The evolution of key functional floral traits in the early divergent angiosperm family Annonaceae

Abstract

Potential key functional floral traits are assessed in the species‐rich early divergent angiosperm family Annonaceae. Pollinators (generally beetles) are attracted by various cues (particularly visual, olfactory and thermogenic), with pollinators rewarded by nectar (generally as stigmatic exudate), heat and protection within the partially enclosed floral chamber. Petals sometimes function as pollinator brood sites, although this may be deceptive. Annonaceae species are self‐compatible, with outcrossing promoted by a combination of protogyny, herkogamy, floral synchrony and dicliny. Pollination efficiency is enhanced by pollen aggregation, changes in anthesis duration, and pollinator trapping involving a close alignment between petal movements and the circadian rhythms of pollinators. Most Annonaceae flowers are apocarpous, with syncarpy restricted to very few lineages; fertilization is therefore optimized by intercarpellary growth of pollen tubes, either via stigmatic exudate (suprastylar extragynoecial compitum, EGC) or possibly the floral receptacle (infrastylar EGC). Although Annonaceae lack a distinct style, the stigmas in several lineages are elongated to form ‘pseudostyles’ that are hypothesized to function as sites for pollen competition. Flowers can be regarded as immature fruits in which the ovules are yet to be fertilized, with floral traits that may have little selective advantage during anthesis theoretically promoting fruit and seed dispersal. The plesiomorphic apocarpous trait may have been perpetuated in Annonaceae flowers since it promotes the independent dispersal of fruit monocarps (derived from separate carpels), thereby maximizing the spatial/temporal distance between seedlings. This might compensate for the lack of genetic diversity among seeds within fruits arising from the limited diversity of pollen donors.
This article is protected by copyright. All rights reserved.

Full text

J
SE Journal of Systematics
and Evolution doi: 10.1111/jse.12645
Review
The evolution of key functional floral traits in the early
divergent angiosperm family Annonaceae
Richard M. K. Saunders*
Division of Ecology and Biodiversity, School of Biological Sciences, The University of Hong Kong, Hong Kong, China
*Author for correspondence. E‐mail: saunders@hku.hk
Received 8 March 2020; Accepted 2 June 2020; Article first published online 4 June 2020
Abstract Potential key functional floral traits are assessed in the species‐rich early divergent angiosperm family
Annonaceae. Pollinators (generally beetles) are attracted by various cues (particularly visual, olfactory, and
thermogenic), with pollinators rewarded by nectar (generally as stigmatic exudate), heat, and protection within
the partially enclosed floral chamber. Petals sometimes function as pollinator brood sites, although this could be
deceptive. Annonaceae species are self‐compatible, with outcrossing promoted by a combination of protogyny,
herkogamy, floral synchrony, and dicliny. Pollination efficiency is enhanced by pollen aggregation, changes in
anthesis duration, and pollinator trapping involving a close alignment between petal movements and the circadian
rhythms of pollinators. Most Annonaceae flowers are apocarpous, with syncarpy restricted to very few lineages;
fertilization is therefore optimized by intercarpellary growth of pollen tubes, either by stigmatic exudate
(suprastylar extragynoecial compitum) or possibly the floral receptacle (infrastylar extragynoecial compitum).
Although Annonaceae lack a distinct style, the stigmas in several lineages are elongated to form “pseudostyles”
that are hypothesized to function as sites for pollen competition. Flowers can be regarded as immature fruits in
which the ovules are yet to be fertilized, with floral traits that may have little selective advantage during anthesis
theoretically promoting fruit and seed dispersal. The plesiomorphic apocarpous trait may have been perpetuated
in Annonaceae flowers as it promotes the independent dispersal of fruit monocarps (derived from separate
carpels), thereby maximizing the spatial/temporal distance between seedlings. This might compensate for the lack
of genetic diversity among seeds within fruits arising from the limited diversity of pollen donors.
Key words: circadian pollinator trapping, extragynoecial compitum, fertilization success, floral synchrony, pollen competition,
pollination efficiency, pollinator attraction.
1 Introduction
Floral morphology is an outstanding model system for
understanding evolutionary processes: the diversity of floral
forms, with their often‐flamboyant visual displays, attractive
scents, and energy‐rich pollinator rewards, present excellent
opportunities for investigating natural selection. Plant–
pollinator interactions have undoubtedly driven many floral
adaptations, with coevolution occurring wherever there is a
high degree of specificity. Evolution is also driven by sexual
selection, which is manifested in angiosperms as pollen
competition between microgametophytes as the pollen
tubes penetrate the carpel and as female choice through
selective pollination and zygote abortion (Moore &
Pannell, 2011). Flowers can furthermore be regarded as
immature fruits in which the ovules are yet to be fertilized:
although floral structure inevitably imposes constraints on
fruit structure (and vice versa), anatomical features of the
flower that might have little selective advantage during
anthesis can theoretically promote fruit and seed dispersal.
Evolutionary considerations of floral form and function must
therefore also accommodate corresponding assessments of
the fruit.
Early divergent angiosperms conspicuously lack two widely
recognized key evolutionary innovations associated with
more derived angiosperm lineages: their carpels are
apocarpous (not fused) and they lack a style. The evolution
of syncarpy (the congenital fusion of carpels) endows
significant advantages by enabling pollen grains deposited
on any stigma to potentially fertilize any ovule in the flower,
hence promoting pollination efficiency by minimizing ovule
wastage and increasing seedset (Endress, 1982; Armbruster
et al., 2002). The evolution of styles is likely to improve the
accuracy of the spatial positioning of stigmas (mirroring
the positioning of anthers conferred by filaments), hence
enhancing the efficiency of pollen transfer to and from
pollinators and providing greater evolutionary lability for
enabling plant–pollinator specificity. Syncarpy and the
evolution of styles have also been explained by reference
to pollen competition (Lora et al., 2016), in which micro-
gametophytes compete as they grow through the style and
are selected by the sporophyte.
July 2020
|
Volume 58
|
Issue 4
|
369–392 © 2020 Institute of Botany, Chinese Academy of Sciences

This review focuses on the species‐rich early divergent
angiosperm family Annonaceae, which is typical of other
magnoliid lineages in possessing an apocarpous gynoecium
composed of carpels that lack a distinct style (Saunders, 2010).
The family is characterized by flowers that are typically
pendent, trimerous, and bisexual, with numerous free
reproductive organs. The family lacks a biochemically
mediated self‐incompatibility mechanism, although there are
various adaptations that enhance outcrossing (Pang &
Saunders, 2014), including protogyny (the temporal separation
of pistillate and staminate function within the flower), which is
prevalent among early divergent angiosperms. Although
the perianth in most early divergent lineages transitions
between sepal‐like outer tepals and petal‐like inner tepals,
Annonaceae flowers possess whorls of distinct sepals and
petals (Saunders, 2010). The two petal whorls are furthermore
generally morphologically distinct, with the inner petals often
convergent or connivent to form a partially enclosed floral
chamber that is functionally important during pollination
(Saunders, 2010).
Several characteristics of the family render it ideal for
drawing inferences of broader relevance: (i) it is a major
component of tropical lowland rainforests both in terms of
biomass and species richness (e.g., Gentry, 1993; Punyasena
et al., 2008); (ii) it is well‐known taxonomically, with many
genera recently monographed (Erkens et al., 2012); (iii)
robust, well‐resolved phylogenies are available for most
lineages (Guo et al., 2017b); and (iv) there are many detailed
empirical assessments of the pollination ecology and
breeding systems of representative taxa (Saunders, 2012).
Early divergent angiosperm families such as Annonaceae
provide invaluable insights into the early evolutionary
diversification of floral morphology. Annonaceae are of
phylogenetic importance as they possess several key
apomorphic characteristics that are hypothesized to have
evolved independently from those in eudicots, including
differentiated perianth whorls, extragynoecial compita,
syncarpy, and the evolution of “pseudostyles”that are
possibly functionally analogous with true styles in derived
angiosperm lineages.
The morphology and function of Annonaceae floral traits
are reviewed here to clarify our understanding of mecha-
nisms that enhance pollinator attraction, maximize the
efficiency of pollination, and optimize fertilization success,
while avoiding self‐pollination and protecting floral repro-
ductive organs against herbivory. Evolutionary innovations
that overcome the limitations of apocarpy are likely to have
been key to this success: these mechanisms are hypothesized
here to include the evolution of extragynoecial compita that
enable intercarpellary growth of pollen tubes, the evolu-
tionary increase of ovule number per carpel, and the
independent origin of syncarpy.
2 Floral Adaptations to Different Pollinator
Guilds
Annonaceae flowers are predominantly pollinated by beetles
(Coleoptera), but also by other insect orders, including thrips
(Thysanoptera), flies (Diptera), bees (Hymenoptera), and
rarely also cockroaches (Blattodea) (Saunders, 2012). There is
little evidence for species‐specific plant–pollinator coevolu-
tion: the flowers are often visited by multiple species
simultaneously and the dominant pollinating species are
likely to vary geographically between populations and
temporally between flowering seasons (Herrera, 2005), as
observed in Annona (Kishore et al., 2012; Costa et al., 2017).
Most Annonaceae species are consistently visited by
representatives of a single pollinator guild, however, and
distinct pollination syndromes are therefore recognizable
(Saunders, 2012).
Two contrasting beetle‐pollination syndromes have
been identified based on the overall size of the beetle
(Saunders, 2012). Most Annonaceae species are pollinated
by small beetles in the families Curculionidae (weevils),
Nitidulidae (sap beetles), and/or Staphylinidae (rove
beetles); these beetles are invariably phytophagous and
often consume nectar and pollen. Some Annonaceae
lineages are pollinated by larger scarab beetles in
Scarabaeidae subfamilies Melolonthinae and Dynastinae;
scarab beetles are also phytophagous and, as they often
consume petals and other floral tissues, the flowers are
accordingly robust, with a thick, fleshy perianth.
Small‐beetle pollination has been inferred as the ancestral
pollination system in Annonaceae, with all other systems
derived (Saunders, 2012). The most important families of
small‐beetle pollinators originated prior to Annonaceae, with
Staphylinidae, Nitidulidae, and Curculionidae having min-
imum stem ages estimated at 198.7, 184.2, and 118.7 Mya,
respectively (McKenna & Farrell, 2009), whereas the
Annonaceae stem lineage originated at least 108.6 Mya
(Xue et al., 2020). Plant–pollinator interactions have
undoubtedly served as significant evolutionary drivers of
species diversity in Annonaceae, influencing floral form and
function.
2.1 Petal color
Although many early studies implicitly assumed that
phytophagous insects and insect pollinators rely on olfactory
cues for locating host plants, subsequent research revealed
the importance of vision, either in isolation or synergistically
with olfaction (Reeves, 2011). Insects possess highly sensitive
visual systems with up to six spectral classes of photo-
receptors (Briscoe & Chittka, 2001). All the major guilds of
Annonaceae pollinators are able to perceive a broad
spectrum of light, ranging from 340 nm (UV) through
650 nm (red), including beetles (Curculionidae: Hausmann
et al., 2004; Nitidulidae: Döring et al., 2012) as well as thrips
(Matteson et al., 1992), flies (Lunau, 2014), and bees
(Avarguès‐Weber et al., 2012). Most Annonaceae flowers
have petals that are cream or yellow and are often flushed
pink, although occasionally with a more intense red
pigmentation. The pervasive hypothesis that bees are not
attracted to red flowers has been refuted (e.g., Chittka &
Waser, 1997), and there are examples of red‐flowered
Annonaceae species that are pollinated by bees (e.g., Uvaria
concava: Silberbauer‐Gottsberger et al., 2003). No published
study has addressed the occurrence of UV reflectance in
Annonaceae flowers, but it is predicted to occur given the
ability of insect pollinators to perceive such short wave-
lengths of light.
370 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

The beetles that pollinate Annonaceae flowers typically
show crepuscular circadian rhythms (Lau et al., 2017a). They
are therefore primarily active at dawn and dusk when
ambient light intensities are low and generally blue‐shifted
(Endler, 1993). Crepuscular beetles have been reported to
possess considerable visual acuity, arising in part from their
refracting superposition compound eyes that enable accu-
rate color perception even at very low light intensities
(Warrant & Dacke, 2011; Honkanen et al., 2017); crepuscular
pollinators are therefore highly likely to use visual cues to
locate Annonaceae flowers despite low ambient light levels.
Flies are often particularly dependent on visual stimuli
(Lunau, 2014), and fly‐pollinated Annonaceae flowers often
have a complex pattern of petal pigmentation with colored
patches (sometimes with translucent “light windows,”e.g.,
Monodora myristica: Fig. 1A; Couvreur, 2009) or distinct
stripes (e.g., Mitrephora vittata: Fig. 1B; Weerasooriya &
Saunders, 2010) that are likely to act as pollinator guides.
Flies rely on different color cues for diverse behavioral
activities, such as locating food and oviposition (Lunau, 2014),
and a characteristic sapromyiophilous pollination syndrome is
evident in which coprophagous, necrophagous, or saproph-
agous flies are deceptively attracted to flowers that mimic
feces or dead/decaying organic matter. Although sapromyio-
philous flowers rely heavily on olfactory attractants
(discussed in Section 2.3), there are also distinctive visual
syndromes, with dull red petals that are often enlarged and
have a corrugated surface (Chen et al., 2015). Examples
within Annonaceae include Asimina parviflora (Norman
et al., 1992), Meiogyne species formerly classified as Fitzalania
(Fig. 1C; Thomas et al., 2012), Monodora tenuifolia
(Gottsberger et al., 2011), Pseuduvaria megalopus (Su
et al., 2005), and Stenanona flagelliflora (Xicohténcatl‐Lara
et al., 2016). A similar saprocantharophilous syndrome has
been reported for Duguetia cadaverica (Teichert et al., 2012),
involving the attraction of nitidulid beetles rather than flies.
2.2 Floral chamber morphology
Annonaceae flowers typically possess a partially enclosed
chamber that is variably formed by the convergence or
connivence of petals around the reproductive organs. Seven
different structural types of floral chamber were identified in
a review of floral evolution in the family (see Saunders, 2010:
fig. 5 for illustrations), namely: Type I, in which the chamber
is formed from suberect petals that remain loosely
contiguous; Type II, formed from petals that are basally
constricted around the reproductive organs; Type III, formed
from inner petals (or rarely outer petals: Guo et al., 2018)
that are apically connivent, with small basal apertures
between the inner petals that are either partially blocked
by the outer petals or remain open throughout the
functional floral phases; Type IV, formed from “boat‐
shaped”inner petals that are convergent along their
margins; Type V, formed from the bowl‐shaped corolla,
with an apical aperture; Type VI, formed from inner petals
that are apically imbricate, with small apertures between the
basal claws of the inner petals; and Type VII, formed from a
corolla tube that results either from basal fusion of petals or
free petals. The breadth of this structural diversity reflects
the independent evolutionary origin of floral chambers in the
family; mapping chamber types onto the molecular
phylogeny unsurprisingly reveals considerable homoplasy,
with multiple origins of each chamber type from an ancestral
floral form that is inferred to have lacked any chamber
(Saunders, 2010).
The independent origin of floral chambers in multiple
lineages in Annonaceae suggests that the chamber is
functionally important in pollinator attraction and/or reward.
The chamber is hypothesized to create a protected environ-
ment for pollinators (often used as a tryst site for copulation,
but also presumably for predator avoidance), as well as
allowing maintenance of a microenvironment with elevated
temperatures (discussed in Section 2.4), filtering out large
and potentially destructive floral visitors that would not be
effective pollinators, and in some cases enabling pollinator
trapping (discussed in Section 4.4). The presence of a floral
chamber is correlated with primary pollinator guild: it is
widespread in beetle‐, thrips‐, and fly‐pollinated species
(Saunders, 2012), but is unknown in species pollinated by
bees (Carvalho & Webber, 2000; Silberbauer‐Gottsberger
et al., 2003; Teichert et al., 2009; Li et al., 2016) or
cockroaches (Nagamitsu & Inoue, 1997).
2.3 Floral scent
Annonaceae flowers are highly aromatic, with Cananga
odorata (ylang‐ylang) widely used in the perfume industry.
Goodrich (2012) reviewed the diversity of floral scents in the
family and used available subjective descriptions to classify
odor types as perceived by human olfaction: 65% of species
were determined as having a “pleasant”smell (including
fragrant, agreeable, fruity, and sweet smells), 10% had a
yeasty‐fungal scent, and 8% were described as “unpleasant”
(including vomit, carrion, or rubbery smells). The remaining
species had miscellaneous scents (including lemongrass,
spearmint, acetone, linseed oil, or balsamic) or else were
reported to lack any noticeable odor. Although this variability
in floral scent is indicative of adaptations to different
pollinators, interpretation is complicated by the highly
dynamic nature of floral scents: as well as differing between
species, they also vary between organs within a flower and
between different phenological stages of the flower (Good-
rich et al., 2006; Goodrich & Raguso, 2009; Goodrich, 2012).
In her review of Annonaceae floral scents, Goodrich (2012)
classified the scent components into five major chemical
classes, namely, aliphatics, aromatics (benzenoids), isopre-
noids (terpenes), and nitrogen‐and sulfur‐containing
compounds. The fruity floral scents emitted by many
Annonaceae flowers often comprise chemically very dispa-
rate volatiles and are indicative of evolutionary convergence
towards beetle pollination in different evolutionary lineages
(Goodrich, 2012). Knowledge of floral scent chemistry in
Annonaceae remains limited, however, with relatively few
detailed studies available and with the taxa that have been
studied often phylogenetically distant (Goodrich, 2012).
Although fruit mimicry has often been highlighted in
interpretations of floral scent, there are very few studies
that directly compare floral scents against possible model
species (Goodrich & Jürgens, 2018).
The floral scent in Asimina species (inclusive of Deer-
ingothamnus) is particularly informative (Goodrich &
Raguso, 2009). There are two distinct floral forms in the
genus: the first group (comprising A. parviflora,A. pygmaea,
371Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

A. tetramera, and A. triloba) has small maroon flowers that
emit a yeasty (sometimes fetid) smell and are pollinated by
small flies and nitidulid beetles or rarely by large scarab
beetles (Willson & Schemske, 1980; Norman & Clayton, 1986;
Norman et al., 1992; Rogstad, 1993); whereas the second
group (all other species) has large, white, pink, or yellow
flowers that emit a sweet, pleasant fragrance and are
pollinated by large beetles or rarely by flies and thrips
(Uphof, 1933; Norman & Clayton, 1986; Norman et al., 1992;
Norman, 2003; Levitt et al., 2013; Barton & Menges, 2018).
Although the pale‐flowered species presumably use a
reward‐based pollination system, the maroon‐flowered
Fig. 1. Floral morphology and flower–pollinator interactions in selected Annonaceae species. A, Monodora myristica flower,
showing patchy petal pigmentation. B, Mitrephora vittata flower, showing striped petal pigmentation. Hairs on the adaxial surface
of the inner petals possibly assist with pollen retention during secondary pollen presentation. C, Meiogyne bidwillii, showing petal
adaptations favoring sapromyiophily. D, Copious stigmatic exudate in Disepalum anomalum.E, Copious stigmatic exudate and
elongated pseudostyles in Goniothalamus palawanensis.F, Inner petal nectaries (arrowed) in Uvaria dulcis.G, Endaenidius beetles
copulating on a Dasymaschalon trichophorum petal; note that the female beetle is boring a hole into the petal with its mouthparts.
H, Female Endaenidius beetle ovipositing into a Dasymaschalon trichophorum petal. I, Inner petal corrugations (arrowed) in Meiogyne
stenopetala. (Photographs: A, R.M.K. Saunders, reproduced from Saunders, 2010; B, J. Beaman, reproduced from Weerasooriya &
Saunders,2001;C,G–I, C.‐C.Pang;D,P.‐S. Li; E, C.C. Tang, reproduced from Tang et al., 2013; F, L. Averyanov).
372 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

species appear to have evolved deceptive pollination. The
floral scents of A. parviflora and A. triloba resemble
fermenting sugars and hence are likely to mimic rotting
fruit, which is the food substrate or brood site for their
pollinators (Goodrich & Raguso, 2009). These fermentation
volatiles are also present in the floral scents of A. pygmaea
and A. tetramera, although these latter species also contain
dimethyl disulfide and indole, respectively, which are by‐
products arising from the microbial degradation of carrion
and feces; this suggests an evolutionary expansion of floral
mimicry in the genus (Goodrich & Raguso, 2009).
There are several well‐documented examples of Annona-
ceae floral scent compounds that serve to attract specific
pollinators –including aliphatic esters and alcohols that
attract nitidulid beetles (e.g., Peña et al., 1999; Jürgens
et al., 2000), 4‐methyl‐5‐vinylthiazole that attracts scarab
beetles (Maia et al., 2012), and trans‐carvone oxide that
attracts male euglossine bees (Whitten et al., 1986; Teichert
et al., 2009) –as well as numerous other studies that
tentatively identify volatiles that have been implicated
elsewhere as semiochemicals but which lack specificfield
bioassays to unequivocally demonstrate function. Floral
scent can also function as a pollinator reward as well as an
attractant: the male euglossine bees that pollinate Unonopsis
flowers (Carvalho & Webber, 2000; Silberbauer‐Gottsberger
et al., 2003; Teichert et al., 2009) are reported to collect floral
volatiles using specialized brushes on their forelegs, and to
use the accumulated scent to attract female bees (Whitten
et al., 1989). Floral volatiles undoubtedly play a key role in
plant–pollinator coevolution, and hence the identification of
pollinator‐specific attractants remains an integral component
of studies determining pollination specificity (e.g., Teichert
et al., 2012).
2.4 Floral thermogenesis
Many angiosperm lineages are reported to have evolutionary
adaptations that enable flowers to maximize the absorption
and retention of exogenous heat, thereby enabling the
optimization of metabolic processes, organ maturation,
pollen germination, and/or pollen tube growth, as well as
the enhanced formation of nutritional rewards for pollinators
(van der Kooi et al., 2019). Such adaptations are particularly
important in cooler climates in which plants and their
generally ectothermic pollinators are subject to thermal
stresses. Although tropical species are unlikely to experience
a significant night‐time fall in ambient temperature, there is
nevertheless widespread evidence for the independent
evolution of endogenous floral heating (thermogenesis)
across 11 angiosperm families (Thien et al., 2000; Luo
et al., 2010). The biochemical basis for floral thermogenesis
in Annonaceae is unknown, although studies of other
thermogenic taxa (Symplocarpus, Araceae: Umekawa
et al., 2016) have highlighted the endothermic production
of NADPH, catalyzed by mitochondrial isocitrate dehydro-
genase.
Floral thermogenesis is likely to promote pollination
efficiency. Floral heat can act as a pollinator attractant,
with some beetles, including Curculionidae (Hausmann
et al., 2004), having infrared (IR) sensors, known as IR
sensilla or IR pit organs, that can detect IR radiation (Schmitz
et al., 1997; Hammer et al., 2001). Floral heat can also directly
serve as an energy reward for ectothermic pollinators,
assisting with the maintenance of body temperature and
stimulating reproduction, feeding, and digestion (Thien
et al., 2000; Seymour et al., 2003; Rands & Whitney, 2008).
The initiation of flight consumes considerable energy, with
beetles reported to require high thoracic temperatures,
often exceeding 30 °C (Seymour & Schultze‐Motel, 1997).
Elevated floral temperatures have also been postulated to
mediate the synthesis and volatilization of floral scents
(Sagae et al., 2008), with floral temperature fluctuations
often correlated with changes in fragrance emission (e.g.,
Ratnayake et al., 2007). Floral thermogenesis does not
necessarily function exclusively to promote pollination,
however: as well as optimizing various metabolic and
developmental processes, there are also examples from
other early divergent angiosperm families of postanthetic
floral thermogenesis in which the heat assists larval
development of gall midge pollinators (Schisandra, Schisan-
draceae: Luo et al., 2010).
Floral temperatures in Annonaceae are often raised up to
3–8 °C above ambient conditions, with considerably larger
increases reported in some species of Annona (c. 15 °C:
Gottsberger & Silberbauer‐Gottsberger, 1988; Gotts-
berger, 1989) and Xylopia (12–13 °C: Küchmeister et al., 1998).
Thermogenesis is widespread in the family (Table 1),
although field studies often fail to include measurements
of floral temperature, and assessments have yet to be made
for any species in subfamily Ambavioideae. Available records
are phylogenetically disparate, with sister genera (e.g.,
Artabotrys and Xylopia) and congeneric species (e.g., within
Anaxagorea,Annona,Monoon, and Xylopia) often differing in
their capacity to generate heat endogenously (Table 1); this
suggests the independent evolutionary origin of (or loss of)
thermogenesis. Floral thermogenesis is likely to be asso-
ciated with the evolution of specific pollination systems in
Annonaceae: all reports of floral thermogenesis are from
flowers that are pollinated by small or large beetles, and
there are no reports from fly‐or bee‐pollinated species.
Significantly, all beetle‐pollinated Xylopia species studied are
thermogenic (Küchmeister et al., 1998; Jürgens et al., 2000;
Silberbauer‐Gottsberger et al., 2003; Ratnayake et al., 2007),
whereas the predominantly thrips‐pollinated congener X.
aromatica lacks any endogenous heat (Jürgens et al., 2000;
Silberbauer‐Gottsberger et al., 2003). The apparent correla-
tion with pollinator type could also be associated with floral
adaptations to minimize heat loss, however: thermogenic
flowers are typically larger, with thicker petals, as the low
surface‐to‐volume ratio minimizes heat loss (Seymour &
Schultze‐Motel, 1997) and the partially enclosed floral
chamber (discussed in Section 2.2) helps retain heat.
2.5 Floral nectar
Floral nectar functions as an important sugar‐rich nutritive
reward for pollinators, with specialized nectary tissues likely
to have evolved independently in many different floral
organs across disparate angiosperm lineages (Bernar-
dello, 2007). Although nectaries have been reported from
the base of the petals in Annonaceae flowers, most nectar in
the family is stigmatic in origin (Figs. 1D, 1E; Lau et al., 2017b).
Stigmatic exudate was previously hypothesized to have been
the ancestral source of nectar in flowering plants (Lloyd &
373Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

Wells, 1992; Endress, 1994), although this interpretation was
based on the assumption that “wet”stigmas (which form a fluid
secretion when receptive) were ancestral to “dry”stigmas
(which have a dry proteinaceous extracellular pellicle layer). The
latter assumption has recently been revised, and the wet stigmas
characteristic of Annonaceae (Lora et al., 2011) are now
interpreted as likely to be derived within the Magnoliales (Lau
et al., 2017b, and references therein).
Nectar is rich in sugars, lipids, and proteins, supplemented
with amino acids, phenols, reactive oxygen/nitrogen species,
and calcium ions (Suárez et al., 2012; Rejón et al., 2014). The
sugars provide an excellent energy source for pollinators
(Galetto et al., 1998) and function as important phagostimu-
lants for insects, including beetles (Mitchell & Gregory, 1979;
Merivee et al., 2008). Floral nectar is dominated by a
combination of sucrose and hexose sugars, with the latter
Table 1 Evidence for thermogenesis in Annonaceae genera
Taxon (No. species sampled/
total)
Evidence for
thermogenesis Reference(s)
Subfamily Anaxagoreoideae
Anaxagorea (5/30) –,+,++ Küchmeister et al., 1998; Braun & Gottsberger, 2011; Teichert
et al., 2011
Subfamily Annonoideae
Tribe Bocageeae
Cymbopetalum (3/27) ++ Murray, 1993; Braun et al., 2011
Tribe Duguetieae
Duguetia (7/94) –,+Küchmeister et al., 1998; Jürgens et al., 2000; Silberbauer‐
Gottsberger et al., 2003
Tribe Xylopieae
Artabotrys (2/105) –Chen et al., 2020
Xylopia (5/164) –,+,+++ Küchmeister et al., 1998; Jürgens et al., 2000; Silberbauer‐
Gottsberger et al., 2003; Ratnayake et al., 2007
Tribe Annoneae
Annona (9/170) –,++,+++ Webber, 1981; Gottsberger, 1989, 1999, 2012; Jürgens et al., 2000,
as “Rollinia”
Asimina (2/17) –Gottsberger, 2012
Disepalum (2/9) –Li et al., 2016
Goniothalamus (3/134) –Silberbauer‐Gottsberger et al., 2003; Lau et al., 2016
Tribe Monodoreae
Isolona (1/20) –Gottsberger et al., 2011
Monodora (1/14) –Gottsberger et al., 2011
Uvariodendron (2/15) +Gottsberger et al., 2011
Uvariopsis (2/19) –Gottsberger et al., 2011
Tribe Uvarieae
Dasymaschalon (1/27) –Pang C‐C & Saunders RMK, unpublished data
Desmos (1/22) –Pang & Saunders, 2015
Fissistigma (1/59) –Lau JYY & Saunders RMK, unpublished data
Friesodielsia (1/38) –Guo X & Saunders RMK, unpublished data
Melodorum (1/11) –Silberbauer‐Gottsberger et al., 2003
Uvaria (2/199) –Attanayake AMAS, Pang C‐C & Saunders RMK, unpublished data
Subfamily Malmeoideae
Tribe Piptostigmateae
Piptostigma (1/13) –Gottsberger et al., 2011
Tribe Malmeeae
Bocageopsis (1/4) –Silberbauer‐Gottsberger et al., 2003
Mosannona (1/14) –Chatrou & Listabarth, 1998
Tribe Miliuseae
Huberantha (1/27) ++ Ratnayake et al., 2006a, as “Polyalthia”
Meiogyne (1/26) –Silberbauer‐Gottsberger et al., 2003
Monoon (1/60) –,++ Silberbauer‐Gottsberger et al., 2003, as “Enicosanthum;”
Ratnayake et al., 2006a, as “Polyalthia”
Pseuduvaria (1/54) –Pang et al., 2013
Supraspecific classification follows Chatrou et al. (2012) and Guo et al. (2017b). –, no thermogenesis recorded (i.e., floral
temperatures <1 °C above ambient conditions); +, weakly thermogenic (1–3 °C above ambient conditions); ++, moderately
thermogenic (3–8 °C above ambient conditions); +++, strongly thermogenic (>8 °C above ambient conditions).
374 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

(primarily consisting of glucose and fructose) particularly
easily digested and rapidly metabolized (Simpson &
Neff, 1983). The proportion of the different sugar types is
often characteristic for the pollinator type, with higher
sucrose‐to‐hexose ratios typical in insect‐pollinated species
(Gottsberger et al., 1984). The stigmatic nectar of Uvaria
grandiflora (Annonaceae) was nevertheless shown to be
dominated by fructose (72.2%), which is readily metabolized
by the beetle pollinators, and with smaller proportions of
glucose (19.3%) and sucrose (8.4%) (Lau et al., 2017b).
The amino acid composition of nectar not only contributes
to its nutritive value (Lord & Webster, 1979) but also alters its
taste as perceived by floral visitors (Mitchell & Gregory, 1979;
Merivee et al., 2008). Ten of the 20 common amino acids
cannot be synthesized by insects and hence must be
obtained through their diet (Haydak, 1970); all these essential
amino acids have significantly been recorded from floral
nectar (Baker & Baker, 1975; Gottsberger et al., 1984),
confirming the importance of nectar as a food source.
Studies of the stigmatic exudate chemistry of Annonaceae
species (Uvaria grandiflora and U. macrophylla: Lau
et al., 2017b; and Anaxagorea javanica: Li & Xu, 2019) report
the presence of most essential amino acids, although
tryptophan is absent from all three species and lysine is
furthermore absent from U. macrophylla. Many of the amino
acids recorded from Annonaceae stigmatic nectar are
reported to be sugar‐sensitive cell stimulants for beetles
and flies (Shiraishi & Kuwabara, 1970; Mitchell &
Gregory, 1979), with phenylalanine particularly abundant.
Phenylalanine has a significant phagostimulatory effect on
honey bees (Inouye & Waller, 1984), which are known to be
important secondary pollinators of U. grandiflora (Pang C‐C&
Saunders RMK, unpublished data).
The Annonaceae genera Anaxagorea (Maas & Westra,
1984; Li & Xu, 2019) and Xylopia (van Heusden, 1992;
Johnson & Murray, 2018) possess inner staminodes that are
morphologically transitional between stamens and carpels.
The inner staminodes in Anaxagorea possess an apical
glandular area (Scharaschkin & Doyle, 2006) that has been
interpreted as homologous with the stigma (Saunders,
2010). The staminode exudate is therefore comparable with
stigmatic exudate and might also function as a pollinator
food reward: the amino acid composition of the exudates of
the two organs in Anaxagorea javanica are very similar (Li &
Xu, 2019), although the staminode exudate appears to have
significantly higher concentrations of most amino acids.
Similar staminode glands have been reported from the
closely related genera Degeneria (Degeneriaceae), Eupo-
matia (Eupomatiaceae), and Galbulimima (Himantandra-
ceae) (Endress, 1984), although structural homology with
the Anaxagorea glands is uncertain.
Nectaries are also common at the base of the adaxial
surface of the inner petals in Annonaceae flowers (Fig. 1F),
with reports from subfamilies Ambavioideae (Tetrameran-
thus), Annonoideae (Asimina,Asteranthe,Diclinanona,Du-
guetia,Monodora,Porcelia,Uvaria, and Xylopia) and
Malmeoideae (Alphonsea,Meiogyne,Monoon,Orophea,
Pseuduvaria, and Sapranthus) (Saunders, 2010). The disparate
phylogenetic occurrence of petal glands implies extensive
homoplasy (Saunders, 2010), although with the caveat that
homology of these “glands”has not been reported and in
many cases their function might not be nectar secretion. The
petal nectaries of Alphonsea glandulosa (Xue et al., 2017) and
Pseuduvaria froggattii (Silberbauer‐Gottsberger et al., 2003;
Su & Saunders, 2006) have received particular attention, and
have been shown to consist of distinct secretory and ground
parenchyma zones, with an epidermal layer in which
modified stomata enable nectar secretion. The evolution of
petal nectaries as an alternative to stigmatic exudate might
have been favored as it would release the plant from the
constraints imposed by protogyny, in which stigmatic
exudate is unlikely to extend into the staminate floral phase.
2.6 Petals as pollinator brood sites
Dasymaschalon trichophorum (Annonaceae) flowers are
pollinated by small Endaenidius beetles (Curculionidae) (Lau
et al., 2017a). The beetles have been observed to copulate on
the flowers, with the female beetles chewing small holes in
the petal surface while mating and then ovipositing into
these holes after the departure of the male (Figs. 1G, 1H;
Pang C‐C & Saunders RMK, unpublished data). Although the
hypothesis remains untested, it seems probable that the
petals might function as a food source for the emerging
beetle larvae, possibly in abscised petals on the forest floor
(cf. Sakai, 2002), thereby contributing to the maintenance of
functional populations of the pollinator.
The Annonaceae genus Meiogyne possesses elaborate
inner petal corrugations (Fig. 1I) that have often been
referred to as “glands”(e.g., van Heusden, 1992), although
without confirmation of glandular function. The corrugations
(strumae) on Meiogyne hainanensis petals (Shao & Xu, 2015,
as “Oncodostigma”) have been shown to be rich in
polysaccharides and to be associated with thrips larvae.
Although this might represent another example of pollinator
brood‐site adaptations of petals in Annonaceae, the study
failed to establish that thrips are the effective pollinator.
Another study (Collier & Armstrong, 2009) reported that
Anaxagorea crassipetala petals are co‐opted as a larval
substrate and pupation site by Diathoneura flies (Drosophi-
lidae). The effective pollinators have previously been shown
to be nitidulid and staphylinid beetles (Armstrong &
Marsh, 1997), however, and drosophilid flies were
never observed to enter the floral chamber (Collier &
Armstrong, 2009).
Although there are very few studies of pollinator brood‐
site adaptations of petals in Annonaceae, it is likely that the
phenomenon may be more widespread. Future studies need
to be much more comprehensive, integrating assessment of
effective pollination with empirical evidence to demonstrate
completion of the insect life cycle, from copulation and
ovipositing through larval development and pupation.
3 Floral Adaptations that Promote
Xenogamy
Numerous studies have reported the capacity of Annonaceae
species to set seed after autogamous and geitonogamous
self‐pollination (see Pang & Saunders, 2014: Table 1 for
references). Although this is unequivocal evidence for the
absence of a biochemically mediated self‐incompatibility
mechanism, Annonaceae species nevertheless appear to
375Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

maintain predominantly xenogamous breeding systems, as
evidenced by population genetic analyses (e.g., Annona
crassiflora: de Almeida‐Júnior et al., 2018; Desmos chinensis:
Pang & Saunders, 2015; Huberantha korinti: Ratnayake
et al., 2006b, as “Polyalthia”; and Monoon coffeoides:
Ratnayake et al., 2006b, as “Polyalthia”) and field‐based
controlled pollination experiments (e.g., Asimina obovata and
A. pygmaea: Norman & Clayton, 1986; Maasia glauca and
M. hypoleuca: Rogstad, 1994, as “Polyalthia”;Popowia
pisocarpa: Momose et al., 1998; Sapranthus palanga:
Bawa, 1974; Uvaria elmeri: Nagamitsu & Inoue, 1997; and
Xylopia championii: Ratnayake et al., 2007). A population
genetic study of Annona crassiflora based on microsatellite
loci (de Almeida‐Júnior et al., 2018) revealed spatially very
restricted dispersal of pollen. The authors furthermore
concluded that the behavior of the Cyclocephala scarab
beetle pollinators, which rarely move between flowers,
results in a considerable proportion of the ovules being
fertilized by the same pollen donor (de Almeida‐Júnior
et al., 2018); although xenogamy dominates, paternal
diversity in the progeny is likely to be very limited.
Annonaceae species promote xenogamy by adopting
various strategies, including temporal separation of sexual
function (protogyny), spatial separation of stamens and
carpels (herkogamy), synchrony of floral reproductive
function (synchronous dichogamy), and partial or complete
separation of sexes (dicliny). These adaptations were
reviewed by Pang & Saunders (2014), and the following
account is largely based on this work, supplemented with
recent additions to the literature.
3.1 Protogyny
Temporal separation of male and female function within
hermaphroditic flowers (intrafloral dichogamy) is widespread
among angiosperms and is an effective strategy for avoiding
autogamous self‐pollination. Protogyny, in which pistillate
function precedes staminate function, is almost ubiquitous
among early divergent lineages, with observations from all
families except Gomortegaceae and Hernandiaceae, which
remain unstudied (Endress, 2010). Protogyny is therefore
undoubtedly the ancestral mechanism in flowering plants,
including Annonaceae.
The pistillate and staminate phases in Annonaceae flowers
are typically separated by a sexually non‐functional interim
phase that reinforces the effectiveness of protogyny. The
duration of the interim phase varies according to overall
anthesis duration: Polyalthia suberosa, which is typical of
many species in the family, has anthesis over 48 h and has a
7‐h interim phase (Fig. 2A; Lau et al., 2017a); whereas species
with abbreviated anthesis (Figs. 2B, 2C) have a shorter
interim phase (e.g., Goniothalamus tapisoides, which has
anthesis over c. 23 h has an interim phase of only 3 h: Lau
et al., 2016).
The evolutionary breakdown of protogyny is observed in
several Annonaceae species, enabling increased reproductive
assurance through autogamy (discussed in Section 5.4).
Although the presence of stigmatic exudate is widely used as
a proxy indicator of stigmatic receptivity, interpretation is
complicated by the occasional retention of exudate into the
interim phase as a nectar reward for pollinators. The sugar
content of the exudate in Uvaria macrophylla changes
temporally (Lau et al., 2017b), however, rising from relatively
low concentrations that are conducive for pollen germination
during the peak pistillate phase to much higher concen-
trations that impede germination toward the end of the
pistillate phase.
3.2 Herkogamy
Herkogamy is the spatial separation of stamens and carpels
within a flower, thereby reducing opportunities for autog-
amous self‐pollination. The significance of this for Annona-
ceae is unclear given the prevalence of protogyny and
perceptions regarding “mess‐and‐soil”pollination (Fægri &
van der Pijl, 1979), in which beetles randomly and
destructively move around the flower. Herkogamy has
nevertheless been hypothesized as a result of inner
staminodes, asymmetric stigmas, conical receptacles, and
elongated pseudostyles.
The Annonaceae genera Anaxagorea (Maas & Westra, 1984;
Li & Xu, 2019) and Xylopia (van Heusden, 1992; Johnson &
Murray, 2018) possess inner staminodes that are effectively
sterile stamens, transitional with carpels (Saunders, 2010). In
Anaxagorea (and possibly also Xylopia) these staminodes
could act as a physical barrier between the anthers and
stigmas, elongating and curving inwards over the carpels as
anthesis progresses (Maas‐van de Kamer, 1993; Webber,
2002; Teichert et al., 2011). Lora et al. (2011) observed that the
stigmas of the outermost carpels in Annona cherimola are
asymmetrical, with a non‐papillate, non‐receptive surface
facing the innermost stamens that presumably also mini-
mizes intrafloral pollen transfer.
Some Annonaceae taxa (e.g., Uvaria buchholzii:Le
Thomas, 1968, as “Balonga”; and Toussaintia: Deroin, 2000)
possess an elongated conical floral receptacle that elevates
the stigmas above the androecium; this spatial separation is
also observed in genera with a convex receptacle (e.g.,
Annona: González & Cuevas, 2011). Several genera further-
more have stigmas that are elongated to form a
“pseudostyle”(e.g., Goniothalamus: Lau et al., 2016, 2017b);
as the pollen tubes of Goniothalamus parallelivenius are only
capable of penetrating the apex of the stigma (Lau
et al., 2017b), the elongated pseudostyles might also
promote herkogamy.
3.3 Floral synchrony
Although protogyny is effective for avoiding autogamy in
Annonaceae, self‐pollination between flowers of the same
individual (geitonogamy) is likely to be common as the family
is self‐compatible. Geitonogamy can be minimized by
reducing the number of co‐occurring flowers, although this
would be detrimental to seedset. Flowering in several
phylogenetically disparate Annonaceae species is reported
to be synchronized, with sexual function aligned within and
between individuals. The most common manifestation of
this is pistillate/staminate‐phase floral synchrony (Pang &
Saunders, 2014), in which all the flowers on an individual
mature in concert, so that co‐occurring flowers are either
pistillate or staminate and hence avoid geitonogamy;
synchrony in different individuals within the population is
furthermore staggered between days, promoting cross‐
pollination. This form of synchronous dichogamy is reported
from several phylogenetically disparate genera within
376 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

Annonaceae, including Annona (Murray & Johnson, 1987, as
“Rollinia”; González & Cuevas, 2011; Lora et al., 2011),
Dasymaschalon (Pang & Saunders, 2014), Desmos (Pang &
Saunders, 2014, 2015), Guatteria (Webber, 2002), Maasia
(Rogstad, 1994, as “Polyalthia”) and Unonopsis (Carvalho &
Webber, 2000). At least one “flower‐free”day is required
between each floral cohort in order to avoid geitonogamous
pollen transfer; as a result, many synchronous species have
abbreviated anthesis, thereby enabling more rapid floral
turnover (discussed further in Section 4.3). Pistillate/
staminate‐phase synchrony is nevertheless imperfect, with
some individuals producing flowers asynchronously within
the population; this presumably enables greater genetic
mixing within the population by avoiding the long‐term
perpetuation of distinct reproductive cohorts.
A more complex form of synchrony, known as hetero-
dichogamy, is observed in Anaxagorea (Teichert et al., 2011)
and Annona (Wester, 1910), involving the formation of
distinct floral forms that are morphologically identical but
with contrasting patterns of sexual function. In Anaxagorea
prinoides, for example, two morphs coexist in equal
proportions: one with a relatively prolonged pistillate phase,
and the other a shorter pistillate phase. The timing of the
sexual phases in the two morphs are aligned so that pollen‐
laden beetles departing from late staminate‐phase flowers
are attracted to flowers of the opposite morph at the onset
of the pistillate phase (Teichert et al., 2011; Pang &
Saunders, 2014). As each individual is consistently of a single
morph with synchronized flowering within the plant,
geitonogamy is minimized and almost uninterrupted flow-
ering is feasible.
Although floral synchrony is relatively uncommon in
angiosperms, it is noteworthy that empirical field studies
that specifically aim to identify synchrony are rare, and the
phenomenon could be more widespread. Temporal syn-
chrony of floral sexual function is more common within
Magnoliales and Laurales than in derived lineages
(Endress, 2020).
3.4 Dicliny
Most Annonaceae species bear hermaphroditic flowers. This
condition is ancestral for the family (Saunders, 2010), with
structurally or functionally unisexual flowers (dicliny)
achieved independently in several lineages as a strategy to
promote xenogamy. Structurally unisexual flowers, arising
following the loss of the androecium or gynoecium, have
evolved in several disparate lineages in subfamilies Annonoi-
deae (Annona,Anonidium,Diclinanona, and Uvariopsis) and
Malmeoideae (Ephedranthus,Greenwayodendron,Klarobelia,
Polyceratocarpus,Pseudephedranthus,Pseudomalmea,Pseu-
duvaria, and Stelechocarpus) (Pang & Saunders, 2014, and
references therein). At the population level, floral unisex-
uality is variously manifested as monoecy (with separate
pistillate and staminate flowers on the same plant) or
andromonoecy (with separate staminate and bisexual
flowers on the same plant), although there are some reports
of dioecy or androdioecy (in which the different floral sexes
are borne on different individuals). Many of the latter reports
are based on incomplete sampling, however, often from
herbarium collections.
Interpretations of floral sex are complicated by the
occurrence of flowers that are structurally bisexual but which
contain relatively few and apparently poorly developed
stamens: these flowers have sometimes been interpreted as
functionally pistillate with sterile staminodes (e.g., Pseuduvaria
mulgraveana: Su & Saunders, 2006, 2009), although subsequent
empirical fieldwork has shown that staminate function is
maintained, albeit limited (Pang et al., 2013). Phylogenetic
reconstructions of Pseuduvaria have revealed evolutionary
reversals from unisexual to bisexual flowers in some New
Guinea lineages (Su et al., 2008); this might have been
selectively advantageous for promoting self‐pollination fol-
lowing the colonization of geographical regions that are
topographically and ecologically complex.
Despite the contradictory interpretations of floral sex in
Pseuduvaria mulgraveana alluded to above, other species in the
genus bear structurally hermaphroditic flowers that are
functionally unisexual. Pseuduvaria macrocarpa, for example, is
reported to have unisexual flowers, with unambiguously
staminate flowers together with structurally hermaphroditic
flowersthatbearfunctionalcarpelsandareducednumberof
staminodes that contain small, incompletely developed pollen
grains that are presumably sterile (Su & Saunders, 2006).
Pseuduvaria mulgraveana has been shown to achieve
functional unisexuality in structurally hermaphroditic flowers
by delaying anther dehiscence until after petal abscission,
after departure of the beetle pollinators (Pang et al., 2013). A
parallel mechanism has been suggested for the congener
P. trimera based on floral ontological investigation (Yang &
Xu, 2016).
Androdioecy/andromonoecy has been reported to be
associated with rapid evolutionary diversification in Annona-
ceae, with the possession of exclusively hermaphroditic
flowers conversely associated with the lowest diversification
rate (Xue et al., 2020). Androdioecy/andromonoecy could be
selectively advantageous in obligately outcrossing species in
which gene flow is limited, with the reduced level of pistillate
function countered by increased seedset arising from the
greater availability of pollen (Lloyd, 1975).
4 Floral Adaptations that Enhance the
Efficiency of Pollinator Use
4.1 Pollen aggregation
Annonaceae pollen is generally dispersed as single grains
(monads), with the formation of aggregate clusters of pollen
grains inferred to have originated independently on multiple
occasions across the family and to be synapomorphic for
tribes Annoneae and Monodoreae within subfamily Anno-
noideae (Doyle & Le Thomas, 2012; Xue et al., 2020:
supplementary data). In most cases, the pollen aggregates
are tetrads (Fig. 3A) that form due to the failure of
microsporocytes to dissociate (e.g., Annona: Tsou &
Fu, 2002; Lora et al., 2009; and Pseuduvaria:Su&
Saunders, 2003); in a small number of genera, however,
larger octads (e.g., Cymbopetalum: Tsou & Fu, 2007; and
Disepalum: Li et al., 2015) and polyads of up to 32 pollen
grains (Porcelia: Le Thomas et al., 1986) arise due to two or
more sporocytes developing in synchrony (Tsou & Fu, 2007).
In addition to the direct cohesion between pollen grains
377Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

Fig. 2. Continued
378 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

within the pollen aggregate, neighboring tetrads in
Pseuduvaria are also linked by non‐sporopollenin pollen‐
connecting threads (Figs. 3A, 3B; Su & Saunders, 2003; Li &
Xu, 2018). Similar threads have since been reported from
Asimina (Zhang et al., 2014), Disepalum (Zhang et al., 2014),
and Monoon (Ratnayake et al., 2006a, as “Polyalthia”) and
could be more widespread in the family as the standard
acetolysis technique that is commonly applied in pollen
preparation destroys threads that lack sporopollenin.
Aggregation of pollen grains –irrespective of whether it is
due to direct cohesion between grains or pollen‐connecting
threads –is hypothesized to function as a mechanism to
enhance siring success by enabling the transfer of multiple
pollen grains following a single pollinator visit (Kress, 1981;
Hesse et al., 2000; Harder & Johnson, 2008). This is likely to
be particularly beneficial for species in which pollinator visits
are infrequent and/or when pollen transfer between flowers
is inefficient. This inevitably comes at a genetic cost,
however, due to the increased proportion of sibling pollen
grains received by stigmas; pollen aggregation will only be
selectively favored if the consequences of aggregation
increase overall reproductive performance (Harder &
Johnson, 2008). In order to optimize pollination efficiency,
pollen tubes growing from pollen aggregates must be able to
access numerous ovules. Many Annonaceae genera with
pollen aggregation are notably either syncarpous (Monodora,
with up to 70 ovules per fused gynoecium) or else have
multi‐ovulate carpels, namely, Cymbopetalum (with up to 25
ovules per carpel), Goniothalamus (up to 10), Meiocarpidium
(up to 20), Porcelia (up to 15), Pseuduvaria (up to 18),
Uvariastrum (up to 25), and Uvariopsis (up to 15). Other
Annonaceae genera showing pollen aggregation have fewer
ovules per carpel (e.g., Annona,Anonidium,Disepalum,
Duckeanthus, and Fusaea), but in these cases intercarpellary
growth of pollen tubes is likely to be promoted by either
partial syncarpy or the presence of an extragynoecial
compitum (discussed further in Section 5).
In a recent study of the influence of specific traits on
evolutionary diversification rates in Annonaceae, Xue et al.
(2020) reported that pollen aggregation is associated with
reduced diversification. This could be associated with the
potentially negative effects of pollen aggregation, such as
the reduction in the number of pollen recipients reached by a
pollen donor (reduced pollen “carryover,”exacerbated when
geitonogamy occurs) and consequently the reduced genetic
variation in the pollen delivered (Harder & Johnson, 2008,
and references therein). Attempts to assess correlations
between evolutionary diversification rates and functional
traits are complicated, however, by possible trait correla-
tions: pollen aggregation, for example, is acknowledged
to be closely aligned with anther septation (Tsou &
Johnson, 2003).
4.2 Stamen abscission and secondary pollen presentation
The stamens in Annonaceae flowers partially abscise as the
anthers begin to dehisce, although each stamen remains
suspended within the pendent flower by the tracheary
elements of its vascular tissue (Fig. 3C; Endress, 1985).
Pollinator movements within the flower are likely to dislodge
the stamens, further assisting pollen release.
Because of the pendent orientation of most Annonaceae
flowers, the apically convergent or connivent pollination
chamber often forms an inverted mitriform structure that
captures fallen pollen grains (Fig. 1B). Dehisced pollen and
abscised stamens have been observed to collect in the floral
chamber in Pseuduvaria mulgraveana (Pang et al., 2013),
where they mix with accumulated nectar; the pollinators
inadvertently collect pollen grains as they consume the
nectar. In some species (e.g., Mitrephora: Weerasooriya &
Saunders, 2010) the adaxial surface of the floral chamber is
adorned with hairs that might assist with the retention of
dehisced pollen (Fig. 1B). These observations are therefore
indicative of secondary pollen presentation, in which pollen is
transferred from the thecae to an intermediary organ prior
to collection by the pollinator (Howell et al., 1993; Yeo, 1993),
and could serve to prolong the period during which pollen is
available.
4.3 Anthesis duration
Anthesis is generally defined as the period during which a
flower is sexually functional, with the onset of anthesis
correlated with the opening of the perianth. This concept is
problematic with respect to Annonaceae, however, as the
sepals and petals often separate very early in development
(although the inner petals sometimes remain apically
convergent or connivent to form a partially enclosed floral
chamber). Annonaceae flowers are protogynous, therefore
the beginning of anthesis is generally determined based on
the formation of stigmatic exudate—used as a proxy for
stigmatic receptivity and hence indicative of the pistillate
phase—and with the end of anthesis (and the staminate
phase) coinciding with petal abscission. The pistillate and
Fig. 2. Floral phenology diagrams for selected Annonaceae species, showing schematic graphs of pollinator circadian rhythms
(in gray) in relation to pistillate and staminate functions (pink and blue bars, respectively). Interfloral pollinator movements
shown by large arrows: movement from pistillate to staminate flowers shown as arrows that transition from pink to blue; and
reverse movement as arrows that transition from blue to pink. A, Polyalthia suberosa, which has anthesis over c. 48 h and
hence potentially has three coetaneous cohorts of flowers: only approximately half of the beetle pollinators departing from
post‐staminate flowers will move directly to pistillate‐phase flowers. B, Desmos chinensis, which has abbreviated anthesis over
c. 27 h and hence only has two coetaneous cohorts of flowers: most beetle pollinators departing from post‐staminate flowers
move directly to pistillate‐phase flowers, although those leaving post‐pistillate flowers cannot directly access staminate‐phase
flowers. C, Goniothalamus tapisoides, which has abbreviated anthesis over c. 23 h and circadian pollinator trapping, with two
coetaneous cohorts of flowers: most of the beetle pollinators departing from post‐staminate flowers are able to move directly
to pistillate‐phase flowers; beetles cannot leave the floral chamber at the end of the pistillate phase, however, due to circadian
trapping.
379Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

Table 2 Prolonged (>3 days) and abbreviated (23–30 h) anthesis in Annonaceae species, with correlated traits (primary pollinator type, floral sex expression, presence/absence
of floral synchrony, and presence/absence of circadian pollinator trapping)
Anthesis type and species
Anthesis
duration Pollinators
Floral sex
expression
Floral
synchrony
Circadian
trapping
†
Reference(s)
Prolonged anthesis
Asimina obovata 5–8 days Scarab beetles Bisexual ? –Norman & Clayton, 1986
Asimina parviflora 6–12 days Flies Bisexual ? –Norman et al., 1992
Asimina pulchella 8–9 days Flies, beetles, thrips Bisexual ? –Norman, 2003
Asimina pygmaea 3–4 days Scarab beetles Bisexual ? –Norman & Clayton, 1986
Asimina rugelii 5–6 days Flies, beetles, thrips Bisexual ? –Norman, 2003
Asimina triloba 6–8 days Flies, beetles Bisexual +–Willson & Schemske, 1980;
Rogstad, 1993
Disepalum anomalum c. 25 days Meliponine bees Bisexual ––Li et al., 2016
Disepalum pulchrum c. 20 days Drosophilid flies, nitidulid
beetles
Bisexual ––Li et al., 2016
Fissistigma oldhamii c. 5 days Drosophilid flies, nitidulid
beetles
Bisexual ––Lau JYY & Saunders RMK,
unpublished data
Monodora myristica 11–13 days Flies? Bisexual ? –Lamoureux, 1975
Popowia pisocarpa c. 4 days Thrips Bisexual ? –Momose et al., 1998
Sapranthus palanga “Many”days Tenebrionid beetles Bisexual ? –Schatz, 1987
Uvariopsis bakeriana c. 4 days Flies Unisexual ––Gottsberger et al., 2011
Uvariopsis congolana 4–5 days Flies Unisexual ––Gottsberger et al., 2011
Abbreviated anthesis
Annona mucosa c. 24 h Beetles Bisexual +–Murray & Johnson, 1987,
as “Rollinia jimenezii”
Artabotrys blumei c. 27 h Beetles? Bisexual –+Chen et al., 2020
Dasymaschalon
trichophorum
c. 26 h Curculionid beetles Bisexual ++Pang & Saunders, 2014;
Lau et al., 2017a
Desmos chinensis c. 27 h Nitidulid beetles Bisexual +–Pang & Saunders, 2015;
Lau et al., 2017a
Friesodielsia borneensis c. 26 h Curculionid beetles, nitidulid
beetles, staphylinid beetles
Bisexual ++Lau et al., 2017a
Goniothalamus suaveolens c. 25 h Curculionid beetles, nitidulid
beetles
Bisexual –+Lau et al., 2016
Goniothalamus tapisoides c. 23 h Curculionid beetles, nitidulid
beetles
Bisexual –+Lau et al., 2016, 2017a
Pseuduvaria froggattii <1 day Flies Unisexual ––Silberbauer‐Gottsberger
et al., 2003
Unonopsis guatterioides
‡
c. 30 h Bees Bisexual ? –Oliveira et al., 2017
Species with standard anthesis duration (36–54 h) are not listed; +, present; ‐, absent; ?, unknown. †Many negative assessments for circadian pollinator trapping are based on
interpretation of floral morphology rather than empirical study. ‡According to Oliveira et al. (2017) and Gottsberger et al. (2018), this is unlikely to be conspecific with the
Unonopsis guatterioides populations studied by Carvalho & Webber (2000).
380 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

staminate phases are often temporally separated by a non‐
sexual interim phase, although this is sometimes obscured by
the continued retention of stigmatic exudate.
Anthesis duration in most hermaphroditic‐flowered Anno-
naceae species is 36–54 h, extending over 2 or 3 days
depending on the timing of the onset of the pistillate phase
(e.g., Polyalthia suberosa: Fig. 2A; Saunders, 2012). Consid-
erably longer anthesis durations, however, are observed in
several phylogenetically disparate genera (Table 2): most of
these examples are pollinated by flies (or a combination of
flies and small beetles), although other examples include
pollination by meliponine bees, scarab beetles, tenebrionid
beetles, and thrips. These changes in flowering rhythm could
represent adaptations to the activity patterns of pollinators
to optimize pollination efficiency (Gottsberger, 2014). Flies,
for example, are irregular floral visitors, typically making very
brief visits to each flower (Fægri & van der Pijl, 1979;
Saunders, 2012) and hence pollination success is likely to be
promoted by prolonged anthesis.
In striking contrast, however, several Annonaceae species
exhibit abbreviated anthesis over 23–30 h (Table 2; e.g.,
Desmos chinensis and Goniothalamus tapisoides: Figs. 2B, 2C).
Although short anthesis is sometimes correlated with
unisexual flowers as there is only a single sexual functional
phase (e.g., within 1 day in Pseuduvaria froggattii:
Silberbauer‐Gottsberger et al., 2003), most examples are of
species with hermaphroditic flowers. Dasymaschalon,
Desmos, and Friesodielsia are phylogenetically closely related
(Guo et al., 2017a) and their abbreviated anthesis is therefore
likely to be synapomorphic for the entire clade (Lau
et al., 2017a); the other four hermaphroditic genera with
abbreviated anthesis (Annona,Artabotrys,Goniothalamus,
and Unonopsis) are phylogenetically disparate (Guo
et al., 2017b), however, and hence there are at least five
independent origins for abbreviated anthesis in the family.
Most Annonaceae species with standard anthesis duration
(36–54 h) have three coetaneous cohorts of flowers over
3 days: flowers entering their pistillate phase will co‐occur
with flowers that are a day older, which are entering their
staminate phase, and flowers that are 2 days older, in which
the staminate phase is ending (e.g., Polyalthia suberosa:
Fig. 2A). Pollen‐laden pollinators departing from post‐
staminate flowers are therefore likely to be attracted to
two different cohorts of flowers. Assuming that the different
cohorts are equally common in the population, pollinators
are equally likely to be attracted towards staminate‐phase
flowers as they are to pistillate‐phase flowers (Fig. 2A).
Species with abbreviated anthesis (e.g., Desmos chinensis,
with anthesis over c. 27 h: Fig. 2B) only have two coetaneous
cohorts within the population, and hence pollinators leaving
post‐staminate flowers are much more likely to be attracted
directly to flowers entering the pistillate phase, thereby
increasing pollination efficiency.
Abbreviated anthesis is furthermore closely allied with
pistillate/staminate‐phase floral synchrony (discussed in
Section 3.3), in which all flowers borne on an individual
undergo the same sexual phase concurrently. Although there
are very few studies of floral synchrony in the family, many of
the species shown to have abbreviated anthesis also display
synchrony (Table 2). Synchronous species showing standard
anthesis duration require a “flower‐free”day every third day
to ensure that pollen from a staminate‐phase flower cannot
pollinate a pistillate‐phase flower on the same plant the
following day; synchronous species with abbreviated an-
thesis can therefore achieve much more rapid turnover of
flowers (with a new flower forming on alternate days rather
than every fourth day) and hence can increase seedset
without undermining xenogamy. The species shown to have
abbreviated anthesis in the absence of floral synchrony
adopt a different strategy to avoid geitonogamy, with
relatively few flowers borne concurrently (e.g., Artabotrys
blumei: Chen et al., 2020; Goniothalamus suaveolens and
G. tapisoides: Lau et al., 2016), although this negatively
impacts seedset.
4.4 Circadian pollinator trapping
Most Annonaceae species have hermaphroditic, protogynous
flowers that are functional over 36–54 h and are typically
pollinated by crepuscular insects that exhibit a bimodal
endogenous circadian rhythm. The Carpophilus beetles
(Nitidulidae) that pollinate Polyalthia suberosa (Lau
et al., 2017a), for example, have an evening activity peak
that coincides with the onset of the pistillate floral phase
(Fig. 2A); the beetles then remain relatively immobile within
the flower until their next activity peak the following
morning, which is aligned with the end of the pistillate
phase. The staminate floral phase is similarly correlated with
beetle activity levels, although pollen‐laden beetles often
remain in the flower until petal abscission, which is aligned
with the onset of the pistillate phase in other flowers.
Goniothalamus species have an abbreviated anthesis of
approximately 23–25 h (Lau et al., 2016) and have firmly
connivent inner petals that form a robust pollination
chamber (Type III sensu Saunders, 2010), in which the
narrow basal apertures between the inner petals are
periodically blocked by movements of the alternately
positioned outer petals. This floral arrangement enables
“circadian trapping”of pollinators within the flower: the
curculionid and nitidulid beetles that pollinate G. tapisoides,
for example, are attracted during their morning activity peak
to open flowers that are entering their pistillate phase
(Fig. 2C); the beetles then remain largely immobile until their
next activity peak, but at this point the beetles are unable to
leave the flower due to closure of the floral apertures (Lau
et al., 2017a). The beetles are finally released from the flower
as the petals abscise after the end of the staminate phase,
with their departure closely aligned with their activity peak
the following morning, coincident with the onset of the
pistillate phase in other flowers. This pollinator trapping
mechanism is therefore heavily dependent on the close
alignment of the timing of petal movements with the
endogenous activity patterns of the pollinators (Lau
et al., 2017a).
Similar patterns of pollinator trapping have been observed
in Artabotrys (Figs. 3D, 3E; Chen et al., 2020) and Friesodielsia
(Lau et al., 2017a), whereas the mechanism reported in
Dasymaschalon differs slightly as the flowers of this genus
have only a single whorl of three petals. The three petals in
Dasymaschalon are inferred to be homologous with the outer
petals of other Annonaceae (Guo et al., 2018), with the floral
chamber opening and closing due to subtle lateral growth of
the petals (Lau et al., 2017a). Although phylogenetic
381Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

reconstructions indicate that Dasymaschalon and Friesodielsia
are sister clades (Guo et al., 2017a), the different mechanisms
underlying circadian trapping in the two genera preclude
inferences regarding homology and it is therefore unclear
whether evolutionary convergence has occurred (Lau
et al., 2017a).
In addition to Artabotrys,Dasymaschalon,Friesodielsia, and
Goniothalamus, assessments of floral structure in other
Annonaceae genera suggest that circadian trapping might also
function in Cyathocalyx,Drepananthus,Mitrella,Neostenanthera,
and Pseudartabotrys,aswellasinTrivalvaria macrophylla (Xue
et al., 2020). Circadian trapping isthereforelikelyto have evolved
multiple times in the family, suggesting that it endows a major
selective advantage. In a recent study of the impact of floral
traits on evolutionary diversification in Annonaceae, Xue et al.
(2020) found that pollinator‐trapping lineages are associated
with significantly accelerated net diversification.
The genera that adopt circadian trapping invariably possess
abbreviated anthesis, with the occurrence of the latter trait in
Desmos –which is sister to the Dasymaschalon‐Friesodielsia clade
(Guo et al., 2017a) but which does not trap pollinators
(Lau et al., 2017a) –suggesting that short anthesis evolved
first and could be a prerequisite for circadian trapping. Species
with abbreviated anthesis but lacking circadian trapping (e.g.,
Desmos chinensis:Fig.2B)areliabletolosepollinatorsthat
exhibit an evening activity peak; pollination exclusively by
morning‐active unimodal pollinators is therefore likely to be
more efficient. Species that trap pollinators (e.g., G. tapisoides:
Fig. 2C) benefit by utilizing an expanded range of pollinators,
irrespective of whether they exhibit unimodal (morning or
evening) or bimodal circadian rhythms (Lau et al., 2017a). Chen
et al. (2020) furthermore noted that species with circadian
trapping can stop scent emission earlier and hence can
potentially avoid detection by floral antagonists (Schiestl, 2015).
5 Floral Adaptations that Optimize
Fertilization Success
5.1 Extragynoecial compita
Unequal deposition of pollen loads onto the stigmas of
apocarpous flowers can result in unbalanced ovule fertilization
Fig. 3. Pollen structure, floral morphology, and pollen–stigma interactions in selected Annonaceae species. A, Pollen tetrads
of Pseuduvaria macrocarpa, with a pollen‐connecting thread linking adjacent tetrads. Scale bar =10 μm. B, Detail of a non‐
sporopollenin pollen‐connecting thread in Pseuduvaria macrocarpa. Scale bar =2μm. C, Stamen abscission in a staminate‐
phase flower of Disepalum anomalum, showing stamens suspended by tracheary threads (arrowed in inset photograph). D,
Artabotrys hexapetalus, with outer petals raised to expose the apertures (arrowed) between the inner petals, opening the
floral chamber. E, Artabotrys hexapetalus, with closed floral chamber. F, Intercarpellary growth of pollen tube through the
extragynoecial compitum of Goniothalamus tapisoides. PT, pollen tube; SE, stigmatic exudate; St, stigma. Scale bar =250 μm.
(Photographs: A, B, Y.C.F. Su; C, P.‐S. Li; D, E, J. Chen; F, J.Y.Y. Lau, reproduced from Lau et al., 2017b).
382 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

and hence the failure of some carpels to bear seed (e.g.,
Liriodendron chinense, Magnoliaceae: Huang & Guo, 2002).
Previous research on the reproductive biology of early
divergent angiosperms has identified several structural
adaptations and mechanisms that have enabled this dis-
advantageofapocarpytobeovercome.Althoughpollentube
growth is typically channeled through specialized routes leading
to the micropyle, the formation of extragynoecial compita
(EGC) enable intercarpellary growth of pollen tubes, thereby
liberating species from the constraints imposed by apocarpy
(Endress, 1982; Wang et al., 2012). Three distinct forms of EGC
have been identified (Wang et al., 2012): (i) suprastylar EGC, in
which pollen tubes grow between stigmas that are either
closely appressed or connected by stigmatic secretions; (ii)
extrastylar EGC, in which the pollen tubes do not penetrate the
stigma but instead approach the ovule through an alternative
route (such as the staminal filament); and (iii) infrastylar EGC, in
which the pollen tubes either extend down the pseudostyle
before growing between carpels through an intercarpellary
exudate, or else they grow to the base of one carpel before
traversing the receptacle in order to penetrate another carpel.
Of these three types, only suprastylar and infrastylar EGC are
potentially applicable to Annonaceae.
A functional suprastylar EGC has previously been observed
in three Annonaceae genera (Annona: Lora et al., 2011;
Asimina: Losada et al., 2017; and Goniothalamus: Lau
et al., 2017b), with intercarpellary growth of pollen tubes
occurring through a stigmatic secretion (Fig. 3F). These three
genera are phylogenetically close (all belong to a single clade
within the tribe Annoneae: Chatrou et al., 2012; Guo
et al., 2017b) and all develop copious exudate when the
stigmas are receptive. Although Lau et al. (2017b) suggested
that suprastylar EGC might be widespread in Annonaceae
(and hence a key functional floral trait for the entire family),
it is also possible that it is phylogenetically restricted within
the tribe Annoneae and/or species with similarly extensive
stigmatic exudate formation. Significantly, however, supra-
stylar EGC has also been reported from Eupomatia (in
the sister family Eupomatiaceae) as well as the closely
related genus Galbulimima (Himantandraceae) (Igersheim &
Endress, 1997) and various Laurales (Endress & Igersheim,
1997), and hence its origin might have predated the origin of
Annonaceae.
In the absence of copious stigmatic exudate, it is
anticipated that a functional EGC could be achieved in
Annonaceae if the stigmas of neighboring carpels are closely
appressed to form a contiguous “stigmatic head.”The
extensive variation in stigma size and shape observed in
Goniothalamus (Saunders, 2002) is likely to represent distinct
adaptations favoring suprastylar EGC, with very broad
stigmas (e.g., G. costulatus: Saunders, 2002) enabling greater
physical contact between neighboring carpels, whereas
spatially divergent stigmas (e.g., G. tapisoides: Lau
et al., 2017b) presumably rely on copious exudate to enable
intercarpellary growth of pollen tubes.
Several Sagittaria (Alismataceae) species have been
reported to possess a functional infrastylar EGC in which
pollen tubes traverse the floral receptacle before entering
the ovule (Wang et al., 2002, 2006, 2012; Huang, 2003;
Endress, 2011). In these cases, ovule placentation is invariably
basal, presumably to enhance pollen tube access through the
micropyle. Basal placentation necessarily limits the number
of ovules per carpel, although this might be compensated for
by an increase in the number of carpels per flower; it can be
hypothesized that such an increase in carpel number would
further reinforce the selective advantage of infrastylar EGC.
The widespread occurrence of major Annonaceae lineages
that exclusively or predominantly show basal placentation
(including several species‐rich genera such as Annona,
Artabotrys,Duguetia,Goniothalamus, and Guatteria: van
Heusden, 1992) raises the possibility that a functional
infrastylar EGC exists in the family. This has never been
investigated, however.
5.2 Increased ovule number per carpel
Increased ovule number per carpel is potentially an
alternative strategy for promoting seedset in apocarpous
flowers. This might be correlated with a reduction in carpel
number per flower and/or flower number per plant to avoid
exceeding the optimal fruit‐carrying capacity of the plant.
Increased ovule numbers are observed in some species of
Fissistigma and Uvaria (Annonaceae), with up to 20 and 30
ovules, respectively (van Heusden, 1992). The functional
significance of this in relation to seedset has never been
evaluated, however.
5.3 Syncarpy
Although syncarpy is widespread among angiosperms and
often cited as a key evolutionary innovation, it is very rare
within Annonaceae: complete syncarpy has previously only
been reported from the sister genera Isolona and Monodora
(Guédès & Le Thomas, 1981; Couvreur et al., 2008; Couvreur,
2009). Two contrasting hypotheses have been developed to
explain the evolution of syncarpy (Endress, 1990): the
“multiplication hypothesis,”in which the gynoecium is
initially reduced to a single carpel, but with a subsequent
increase in carpel number due to branching of the carpel
primordium; and the “fusion hypothesis,”involving congen-
ital fusion of separate carpels. Developmental studies of the
gynoecium in Monodora crispata (Leins & Erbar, 1982)
revealed similarities with that of a single carpel, supporting
the former hypothesis. Anatomical interpretations subse-
quently led Deroin (1997) to favor the fusion hypothesis,
however, as he observed gynoecial vascular patterns
consistent with an origin from 6 to 14 fused carpels (see
also Deroin, 1985). Molecular phylogenetic reconstructions of
the tribe Monodoreae furthermore indicate that the
Isolona–Monodora lineage is nested within a multicarpellate
clade (Couvreur et al., 2008), adding further support to
Deroin's interpretation.
The Annonaceae genus Cyathocalyx s.str. is weakly
supported as sister to Drepananthus (formerly included in
Cyathocalyx), and collectively sister to Cananga (Surveswaran
et al., 2010; Guo et al., 2017b). Cyathocalyx is characterized by
an inferred synapomorphy of a large peltate stigma, whereas
Drepananthus and closely related genera such as Cananga
possess smaller ellipsoid or obconical stigmas (Wang &
Saunders, 2006; Surveswaran et al., 2010) that are more
typical for Annonaceae. The large peltate stigma in
Cyathocalyx resembles a cluster of fused stigmas, raising
the possibility of an independent origin of syncarpy:
significantly, the solitary pistil in Cyathocalyx flowers
383Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

possesses up to 38 ovules, arranged in up to four rows
(Wang & Saunders, 2006); Drepananthus species, in contrast,
have up to 32 carpels per flower, with carpels possessing
fewer ovules (generally up to six, but rarely 20) that are
uniseriate or biseriate (Surveswaran et al., 2010; Wang RJ &
Saunders RMK, unpublished data). Anatomical preparations
of Cyathocalyx flowers of several species (Endress, 2008;
Chen Y & Saunders RMK, unpublished data) invariably reveal
multiple vascular traces in transverse sections through the
pistil; although this is consistent with a compound structure
derived from the fusion of several carpels, multiplication of
vascular traces might have evolved to provide an adequate
xylem and phloem supply to the developing fruit after
fertilization (Endress, 2019).
Partial syncarpy has been reported in the Annonaceae
genera Annona,Anonidium,Cananga,Fusaea, and Pseudarta-
botrys (Deroin, 1988, 1997; van Heusden, 1992; Deroin &
Bidault, 2017). Although the hypothesis remains untested, it
is possible that this incomplete syncarpy endows the same
benefits as full syncarpy –allowing pollen grains deposited
on any stigma to fertilize any ovule in the flower –while
enabling the development of separate fruit monocarps (fruit
units derived from individual carpels) that can be dispersed
individually, and hence avoiding the dispersal constraints
imposed by large fruits (discussed further in Section 8).
5.4 Self‐pollination
Annonaceae species lack any biochemically mediated self‐
incompatibility mechanism and hence are capable of
autogamous and geitonogamous self‐fertilization (Pang &
Saunders, 2014). As discussed in Section 3, Annonaceae have
various adaptations that promote xenogamy, and conse-
quently most species predominantly out‐cross. The most
widespread of these adaptations is protogyny, in which
pistillate floral function precedes stamen dehiscence. Several
studies, however, have reported temporal overlap between
the pistillate and staminate phases (e.g., several Asimina
species: Norman & Clayton, 1986; Norman et al., 1992;
Norman, 2003, as “Deeringothamnus”;Disepalum anomalum:
Li et al., 2016; and Uvaria concava: Silberbauer‐Gottsberger
et al., 2003). In some cases, the overlap between phases
might represent an adaptation to specific pollination
systems: the meliponine bees that pollinate D. anomalum,
for example, are rewarded with pollen and hence would only
be attracted to pistillate‐phase flowers after anther
dehiscence (Li et al., 2016). In other species, however, the
overlap between the two phases might be an adaptation to
ensure seedset when cross‐pollination is limited, possibly
because of a population bottleneck or because of the
inadequate availability of effective pollinators. Population
genetic studies of such species are lacking, however.
5.5 Apomixis
Controlled pollination experiments involving the Annonaceae
species Cymbopetalum brasiliense (Braun et al., 2011) revealed
significant fruitset in tests for spontaneous selfing (in which
preanthetic flowers were bagged to exclude pollinators) and
apomixis (in which preanthetic flowers were emasculated
and then bagged). Although the results of the latter
experiment convincingly indicate apomixis, fruitset arising
from the former experiment was also probably due to
apomixis as autogamy was precluded by the abscission of
stigmas at the end of the pistillate phase prior to anther
dehiscence. Cymbopetalum brasiliense is the only known
example of apomixis in Annonaceae, although in the absence
of comparable studies it is unclear how widespread this
phenomenon could be. It is probably significant that
C. brasiliense is a triploid (Morawetz, 1986) and hence likely
to be incapable of forming functional gametes. Gameto-
phytic apomicts are almost invariably polyploids (Bicknell &
Koltunow, 2004).
6 Floral Adaptations that Enhance Pollen
Competition
Pollen competition often occurs in angiosperms as the pollen
tubes penetrate the carpel, enabling favored genotypes to
be selected by stigmatic and stylar tissues (Moore &
Pannell, 2011). Although it could be argued that the minimal
pollen loads typically carried by the beetles that pollinate
most Annonaceae flowers are unlikely to generate conditions
that would promote intense pollen competition, no attempt
has previously been made to assess the significance of pollen
competition in the family.
Many Annonaceae genera are characterized by enlarged
stigmas. Although this might promote suprastylar EGC by
enhancing physical contact between neighboring stigmas
and/or be associated with increased exudate formation, it is
also possible that enlarged stigmas directly increase
opportunities for pollen deposition. Large stigmas would
probably minimize pollen selection, however, as fertilization
would be heavily influenced by the random location of pollen
deposition on the stigma (Armbruster, 1996): pollen grains
deposited close to the base of the stigma would be at a
selective advantage as their pollen tubes would have a
shorter distance to grow, and hence successful fertilization
would not be as dependent on paternal vigor.
Stigma morphology is very diverse in Annonaceae,
however, with many taxa, including Fissistigma and
Goniothalamus (Fig. 1E), often possessing highly elongated
stigmas. Lau et al. (2017b) reported that entry of pollen tubes
is restricted to the apex of the stigma in G. parallelivenius.In
such cases, the elongated stigmatic stalk closely resembles
the style in derived angiosperm lineages; this pseudostyle
can therefore be hypothesized to function as a true style and
possibly enhance pollen competition. The most extreme
pseudostylar adaptation in Annonaceae is likely to be
Goniothalamus flagellistylus (Tagane et al., 2015), which has
a stigma/pseudostyle that is c. 8.5 mm long, greatly
exceeding the length of the ovary (1.4–1.7 mm).
Pollen competition in the pseudostyle might be intensified
in cases where the pistillate phase is prolonged (e.g., c. 44 h
duration in Fissistigma oldhamii: Lau JYY & Saunders RMK,
unpublished data). Pollen competition might also be
promoted in Annonaceae if the pollen tubes are directed
through an alternative route to the ovule, as in species with
putative infrastylar EGC (discussed in Section 5.1).
As discussed in Section 3.1, most Annonaceae species have
bisexual, protogynous flowers, in which the pistillate and
staminate phenological phases are separated by a non‐sexual
interim phase that effectively precludes autogamy (Pang &
384 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

Saunders, 2014). There are several recorded cases (discussed
in Section 5.4), however, in which the pistillate phase
appears to overlap temporally with the staminate phase.
This overlap might be an adaptation to ensure seedset when
cross‐pollination is limited. As xenogamous pollen would be
deposited on the stigma earlier than autogamous pollen,
however, xenogamous pollen would possibly be selected
passively simply by virtue of its earlier germination.
7 Floral Adaptations that Protect Against
Herbivory
Annonaceae species show several morphological and
chemical adaptations that can most easily be explained by
reference to protection against herbivory rather than as
adaptations to promote pollination or fertilization. Although
the partially enclosed floral chambers alluded to earlier
(Section 2.2) are generally weakly formed from convergent
or connivent petals that can easily be separated by
phytophagous insects, several Annonaceae genera are
characterized by much more robust chambers. In these
genera (e.g., Artabotrys: Chen et al., 2020; Dasymaschalon:
Guo et al., 2018; Friesodielsia: Guo et al., 2017a; and
Goniothalamus: Lau et al., 2016) the floral chambers function
as pollinator traps (discussed in Section 4.4) and have petals
that are very intimately connected, with narrow apertures.
These apertures are sufficiently small to prevent access by
larger and potentially destructive floral visitors, while
allowing entry of smaller effective pollinators.
Annonaceae stamens show polycyclic phyllotaxis (Endress
& Doyle, 2007), with the stamens closely appressed prior to
anther dehiscence. The staminal connectives are often
elongated and are typically either dilated laterally to cover
the apex of the thecae or else lack such an extension
(“uvarioid”and “miliusoid”stamens, respectively, sensu
Prantl, 1891). The uvarioid connective presumably protects
the thecae from phytophagous insects prior to anther
dehiscence, with the staminal connectives forming a closely
appressed tessellated protective shield (Saunders, 2010).
Some Xylopia species (in sections Stenoxylopia p.p. and
Xylopia: Johnson & Murray, 2018) are unusual in Annonaceae
in possessing a deeply invaginated structure that surrounds
the ovary. This structure has variously been interpreted as an
extension of the floral receptacle (Kramer, 1969;
Deroin, 1989) or as a “staminal cone,”primarily formed
from the fusion of staminal filaments, with a limited
contribution from sepal and petal tissues (Verdcourt, 1971;
Dias et al., 1998). Irrespective of the anatomical interpreta-
tion, the cone is likely to function by protecting the ovary
from insects, mirroring the role of the hypanthium in derived
angiosperm lineages.
Floral scents not only function as olfactory cues to attract
pollinators, but often also incorporate volatiles that
discourage potential herbivores (Pichersky & Gershenzon,
2002), with pollinator‐attracting scents hypothesized to have
evolved from ancestral defense volatiles (Pellmyr &
Thien, 1986). Naphthalene has been widely reported as a
component of floral scent in Annonaceae (e.g., Annona:
Jürgens et al., 2000, as “Rollinia”;Desmos: Pang &
Saunders, 2015; Uvariopsis: Gottsberger et al., 2011; and
Xylopia: Jürgens et al., 2000; Ratnayake et al., 2007),
although there is disagreement whether it protects flowers
against herbivores (Azuma et al., 1996) or whether it is of
anthropogenic origin (Jürgens et al., 2000). Limonene, which
possibly functions as an insect repellent (El‐Sayed, 2019, and
references therein), has similarly been reported from the
floral scents of Desmos (Pang & Saunders, 2015), Duguetia
(Jürgens et al., 2000), and Unonopsis (Teichert et al., 2009;
Oliveira et al., 2017; Gottsberger et al., 2018). Annonaceae
floral volatiles that potentially function as herbivore
deterrents require further study.
8 Floral Constraints Imposed by Adapta-
tions that Promote Seed Dispersal
Although Annonaceae adopt various mechanisms to pro-
mote xenogamy, paternal diversity amongst seeds within a
single fruit is nevertheless likely to be limited due to the
restricted number of pollinator visits to staminate‐phase
flowers prior to the pollinator's arrival at a pistillate‐phase
flower (e.g., Annona crassiflora: de Almeida‐Júnior et al., 2018)
and pollen aggregation (discussed in Section 4.1). The
retention of separate, unfused carpels in flowers could
compensate for this by enhancing the geographical spread of
seed paternal types by enabling the independent dispersal of
fruit monocarps.
Flowers can be regarded as immature fruits in which the
ovules are yet to be fertilized. Floral structure inevitably
imposes constraints on fruit structure and vice versa:
anatomical features of the flower that might have little
selective advantage during anthesis can theoretically
promote fruit and seed dispersal. Although there are clear
disadvantages associated with floral apocarpy, as detailed
above, it can be hypothesized that the retention of separate,
unfused carpels in flowers might subsequently serve to
enhance patterns of seed dispersal. Four main functional
fruit morphologies are identified in Annonaceae, discussed
below.
Whole fruits as dispersal units—Some Annonaceae species
develop relatively large fruits in which the anatomically
separate monocarps (fruit units derived from individual
carpels) become closely appressed (functionally syncarpous)
due to the limited development of monocarp stalks; in these
cases, the dispersal unit is the entire fruit, with dispersal by
relatively large frugivores such as primates (e.g., Duguetia,
p.p.: Maas et al., 2003; Goniothalamus, p.p.: Tang et al., 2015).
Some genera, including Annona and Duguetia, p.p. (van
Setten & Koek‐Noorman, 1992), undergo differing degrees of
post‐fertilization fusion of carpels, resulting in “pseudosyn-
carpy,”in which the entire fruit is again dispersed as a
single unit.
Monocarps as dispersal units—Most Annonaceae species
have fruits in which the base of each monocarp is elongated
to form a stipe that ensures separation of monocarps at
maturity; in these cases, individual monocarps sometimes
mature at different rates and are dispersed separately, either
by birds that swallow the monocarps whole and defecate the
seed intact, or by primates that spit out the seeds. A few
genera (e.g., Disepalum: Li et al., 2015) possess “carpo-
phores”that closely resemble stipes, but which are
385Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

extensions of the fruit receptacle and hence are not
homologous; the convergent evolution of carpophores and
stipes presumably reflects the functional importance of
monocarp stalks.
Single‐seeded monocarp segments as dispersal units—Mono-
carps of many species in the Monanthotaxis–Dasymaschalon–-
Desmos clade are elongated and moniliform, with constrictions
between seeds (Guo et al., 2017a); each monocarp
ripens progressively from apex to base, with single‐seeded
segments removed sequentially by avian frugivores as the fruit
ripens (Wang et al., 2012). Increasing the temporal separation
between seed dispersal events possibly minimizes potential
density‐dependent seedling mortality due to over‐crowding,
fungal infection, seed predation, or unfavorable germination
site. Although such dispersal units are unique to the
Monanthotaxis–Dasymaschalon–Desmos clade within Annona-
ceae, it parallels the lomentum observed in some legumes
(Spjut, 1994).
Seeds as dispersal units—The Annonaceae genera Cardiope-
talum,Cymbopetalum,Trigynaea (Johnson & Murray, 1995),
and Xylopia (Johnson & Murray, 2018) possess dehiscent
monocarps that split along a dorsal suture to expose bird‐
dispersed seeds with a brightly colored aril or sarcotesta (e.g.,
Coates‐Estrada & Estrada, 1988). Some Anaxagorea species
have dehiscent monocarps with ballistic dispersal of non‐
arillate seeds over distances up to 5m (Gottsberger, 2016).
Although it can be hypothesized that many of the floral
adaptations highlighted in this review (such as EGC,
increased ovule number per carpel, syncarpy, and the
formation of pollen polyads) would promote pollination
efficiency by enhancing seedset, it is also likely that this
would be achieved at the expense of paternal genetic
diversity. This lack of genetic diversity in seeds might be
compensated for by ensuring maximum spatial distance
between seedlings after dispersal. In species in which
the individual monocarp is the dispersal unit, single‐seeded
monocarps would be most effective in maximizing distance
between seedlings. Although seeds in multiseeded mono-
carps would probably be dispersed together, such mono-
carps are likely to develop from flowers that are pollinated
by insects that undertake more extensive interfloral move-
ments, resulting in seeds with greater paternal diversity:
Fissistigma oldhamii, for example, has c. 10 seeds per
monocarp but is predominantly pollinated by drosophilid
flies (Lau JYY & Saunders RMK, unpublished data). Beetle‐
pollinated species with multiseeded monocarps often
disperse their seeds separately, either as single‐seeded
monocarp segments (in the Monanthotaxis–Dasymaschalon–-
Desmos clade: Wang et al., 2012), or as seeds in dehiscent
monocarps (e.g., Xylopia: Johnson & Murray, 2018). Large
fruits that are dispersed as a single unit (e.g., Annona and
Goniothalamus, p.p.) are hypothesized to be less likely to
evolve in lineages with multiple ovules per carpel.
In a recent analysis of evolutionary diversification in
Annonaceae, Xue et al. (2020) used comparative phyloge-
netic methods to identify significant rate shifts and their
correlations with specific traits. They revealed a strong
correlation with seed dispersal unit: dispersal by single‐
seeded monocarp segments was associated with the highest
diversification rate, followed by single monocarps, and then
entire fruits. Although these results conform to the
predictions discussed above regarding adaptations that
maximize the spatial distance between seedlings, Xue et al.
(2020) also showed that direct seed dispersal from dehiscent
monocarps was surprisingly correlated with the lowest
diversification rate. These analyses merely identify correla-
tions without determining causality, however, and might also
reflect correlations between traits that are subject to very
different selective pressures.
9 Conclusions
Annonaceae possess visual, olfactory, and thermogenic floral
traits that attract pollinators, often either with a reward or
by deceit. The petals provide a key visual cue, with
crepuscular insect pollinators able to perceive a broad
spectral range despite the low ambient light intensities at
dawn and dusk. The flowers are typically highly fragrant,
emitting scents that are either sweet (advertising a nectar
reward) or deceptively mimicking the pollinator's food
substrate or brood site. Many Annonaceae flowers are
thermogenic, with beetle pollinators able to locate the
flowers by sensing the heat using their IR sensilla.
Annonaceae flowers offer various pollinator rewards,
including enclosed environments within the floral chamber
(providing a reproductive tryst site and protection from
predators) and stigmatic or petal‐derived nectar. Other
rewards include floral heat (providing the energy required
by ectothermic pollinators), floral scent (with male euglos-
sine bees reported to collect floral scent volatiles from
Unonopsis flowers to attract female bees), and petals as
brood sites (directly or indirectly providing a food source for
emerging larvae). In some cases, flowers that appear to
serve as potential brood sites are likely to be deceptive.
Annonaceae species lack a biochemically mediated self‐
incompatibility mechanism and hence are capable of self‐
fertilizing. Most reproduction in the family is nevertheless
xenogamous, with out‐crossing promoted by various means,
including: (i) temporal separation of sexual function (as
protogyny); (ii) spatial separation of anthers and stigmas
(herkogamy); (iii) synchrony of floral reproductive function
(synchronous dichogamy); and (iv) partial or complete
separation of sexes (dicliny). Despite the prevalence of
xenogamy, most ovules in Annonaceae flowers are likely to
be fertilized by the same pollen donor, significantly
constraining paternal genetic diversity.
Various strategies have evolved to enhance the efficiency
of pollinator use. The aggregation of pollen grains as
tetrads or larger polyads occurs in several clades, with non‐
sporopollenin pollen‐connecting threads further promoting
the transfer of multiple pollen grains following a single
pollinator visit. This is likely to be beneficial when pollinator
visits are infrequent and/or when interfloral pollen transfer
is inefficient; because of the increased proportion of sibling
pollen grains received by stigmas, however, pollen
aggregation will only be selectively favored if it increases
overall reproductive performance. Pollen release is pro-
moted by the partial abscission of stamens (which remain
suspended in the pendent flower) and in some cases by
secondary pollen presentation within the inverted floral
chamber.
386 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

Anthesis usually occurs over 36–54 h, although it is
sometimes extended up to 25 days (especially in fly‐or
bee‐pollinated species) or abbreviated to only 23–30 h
(in those showing floral synchrony and/or pollinator
trapping). Species with abbreviated anthesis have fewer
coetaneous cohorts of flowers, enhancing pollination
efficiency by promoting staminate‐to‐pistillate interfloral
movement of pollinators. Many Annonaceae lineages with
abbreviated anthesis have also evolved circadian pollinator
trapping, in which petal movements controlling the opening
or closing of the floral chamber are closely aligned with the
endogenous circadian rhythms of the pollinators. This
mechanism, which appears to be unique to Annonaceae,
serves to broaden the range of potential pollinators,
including those with either unimodal or bimodal circadian
rhythms.
Various strategies exist or are hypothesized to optimize
fertilization success. Almost all Annonaceae species have
unfused carpels (apocarpy); the reproductive limitations of
this are circumvented by the formation of stigmatic
exudate, which functions as a suprastylar EGC, enabling
intercarpellary growth of pollen tubes. Suprastylar EGC has
been confirmed for Annona,Asimina,andGoniothalamus
species, but is likely to be more widespread in the family
and might represent a key evolutionary innovation.
Observations of floral anatomy (especially the predom-
inance of basal placentation in many species‐rich genera)
suggest that infrastylar EGC might also function in the
family, with pollen tubes traversing the floral receptacle to
access ovules. Other evolutionary strategies for optimizing
fertilization success include: increased ovule number per
carpel, carpel fusion (syncarpy), which is known from the
Isolona–Monodora clade and hypothesized here for
Cyathocalyx,self‐pollination, primarily due to breakdown
of protogyny, and rarely also apomixis.
Other floral adaptations are likely to promote pollen
competition. Many Annonaceae species possess elongated
stigmas that resemble styles; these “pseudostyles”might
enable sporophytic selection of microgametophytes based
on fitness, with xenogamous pollen potentially selected over
geitonogamous pollen.
Some floral adaptations provide protection against
herbivory. These adaptations include the exclusion by the
floral chamber of potentially destructive floral visitors, the
protection of undehisced anthers by the closely appressed
and tessellated stamen connectives, the staminal cone that
surrounds and protects the carpels of some Xylopia species,
and the emission of floral defense volatiles.
Flowers can be regarded as immature fruits in which
the ovules are yet to be fertilized, with plesiomorphic
floral structures perpetuated due to their benefits during
fruit and/or seed dispersal rather than during anthesis.
Apocarpy might have been maintained in flowers because
it enables the independent dispersal of monocarps and
hence promotes the geographical spread of seed
genotypes. This would be particularly important given
the limited genetic diversity among seeds within each
fruit due to the restricted interfloral movements of
pollinators.
Acknowledgements
I am indebted to my research postgraduates, past and
present, who have contributed to the development of many
of the ideas presented here. I am particularly grateful to Hazel
Chen, Jenny Lau, and Pang Chun‐Chiu for providing feedback
on the text. Leonid Averyanov, Chen Junhao, Jenny Lau, Li Pui
Sze, Pang Chun‐Chiu, Yvonne Su, and Tang Chin Cheung kindly
agreed to allow their photographs to be reproduced.
References
Armbruster WS. 1996. Evolution of floral morphology and function:
An integrative approach to adaptation, constraint, and com-
promise in Dalechampia (Euphorbiaceae). In: Lloyd DG, Barrett
SCH eds. Floral biology: Studies on floral evolution in animal‐
pollinated plants. London: Chapman & Hall. 241–272.
Armbruster WS, Debevec EM, Willson MF. 2002. Evolution of
syncarpy in angiosperms: Theoretical and phylogenetic analyses
of the effects of carpel fusion on offspring quantity and quality.
Journal of Evolutionary Biology 15: 657–672.
Armstrong JE, Marsh D. 1997. Floral herbivory, floral phenology,
visitation rate, and fruit set in Anaxagorea crassipetala
(Annonaceae), a lowland rain forest tree of Costa Rica. Journal
of the Torrey Botanical Society 124: 228–235.
Avarguès‐Weber A, Mota T, Giurfa M. 2012. New vistas on honey bee
vision. Apidologie 43: 244–268.
Azuma H, Toyota M, Asakawa Y, Kawano S. 1996. Naphthalene: A
constituent of Magnolia flowers. Phytochemistry 42: 999–1004.
Baker HG, Baker I. 1975. Studies of nectar‐constitution and pollinator‐
plant coevolution. In: Gilbert LE, Raven PH eds. Coevolution of
animals and plants. Austin: University of Texas Press. 591–600.
Barton LK, Menges ES. 2018. Effects of fire and pollinator visitation
on the reproductive success of Asimina reticulata (Annonaceae),
the netted pawpaw. Castanea 83: 323–333.
Bawa KS. 1974. Breeding systems of tree species of a lowland
tropical community. Evolution 28: 85–92.
Bernardello G. 2007. A systematic survey of floral nectaries. In:
Nicolson SW, Nepi M, Pacini E eds. Nectaries and nectar.
Dordrecht: Springer. 19–128.
Bicknell RA, Koltunow AM. 2004. Understanding apomixis: Recent
advances and remaining conundrums. The Plant Cell 16: 228–245.
Braun M, Dötterl S, Gottsberger G. 2011. Absence of pollinators and
apomictic fruit production in an Atlantic rainforest population of
Cymbopetalum brasiliense (Annonaceae). Plant Systematics and
Evolution 296: 265–273.
Braun M, Gottsberger G. 2011. Floral biology and breeding system of
Anaxagorea dolichocarpa (Annonaceae), with observations on the
interval between anthesis and fruit formation. Phyton 51: 315–327.
Briscoe AD, Chittka L. 2001. The evolution of color vision in insects.
Annual Review of Entomology 46: 471–510.
Carvalho R, Webber AC. 2000. Biologia floral de Unonopsis
guatterioides (A.D.C.) R.E. Fr., uma Annonaceae polinizada por
Euglossini. Revista Brasileira de Botánica 23: 421–425.
Chatrou LW, Listabarth C. 1998. Studies in Annonaceae XXXV. Facets
of reproductive biology of Mosannona. In: Changing genera:
Systematic studies in Neotropical and West African Annonaceae.
Ph.D. thesis. Netherlands: University of Utrecht. 195–205.
387Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

Chatrou LW, Pirie MD, Erkens RHJ, Couvreur TLP, Neubig KM, Abbott
JR, Mols JB, Maas JW, Saunders RMK, Chase MW. 2012. A new
subfamilial and tribal classification of the pantropical flowering
plant family Annonaceae informed by molecular phylogenetics.
Botanical Journal of the Linnean Society 169: 5–40.
Chen G, Ma X‐K, Jürgens A, Lu J, Liu E‐X, Sun W‐B, Cai X‐H. 2015.
Mimicking livor mortis: A well‐known but unsubstantiated color
profile in sapromyiophily. Journal of Chemical Ecology 41: 808–815.
Chen J, Liu M‐F, Saunders RMK. 2020. Contrasting floral biology of
Artabotrys species (Annonaceae): Implications for the evolution
of pollinator trapping. Plant Species Biology (early view).
Chittka L, Waser NM. 1997. Why red flowers are not invisible to bees.
Israel Journal of Plant Sciences 45: 169–183.
Coates‐Estrada R, Estrada A. 1988. Frugivory and seed dispersal in
Cymbopetalum baillonii (Annonaceae) at Los Tuxtlas, Mexico.
Journal of Tropical Ecology 4: 157–172.
Collier GE, Armstrong JE. 2009. Sequential florivory/saproflorivory of
Anaxagorea crassipetala (Annonaceae) by Diathoneura tessellata
(Drosophilidae). Annals of the Entomological Society of America
102: 492–497.
Costa MS, Silva RJ, Paulino‐Neto HF, Pereira MJB. 2017. Beetle
pollination and flowering rhythm of Annona coriacea Mart.
(Annonaceae) in Brazilian cerrado: Behavioral features of its
principal pollinators. PLoS One 12: e0171092.
Couvreur TLP. 2009. Monograph of the syncarpous African genera
Isolona and Monodora (Annonaceae). Systematic Botany Mono-
graphs 87: 1–150.
Couvreur TLP, Richardson JE, Sosef MSM, Erkens RHJ, Chatrou LW.
2008. Evolution of syncarpy and other morphological characters
in African Annonaceae: A posterior mapping approach. Molecular
Phylogenetics and Evolution 47: 302–318.
de Almeida‐Júnior EB, Collevatti RG, Telles MP de C, Chaves LJ, Neres
DF, Soares TN. 2018. Short‐distance pollen dispersal in a
protogynous Annonaceae tree species from the Brazilian
Cerrado. Plant Systematics and Evolution 304: 1091–1099.
Deroin T. 1985. Contribution à la morphologie comparée du gynécée
des Annonaceae‐Monodoroideae. Bulletin du Muséum d'Histoire
Naturelle, Paris. B, Adansonia 2: 167–176.
Deroin T. 1988. Biologie florale d'une Annonacée introduite en Côte
d'Ivoire: Cananga odorata (Lam.) Hook. f. & Thoms. Bulletin du
Muséum d'Histoire Naturelle, Paris. B, Adansonia 10: 377–393.
Deroin T. 1989. Définition et signification phylogénique des systèmes
corticaux floraux: L'exemple des Annonacées. Comptes Rendus
Hebdomadaires des Séances de l'Académie des Sciences, Paris,
Série III 308: 71–75.
Deroin T. 1997. Confirmation and origin of the paracarpy in
Annonaceae, with comments on some methodological aspects.
Candollea 52: 45–52.
Deroin T. 2000. Floral anatomy of Toussaintia hallei Le Thomas, a
case of convergence of Annonaceae with Magnoliaceae. In: Liu
YH, Fan HM, Chen ZY, Wu QG, Zeng QW eds. Proceedings of the
International Symposium on the family Magnoliaceae. Beijing:
Science Press. 168–176.
Deroin T, Bidault E. 2017. Floral anatomy of Pseudartabotrys Pellegrin
(Annonaceae), a monospecific genus endemic to Gabon.
Adansonia, Série III 39: 111–123.
Dias MC, Castro M, de M, Kinoshita LS, Semir J. 1998. Hipanto e tubo
estaminal em Xylopia aromatica (Lam) Mart. (Annonaceae).
Revista Brasileira de Botánica 21: 81–88.
Döring TF, Skellern M, Watts N, Cook SM. 2012. Colour choice
behaviour in the pollen beetle Meligethes aeneus (Coleoptera:
Nitidulidae). Physiological Entomology 37: 360–368.
Doyle JA, Le Thomas A. 2012. Evolution and phylogenetic significance
of pollen in Annonaceae. Botanical Journal of the Linnean Society
169: 190–221.
El‐Sayed AM. 2019. The Pherobase: Database of Pheromones and
Semiochemicals [online]. Available from https://www.pherobase.
com [accessed 20 June 2020].
Endler JA. 1993. The color of light in forests and its implication.
Ecological Monographs 63: 1–27.
Endress PK. 1982. Syncarpy and alternative modes of escaping
disadvantages of apocarpy in primitive angiosperms. Taxon
31: 48–52.
Endress PK. 1984. The role of the inner staminodes in the floral
display of some relic Magnoliales. Plant Systematics and
Evolution 146: 269–282.
Endress PK. 1985. Stamenabszission und Pollenpräsentation bei
Annonaceae. Flora 176: 95–98.
Endress PK. 1990. Evolution of reproductive structures and functions
in primitive angiosperms (Magnoliidae). Memoirs of the New York
Botanical Garden 55: 5–34.
Endress PK. 1994. Floral structure and evolution of primitive
angiosperms: Recent advances. Plant Systematics and Evolution
192: 79–97.
Endress PK. 2008. The whole and the parts: Relationships between
floral architecture and floral organ shape, and their repercus-
sions on the interpretation of fragmentary floral fossils. Annals
of the Missouri Botanical Garden 95: 101–120.
Endress PK. 2010. The evolution of floral biology in basal
angiosperms. Philosophical Transactions of the Royal Society of
London. Series B, Biological Sciences 365: 411–421.
Endress PK. 2011. Evolutionary diversification of the flowers in
angiosperms. American Journal of Botany 98: 370–396.
Endress PK. 2019. The morphological relationship between carpels
and ovules in angiosperms: Pitfalls of morphological interpreta-
tion. Botanical Journal of the Linnean Society 189: 201–227.
Endress PK. 2020. Structural and temporal modes of hetero-
dichogamy and similar patterns across angiosperms. Botanical
Journal of the Linnean Society 193: 5–18.
Endress PK, Doyle JA. 2007. Floral phyllotaxis in basal angiosperms:
Development and evolution. Current Opinion in Plant Biology
10: 52–57.
Endress PK, Igersheim A. 1997. Gynoecium diversity and systematics
of the Laurales. Botanical Journal of the Linnean Society 125:
93–168.
Erkens RHJ, Mennega EA, Westra LYT. 2012. A concise bibliographic
overview of Annonaceae. Botanical Journal of the Linnean Society
169: 41–73.
Fægri K, van der Pijl L. 1979. The principles of pollination ecology.
Oxford: Pergamon.
Galetto L, Bernardello G, Sosa CA. 1998. The relationship between
floral nectar composition and visitors in Lycium (Solanaceae)
from Argentina and Chile: What does it reflect? Flora 193:
303–314.
Gentry A. 1993. Diversity and floristic composition of lowland tropical
forest in Africa and South America. In: Goldblatt P ed. Biological
relationships between Africa and South America. New Haven: Yale
University Press. 500–547.
González M, Cuevas J. 2011. Reproductive barriers in Annona
cherimola (Mill.) outside of its native area. Plant Systematics
and Evolution 297: 227–235.
Goodrich KR. 2012. Floral scent in Annonaceae. Botanical Journal of
the Linnean Society 169: 262–279.
388 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

Goodrich KR, Jürgens A. 2018. Pollination systems involving floral
mimicry of fruit: Aspects of their ecology and evolution. New
Phytologist 217: 74–81.
Goodrich KR, Raguso RA. 2009. The olfactory component of floral
display in Asimina and Deeringothamnus (Annonaceae). New
Phytologist 183: 457–469.
Goodrich KR, Zjhra ML, Ley CA, Raguso RA. 2006. When flowers
smell fermented: The chemistry and ontogeny of yeasty floral
scent in pawpaw (Asimina triloba: Annonaceae). International
Journal of Plant Sciences 167: 33–46.
Gottsberger G. 1989. Beetle pollination and flowering rhythm of
Annona spp. (Annonaceae) in Brazil. Plant Systematics and
Evolution 167: 165–187.
Gottsberger G. 1999. Pollination and evolution in Neotropical
Annonaceae. Plant Species Biology 14: 143–152.
Gottsberger G. 2012. How diverse are Annonaceae with regard to
pollination? Botanical Journal of the Linnean Society 169: 245–261.
Gottsberger G. 2014. Evolutionary steps in the reproductive biology
of Annonaceae. Revista Brasileira de Fruticultura 36: 32–43.
Gottsberger G. 2016. Generalist and specialist pollination in basal
angiosperms (ANITA grade, basal monocots, magnoliids,
Chloranthaceae and Ceratophyllaceae): What we know now.
Plant Diversity and Evolution 131: 263–362.
Gottsberger G, Meinke S, Porembski S. 2011. First records of flower
biology and pollination in African Annonaceae: Isolona,
Piptostigma,Uvariodendron,Monodora and Uvariopsis.Flora
206: 498–510.
Gottsberger G, Schrauwen J, Linskens HF. 1984. Amino acids and
sugars in nectar, and their putative evolutionary significance.
Plant Systematics and Evolution 145: 55–77.
Gottsberger G, Silberbauer‐Gottsberger I. 1988. Pollination strategies
of Annona species from the cerrado vegetation in Brazil.
Lagascalia 15: 665–672.
Gottsberger G, Silberbauer‐Gottsberger I, Webber AC, Dötterl S.
2018. Populations of Unonopsis guatterioides (Annonaceae) in
Amazonas and Minas Gerais, Brazil, potentially represent
different species: Floral scent, flower characters and pollinators
revisited. Biochemical Systematics and Ecology 78: 17–20.
Guédès M, Le Thomas A. 1981. Le gynécées syncarpe de Monodora
(Annonacées‐Monodoroidées). Comptes Rendus Hebdomadaires
des Séances de l'Académie des Sciences, Paris, Série III 292:
1025–1028.
Guo X, Hoekstra PH, Tang CC, Thomas DC, Wieringa JJ, Chatrou LW,
Saunders RMK. 2017a. Cutting up the climbers: Evidence for
extensive polyphyly in Friesodielsia (Annonaceae) necessitates
generic realignment across the tribe Uvarieae. Taxon 66: 3–19.
Guo X, Tang CC, Thomas DC, Couvreur TLP, Saunders RMK. 2017b. A
mega‐phylogeny of the Annonaceae: Taxonomic placement of
five enigmatic genera and recognition of a new tribe,
Phoenicantheae. Scientific Reports 7: 7323.
Guo X, Thomas DC, Saunders RMK. 2018. Organ homologies and
floral evolution in the Dasymaschalon Alliance (Annonaceae):
Inner petal loss and its functional consequences. Frontiers in
Plant Science 9: 174.
Hammer DX, Schmitz H, Schmitz A, Rylander HG, Welch AJ. 2001.
Sensitivity threshold and response characteristics of infrared
detection in the beetle Melanophila acuminata (Coleoptera:
Buprestidae). Comparative Biochemistry and Physiology, A 128:
805–819.
Harder LD, Johnson SD. 2008. Function and evolution of aggregated
pollen in angiosperms. International Journal of Plant Sciences
169: 59–78.
Hausmann C, Samietz J, Dorn S. 2004. Visual orientation of
overwintered Anthonomus pomorum (Coleoptera: Curculio-
nidae). Environmental Entomology 33: 1410–1415.
Haydak MH. 1970. Honey bee nutrition. Annual Review of Entomology
15: 143–156.
Herrera CM. 2005. Plant generalizations on pollinators: Species
property or local phenomenon? American Journal of Botany
92: 13–20.
Hesse M, Vogel S, Halbritter H. 2000. Thread‐forming structures in
angiosperm anthers: Their diverse role in pollination ecology.
Plant Systematics and Evolution 222: 281–292.
Honkanen A, Immonen E‐V, Salmela I, Heimonen K, Weckström M.
2017. Insect photoreceptor adaptations to night vision. Philo-
sophical Transactions of the Royal Society of London. Series B,
Biological Sciences 372: 20160077.
Howell GJ, Slater AT, Knox RB. 1993. Secondary pollen presentation
in angiosperms and its biological significance. Australian Journal
of Botany 41: 417–438.
Huang SQ. 2003. Reproductive success by unusual growth of pollen
tubes to ovules. New Phytologist 158: 232–234.
Huang S‐Q, Guo Y‐H. 2002. Variation of pollination and resource
limitation in a low seed‐set tree, Liriodendron chinense
(Magnoliaceae). Botanical Journal of the Linnean Society
140: 31–38.
Igersheim A, Endress PK. 1997. Gynoecium diversity and systematics
of the Magnoliales and winteroids. Botanical Journal of the
Linnean Society 124: 213–271.
Inouye DW, Waller GD. 1984. Responses of honey bees (Apis
mellifera) to amino acid solutions mimicking floral nectars.
Ecology 65: 618–625.
Johnson DM, Murray NA. 1995. Synopsis of the tribe Bocageeae
(Annonaceae), with revisions of Cardiopetalum,Froesiodendron,
Trigynaea,Bocagea, and Hornschuchia.Brittonia 47: 248–319.
Johnson DM, Murray NA. 2018. A revision of Xylopia L. (Annonaceae):
The species of tropical Africa. PhytoKeys 97: 1–252.
Jürgens A, Webber AC, Gottsberger G. 2000. Floral scent compounds
of Amazonian Annonaceae species pollinated by small beetles
and thrips. Phytochemistry 55: 551–558.
Kishore K, Shukla AK, Babu N, Sarangi DN, Patanayak S. 2012.
Pollination biology of Annona squamosa L. (Annonaceae):
Evidence for pollination syndrome. Scientia Horticulturae 144:
212–217.
Kramer D. 1969. Comparative anatomical studies in the Annonaceae,
section Xylopieae. Ph.D. thesis. Bloomington: Indiana University.
Kress WJ. 1981. Sibling competition and evolution of pollen unit,
ovule number, and pollen vector in angiosperms. Systematic
Botany 6: 101–112.
Küchmeister H, Webber AC, Silberbauer‐Gottsberger I, Gottsberger
G. 1998. A polinização e sua relação com a termogênese em
espécies de Arecaceae e Annonaceae da Amazônia central. Acta
Amazonica 28: 217–245.
Lamoureux CH. 1975. Phenology and floral biology of Monodora
myristica (Annonaceae) in Bogor, Indonesia. Annales Bogoriense
6: 1–25.
Lau JYY, Guo X, Pang C‐C, Tang CC, Thomas DC, Saunders RMK.
2017a. Time‐dependent trapping of pollinators driven by the
alignment of floral phenology with insect circadian rhythms.
Frontiers in Plant Science 8: 1119.
Lau JYY, Pang C‐C, Ramsden L, Saunders RMK. 2016. Reproductive
resource partitioning in two sympatric Goniothalamus species
389Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

(Annonaceae) from Borneo: Floral biology, pollinator trapping
and plant breeding system. Scientific Reports 6: 35674.
Lau JYY, Pang C‐C, Ramsden L, Saunders RMK. 2017b. Stigmatic
exudate in the Annonaceae: Pollinator reward, pollen germination
medium or extragynoecial compitum? Journal of Integrative Plant
Biology 59: 881–894.
Le Thomas A. 1968. Balonga Le Thomas nouveau genre Africain de la
famille des Annonacées. Adansonia, Série II 8: 105–111.
Le Thomas A, Morawetz W, Waha M. 1986. Pollen of palaeo‐and
neotropical Annonaceae: Definition of the aperture by morpho-
logical and functional characters. In: Blackmore S, Ferguson IK
eds. Pollen and spores: Form and function. London: Academic
Press. 375–388.
Leins P, Erbar C. 1982. Das monokarpellate Gynoecium von Monodora
crispata (Annonaceae). Beiträge zur Biologie der Pflanzen 57: 1–13.
Levitt MM, Weekley CW, Menges ES. 2013. Reproductive ecology of
Asimina obovata (Annonaceae) in Florida scrub. Florida Scientist
76: 401–408.
Li B, Xu F. 2018. Formation pattern in five types of pollen tetrad in
Pseuduvaria trimera (Annonaceae). Protoplasma 256: 53–68.
Li B, Xu F. 2019. Homology and functions of inner staminodes in the
beetle‐pollinated flowers of Anaxagorea javanica. In: Timonin AC,
SokoloffDD eds. Plant anatomy: Traditions and perspectives.
Moscow: MAKS Press. 152–158.
Li P‐S, Pang C‐C, Saunders RMK. 2016. Contrasting pollination ecology
of Disepalum species (Annonaceae): Evolutionary loss of the
floral chamber and partial breakdown of protogyny associated
with a shift in pollination system. Botanical Journal of the Linnean
Society 182: 708–718.
Li P‐S, Thomas DC, Saunders RMK. 2015. Phylogenetic reconstruction,
morphological diversification and generic delimitation of
Disepalum (Annonaceae). PLoS One 10: e0143481.
Lloyd DG. 1975. The maintenance of gynodioecy and androdioecy in
angiosperms. Genetica 45: 325–339.
Lloyd DG, Wells MS. 1992. Reproductive biology of a primitive
angiosperm, Pseudowintera colorata (Winteraceae), and the
evolution of pollination systems in the Anthophyta. Plant
Systematics and Evolution 181: 77–95.
Lora J, Herrero M, Hormaza JI. 2011. Stigmatic receptivity in a
dichogamous early‐divergent angiosperm species, Annona
cherimola (Annonaceae): Influence of temperature and humidity.
American Journal of Botany 98: 265–274.
Lora J, Hormaza JI, Herrero M. 2016. The diversity of the pollen tube
pathway in plants: Toward an increasing control by the
sporophyte. Frontiers in Plant Science 7: 107.
Lora J, Testillano PS, Risueña MC, Hormaza JI, Herrero M. 2009.
Pollen development in Annona cherimola Mill. (Annonaceae).
Implications for the evolution of aggregated pollen. BMC Plant
Biology 9: 129.
Lord EM, Webster BD. 1979. The stigmatic exudate of Phaseolus
vulgaris L. Botanical Gazette 140: 266–271.
Losada JM, Hormaza JI, Lora J. 2017. Pollen‐pistil interaction in
pawpaw (Asimina triloba), the northernmost species of the
mainly tropical family Annonaceae. American Journal of Botany
104: 1891–1903.
Lunau K. 2014. Visual ecology of flies with particular reference to
colour vision and colour preferences. Journal of Comparative
Physiology A 200: 497–512.
Luo SX, Chaw SM, Zhang D, Renner SS. 2010. Flower heating
following anthesis and the evolution of gall midge pollination in
Schisandraceae. American Journal of Botany 97: 1220–1228.
Maas PJM, Westra LYT. 1984. Studies in Annonaceae. II. A
monograph of the genus Anaxagorea A. St. Hil. Part 1. Botanische
Jahrbücher für Systematik, Pflanzengeschichte und Pflanzengeog-
raphie 105: 73–134.
Maas PJM, Westra LYT, Chatrou LW. 2003. Duguetia (Annonaceae),
Flora Neotropica monograph. New York: Organization for Flora
Neotropica. 88.
Maas‐van de Kamer H. 1993. Floral biology of Anaxagorea
dolichocarpa, and some notes on flower biology in other
Annonaceae. Annonaceae Newsletter 9: 19–24.
Maia ACD, Dötterl S, Kaiser R, Silberbauer‐Gottsberger I, Teichert H,
Gibernau M, do Amaral Ferraz Navarro DM, Schlindwein C,
Gottsberger G. 2012. The key role of 4‐methyl‐5‐vinylthiazole in
the attraction of scarab beetle pollinators: A unique olfactory
floral signal shared by Annonaceae and Araceae. Journal of
Chemical Ecology 38: 1072–1080.
Matteson N, Terry I, Ascoli‐Christensen A, Gilbert C. 1992. Spectral
efficiency of the western flower thrips, Frankliniella occidentalis.
Journal of Insect Physiology 38: 453–459.
McKenna DD, Farrell BD. 2009. Beetles (Coleoptera). In: Hedges SB,
Kumar S eds. The timetree of life. Oxford: Oxford University
Press. 278–289.
Merivee E, Märtmann H, Must A, Milius M, Williams I, Mänd M. 2008.
Electrophysiological responses from neurons of antennal taste
sensilla in the polyphagous predatory ground beetle Pterostichus
oblongopunctatus (Fabricius 1787) to plant sugars and amino
acids. Journal of Insect Physiology 54: 1213–1219.
Mitchell BK, Gregory P. 1979. Physiology of the maxillary sugar
sensitive cell in the red turnip beetle, Entomoscelis americana.
Journal of Comparative Physiology 132: 167–178.
Momose K, Nagamitsu T, Inoue T. 1998. Thrips cross‐pollination of
Popowia pisocarpa (Annonaceae) in a lowland dipterocarp forest
in Sarawak. Biotropica 30: 444–448.
Moore JC, Pannell JR. 2011. Sexual selection in plants. Current Biology
21: R176–R182.
Morawetz W. 1986. Remarks on karyological differentiation patterns
in tropical woody plants. Plant Systematics and Evolution 152:
49–100.
Murray NA. 1993. Revision of Cymbopetalum and Porcelia (Annona-
ceae). Systematic Botany Monographs 40: 1–121.
Murray NA, Johnson DM. 1987. Synchronous dichogamy in a Mexican
anonilla, Rollinia jimenezii var. nelsonii (Annonaceae). Contribu-
tions from the University of Michigan Herbarium 16: 173–178.
Nagamitsu T, Inoue T. 1997. Cockroach pollination and breeding
system of Uvaria elmeri (Annonaceae) in a lowland mixed‐
dipterocarp forest in Sarawak. American Journal of Botany 84:
208–213.
Norman EM. 2003. Reproductive biology of Deeringothamnus rugelii
and D. pulchellus (Annonaceae). Taxon 52: 547–555.
Norman EM, Clayton D. 1986. Reproductive biology of two Florida
pawpaws: Asimina obovata and A. pygmaea (Annonaceae).
Bulletin of the Torrey Botanical Club 113: 16–22.
Norman EM, Rice K, Cochran S. 1992. Reproductive biology of
Asimina parviflora (Annonaceae). Bulletin of the Torrey Botanical
Club 119: 1–5.
Oliveira PE, Tomé CER, Torezan‐Silingardi HM, Dötterl S, Silberbauer‐
Gottsberger I, Gottsberger G. 2017. Differential pollination modes
between distant populations of Unonopsis guatterioides (Anno-
naceae) in Minas Gerais and Amazonas, Brazil. Flora 232: 39–46.
Pang C‐C, Saunders RMK. 2014. The evolution of alternative
mechanisms that promote out‐crossing in Annonaceae, a
390 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn

self‐compatible family of early‐divergent angiosperms. Botanical
Journal of the Linnean Society 174: 93–109.
Pang C‐C, Saunders RMK. 2015. Floral biology and pollination ecology
of Desmos chinensis (Annonaceae): Assessing the efficacy of
floral synchrony for promoting xenogamy. International Journal
of Plant Sciences 176: 333–345.
Pang C‐C, Scharaschkin T, Su YCF, Saunders RMK. 2013. Functional
monoecy due to delayed anther dehiscence: A novel mechanism
in Pseuduvaria mulgraveana (Annonaceae). PLoS One 8: e59951.
Pellmyr O, Thien LB. 1986. Insect reproduction and floral fragrances:
Keys to the evolution of the angiosperms? Taxon 35: 76–85.
Peña JE, Castiñeiras A, Bartelt R, Duncan R. 1999. Effect of
pheromone bait stations for sap beetles (Coleoptera: Nitidu-
lidae) on Annona spp. fruit set. The Florida Entomologist 82:
475–480.
Pichersky E, Gershenzon J. 2002. The formation and function of plant
volatiles: Perfumes for pollinator attraction and defence. Current
Opinion in Plant Biology 5: 237–243.
Prantl K. 1891. Anonaceae. In: Engler E, Prantl K eds. Die natürlichen
Pflanzenfamilien. Div. III. Leipzig: Wilhelm Engelmann. 2: 23–39.
Punyasena SW, Eshel G, McElwain JC. 2008. The influence of climate
on the spatial patterning of Neotropical plant families. Journal of
Biogeography 35: 117–130.
Rands SA, Whitney HM. 2008. Floral temperature and optimal
foraging: Is heat a feasible floral reward for pollinators? PLoS
One 3: e2007.
Ratnayake RMCS, Gunatilleke IAUN, Wijesundara DSA, Saunders
RMK. 2006a. Reproductive biology of two sympatric species of
Polyalthia (Annonaceae) in Sri Lanka. I. Pollination by curculionid
beetles. International Journal of Plant Sciences 167: 483–493.
Ratnayake RMCS, Su YCF, Gunatilleke IAUN, Wijesundara DSA,
Saunders RMK. 2006b. Reproductive biology of two sympatric
species of Polyalthia (Annonaceae) in Sri Lanka. II. Breeding
systems and population genetic structure. International Journal
of Plant Sciences 167: 495–502.
Ratnayake RMCS, Gunatilleke IAUN, Wijesundara DSA, Saunders
RMK. 2007. Pollination ecology and breeding system of Xylopia
championii (Annonaceae): Curculionid beetle pollination, pro-
moted by floral scents and elevated floral temperatures.
International Journal of Plant Sciences 168: 1255–1268.
Reeves JL. 2011. Vision should not be overlooked as an important
sensory modality for finding host plants. Environmental
Entomology 40: 855–863.
Rejón JD, Delalande F, Schaeffer‐Reiss C, Carapito C, Zienkiewicz K,
de Dios Alché J, Rodríguez‐García MI, Van Dorsselaer A, Castro
AJ. 2014. The plant stigma exudate: A biochemically active
extracellular environment for pollen germination? Plant Signaling
and Behavior 9: e28274.
Rogstad SH. 1993. Variability in timing of sexual phases in the North
American pawpaw, Asimina triloba (L.) Dunal (Annonaceae).
Trillium 2: 5–8.
Rogstad SH. 1994. The biosystematics and evolution of the Polyalthia
hypoleuca species complex (Annonaceae) of Malesia. III. Floral
ontogeny and breeding systems. American Journal of Botany 81:
145–154.
Sagae M, Oyama‐Okubo N, Ando T, Marchesi E, Nakayama M. 2008.
Effect of temperature on the floral scent emission and
endogenous volatile profile of Petunia axillaris.Bioscience,
Biotechnology, and Biochemistry 72: 110–115.
Sakai S. 2002. A review of brood‐site pollination mutualism: Plants
providing breeding sites for their pollinators. Journal of Plant
Research 115: 161–168.
Saunders RMK. 2002. The genus Goniothalamus (Annonaceae) in
Sumatra. Botanical Journal of the Linnean Society 139: 225–254.
Saunders RMK. 2010. Floral evolution in the Annonaceae: Hypoth-
eses of homeotic mutations and functional convergence.
Biological Reviews 85: 571–591.
Saunders RMK. 2012. The diversity and evolution of pollination
systems in Annonaceae. Botanical Journal of the Linnean Society
169: 222–244.
Scharaschkin T, Doyle JA. 2006. Character evolution in Anaxagorea
(Annonaceae). American Journal of Botany 93: 36–54.
Schatz GE. 1987. Systematic and ecological studies of Central American
Annonaceae. Ph.D. thesis. Madison: University of Wisconsin.
Schiestl FP. 2015. Ecology and evolution of floral volatile‐mediated
information transfer in plants. New Phytologist 206: 571–577.
Schmitz H, Bleckmann H, Murtz M. 1997. Infrared detection in a
beetle. Nature 386: 773–774.
Seymour RS, Schultze‐Motel P. 1997. Heat‐producing flowers.
Endeavour 21: 125–129.
Seymour RS, White CR, Gibernau M. 2003. Heat reward for insect
pollinators. Nature 426: 243–244.
Shao Y‐Y, Xu F‐X. 2015. Development and function of the strumae on
the adaxial sides of the inner petals of Oncodostigma hainanense.
Plant Science Journal 33: 595–601. (In Chinese, with English
abstract)
Shiraishi A, Kuwabara M. 1970. The effects of amino acids on the
labellar hair chemosensory cells of the fly. The Journal of General
Physiology 56: 768–782.
Silberbauer‐Gottsberger I, Gottsberger G, Webber AC. 2003.
Morphological and functional flower characteristics of New
and Old World Annonaceae with respect to their mode of
pollination. Taxon 52: 701–718.
Simpson B, NeffJ. 1983. Evolution and diversity of floral rewards. In:
Jones CE, Little RJ eds. Handbook of experimental pollination
biology. New York: Van Nostrand Reinhold. 277–293.
Spjut R. 1994. A systematic treatment of fruit types. Memoirs of the
New York Botanical Garden 70: 1–182.
Su YCF, Mols JB, Takeuchi W, Keßler PJA, Saunders RMK. 2005.
Reassessing the generic status of Petalolophus (Annonaceae):
Evidence for the evolution of a distinct sapromyophilous lineage
within Pseuduvaria.Systematic Botany 30: 494–502.
Su YCF, Saunders RMK. 2003. Pollen structure, tetrad cohesion, and
pollen‐connecting threads in Pseuduvaria (Annonaceae). Bota-
nical Journal of the Linnean Society 143: 69–78.
Su YCF, Saunders RMK. 2006. Monograph of Pseuduvaria (Annona-
ceae). Systematic Botany Monographs 79: 1–204.
Su YCF, Saunders RMK. 2009. Evolutionary divergence times in
Annonaceae: Evidence of a late Miocene origin of Pseuduvaria in
Sundaland with subsequent diversification in New Guinea. BMC
Evolutionary Biology 9: 153.
Su YCF, Smith GJD, Saunders RMK. 2008. Phylogeny of the basal
angiosperm genus Pseuduvaria (Annonaceae) inferred from five
chloroplast DNA regions, with interpretation of morphological
character evolution. Molecular Phylogenetics and Evolution 48:
188–206.
Suárez C, Castro A, Rapoport H, Rodríguez‐García M. 2012.
Morphological, histological and ultrastructural changes in the
olive pistil during flowering. Sexual Plant Reproduction 25:
133–146.
Surveswaran S, Wang RJ, Su YCF, Saunders RMK. 2010. Generic
delimitation and historical biogeography in the early‐divergent
391Functional floral traits in Annonaceae
J. Syst. Evol. 58(4): 369–392, 2020www.jse.ac.cn

‘ambavioid’lineage of Annonaceae: Cananga,Cyathocalyx and
Drepananthus.Taxon 59: 1721–1734.
Tagane S, Dang VS, Yahara T, Toyama H, Tran H. 2015. Goniothalamus
flagellistylus Tagane & V.S. Dang (Annonaceae), a new species
from Mt. Hon Ba, Vietnam. PhytoKeys 50: 1–8.
Tang CC, Thomas DC, Saunders RMK. 2015. Molecular phylogenetics
of the species‐rich angiosperm genus Goniothalamus (Annona-
ceae) inferred from nine chloroplast regions: Synapomorphies
and putative correlated evolutionary changes in fruit and seed
morphology. Molecular Phylogenetics and Evolution 92: 124–139.
TangCC,XueB,SaundersRMK.2013.AnewspeciesofGoniothalamus
(Annonaceae) from Palawan, and a new nomenclatural combina-
tion in the genus from Fiji. PhytoKeys 32: 27–35.
Teichert H, Dötterl S, Frame D, Kirejtshuk A, Gottsberger G. 2012. A
novel pollination mode, saprocantharophily, in Duguetia cadav-
erica (Annonaceae): A stinkhorn (Phallales) flower mimic. Flora
207: 522–529.
Teichert H, Dötterl S, Gottsberger G. 2011. Heterodichogamy and
nitidulid beetle pollination in Anaxagorea prinoides, an early
divergent Annonaceae. Plant Systematics and Evolution
291: 25–33.
Teichert H, Dötterl S, Zimma B, Ayasse M, Gottsberger G. 2009.
Perfume‐collecting male euglossine bees as pollinators of a basal
angiosperm: The case of Unonopsis stipitata (Annonaceae). Plant
Biology 11: 29–37.
Thien LB, Azuma H, Kawano S. 2000. New perspectives on the
pollination biology of basal angiosperms. International Journal of
Plant Sciences 161: S225–S235.
Thomas DC, Surveswaran S, Xue B, Sankowsky G, Mols JB, Keßler
PJA, Saunders RMK. 2012. Molecular phylogenetics and historical
biogeography of the Meiogyne‐Fitzalania clade (Annonaceae):
Generic paraphyly and late Miocene‐Pliocene diversification in
Australasia and the Pacific. Taxon 61: 559–575.
Tsou C‐H, Fu Y‐L. 2002. Tetrad pollen formation in Annona
(Annonaceae): Proexine formation and binding mechanism.
American Journal of Botany 89: 734–747.
Tsou C‐H, Fu Y‐L. 2007. Octad pollen formation in Cymbopetalum
(Annonaceae): The binding mechanism. Plant Systematics and
Evolution 263: 13–23.
Tsou C‐H, Johnson DM. 2003. Comparative development of aseptate
and septate anthers of Annonaceae. American Journal of Botany
90: 832–848.
Umekawa Y, Seymour RS, Ito K. 2016. The biochemical basis for
thermoregulation in heat‐producing flowers. Scientific Reports 6:
24830.
Uphof JCT. 1933. Die nordamerikanischen Arten der Gattung Asimina.
Mitteilungen der Deutschen Dendrologischen Gesellschaft 45: 61–76.
van der Kooi CJ, Kevan PG, Koski MH. 2019. The thermal ecology of
flowers. Annals of Botany 124: 343–353.
van Heusden ECH. 1992. Flowers of Annonaceae: Morphology,
classification, and evolution. Blumea Suppl. 7: 1–218.
van Setten AK, Koek‐Noorman J. 1992. Fruits and seeds of
Annonaceae: Morphology and its significance for classification.
Bibliotheca Botanica 142: 1–101.
Verdcourt B. 1971. Annonaceae. In: Milne‐Redhead E, Polhill RM eds.
Flora of Tropical East Africa. London: Crown Agents. 1–132.
Wang J, Thomas DC, Su YCF, Meinke S, Chatrou LW, Saunders RMK.
2012. A plastid DNA phylogeny of Dasymaschalon (Annonaceae)
and allied genera: Evidence for generic non‐monophyly and the
parallel evolutionary loss of inner petals. Taxon 61: 545–558.
Wang RJ, Saunders RMK. 2006. A synopsis of Cyathocalyx species
(Annonaceae) in Peninsular Malaysia, Sumatra and Borneo, with
descriptions of two new species. Botanical Journal of the Linnean
Society 152: 513–532.
Wang XF, Armbruster WS, Huang SQ. 2012. Extra‐gynoecial pollen‐
tube growth in apocarpous angiosperms is phylogenetically
widespread and probably adaptive. New Phytologist 193:
253–260.
Wang X‐F, Tan Y‐Y, Chen J‐H, Lu Y‐T. 2006. Pollen tube reallocation in
two preanthesis cleistogamous species Ranalisma rostratum and
Sagittaria guyanensis ssp. lappula (Alismataceae). Aquatic Botany
85: 233–240.
Wang XF, Tao YB, Lu YT. 2002. Pollen tubes enter neighbouring
ovules by way of receptacle tissue, resulting in increased fruit‐set
in Sagittaria potamogetifolia Merr. Annals of Botany 89: 791–796.
Warrant E, Dacke M. 2011. Vision and visual navigation in nocturnal
insects. Annual Review of Entomology 56: 239–254.
Webber AC. 1981. Alguns aspectos da biologia floral de Annona
sericea Dun. (Annonaceae). Acta Amazonica 11: 61–65.
Webber AC. 2002. Floral biology and pollination of some Neotropical
Annonaceae. Annonaceae Newsletter 13: 18–21.
Weerasooriya AD, Saunders RMK. 2001. Three new species of
Mitrephora (Annonaceae) from Sabah, Malaysia. Botanical
Journal of the Linnean Society 135: 305–314.
Weerasooriya AD, Saunders RMK. 2010. Monograph of Mitrephora
(Annonaceae). Systematic Botany Monographs 90: 1–167.
Wester P. 1910. Pollination experiments with anonas. Bulletin of the
Torrey Botanical Club 37: 529–539.
Whitten WM, Williams NH, Armbruster WS, Battiste MA, Strkowski L,
Lindquist N. 1986. Carvone oxide: An example of convergent
evolution in euglossine pollinated plants. Systematic Botany 11:
222–228.
Whitten WM, Young AM, Williams NH. 1989. Function of glandular
secretions in fragrance collection by male euglossine bees
(Apidae: Euglossini). Journal of Chemical Ecology 15: 1285–1295.
Willson MF, Schemske DW. 1980. Pollinator limitation, fruit
production, and floral display in pawpaw (Asimina triloba).
Bulletin of the Torrey Botanical Club 107: 401–408.
Xicohténcatl‐Lara L, Figueroa‐Castro D, Andrés‐Hernández AR,
Campos‐Villanueva Á. 2016. Aspects of the reproductive biology
of Stenanona flagelliflora (Annonaceae). Pakistan Journal of
Botany 48: 211–221.
Xue B, Guo X, Landis JB, Sun M, Tang CC, Soltis PS, Soltis DE,
Saunders RMK. 2020. Accelerated diversification correlated with
functional traits shapes extant diversity of the early‐divergent
angiosperm family Annonaceae. Molecular Phylogenetics and
Evolution 142: 106659.
Xue B, Shao Y‐Y, Saunders RMK, Tan Y‐H. 2017. Alphonsea glandulosa
(Annonaceae), a new species from Yunnan, China. PLoS One 12:
e0170107.
Yang G, Xu F. 2016. Floral development of monoecious Pseuduvaria
trimera (Annonaceae) and comparative morphology and
structure of its stamens and indehiscent staminodes. Interna-
tional Journal of Plant Sciences 177: 736–748.
Yeo PF. 1993. Secondary pollen presentation: Form, function and
evolution,Plant systematics and evolution, supplement 6. Vienna
and New York: Springer‐Verlag.
Zhang YM, Yang GF, Xu FX. 2014. Pollen morphology of three
selected species from Annonaceae. Plant Science Journal 32:
315–319. (In Chinese, with English abstract)
392 Saunders
J. Syst. Evol. 58(4): 369–392, 2020 www.jse.ac.cn