. The maximum likelihood approach to reconstructing ancestral character states of discrete characters on
phylogenies. Syst. Biol. 48: 612622.
Pagel, M. & Meade, A. 2006. Bayesian analysis of correlated
evolution of discrete characters by reversible-jump Markov
Chain Monte Carlo. Am. Nat. 167: 808825.
Penet, L., Nadot, S., Ressayre, A., Forchioni, A., Dreyer, L. &
Gouyon, P.-H. 2005. Multiple developmental pathways leading to a single morph: monosulcate pollen (examples from the
Asparagales). Ann. Bot. - London 95: 331343.
Pereira Nunes, E.L., Bona, C., Chiara Moco, M.C. & Ike Coan, A.
2009. Microsporogenesis features and inaperturate pollen
formation in three species of Potamogeton L. (Potamogetonaceae, Alismatales). Ann. Bot. - London 104: 10111015.
Posada, D. & Crandall, K.A. 1998. MODELTEST: testing the
model of DNA substitution. Bioinformatics 14: 817818.
Punt, W. 1962. Pollen morphology of the Euphorbiaceae with
special reference to taxonomy. Wentia 7: 1116.
1095
R Development Core Team 2008. R: A Language and Environment

for Statistical Computing. R Development Core Team, Vienna,
Austria.
Raff, R.A. 1996. The Shape of Life : Genes, Development, and the
Evolution of Animal Form. University of Chicago Press, Chicago.
Rao, P.N. 1962. A note on the embryology of Micrococca
mercurialis Benth. Curr. Sci. India 10: 426427.
Rao, A.N. 1964. Notes on the embryology of Hevea brasiliensis
Muell. Curr. Sci. India 24: 739740.
Rao, P.S. & Devi, K.R. 1974. Some embryological features of
Euphorbia vermiculata Raf. Curr. Sci. India 43: 257259.
Raup, D.M. 1966. Geometric analysis of shell coiling: general
problems. J. Paleontol. 40: 11781190.
Ressayre, A. 2001. Equatorial aperture pattern in monocots:
same definition rules as in eudicots? The example of two
species of Pontederiaceae Int. J. Plant Sci. 162: 12191224.
Ressayre, A., Godelle, B., Mignot, A. & Gouyon, P.-H. 1998. A
morphogenetic model accounting for pollen aperture pattern
distribution in flowering plants. J. Theor. Biol. 193: 321334.
Ressayre, A., Godelle, B., Raquin, C. & Gouyon, P.-H. 2002a.
Aperture pattern ontogeny in angiosperms. J. Exp. Zool. Part B
294: 122135.
Ressayre, A., Raquin, C., Mignot, A., Godelle, B. & Gouyon, P.H. 2002b. Correlated variation in microtubule distribution,
callose deposition during male post-meiotic cytokinesis, and
pollen aperture number across Nicotiana species (Solanaceae).
Am. J. Bot. 89: 393400.
Ressayre, A., Triky-Teurtroy, S., Forchioni, A., Dreyer, L. &
Nadot, S. 2005. Post-meiotic cytokinesis and pollen aperture
pattern ontogeny : comparison of development in four species
differing in aperture pattern. Am. Nat. 92: 576583.
Ronquist, F. & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian
phylogenetic inference under mixed models. Bioinformatics 19:
15721574.
Sagun, V.G. 2008. Systematics of Malesian Acalypha (Euphorbiaceae). Ph.D. diss., University of Illinois at Urbana-Champaign.
Sagun, V.G., Levin, G.A. & van der Ham, R. 2006. Pollen
morphology and ultrastructure of Acalypha (Euphorbiaceae).
Rev. Palaeobot. Palyno. 140: 123143.
Salazar-Ciudad, I. 2006. On the origins of morphological
disparity and its diverse developmental bases. BioEssays 28:
11121122.
Sannier, J., Nadot, S., Forchioni, A., Harley, M.M. & Albert, B.
2006. Variations in the microsporogenesis of monosulcate
palm pollen. Bot. J. Linn. Soc. 151: 93102.
Schlosser, G. 2007. Functional and generative constraints on life
cycle evolution: an attempt on the architecture of constraints.
In: Integrating Evolution and Development. From Theory to Practice
(R. Sansom & R.N. Brandon, eds), pp. 113173. MIT Press,
Cambridge, MA.
Scott, R.J., Spielman, M. & Dickinson, H.G. 2004. Stamen
structure and function. Plant Cell 16: S46S60.
Sheldon, J.M. & Dickinson, H.G. 1983. Determination of
patterning in the pollen wall of Lilium henryi. J. Cell Sci. 63:
191208.
Sheldon, J.M. & Dickinson, H.G. 1986. Pollen wall formation in
Lilium: the effect of chaotropic agents, and the organisation of
the microtubular cytoskeleton during pattern development.
Planta 168: 1123.
Stoltzfus, A. & Yampolsky, L.Y. 2009. Climbing mount probable:
mutation as a cause of nonrandomness in evolution. J. Hered.
100: 637647.

1096
Suarez-Cervera, M., Gillespie, L., Arcalis, E., Le Thomas, A.,

Lobreau-Callen, D. & Seoane-Camba, J.A. 2001. Taxonomic
significance of sporoderm structure in pollen of Euphorbiaceae: tribes Plukenetieae and Euphorbieae. Grana 40: 78104.
Takahashi, M., Nowicke, J.W., Webster, G.L., Orli, S.S. &
Yankowski, S. 2000. Pollen morphology, exine structure,
and systematics of acalyphoideae (Euphorbiaceae), part 3. Rev.
Palaeobot. Palyno. 110: 166.
Teotonio, H. & Rose, M.R. 2001. Perspective: reverse evolution.
Evolution 55: 653660.
Thanikaimoni, G., Caratini, C., Nilsson, S. & Grafstrom, E. 1984.
Omniaperturate Euphorbiaceae pollen with striate spines. Bull.
Jard. Bot. Natl Belg. 54: 105125.
Thathachar, T. 1952. Morphological studies in the Euphorbiaceae: I. Acalypha lanceolata Willd. Phytomorphology 2: 197201.
Till-Bottraud, I., Vincent, M., Dajoz, I. & Mignot, A. 1999. Pollen
aperture heteromorphism. Variation in pollen-type proportions along altitudinal transects in Viola calcarata. CR Acad. Sci.
III - Vie 322: 579589.
Tokuoka, T. 2007. Molecular phylogenetic analysis of Euphorbiaceae sensu stricto based on plastid and nuclear DNA
sequences and ovule and seed character evolution. J. Plant.
Res. 120: 511522.
Van Ee, B., Berry, P.E., Riina, R. & Gutierez-Amaro, J. 2008.
Molecular phylogenetics and biogeography of the Caribbeancentered Croton subgenus Moacroton (Euphorbiaceae s.s.). Bot.
Rev. 74: 132165.
Wagner, A. 2011. Genotype networks shed light on evolutionary
constraints. Trends Ecol. Evol. 26: 577584.
Wagner, G.P., Pavlicev, M. & Cheverud, J.M. 2007. The road to
modularity. Nat. Rev. Genet. 8: 921931.
Wake, D.B. & Larson, A. 1987. Multidimensional analysis of an
evolving lineage. Science 238: 4248.
Webster, G.L. 1994. Classification of the Euphorbiaceae. Ann.
Mo. Bot. Gard. 81: 332.
Weill, M., Berthomieu, A., Berticat, C., Lutfalla, G., Negre, V.,
Pasteur, N. et al. 2004. Insecticide resistance: a silent base
prediction. Curr. Biol. 14: R552R553.
Wiens, J.J., Brandley, M.C. & Reeder, T.W. 2006. Why does a trait
evolve multiple times within a clade? Repeated evolution of
snakelike body form in squamate reptiles Evolution 60: 123141.
Williams, G.C. 1992. Natural Selection. Domains, Levels and Challenges. Oxford University Press, Oxford, New York, Toronto.
Wodehouse, R. 1935. Pollen Grains, their Structure, Identification
and Significance in Science and Medicine. McGraw-Hill, New York.
Wurdack, K.J. & Davis, C.C. 2009. Malpighiales phylogenetics:
gaining ground on one of the most recalcitrant clades in the
angiosperm tree of life. Am. J. Bot. 96: 15511570.
Wurdack, K.J., Hoffmann, P. & Chase, M.W. 2005. Molecular

phylogenetic analysis of uniovulate Euphorbiaceae (Euphorbiaceae sensu stricto) using plastid rbcL and trnL-F DNA
sequences. Am. Nat. 92: 13971420.
Supporting information
Additional Supporting Information may be found in the
online version of this article:
Table S1 Genbank accession numbers (or source) of the
sequences used for the phylogenetic analysis; and matrix
used for the optimization and for the test of correlated
evolution
Table S2 Number of tetrads for each type of tetrad
observed in the heteromorphic species, and results of the
chi-square tests, testing if each type of tetrad is produced
in similar proportions within a plant
Figure S1 Number and position of callose cleavage
planes, relative to the arrangement of the microspores in
the tetrad and to the cytokinesis type
Figure S2 Mirror tree of Euphorbiaceae s.s. comparing
optimizations of the aperture pattern (a) and mode of
callose deposition (b)
optimizations of the aperture pattern (a) and cytokinesis
type (b)
optimizations of the aperture pattern (a) and tetrad form
(b) using the topology of Wurdack et al. (2005)
optimizations of the aperture pattern (a) and mode of
callose deposition (b) using the topology of Wurdack
et al. (2005)
As a service to our authors and readers, this journal
provides supporting information supplied by the authors.
Such materials are peer-reviewed and may be reorganized for online delivery, but are not copy-edited
or typeset. Technical support issues arising from supporting information (other than missing files) should be
addressed to the authors.
Received 2 December 2011; revised 22 February 2012; accepted 24
February 2012


Evolutionary Stasis in Euphorbiaceae Pollen

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Evolutionary Stasis in Euphorbiaceae Pollen

Uploaded by

Copyright:

Available Formats

doi: 10.1111/j.1420-9101.2012.02494.

Evolutionary stasis in Euphorbiaceae pollen: selection

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Neo-Darwinians consider two classes of mechanisms

the tetrad and centrifugal according to the cleavage plane)

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Stasis in Euphorbiaceae pollen

Mode of callose deposition

morphs with higher aperture numbers obtained through

As pointed out by Mayr (1961), evolutionary changes

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

have been removed should allow testing for the action of

Material and methods

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Stasis in Euphorbiaceae pollen

Table 1 Studied specimens, with origin of species for which new

Origin (accession number)

Baloghia inophylla (G. Forst.)

Museum national dHistoire naturelle,

Caperonia serrata (Turcz.)

Dalechampia roezliana Mull. Arg.

Cultivated, seeds from Hedgerow

Species for which data were obtained from the literature

Acalypha alnifolia Klein ex Willd.

that were available in GenBank (i.e. from Tokuoka,

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Tetrad form (percentage of each)

coding of microsporogenesis was the same as that used

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Stasis in Euphorbiaceae pollen

implemented in BA Y E S DI S C R E T E [ML and Markov chain

with Crotonoideae s.l., whereas both Tokuokas study

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Stasis in Euphorbiaceae pollen

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

inaperturate or a global pattern: all the species with

callose deposition. Detailed microsporgenesis sequences

Species with equatorial apertures

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Stasis in Euphorbiaceae pollen

serrata, which produces pollen with six apertures

Species with inaperturate or global aperture patterns

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

deposition were found. This variation was observed

found to follow the CpP mode (Fig. 5p). As in the case

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Stasis in Euphorbiaceae pollen

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Microsporogenesis in Croton californicus (Fig. 6ad),

simultaneous cytokinesis through the synthesis of six

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Stasis in Euphorbiaceae pollen

equatorial clade and that variation may have appeared

palynological data are available) produce pollen grains

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

clades producing pollen with equatorial apertures are

nesis type seems to be less labile, as variation in this

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Stasis in Euphorbiaceae pollen

esis features were rare or absent, respectively, whereas

and S. Siljak-Yakovlev for technical help; and staff of

2012 THE AUTHORS. J. EVOL. BIOL. 25 (2012) 10771096

Dajoz, I., Till-Bottraud, I. & Gouyon, P.-H. 1993. Pollen aperture