J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

A Morphological and Histological Characterization

of Bisexual and Male Flower Types in Pomegranate
Hazel Y. Wetzstein1
University of Georgia, Department of Horticulture, 1111 Miller Plant Science Building, Athens,
GA 30602
Nadav Ravid and Erik Wilkins
Paramount Farming Company, 33141 E. Lerdo Highway, Bakersfield, CA 93308
Adriana Pinheiro Martinelli
University of Sao Paulo, CENA, Av. Centenario 303, 13400-970, Piracicaba, Sao Paulo, Brazil
ADDITIONAL INDEX WORDS. Punica granatum, hermaphroditic, staminate, ovules, pollen, stigma
ABSTRACT. Pomegranate [Punica granatum (Punicaceae)] is characterized by having two types of flowers on the same
tree: hermaphroditic bisexual flowers and functionally male flowers. This condition, defined as functional
andromonoecy, can result in decreased yields resulting from the inability of male flowers to set fruit. Morphological
and histological analyses of bisexual and male flowers were conducted using light and scanning electron microscopy
(SEM) to characterize the different flower types observed in pomegranate plants and to better understand their
developmental differences. Bisexual flowers had a discoid stigma covered with copious exudate, elongated stigmatic
papillae, a single elongate style, and numerous stamens inserted on the inner wall of the calyx tube. Using fluorescence
staining, high numbers of pollen tubes were observed growing through a central stylar canal. Ovules were numerous,
elliptical, and anatropous. In contrast, male flowers had reduced female parts and exhibited shortened pistils of
variable heights. Stigmatic papillae of male flowers had little exudate yet supported pollen germination. However,
pollen tubes were rarely observed in styles. Ovules in male flowers were rudimentary and exhibited various stages
of degeneration. Pollen from both types of flowers was of similar size, 20 mm, and exhibited similar percent
germination using in vitro germination assays. Pollen germination was strongly influenced by temperature. Maximal
germination (greater than 74%) was obtained at 25 and 35 8C; pollen germination was significantly lower at 15 8C
(58%) and 5 8C (10%).

Pomegranate has been grown as a fruit crop since ancient

times. Native to central Asia, this tree/shrub is highly adaptive
to a wide range of climates and soil conditions and is grown in
many geographical regions, including the Mediterranean basin,
Asia, and the United States, particularly in California (Holland
et al., 2009). It has long been valued for its flavorful, juicy aril
sacs and more lately for commercial juice production. Recently,
numerous studies have verified the health benefits associated
with pomegranates. Pomegranate juice and products are purported
to show efficacy against a wide range of conditions, including
cancer, coronary heart disease, atherosclerosis, hypercholesterolemia, hyperlipidemia, hypertension, HIV, infectious diseases,
aging, and brain disorders (Basu and Penugonda, 2008; Holland
et al., 2009; Lansky and Newman, 2007; Seeram et al., 2006). This
has led to a higher awareness of the public to the benefits of
pomegranate and a prominent increase in the consumption of the
fruit and juice (Holland et al., 2009).
Flowering in pomegranate is characterized as having both
hermaphroditic (bisexual) flowers and functionally male
flowers on the same plant, a condition referred to as andromonoecy. The hermaphroditic flowers have well-formed female
(stigma, style, ovary) and male (filaments and anthers) parts
and have been referred to as fertile, vase-shaped, and
Received for publication 3 Nov. 2010. Accepted for publication 6 Dec. 2010.
We acknowledge Paramount Farming Company for field, research, and
financial support; CNPq, Conselho Nacional de Desenvolvimento Cientfico e
Tecnologico, for a research fellowship to A.P.M. and J. G. Bracalion for
assistance with micrographs.
1
Corresponding author. E-mail: [email protected].

J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

bisexual flowers. Because the hermaphroditic flowers are the

type that set fruit, they are commonly referred to as female
flowers, albeit with some inaccuracy. The male flowers produce
well-developed male parts, but on closer examination of the pistil
contain reduced female parts. Thus, their role is more accurately
depicted as functionally male flowers (i.e., flowers are not strictly
male), but rather have degenerated female parts. Male flowers
typically drop and fail to set fruit (Holland et al., 2009; Shulman
et al., 1984). Chaudhari and Desai (1993) classified pomegranate
flowers into three types: male, hermaphroditic, and intermediate.
Observations of gradients of flower types in some pomegranate
genotypes support this concept. Synonyms for male flowers
include infertile and bell flowers. Herein, we refer to the
flower types as bisexual and male for hermaphroditic and
functionally male flower types, respectively.
Andromonoecy as well as other fluctuating sexual expression
types is proposed to allow a species to optimize the allocation of
limited resources to male and female function (Bertin, 1982;
Wilson, 1983). Manipulating the relative ratio of flower types to
environmental conditions can be very advantageous. As a result
of the high costs associated with female expression, repression of
this flower type under poor environmental conditions could be
a means to conserve limited resources at a time when maturation
of a high fruit/seed yield is not possible. Furthermore, having high
numbers of male flowers can be a way to spread genes, because
pollen spread is more efficient with more male flowers (Herlihy
and Eckert, 2002; Tanurdzic and Banks, 2004).
Under agricultural production conditions, male:female
flower ratios in pomegranate can impact crop productivity
83

and yield. Male flowers drop and generally fail to set; thus,
fruits develop exclusively from bisexual flowers. A positive
relation was found between the percentage of bisexual flowers
and bearing capacity (Chaudhari and Desai, 1993; El Sese,
1988). The percentage of flowers that are male in pomegranate
can be significant and more than 60% to 70% depending on
variety and season (Chaudhari and Desai, 1993; Mars, 2000).
Furthermore, the ratio of bisexual and functionally male flowers
can vary with season (N. Ravid, personal communication),
cultural location, and genotype. Fluctuations of flower types
within a season can proceed with a predominant appearance of
bisexual flowers followed by male flowers or vice versa.
Although this crop has been grown as an agricultural crop
since antiquity, scientific literature on many fundamental
aspects of pomegranate development and physiology is lacking,
including basic aspects of floral biology. A clear understanding
of male and female flowering is lacking. As part of an ongoing
project on pomegranate reproductive biology, the present work
aims to describe the morphology and anatomy of bisexual and
functionally male flower types in pomegranate. Morphological
and histological evaluations of hermaphroditic and male
flowers were conducted using light microscopy (LM) and SEM
to better understand developmental differences between the
flower types. Sex expression is labile in this species. A comparison of the form and function of flower parts in the two morphs,
including pollen viability, stigma development, and pollenpistil
interactions, would provide information useful in the development of crop production protocols to enhance fruit production.
Material and Methods
PLANT MATERIAL. Pomegranate flowers were collected from
8-year-old Wonderful trees grown in a commercial orchard
block located near Delano, CA, that was managed under
conventional methods. Bisexual and functionally male flowers
were separated based on the size of the pistil, which in male
flowers is expressed as a shortened style. One hundred newly
opened flowers of each type were collected from several trees
and taken back to the laboratory for measurements; flowering
was early to midseason. Some tissues were fixed and processed
for LM and SEM as described subsequently. In addition, a set
of 20 female and 20 male flowers was emasculated and bagged
in Delnet pollination bags (DelStar Technologies, Middletown,
DE) before anther dehiscence and petal opening so that
histological evaluations could be made of unpollinated flowers
to facilitate stigma and stylar histological studies.
LIGHT MICROSCOPY. Samples were prepared for LM as we
have previously described (Yi et al., 2006). Tissues were fixed
in a mixture of 3% paraformaldehyde (w/v) and 2% glutaraldehyde (v/v) in 0.2 M cacodylate buffer, pH 7.2 at 4 C;
dehydrated in a series of methyl cellosolve, ethanol, propanol,
and n-butanol; and then infiltrated and embedded in Historesin
(Leica Instruments, Heidelberg, Germany), which consists of
hydroxyethylmethacrylate. Sections 5 to 6 mm thick were cut
with a tungsten knife using a rotary microtome (HM350;
Microm, Heidelberg, Germany). Serial sections were made to
aid interpretations. Sections were stained successively in 1%
acid fuchsin (w/v) and 0.05% toluidine blue O (w/v) and then
mounted in Mount-quick (EMS, Fort Washington, PA). Starch
was visualized by staining sections in a 50% IKI solution
(Jensen, 1962). Sections were examined under bright field using
a light microscope (BX51; Olympus America, Center Valley,
84

PA) and photographed digitally. To visualize pollen tube

growth in the style, pistils were dissected from flowers and
then fixed in ethanol:acetic acid (3:1, v/v). To clear and soften
tissues, stigma-style portions of each pistil were transferred into
10% sodium sulfite (w/v) and then autoclaved at 120 C for
10 min. Samples were rinsed in water and then stained with
aniline blue in K3PO4 buffer for 18 to 24 h. Whole mounts of
styles were observed under both ultraviolet and transmitted light.
SCANNING ELECTRON MICROSCOPY. Bisexual and male flowers
were dissected and prepared using methods described by Woo
and Wetzstein (2008) for SEM. Flowers were fixed in 2% (v/v)
glutaraldehyde in 0.1 M cacodylate buffer, pH 7.2, dehydrated
through an ethanol series, and critical point dried through carbon
dioxide using a critical point drier (Samdri-790; Tousimis
Research, Rockville, MD). Dried samples were mounted on
aluminum stubs using carbon conductive tabs and sputter coated
(SPI-Module; SPI Supplies, West Chester, PA) with gold.
Samples were examined using a SEM (JSM 5800; JEOL, Tokyo,
Japan) at 15 kV.
IN VITRO POLLEN GERMINATION. In vitro germination assays
were conducted according to Yi et al. (2003). Freshly collected
pollen was inoculated into 96-well plates (Microtest III; Falcon,
Lincoln Park, NJ) containing a pollen germination medium
consisting of 0.062% CaNO3 (w/v), 0.024% boric acid (w/v),
and 12% sucrose (w/v). Cultured pollen was maintained at one
of four temperatures (5, 15, 25, or 35 C), and percent germination assessed at 1-, 3-, and 5-h incubation times using an
inverted microscope (DIAPHOT; Nikon, Garden City, NY). A
pollen grain was considered germinated if tube length was
longer than the diameter of the pollen grain. Eight hundred
pollen grains were evaluated for each temperature (800 pollen
grains four temperatures).
STATISTICAL ANALYSIS. Statistical analyses were conducted
using the General Linear Model (SAS Version 9.1; SAS Institute,
Cary, NC). Differences in flower measurements between bisexual and male flowers were compared using Students t test
at a = 0.05. Pollen germination under different temperatures was
compared by Duncans multiple range test at a = 0.05.
Results
Pomegranate trees produce large showy flowers, which in
Wonderful are a vivid orangered color (Fig. 1A). Flowers are
near sessile and are born either singly (Fig. 1B) or in clusters
(Fig. 1C) of commonly one terminal flower subtended by two
to four lateral flowers. The calyx is red, thick, leathery, fused
at the base, and usually pentamerous, although infrequently
tetramerous. At the distal portion of unopened buds (Figs. 1A
and 1C), indentations are evident showing the region where the
calyx will separate and form acuminate triangular lobes (Fig.
1B), which will persist and form the crown of the mature fruit.
The normally five antesepalous petals are ovate, delicate,
slightly crenulated, and may show streaks of darker pigmentation. The ovary is inferior, and the numerous stamens are
composed of yellow anthers attached to long red filaments,
which are inserted on the inner surface of the calyx tube (Fig.
1DE). Anthers surround an elongated style, which at the base
broadens into a conical-shaped stylopodium (Fig. 1D). When
flowers are bisected longitudinally, the two flower types are
readily discernable based on differences in pistil development.
Bisexual flowers (Fig. 1D) possess a well-developed pistil with
an elongated style that extends at or above the height of anthers.
J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

Fig. 1. Pomegranate flowers. (A) Fully open flower showing vivid orangered petals and numerous anthers attached to long, red filaments. (B) Single near sessile
flower. The leathery calyx is fused at the base. (C) Flower cluster with an advanced central flower subtended by closed buds. (D) Longitudinal section of a bisexual
flower showing anthers with filaments inserted on the inner surface of the calyx tube. (E) Longitudinal section of a functionally male flower showing welldeveloped stamens but an underdeveloped pistil. (F) Details of well-formed ovules from a hermaphroditic flower. (G) Underdeveloped ovules from a functionally male flower. Scale bars: DE = 5.0 mm, FG = 500 mm.

Flowers have a prominent U-shaped ovary (Fig. 1D) containing

numerous elliptical ovules (Fig. 1F). In contrast, male flowers
(Fig. 1E) have a style that is shorter than the height of anthers
and an ovary with a more V-shaped configuration. Ovules are
rudimentary, much smaller in size, and have an irregular surface
(Fig. 1G) compared with the rounded, cream-colored, and glistening appearance of ovules in bisexual flowers (Fig. 1F).
Male flowers are significantly smaller than bisexual flowers
(Table 1). In male flowers, the length of flowers from the tip of
sepals to the base of flowers was 75% of the mean length found
on bisexual flowers. The mean pistil length in male flowers was
also significantly shorter (70% the length of bisexual flowers).
Longer pistil heights in bisexual flowers are primarily the result
of more extensive growth of the ovary region (which was 300%
longer in bisexual versus male flowers) and to a lesser extent
the result of increases in the height of the stigma + style +
stylopodium (7% taller in bisexual versus male flowers).
Bisexual flowers have stigmas that are close to or greater than
the height of the anthers. In contrast, the height of stigmas in
male flowers is generally well below the height of the tallest
anthers with the mean height 14 mm below anther height.
SEM observations of bisexual flowers show a prominent,
slightly irregular disc-shaped stigma ending on a smooth
elongated style (Fig. 2A). Elongated stigmatic papillae extend
on the surface of the stigma. At anthesis, the stigmatic surface
is covered with a copious exudate that partially obscures the
basal portion of papillae (Fig. 2B). The style has a uniseriate
epidermis and is composed of several layers of elongate cells
evident in longitudinal sections (Fig. 2C). Micrographs taken of
J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

Table 1. Flower characteristics of bisexual and male pomegranate

flowers showing differences in the sizes of flower organs.z
Bisexual
Male
flowers
flowers
Flower characteristicy
(mean SD)
26.7 2.9 b
Length of pedicel to tip of
36.0 2.7 ax
sepals (mm)
Pistil length (mm)
27.5 2.1 a
19.4 2.9 b
17.1 2.0 a
15.9 2.5 b
Stigma + style length (mm)w
Ovary height (mm)
10.5 1.8 a
3.5 2.1 b
Height of stigma exsertion (mm)
1.4 1.6 a
13.8 1.8 b
Anthers (no.)
137 34 b
193 53 a
Sepals (no.)
5.2 0.7 a
5.4 0.7 a
z

Values are means taken from 100 bisexual and 100 male flower types.
Characteristics correspond to letters as noted on the accompanying
figure.
x
Means within a row (comparing bisexual versus male flowers)
followed by different letters are significantly different at P # 0.05
using Students t test.
w
Measurement was taken from the tip of the stigma to the base of the
stylopodium.
y

flowers that were bagged to prevent pollination (Fig. 2D) show

an open, irregularly shaped, centrally located stylar canal (Fig.
2D). In cross-section, circular groups of vascular strands that
transverse the length of the stigma are peripherally oriented
around the stylar canal (Fig. 2E). Exudate is present lining the
central stylar canal. Germinating pollen and tubes are evident
85

exudate. Tubes transverse the stigma

and then grow within the stylar
canal. An SEM cross-section shows
a tuft-like collection of filamentous
pollen tubes growing through the
stylar canal (Fig. 2G). Styles fractured longitudinally show hundreds
of pollen tubes growing within the
stylar canal (Fig. 2H). Pollen tubes
are numerous, forming clusters of
chord-like aggregations within the
canal.
The stigmas of male flowers are
well developed and have a discshaped configuration with elongated
stigmatic papillae (Fig. 3A) similar
to that seen in bisexual flowers.
However, at the time of petal opening, stigmas of male flowers have
markedly less stigmatic exudate so
that the smooth surfaces as well as
the basal portions of the stigmatic
papillae are clearly evident (Fig.
3B). This is in contrast to the partially submerged stigmatic cells in
bisexual flowers (Fig. 2B). The
shorter styles of male flowers have
a dominant central canal clearly
evident in longitudinal sections
(Fig. 3CD). Although stigmas on
male flowers possess reduced
amounts of exudates, pollen germination and tube extension were observed (Fig. 3E). However, pollen
tubes often lacked directional growth
and were observed growing in the
air (Fig. 3E) in contrast to pollen
tube growth in bisexual flowers,
which was more confined to along
the stigmatic surface (Fig. 2F). Pollen
tubes in male flowers generally failed
to penetrate the stigma and were not
observed within the stylar canal even
in hand-pollinated flowers (Fig. 3F).
Ovules in bisexual flowers are
numerous, elliptical, and tightly
packed (Fig. 4A). Ovules are anatropous and at maturity have completely inverted so that the opening
to the micropyle (point) is positioned
Fig. 2. Stigmas and styles of bisexual pomegranate flowers. (A) Scanning electron microscopy (SEM) of an
unpollinated flower showing a rounded, disc-shaped stigma (st) ending on a smooth elongated style (sy). (B) adjacent to the funiculus (Fig. 4C
Higher magnification of the stigmatic surface showing elongated papillary cells (pc) with abundant exudate D). Removed ovules reveal the ataccumulation characteristic of bisexual flowers. (C) Longitudinal section of the stigma (st) and style (sy) tachment site of the funicle, which is
showing uniseriate epidermis (e) and elongated stylar cells. (D) Cross-section of the style in an unpollinated shown as a scar. The outer surface of
flower showing vascular strands (vs) and stylar canal. (E) Higher magnification of the open central stylar canal
ovules in bisexual flowers has regu(sc). (F) SEM of the stigmatic surface of a pollinated flower showing numerous pollen grains (pg) and some
germinated pollen and tubes (pt). (G) Cross-section of the style of a pollinated flower showing massive numbers larly shaped epidermal cells that
of pollen tubes within the stylar canal. (H) Longitudinal fracture of a style showing the canal region with form a smooth surface. In contrast,
aggregates of pollen tubes (pt). Scale bars: AD, GH = 100 mm, EF = 50 mm.
ovules in male flowers are considerably less developed (Fig. 4B) and
on the stigmas of pollinated flowers (Fig. 2F). Pollen grains can be noticeably smaller than those in bisexual flowers (ovules
varied from those showing various degrees of hydration with in Fig. 4A versus Fig. 4B are 10 larger). Ovule surfaces in male
some pollen grains becoming partially submerged by stigmatic flowers have epidermal cells that appeared sunken and exhibit
86

J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

structures could be observed with

ovules attached at a funicular region
and enclosed within a differentiated
bi-integument (Fig. 5CD). However,
integuments had incomplete development of cell layers and exhibited cell
collapse and separation of the inner
and outer integument layers (Fig.
5D). Ovules could be partially anatropous (Fig. 5D) or lack curvature
and remain in an upright, orthotropous orientation (Fig. 4E). Some
flowers exhibited a drastic failure in
the developmental program (Fig. 5E
F). Undifferentiated cellular protrusions formed from limited mitotic
activity (Figs. 4F and 5F). Structures
emerging from placental regions
lacked a differentiated embryo sac,
funiculus, or integuments.
Stamens are numerous in both
flower types and are composed of
red-colored filaments ended by light
yellow anthers (Fig. 1AB). Bright
yellow pollen is released from welldifferentiated anthers, which dehisce
along two abscission zones (Fig.
6A). Pollen is spherical and tricolpate with a smooth exine surface
(Fig. 6B). Pollen from both bisexual
and male flower types is similar in
size (20 mm). However, male
flowers have significantly more stamens per flower (i.e., 1.4 times more
numerous) than bisexual flowers
(Table 1). Pollen viability in both
bisexual and male flowers is relaFig. 3. Stigmas and styles of functionally male pomegranate flowers. (A) Scanning electron microscopy (SEM) of tively high as assessed by in vitro
rounded stigma (st) from a male flower showing an elongated style (sy). (B) Stigmatic surface with elongated germination assays with percent gerstigmatic papillary cells (pc); no accumulation of stigmatic exudates is observed. (C) Longitudinal light mination routinely at 75% to 80%.
microscopy section of stigma and stylar region showing stylar canal (sc). (D) Higher magnification of the stylar
Whole mounts of a cleared pistil
canal. (E) SEM of the stigmatic surface of a pollinated flower. Some pollen has germinated; however, aerial
pollen tube growth (pt) is observed. (F) Longitudinal fracture of stylar region showing an absence of pollen tube from a bisexual flower show an
expanded stigmatic region and vasgrowth within the stylar canal (sc). Scale bars: A, C = 200 mm, B, DE = 50 mm, F = 100 mm.
cular stands (linear darker areas) that
extend from the stigma (Fig. 6C) and
cellular collapse. Ovules show variable stages of arrested de- continue down the length of the style (Fig. 6E). Aniline blue
velopment ranging from those that show collapse primarily in staining of tissues observed under ultraviolet light shows the path
apical regions (Fig. 4E) to those that exhibit more extensive of pollen tubes throughout much of the stigma (Fig. 6D) but
abortive development so that ovules are rudimentary, shriveled, which become directed to grow only within the central stylar
severely degenerated, and appear as collapsed emergences canal (Figs. 6D and 6F). Figures 6C versus 6D and Figures 6E
(Fig. 4F).
versus 6F are companion micrographs taken of identical locaThe orientation of ovules along the placental surface is tions but viewed under regular transmitted versus ultraviolet
shown in LM sections of the two flower types (Fig. 5). In light, respectively. The paths of hundreds of pollen tubes with
bisexual flowers (Fig. 5AB), ovules are elliptical in longitu- callose plugs are shown in a higher magnification of the stylar
dinal section and are bitegmic with a two-layered integument. canal in a bisexual flower (Fig. 6G). This is in contrast to male
The inner integument is composed of two to three cell layers, flowers, in which pollen tubes were seldom observed in stylar
whereas the outer integument is thicker. At floral maturity, the canals (micrographs not shown). A lack of pollen tubes in the
axis of the ovule is inverted and curved downward so that the stylar canal of male flowers is in agreement with observations
micropylar opening faces the placenta and base of the funiculus made with LM-sectioned and SEM material (Figs. 3C and 3E).
(Fig. 5B).
Pollen germination assays conducted at different temperaVariations in ovule development were observed in male tures showed that pomegranate pollen germination is markedly
flowers (Fig. 5CF). In some cases, some partially developed influenced by temperature (Fig. 7). At 25 and 35 C, percent
J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

87

male and female flower parts occurs

in all flowers, the degree of development in pistils fluctuates. Male
flowers have decreased pistil development characterized with shortened stylar length and abortive
ovules. Ovule degeneration can be
slight or severe with structures from
the placenta being rudimentary and
shriveled. As a consequence of degenerated pistil development, only
bisexual flowers with well-developed
ovules are capable of setting fruit. Sex
determination similarly results from
floral organ degeneration in a number
of other taxa. In dioecious date palm
(Phoenix dactylifera), female flowers
have functional carpels but vestigial
sterile staminodes; male flowers have
functional stamens but underdeveloped carpels (DeMason and Stolte,
1982). Likewise, pistillate and staminate flower development in dioecious
pistachio (Pistacia vera) proceeds
with androecial and gynoecial primordia initiation in both floral types,
which is followed by unisexual development resulting from arrested
development of male or female development (Hormaza and Polito,
1996). In mahogany (Swietenia macrophylla), a monoecious species, the
differentiation of male and female
flowers takes place during the later
stages of floral development leading
to arrested ovule development in
male flowers and aborted microspore
Fig. 4. Morphology of pomegranate ovules from bisexual (A, C, D) and functionally male (B, E, F) flowers. (A) development in female flowers
Ovules (ov) in bisexual flowers are elliptical and tightly packed. Some ovules have been removed to show the (Gouvea et al., 2008). Olive (Olea
point of attachment to the placenta (pl). (B) Ovules in male flowers are markedly small and underdeveloped. (C) europaea) produces bisexual and
Ovule from a bisexual flower showing a smooth outer surface; the attachment site of the funicule is shown as
male flower types. Functionally
a scar. (D) Higher magnification of ovules from bisexual flowers. The micropylar (m) opening where pollen will
enter to fertilize the zygote is adjacent to the funiculus (f). (E) Ovules (ov) from a male flower showing male, staminate flowers result from
underdeveloped and collapsed regions. (F) Ovules (ov) from a male flower with more extensive abortive the abortion of pistils (Cuevas et al.,
development so that ovules are rudimentary, shriveled, and severely degenerated. Scale bars: AB = 200 mm, C 1999). Unisexual flower formation is
E = 100 mm, F = 50 mm.
not always formed from the degeneration of floral parts. For example,
pollen germination was similar and reached 74% to 79%, two monoecious tree species, Juglans (Lin et al., 1977; Sattler,
respectively, at 5 h incubation. Germination was rapid with 1973) and Carya (Wetzstein and Sparks, 1983, 1984), are
a T50 (time to 50% germination in the population of germinat- structurally unisexual from initiation throughout development.
ing pollen grains) of 45 min. In contrast, germination was
A difference in stigmatic exudate was another distinction
significantly reduced at the two lower temperatures. Percent observed between male and bisexual pomegranate flowers.
germination at 1 h was only 6.6% and 0.5% at 15 and 5 C Critical stages of pollenstigma interactions include pollen
temperatures, respectively, compared with 50% at the two capture, adhesion, hydration, germination, and tube growth
higher temperatures. The T50 at 15 C was 2 h. Maximum (Hiscock and Allen, 2008) of which stigmatic secretions play an
percent germination at 15 C was 73% of that obtained at 25 C. important role. In bisexual pomegranate flowers, stigmatic
Germination was reduced by over 90% when pollen assays exudate was copious, well-developed, and covered the base
were conducted at 5 C.
of stigmatic papillae. In contrast, exudate in male flowers was
much reduced. Although male flowers lacked abundant exudate, pollen germination and extensive tube growth were still
Discussion
noted. However, pollen tube growth appeared to lack direcPomegranate flowers are large, showy flowers with vibrant tional growth, failed to penetrate the stigma, and was rarely
petals and numerous anthers. Although organogenesis of both observed in the style. Stigmatic exudate has been shown to be
88

J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

tubes needed for fertilization of

ovules required for subsequent aril
development.
In contrast to the divergent pistil
development observed between bisexual and male pomegranate flowers,
no differences in pollen morphology
or performance were observed. Pollen
from bisexual and male flowers both
exhibited high germination values of
greater than 70%, which showed no
statistical differences. A similar lack
of differences in pollen viability
between sexual morphs has been
reported in other andromonoecious
species, including olive (Cuevas and
Polito, 2004), Euphorbia nicaeenis
(Narbona et al., 2008), and Euphorbia boetica (Narbona et al., 2005). In
pomegranate, male flowers produced
significantly greater numbers of anthers suggesting that the development
of staminate flowers may serve to
enhance pollination by producing
greater amounts of pollen. Pollen
surplus has been proposed to enhance
male fitness and be a selective advantage (Connolly and Anderson, 2003;
Podolsky, 1993). Male flowers of
Capparis sinosa, an andromonoecious shrub, produced larger anthers
and smaller ovaries than bisexual
flowers indicating that the role of
male flowers is to save resources for
female function and serve primarily
as a pollen donor (Zhang and Tan,
2009).
Temperatures during the flowering
Fig. 5. Histological details of pomegranate ovules from mature hermaphroditic (AB) and functionally male
flowers (CF). (A) Longitudinal section of a bisexual flower showing elliptical well-developed ovules (ov). (B) period can have a pronounced effect
Closer detail of an ovule showing the funiculus (f), outer integument (oi), inner integument (ii), micropyle (m), on reproductive processes. The effects
nucellus (n), chalaza (c), and embryo sac (es). (CF) Different degrees of ovule degeneration found in male of temperature on pollen germination
flowers. (C) Ovules from a male flower, which has partially degenerated structures. Integuments and the have been documented in a number
funniculus show collapse and separation. (D) Higher magnification of ovules similar to those in C. (E) Severely
degenerated ovules from a male flower; ovule development was arrested at an early stage and appears as limited of fruit tree crops. In the current study,
cellular protrusions. (F) Higher magnification of a severely degenerated ovule. Scale bars: A, C = 100 mm; pomegranate pollen was markedly
B, E = 50 mm; D, F = 25 mm.
affected by temperature. Pollen germination was significantly reduced at
5 and 15 C, whereas the highest and
important in providing directional cues for pollen tubes to grow similar rates of germination were obtained at 25 and 35 C. That
toward and into the stigma by providing directional gradients of pomegranate pollen germination was not inhibited at 35 C is
water potential (Lush et al., 2000) and in acting as a chemo- unusual in that germination and tube growth are often reduced
tropic attractant directing pollen tubes to grow toward stylar at such an elevated temperature. In comparisons of nine Pistacia
canals (Kim et al., 2003).
species and cultivars over a range of temperatures from 4 to
The hollow stylar canal in pomegranate can support the 40 C, the mean temperature optimum for pollen germination
growth of large numbers of pollen tubes. Styles stained with was 24 C but was severely reduced under both higher and lower
aniline blue confirmed that massive numbers of pollen tubes temperatures; no pollen germination was observed for any of the
grow through the stylar canal in pollinated bisexual flowers. genotypes at 40 C (Acar and Kakani, 2010). Similarly in mango,
Pomegranate fruits are composed of a leathery pericarp that optimum pollen germination occurred at 15 and 25 C, whereas
contains hundreds of arils. Each aril is derived from an ovule a decrease in germination occurred at 30 C (Sukhvibul et al.,
and is the result of an independent fertilization event. The juice- 2000). In the Central Valley of California, which is a major
containing translucent layer surrounds a sclarified seedcoat that production area for pomegranates in the United States, flowering
encloses an embryo. The presence of a large central stylar canal occurs from May to November (Stover and Mercure, 2007); thus,
would facilitate the transport of the large numbers of pollen a wide range of temperatures during bloom may occur. Most
J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

89

in temperature. In peach, germination assays conducted at 10, 20, or

30 C affected primarily the rate of
pollen germination with final germination only slightly affected (Hedhly
et al., 2005).
Among angiosperms, labile sex
determination appears to be more
common among dioecious and monoecious plants than among hermaphrodites (Korpelainen, 1998).
Andromonoecy or the appearance
of bisexual and functionally male
flowers on an individual plant can
provide ecological and evolutionary
advantages. Changes in sex allocation are considered a strategy to enhance fitness over the lifetime of
the plant (Policansky, 1981) and can
allow control of male and female
function by modifying the emphasis
of ovules or pollen development.
With andromonoecy, the ratio of bisexual and male flowers can change
with season, plant age, position
within the plant, and environment.
Gender diphasy or sex choice involves a developmental decision
between two modes of prefertilization investment in femaleness (i.e.,
ovule production) versus maleness
(Schlessman, 1988). In general,
female:male flower ratios decline
as environmental resources become
more limiting (Korpelainen, 1998).
Low nutrition, low light, limiting
photosynthate, and stress from
weather or water conditions can
promote decreased relative numbers
of female to male flowers. This is
observed in a number of taxa such
Fig. 6. Pomegranate anther, pollen, and growth of pollen tubes within the pistil. (A) Anther (a) after dehiscence
showing release of pollen grains (pg). (B) Spherical and tricolpate pollen grain with a smooth exine surface (p = as Solanum (Miller and Diggle,
pore). (C) Cleared whole mount of a stigma stained with decolorized aniline blue and viewed under normal 2007), Euphorbiaceae (Narbona
transmitted light. Stigmatic papillae (point) are on the surface; darkened regions are vascular strands
et al., 2008), Myrtaceae (Primack
(arrowhead). (D) The same region as in C but viewed under ultraviolet light showing fluorescing pollen tubes and Lloyd, 1980), and Proteaceae
growing in the stigma and through the stylar canal. (E) Transmitted light view of a cleared portion of the style,
also stained with decolorized aniline blue, showing darker vascular strands. (F) Identical region as in E viewed (Walker and Whelan, 1991).
The timing of flower types can
under ultraviolet light to show central fluorescing band composed of hundreds of pollen tubes growing within the
stylar canal. (G) Higher magnification showing individual callose plugs (cp) of pollen tubes growing within the show seasonal and yearly variations.
style. Scale bars: A, G = 200 mm; B = 5 mm; CF = 100 mm.
In studies conducted in Israel with
Mules Head and Wonderful
flowering occurs in May and June, when high temperatures can pomegranate trees, Shulman et al. (1984) observed normal
be an issue. The high germination rates at elevated temperatures fertile flowering to proceed for 5 weeks before abnormal
observed with the Wonderful pollen used in this study suggest unfruitful flowering. Yet, in another season, flowering of both
that inhibition of pollen at high temperatures during bloom may flower types occurred together. Also in pomegranate, Martinez
be less of a concern than in other tree crops. Temperature effects et al. (2000) found that the proportion of the two flower types
can exhibit genotypic differences. In a comparison of several varied among cultivars and year to year. In Leptospermum
olive cultivars, the optimum temperatures for pollen germination scoparium, an andromonoecious montane shrub, hermaphrodite
were generally 20 to 25 C; however, Amigdalolia pollen flowers tended to open in the first flush of flowering with the
germinated similarly well at 25 and 30 C (Koubouris et al., proportion of male flowers increasing later in the season
2009). Considering the importance of pollination for aril and (Primack and Lloyd, 1980). Chaudhari and Desai (1993) showed
fruit development, studies comparing different pomegranate that applications of plant growth regulators can influence the
cultivars may be warranted, particularly in areas with extremes sex expression distribution of flower types in pomegranate.
90

J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

Fig. 7. Percent germination of pomegranate pollen over time using in vitro

germination assays conducted at different temperatures. Mean separations
compare different temperatures within an assay time using Duncans multiple
range test; means with a different letter are significantly different at P # 0.05.
Values are means of eight replications with 100 pollen grains each.

Gibberellic acid induced more male flowers and reduced

hermaphrodite flowers, whereas ethrel and maleic hydrazide
induced more hermaphrodite and fewer male flowers.
Although pomegranate flowers are commonly described as
falling into either male or bisexual flower types, sex expression
appears to follow a spectrum where flowers range from those
that have strong pistil development with elongated styles and
well-developed ovules to those with severely degenerate ovules
having only vestigial development. Intermediate forms exist
within this range in which male flowers can express different
degrees of ovule degeneration.
Furthermore, our more recent work evaluating flower
quality indicates that female flowers vary in size and number
of ovules (H.Y. Wetzstein, unpublished data). Verification of
this gradient condition of flower quality can have important
implications in crop production applications. Sex expression in
pomegranate can vary with season and time of bloom. Studies
directed toward determining when male and bisexual flower
determination occurs are warranted and would provide insight
into what environmental and physiological factors may be
contributing to sex determination in pomegranate. Optimizing
cultural conditions may be a means to promote the development
of greater numbers of bisexual flowers with high vigor to obtain
maximum fruit set and yield.
Literature Cited
Acar, I. and V.G. Kakani. 2010. The effects of temperature on in vitro
pollen germination and pollen tube growth of Pistacia spp. Sci. Hort.
125:569572.
Basu, A. and K. Penugonda. 2008. Pomegranate juice: A heart healthy
fruit juice. Nutr. Rev. 67:4956.
Bertin, R.I. 1982. The evolution and maintenance of andromonoecy.
Evol. Theory 6:2532.
Chaudhari, S.M. and U.T. Desai. 1993. Effects of plant growth
regulators on flower sex in pomegranate (Punica granatum L.).
Indian J. Agr. Sci. 63:3435.
Connolly, B.A. and G.J. Anderson. 2003. Functional significance of
the androecium in staminate and hermaphroditic flowers of Solanum
carolinense (Solanaceae). Plant Syst. Evol. 240:235243.
Cuevas, J., K. Pinney, and V.S. Polito. 1999. Flower differentiation,
pistil development and pistil abortion in olive (Oleo europaea L.).
Acta Hort. 474:293296.
Cuevas, J. and V.S. Polito. 2004. The role of staminate flowers in the
breeding system of Oleo europaea (Oleraceae): An andromonoecious, wind-pollinated taxon. Ann. Bot. (Lond.) 93:547553.

J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.

DeMason, D.A. and K.W. Stolte. 1982. Floral development in Phoenix

dactylifera. Can. J. Bot. 60:14371446.
El Sese, A.M. 1988. Physiological studies on flowering and fruiting
habits of some pomegranate cultivars under Assiut conditions. Assuit
J. Agr. Sci. 19:320336.
Gouvea, C.F., M.C. Dornelas, and A.P. Martinelli. 2008. Characterization of unisexual flower development in the endangered mahogany tree Swietenia macrophylla King. (Meliaceae). Bot. J. Linn. Soc.
156:529535.
Hedhly, A., J.I. Hormaza, and M. Herrero. 2005. The effect of
temperature on pollen germination, pollen tube growth, and stigmatic
receptivity in peach. Plant Biol. 7:476483.
Herlihy, C.R. and C.G. Eckert. 2002. Genetic cost of reproductive
assurance in a self-fertilizing plant. Nature 416:320323.
Hiscock, S.J. and A.M. Allen. 2008. Diverse cell signaling pathways
regulate pollen-stigma interactions: The search for consensus. New
Phytol. 179:286317.
Holland, D., K. Hatib, and I. Bar-Yaakov. 2009. Pomegranate:
Botany, horticulture, breeding. Hort. Rev. (Amer. Soc. Hort. Sci.) 35:
127191.
Hormaza, J.I. and V.S. Polito. 1996. Pistillate and staminate flower
development in dioecious Pistacia vera (Anacardiaceae). Amer. J. Bot.
83:759766.
Jensen, W.A. 1962. Botanical histochemistry. Freeman, San Francisco, CA.
Kim, S., J.-C. Mollet, J. Dong, K. Zhang, S.-Y. Park, and E.M. Lord.
2003. Chemocyanin, a small basic protein from the lily stigma, induces
pollen tube chemotropism. Proc. Natl. Acad. Sci. USA 100:61256130.
Korpelainen, H. 1998. Labile sex expression in plants. Biol. Rev.
Camb. Philos. Soc. 73:157180.
Koubouris, G.C., I.T. Metzidakis, and M.D. Vasilakakis. 2009. Impact of
temperature on olive (Olea europaea L.) pollen performance in relation
to relative humidity and genotype. Environ. Exp. Bot. 67:209214.
Lansky, E.P. and R.A. Newman. 2007. Punica granatum (pomegranate) and its potential for prevention and treatment of inflammation
and cancer. J. Ethnopharmacol. 109:177206.
Lin, J., B. Shabany, and D. Ramos. 1977. Pistillate flower development
and fruit growth in some english walnut cultivars. J. Amer. Soc. Hort.
Sci. 102:702703.
Lush, W.M., T. Spurck, and R. Joosten. 2000. Pollen tube guidance by
the pistil of a solanaceous plant. Ann. Bot. 85:3947.
Mars, M. 2000. Pomegranate plant material: Genetic resources and
breeding, a review. Options Mediterraneennes Serie A 42:5562.
Martinez, J.J., P. Melfagejo, and F. Martinez. 2000. Study of the floral
morphology of the pomegranate clones: PTO8, CRO1 and ME14.
Options Mediterraneenees Serie A 42:105113.
Miller, J.S. and P.K. Diggle. 2007. Correlated evolution of fruit size
and sexual expression in andromonoecious Solanum sections Acanthophora and Lasiocarpa (Solanaceae). Amer. J. Bot. 94:17061715.
Narbona, E., P.L. Ortiz, and M. Arista. 2005. Dichogamy and sexual
dimorphism in floral traits in the andromonoecious Euphorbia
boetica. Ann. Bot. (Lond.) 95:779787.
Narbona, E., P.L. Ortiz, and M. Arista. 2008. Sexual dimorphism in the
andromonoecious Euphorbia nicaeensis: Effects of gender and
inflorescence development. Ann. Bot. (Lond.) 101:717726.
Podolsky, R.D. 1993. Evolution of a flower dimorphismHow
effective is pollen dispersal by male flowers. Ecology 74:22552260.
Policansky, D. 1981. Sex choice and the size advantage model in jackin-the-pulpit (Arisaema triphyllum). Proc. Natl. Acad. Sci. USA 78:
13061308.
Primack, R.B. and D.G. Lloyd. 1980. Andromonoecy in the New Zealand
montane shrub manuka, Leptospermum scoparium (Myrtaceae). Amer.
J. Bot. 67:361368.
Sattler, R. 1973. Organogenesis of flowers. Univ. Toronto Press, Toronto,
Canada.
Schlessman, M.A. 1988. Gender diphasy (sex choice), p. 139153.
In: J.L. Doust and L.L. Doust (eds.). Plant reproductive ecology
patterns and strategies. Oxford Univ. Press, New York, NY.

91

Seeram, N.P., R.N. Schulman, and D. Heber (eds.). 2006. Pomegranates: Ancient roots to modern medicine. CRC Press, Boca Raton, FL.
Shulman, Y., L. Fainberestein, and S. Lavee. 1984. Pomegranate fruit
development and maturation. J. Hort. Sci. 59:265274.
Stover, E. and E.W. Mercure. 2007. The pomegranate: A new look at
the fruit of paradise. HortScience 42:10881092.
Sukhvibul, N., A.W. Whiley, V. Vithanage, M.K. Smith, V.J. Doogan,
and S.E. Hetherington. 2000. Effect of temperature on pollen
germination and pollen tube growth of four cultivars of mango
(Mangifera indica L.). J. Hort. Sci. Biotechnol. 75:214222.
Tanurdzic, M. and J.A. Banks. 2004. Sex-determining mechanisms in
land plants. Plant Cell 16:S61S71.
Walker, B.A. and R.J. Whelan. 1991. Can andromonoecy explain low
fruit:flower ratios in the Proteaceae? Biol. J. Linnaean Soc. 44:41
46.
Wetzstein, H.Y. and D. Sparks. 1983. The morphology of pistillate
flower differentiation in pecan. J. Amer. Soc. Hort. Sci. 108:997
1003.

92

Wetzstein, H.Y. and D. Sparks. 1984. The morphology of staminate

flower differentiation in pecan. J. Amer. Soc. Hort. Sci. 109:245
252.
Wilson, M.F. 1983. Plant reproductive ecology. Wiley, New York,
NY.
Woo, S.M. and H.Y. Wetzstein. 2008. Morphological and histological
evaluations of in vitro regeneration in Elliottia racemosa leaf
explants induced on media with thidiazuron. J. Amer. Soc. Hort. Sci.
133:167172.
Yi, W., S.E. Law, D. McCoy, and H.Y. Wetzstein. 2006. Stigma
development and receptivity in almond (Prunus dulcis). Ann. Bot.
(Lond.) 97:5763.
Yi, W., S.E. Law, and H.Y. Wetzstein. 2003. An in vitro study of
fungicide effects on pollen germination and tube growth in almond.
HortScience 38:10861088.
Zhang, T. and D.Y. Tan. 2009. An examination of the function of male
flowers in an andromonoecious shrub Capparis spinosa. J. Integr.
Plant Biol. 51:316324.

J. AMER. SOC. HORT. SCI. 136(2):8392. 2011.