Content uploaded by Paula J. Rudall
Author content
All content in this area was uploaded by Paula J. Rudall on Dec 19, 2013
Content may be subject to copyright.
EVOLUTIONARY TRANSITIONS AMONG FLOWERS OF PERIANTHLESS PIPERALES:
INFERENCES FROM INFLORESCENCE AND FLOWER DEVELOPMENT IN THE
ANOMALOUS SPECIES PEPEROMIA FRASERI (PIPERACEAE)
Margarita Remizowa,* Paula J. Rudall,1
,
yand Dmitry Sokoloff*
*Higher Plants Department, Biological Faculty, Moscow State University, 119992 Moscow, Russia; and
yJodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey, United Kingdom
Floral and inflorescence structure and ontogeny are described in detail in Peperomia fraseri, an anomalous
species of Piperaceae that differs in several respects from other species of Peperomia and other perianthless
Piperales (Piperaceae and Saururaceae). Inflorescence structure is atypical in this species, with numerous spikes
arranged spirally on a long raceme. There is a gender distribution of flowers along each spike, with the
proximal (lower) part bearing bisexual flowers and the distal (upper) region bearing female flowers.
Furthermore, there is a high degree of polymorphism in the structure (including presence or absence) of flower-
subtending bracts in the same inflorescence. Bract suppression at the basalmost abaxial position in P. fraseri
could be explained by a strong inhibitory activity of the spike-subtending bract or its primordium. Bract
suppression among scattered flowers in the lower (bisexual) region is interpreted as an example of somatic
instability. By contrast, more distal bract and stamen suppression seems to be part of a general tendency to
flower reduction toward the spike apex in P. fraseri, where spatial constraints in the inflorescence control the
size and shape of the bract and floral primordia, resulting in stamen suppression in smaller primordia.
However, it is difficult to understand why an abracteate flower is formed in P. fraseri rather than a sterile bract,
as in Peperomia hispidula. Furthermore, in contrast with most other Piperaceae, P. fraseri has no apical
residuum on mature spikes; at least some partial inflorescences bear a terminal flowerlike structure, although
this more closely resembles those of some Saururaceae than the ‘‘true’’ terminal flower of perianth-bearing
Piperales, such as Aristolochiaceae. Peperomia fraseri and other atypical members of Piperaceae represent
potentially useful models for examination of patterns of bract and stamen loss in perianthless Piperales, in
which spatial constraints in the developing spike are important for floral construction.
Keywords: basal angiosperms, bract, floral evolution, ontogeny, Piperales, stamen suppression, terminal flower.
Introduction
The monophyly of the five families of the basal angiosperm
order Piperales sensu Angiosperm Phylogeny Group (1998,
2003) has been demonstrated by molecular phylogenetic
analyses (Qiu et al. 2000). The order displays two distinct
and contrasting patterns of flower and inflorescence mor-
phology: perianthless/indeterminate and perianth-bearing/
determinate. These two patterns are clearly correlated with
infraordinal relationships based on recent molecular analyses
(Jaramillo et al. 2004). Flowers of one subclade (Aristolo-
chiaceae, Lactoridaceae, and possibly Hydnoraceae) possess
a distinct perianth and are usually arranged in cymose partial
inflorescences (i.e., with a terminal flower; Gonza
´lez 1999;
Gonza
´lez and Rudall 2001). In the other subclade (Saurura-
ceae and Piperaceae), flowers are entirely perianthless, and
partial inflorescences are dense and racemose (spikes, spadi-
ces, or rarely racemes; Cheng-Yih and Kubitzki 1993). No
morphological transitions have been described between peri-
anthless and perianth-bearing Piperales, although some Sau-
ruraceae reportedly possess a terminal flower (Nozeran 1955;
Rohweder and Treu-Koene 1971; Tucker 1979, 1981).
The relatively simple floral structure of the perianthless Pi-
perales has facilitated ontogenetic studies that have been
used as a model for examining patterns of floral evolution
(Tucker et al. 1993; Hufford 1997; Jaramillo et al. 2004).
Here we examine reproductive structure and development in
Peperomia fraseri, which is highly unusual in Piperaceae be-
cause individual partial inflorescences (spikes) are aggregated
into a dense raceme (figs. 1, 2) in contrast to the condition in
most other Piperaceae, which possess either a single spike or
a group of spikes clustered into an umbel (Yuncker 1958).
Reproductive development in P. fraseri has not previously
been examined in detail; Murty (1959) included it in his sur-
vey of 20 species but did not report detailed structure of the
inflorescence. Our preliminary examination of mature repro-
ductive structures in this species revealed that some spikes
apparently possess a terminal flower (fig. 2G,2H) but others
do not. Furthermore, there is some variation in bract pres-
ence and floral gender in this species. Both of these factors
are normally stable in perianthless Piperales (Tucker 1975,
1979, 1980, 1981, 1982a, 1982b, 1985; Liang and Tucker
1989, 1990, 1995; Liang 1994; Tucker and Douglas 1996;
Lei and Liang 1999). Peperomia species normally have
1Author for correspondence; e-mail p.rudall@rbgkew.org.uk.
Manuscript received December 2004; revised manuscript received June 2005.
925
Int. J. Plant Sci. 166(6):925–943. 2005.
Ó2005 by The University of Chicago. All rights reserved.
1058-5893/2005/16606-0004$15.00
a strongly conserved floral morphology, almost invariably
with a well-developed peltate (or hypopeltate; Leinfellner
1953) bract subtending each flower (Tucker 1980). Pepero-
mia flowers are usually bisexual, possessing two stamens in
the lateral position and a single, probably monocarpellary,
gynoecium (Tucker 1980; Igersheim and Endress 1998). In
other Piperaceae and Saururaceae, stamen number, position,
and sequence of initiation are relatively variable, but bract
presence is more stable, though in two genera of Saururaceae
(Houttuynia and Anemopsis), the subtending bracts of the
proximal (lower) flowers in a spike are showy and petaloid,
and the whole inflorescence has a flowerlike appearance
(Wood 1971). Thus, a detailed ontogenetic study of P. fraseri
could help to elucidate evolutionary transitions in floral and
inflorescence morphology in perianthless Piperales.
Material and Methods
Peperomia fraseri C.DC. (Peperomia resedaeflora Linden
& Andre) is native to western tropical South and Central
Fig. 1 Herbarium specimen of Peperomia fraseri (K).
926 INTERNATIONAL JOURNAL OF PLANT SCIENCES
Fig. 2 Peperomia fraseri, inflorescence morphology. A–C, General view of inflorescences. A, Nonfasciated main inflorescence (raceme of
spikes) and three paracladia. Paracladia are much delayed in development compared with main inflorescence. Two paracladia are situated in axils
of very small leaves (arrowheads); a third paracladium occurs in the axil of a foliage leaf. B, Paracladium (yellow arrow) inserted slightly above the
lowest spike (red arrow) of the main inflorescence. C, Fasciated main inflorescence. D, Distal portion of young main inflorescence, general view.
Spike-subtending bracts are longer than spikes at this stage. E, Portion of main inflorescence at anthesis. Numerous stalked spikes are borne along
the main inflorescence axis. Each spike is subtended by a nonpeltate bract. F, Spike at anthesis showing structure of bisexual flowers. G,H, Spike
apices at anthesis. Flowers are female in the distal portion of the spike. In both spikes illustrated, there is a single abracteate female flower that
occupies an almost terminal position on the spike. BR ¼bract, G¼gynoecium, G surrounded by circle ¼gynoecium of abracteate female flower,
ST ¼stamen.
Fig. 3 Peperomia fraseri, mature spike morphology. A–C, Top view of distal region of spike, which bears female flowers. A, Spike terminated
by four bracteate female flowers. B, Single abracteate female flower in an almost terminal position (marked by black letters fm). One of the upper
female flowers has a reduced bract. C, Two abracteate female flowers (black letters fm) at the top of the spike. D, General view of a spike with
lower region of bracteate bisexual flowers and upper region of female flowers; the female flowers are bracteate with the exception of a single
flower (black letters fm), which occupies an almost terminal position on the spike. E, Typical structure of a bisexual flower with a peltate bract,
two bisporangiate stamens, and a gynoecium. F,G, Rare example of a spike peduncle with two flowers at the very base, so that an elongated spike
stalk is present between the basal flowers and the spike. One of the two basal flowers is abracteate (illustrated in fig. 2F); the other one is bracteate
(not visible in fig. 2F, as it is situated on the opposite side of the spike peduncle; illustrated in fig. 2G). BR ¼flower-subtending bract of normal
size, br ¼reduced flower-subtending bract, fm ¼female flower, G¼gynoecium of bisexual flower, MIA ¼main inflorescence axis, SS ¼spike
stalk, SSB ¼spike-subtending bract, ST ¼stamen. Scale bars: A–D¼500 mm, E¼200 mm, F¼1 mm, G¼300 mm.
America (Ecuador, Peru, and Mexico). It is a greenhouse
ornamental plant in various countries. We examined both
wild-source herbarium specimens (K) (fig. 1) and living
material of Ecuadorian origin cultivated in the Princess of
Wales Conservatory at the Royal Botanic Gardens, Kew
(voucher specimen deposited at K). Five different well-
branched living plants were available, probably from the
same accession, with each individual bearing hundreds of
spikes, thus allowing sufficient sampling to examine spike
variability. No significant morphological differences were
observed between herbarium and living material. In addi-
tion, three good-resolution images of P. fraseri were found
Fig. 4 Peperomia fraseri, bract initiation and early development. A,B, Top view of apical meristem of main inflorescence (i.e., raceme of
spikes), showing various ontogenetic stages of spike-subtending bracts. A, Nonfasciated inflorescence with 8 þ13 contact parastiches. B,
Fasciated inflorescence. C, Portion of a young raceme of spikes showing different stages of spike initiation and early development. Most spike-
subtending bracts are removed to show spikes. D, Spike primordium before initiation of flower-subtending bract. E, First flower-subtending bracts
initiated on spike apical meristem. F, Small portion of young raceme of spikes showing seven spikes at different developmental stages; spike-
subtending bracts have been removed. G, Very young spike, abaxial view. Note flower-subtending bract primordia are suppressed in the lower
abaxial side of the spike. BR ¼flower-subtending bract, SA ¼spike apex, SSB ¼spike-subtending bract. Scale bars: A,C¼500 mm; B¼1 mm;
D¼80 mm; E,G¼60 mm; F¼120 mm.
929
REMIZOWA ET AL.—REPRODUCTIVE DEVELOPMENT IN PEPEROMIA FRASERI
on the Web, and they are very similar to the material
studied.
For developmental investigation, living material was fixed
in FAA and transferred into 70% ethanol. For scanning elec-
tron microscope (SEM) examination, material was dissected
and then dehydrated in an ethanol series. Dehydrated mate-
rial was then critical-point dried, mounted onto SEM stubs
using double-sided adhesive tape, coated with platinum or
gold using a sputter coater, and examined using a Hitachi
cold field emission SEM S-4700-II (2 kV) and a Hitachi SEM
S-405A (15 kV). In total, more than 200 digital images were
saved.
Results
Organography
Inflorescences. Peperomia fraseri possesses a polytelic
synflorescence. The main inflorescence (Hauptfloreszenz sensu
Weberling 1981, 1989) is a raceme of spikes (figs. 1, 2). Indi-
vidual spikes also may be described as spadices because their
primary axis is relatively thick, especially at its distal part.
Each raceme contains numerous (up to 300) spikes (fig. 2A–
2C), whereas each spike contains relatively few (25–30) flow-
ers (figs. 2F,6D). All spikes are situated in the axils (or slightly
above the axils) of bilateral (nonpeltate) and lanceolate sub-
tending bracts (fig. 2D,2E;fig.3F). A terminal spike is always
absent. Spike-subtending bracts are spirally arranged along the
main axis of the raceme. For clarity and brevity, in the descrip-
tion of P. fraseri below, ‘‘raceme’’ always means a compound
inflorescence (raceme of spikes) rather than a simple raceme,
which was never observed in this species.
Stem fasciation is relatively common but is usually con-
fined to the primary axis of the main inflorescence (fig. 2C),
in which case the raceme axis may be distally bifurcated or
flattened or even split into several separate racemes. In race-
mes with a fasciated main axis, the lateral spikes have a nor-
mal structure.
In addition to the main inflorescence, most plants of P. fra-
seri possess one or several lateral branches (paracladia) that
are structurally similar to the main inflorescence but have
fewer spikes per raceme. First-order paracladia occur in the
axils of the leaves of the main axis below the main inflores-
cence. Leaves that subtend paracladia can possess a well-
developed blade similar to leaves of the vegetative region, or
they can have a reduced blade. They are often arranged in
verticels of three or in pairs, echoing the arrangement of
leaves in the vegetative region of the plant. Paracladium-
subtending leaves of the same node can contrast strongly in
size (fig. 2A). Sometimes, paracladia are inserted between lat-
eral spikes along the main inflorescence axis. For example,
figure 2Bshows a paracladium that is inserted higher on the
axis than the lowest spike in the main inflorescence. Each
paracladium has two small lateral prophylls. Second-order
paracladia are usually initiated in their axils, each also bear-
ing two prophylls in which the next order paracladia are ini-
tiated, and so on. Second- and higher-order paracladia are
not always fully developed.
Each spike can be divided into three distinct regions: (1)
a naked stalk, (2) a lower (proximal) flowering region con-
taining bisexual flowers, and (3) an upper (distal) flowering
region containing unisexual (female) flowers. In regions 2
and 3, constituting the flowering zone, flowers are spirally ar-
ranged along the spike axis. During and especially after an-
thesis, the spike axis becomes elongated basally (proximally)
and remains relatively short distally so that lower bisexual
flowers are well spaced from each other whereas female flow-
ers are densely packed.
Spike stalk. The basal portion of the lateral spike pedun-
cle is elongated (ca. 1 mm long just before anthesis) and usu-
ally lacks flowers, prophylls, or bracts. This naked elongated
portion (the spike stalk) is about twice as long as the spike at
anthesis and elongates further after anthesis. In rare exam-
ples, the peduncle possesses two flowers at the very base, in
which case an elongated spike stalk is present between the
basal flowers and the spike. In these unusual examples, the
basal flowers are either bracteate (fig. 3G) or abracteate (fig.
3F) but always bisexual. The basal flowers are situated on
the peduncle in transverse orientation, and thus their bracts
(which were apeltate in our material; fig. 3G) should be in-
terpreted as prophylls.
Proximal, bisexual flowering region. In the lower (proxi-
mal) half of the flowering zone, the flowers are bisexual and
mostly subtended by well-developed peltate bracts. However,
in one or a few of the lower flowers, the bract is reduced or
absent. Typically, in the lowermost flower that is situated in
an abaxial position on the spike axis, i.e., the flower closest
to the spike-subtending bracts, the bract is reduced or absent.
However, abracteate flower(s) can also occur on the adaxial
side of the peduncle; indeed, abracteate flowers are not neces-
sarily the lowermost flowers on the whole spike.
Bisexual flowers lack any vestige of a perianth and consist
of only two stamens and a gynoecium (fig. 3E). The gynoe-
cium is unilocular, containing a single ovule and terminating
in a sessile stigma covered by unicellular papillae. Just before
anthesis, when papillae are short, the stigma often appears
bilabiate, with left and right lobes. Sometimes, however, the
stigma appears to be entire. The stamens typically resemble
each other in both size and shape. They are inserted in
a transverse-abaxial position in the flower. Anthers are bi-
sporangiate. Exceptionally, some bisexual flowers with un-
equal stamens are found. Rarely, one of the upper bisexual
flowers bears only a single stamen. Bisexual flowers are pro-
togynous. One of the basalmost flowers was female in only
a single example of several hundred spikes examined; in the
basal part of the same spike, it was followed by bisexual
flowers.
Distal, unisexual flowering region. Flowers in the upper
(distal) region of the spike are entirely unisexual and female.
Female flowers also wholly lack a perianth and consist only
of a gynoecium. Most female flowers are subtended by well-
developed peltate bracts. In most spikes, one or two of the
uppermost female flowers are abracteate or possess a minute
reduced bract, but in some spikes, all the flowers are sub-
tended by bracts. There is no residual apex at the spike apex.
In some cases, a single abracteate female flower appears to
terminate the spike axis at anthesis. The gynoecium of a fe-
male flower is morphologically similar to the gynoecium of
a bisexual flower. An unusual feature found in a few of the
abracteate female flowers is a three-lobed stigma.
930 INTERNATIONAL JOURNAL OF PLANT SCIENCES
Fig. 5 Peperomia fraseri, development of bisexual flowers. A,B, Horseshoe-shaped flower primordia. C–E, Stamen and gynoecium initiation.
In E, stamen primordia are already initiated, but the gynoecium primordium is still not visible. BR ¼flower-subtending bract of normal size,
br ¼reduced flower-subtending bract, FP ¼floral primordium, G¼gynoecium primordium, ST ¼stamen primordium. Scale bars: A¼30 mm;
B¼20 mm; C,F¼100 mm; D¼40 mm; E¼60 mm.
Fig. 6 Peperomia fraseri, variation in degree of bract suppression in some flowers in the basalmost region of the spike. A,B, General views of
young spikes from abaxial side. A, Basalmost flower in abaxial position, with highly reduced bract. The second flower in an abaxial position also
Indumentum. Stems, spike-subtending and flower-
subtending bracts, and gynoecia are all covered by small
glandular hairs. The raceme axis also bears unicellular non-
glandular simple hairs, though these are rare on spike pedun-
cles. All organs, including the stamens, bear ethereal oil cells
in the epidermis. Stomata were observed only on bracts.
Organogenesis
Inflorescence. The apical meristem of the raceme, if not
fasciated, is 120–250 mm in diameter and dome shaped. The
height of the meristem above the level of formation of
the first primordia is about half its diameter at this level. The
apical meristem produces spirally arranged spike-subtending
bract primordia in an acropetal sequence. Spike-subtending
bract primordia typically form 8 þ13 contact parastichies
(fig. 4A). The primordia are hemispherical and ca. 30 mmin
diameter at initiation. They soon elongate in a transverse di-
rection to become ellipsoidal.
In fasciated racemes, the apical meristem is almost linear
when viewed from above. It may be more than 1.5 mm long
and 50–250 mm wide (the width is variable along the meri-
stem length; fig. 4B). When the spike-subtending bract is ca.
100 mm long and 60 mm wide, a small (ca. 20 mm) hemi-
spherical spike primordium is initiated in its axil (fig. 4C,
4D). The first flower-subtending bract primordia (fig. 4E) are
initiated when the spike primordium (now spike apical meri-
stem) reaches 70–90 mm in diameter. At that point, the
spike-subtending bract is ca. 600 mm long. The spike apical
meristem is dome shaped when the first flower-subtending
bract primordia are initiated on it but almost flat during its
final stages of activity.
At initiation, flower-subtending bract primordia are
rounded in outline and 20–30 mm in diameter (fig. 4F,4G).
They are initiated in a rapid sequence, rendering details dif-
ficult to discern. However, at least in some cases, the first
two flower-subtending bract primordia are initiated in a
transverse-abaxial position and the third one in an adaxial
position. At later stages of spike development, the flower-
subtending bracts are spirally arranged, with 5 þ8 contact
parastichies often recognizable. The peltate blade of the bract
is formed first, and its stalk appears later, following interca-
lary growth.
Proximal, bisexual flowering region. Typically, primordia
of bisexual flowers are initiated when flower-subtending
bracts are already well differentiated. However, a few pri-
mordia of bisexual flowers are initiated in the axils of small
undifferentiated bracts (fig. 5B,5E; fig. 6A,6B); in the ma-
ture spike, these bracts apparently remain smaller than the
other bracts (e.g., reduced bract in fig. 6D). The lowermost
abaxial flower and/or some other flowers can be initiated in
the absence of their subtending bract (fig. 6C), or the bract
can be represented by a scarcely visible bulge (fig. 6E,6F).
Young primordia of bisexual flowers are ca. 30–40 mm3
15–20 mm in size and are horseshoe shaped, the concavity
denoting the abaxial side of the primordium (fig. 5A,5B;
fig. 6A). When the floral primordium reaches 65–80 mmin
width, two stamen primordia are initiated at each transverse-
adaxial end (fig. 5E), followed by the gynoecium primordium
in the center (fig. 5C,5D,5F). Stamen and gynoecium pri-
mordia are almost hemispherical at initiation. During early
development, the gynoecium is much smaller than the sta-
mens. The gynoecium develops as an ascidiate structure (fig.
8D). When it reaches ca. 60 mm, its ascidiate nature becomes
apparent because of a small depression at the center. Growth
of the meristematic ring around this depression soon closes
the gynoecial opening so that ovule initiation is not visible
without dissection.
Distal, unisexual flowering region (figs. 7, 8). In the up-
per half of the spike (the female-flowered region), the lower
flowers are initiated in the axils of their subtending bracts.
Like the bisexual flowers in the proximal region of the spike,
they are usually initiated when the bract is already quite
large and differentiated as a peltate structure (fig. 7A). Some-
times the flower primordium deviates slightly to the right or
left from the median plane of its subtending bract (fig. 8A).
At initiation, the female flower primordium is typically ellip-
soidal, being elongated in a transverse plane. Primordia of
lower female flowers are intermediate between ellipsoidal
and horseshoe shaped. The entire primordium produces the
gynoecium, which develops in the same way as in bisexual
flowers, though there is much variability in gynoecial shape.
The uppermost female flowers usually are not subtended
by bracts at initiation, or the bract is very small. We did
not find examples of bract initiation after formation of
flower primordia. There is no residual apex; at the top of the
spike apical meristem, one or two abracteate female flowers
are initiated that fill almost all the available space. Normally,
no single female primordium occupies an exact terminal
position on the spike apical meristem, but we observed a
few cases of initiation of a female flower in a terminal posi-
tion. The uppermost female flowers are delayed in initiation
and development compared with other female flowers (fig.
8B–8F).
During development of the lowermost female flowers (i.e.,
those that are situated immediately distal to the bisexual
flowers along the spike axis), one or two stamen primordia
can be initiated (fig. 7F). These stamen primordia are then ar-
rested in development and remain as small, undifferentiated
structures during late developmental stages. They are not vis-
ible in mature flowers.
has a reduced bract. B, Basalmost flower in abaxial position with a reduced bract. C, General view of young spike from adaxial side with
basalmost adaxial flower abracteate. D, General view of mature spike. One of the bisexual flowers has a reduced bract. E, Adaxial view of a rare
spike type in which one of the lowermost flowers is female. Two of the bisexual flowers (represented by horseshoe-shaped primordia at this stage)
are abracteate. F, Young abracteate bisexual flower situated in the basal region of the adaxial side of the spike. Stamen primordia are well formed
at this stage, but the gynoecium primordium is still very small. BR ¼flower-subtending bract (or its primordium) of normal size, br ¼reduced
flower-subtending bract (or its primordium), fm ¼female flower, FP ¼floral primordium, G¼gynoecium or its primordium, ST ¼stamen
or its primordium. Arrowhead ¼position of absent flower-subtending bract. Scale bars: A,B,E¼60 mm; C¼120 mm; D¼500 mm;
F¼30 mm.
933
REMIZOWA ET AL.—REPRODUCTIVE DEVELOPMENT IN PEPEROMIA FRASERI
Fig. 7 Peperomia fraseri, development of female flowers and spike apex. A–G, Young spike apices with some flower-subtending bracts
removed. A, Primordia of female flowers subtended by bracts. B, Youngest primordium (arrow), which will probably produce an abracteate
female flower. C, Primordia of female flowers, all subtended by bracts. The fate of the apical portion of the spike apex is still unclear at this
Indumentum. Glandular hairs (trichomes) are initiated
and differentiated at a surprisingly early developmental stage
(fig. 5A,5C; figs. 7H,8E). They can be initiated simulta-
neously with, or even slightly before, floral primordium for-
mation, as well as on small primordia. They differentiate
rapidly; e.g., we found some gynoecial primordia that were
almost the same size as a neighboring trichome. Trichome
position seems to be random. It is not clear whether these tri-
chomes are always present on mature structures; some tri-
chomes were broken in older flowers examined.
Discussion
Comparison with Closely Related Taxa
Peperomia fraseri has many features in common with
other Peperomia species (Johnson 1914; Leinfellner 1953;
Murty 1959; Sastrapradja 1968; Tucker 1980; Takhtajan
1997; Lei and Liang 1999), including flowers arranged in
spikes, peltate bracts, horseshoe-shaped primordia of bisex-
ual flowers, presence of only two stamens in bisexual flowers,
bisporangiate anthers, and a gynoecium that can be inter-
preted as a single ascidiate carpel (fig. 9). However, it differs
in several respects from other species of Peperomia and most
other Piperaceae, distinguished by the unusual structure of
the inflorescence (a raceme of spirally inserted spikes), the
unusual gender distribution of flowers along each spike, the
absence of a naked region at the apex of the mature spike
(fig. 10), and the absence of a subtending bract in some flow-
ers, together with variation in bract size and shape.
Peperomia fraseri is apparently unusual among Piperaceae
in bearing five different flower types plus intermediate forms,
often on the same spike (fig. 11). These are (A) bracteate bi-
sexual flowers, (B) abracteate bisexual flowers, (C) bracteate
functionally female flowers with abortive stamen primordia,
(D) bracteate female flowers without any trace of stamens,
and (E) abracteate entirely female flowers. Intermediate
forms exist between these types; e.g., types A and B and
types D and E are both linked by intermediate flowers with
very small bracts. Type C is intermediate between types A
and D.
Bract and Stamen Loss
Flower-subtending bracts are present in most other Pipera-
ceae, in which initiation of the flower primordium is typically
delayed compared with the differentiation of its subtending
bract, so that the bract is quite large and differentiated by
the time the flower primordium is initiated. Peperomia metal-
lica is exceptional in that it shows a tendency for accelerated
production of floral buds closer to the inflorescence apex
in the last-formed bracts of an older inflorescence bearing
numerous flowers (Tucker 1980). Among other perianthless
Piperales, bract size and shape are more variable in Saurura-
ceae than in Piperaceae. In Saururaceae, the bract and flower
are normally produced from a shared (common) primordium,
except for the most proximal (first-formed) six to eight bracts
at the base of the spike, which are showy and petallike (e.g.,
in Houttuynia; Tucker 1981). Because a perianth is entirely
absent from Saururaceae, the bract could be interpreted as
a perianth organ (Buzgo et al. 2005), especially because it is
adnate to the pedicel in Saururus and adnate to the ovary in
Anemopsis, although in Houttuynia it is free (Wood 1971).
Complete suppression of the bract, as in some flowers of
P. fraseri, is otherwise extremely rare in perianthless Piperales.
However, polymorphism in presence or absence of a well-
developed subtending bract in flowers of the same inflores-
cence has been reported in several other angiosperms; e.g., in
racemes of some Brassicaceae and umbels of some Apiaceae
(Weberling 1981, 1989; Kusnetzova 1988). However, in most
of these cases, bract reduction increases (either gradually or
abruptly) toward the distal end of the main inflorescence
axis. By contrast, in P. fraseri, when some of the basal flow-
ers are abracteate, they are surrounded by bracteate flowers.
Importantly, bracteate flowers can occur above abracteate
flowers on the spike axis. If all flowers are bracteate, the
bracts may be unequal in size; even if the basalmost abaxial
flower is bracteate, its bract is smaller than those of sur-
rounding flowers. There may be different reasons for bract
loss in different positions on the spike.
Bract loss in the lower (bisexual) flowering region of the
spike. Bract suppression and loss at the basalmost abaxial
position could be explained by a strong inhibitory activity of
the spike-subtending bract or its primordium, although this
does not explain bract loss in some basal adaxial flowers of
the spike. Interestingly, shape and size of the abracteate
flower primordia that occur in the lower flowering region of
the spike are not variable in P. fraseri in contrast with the
variability of their subtending bract primordia. The flower
primordium has a horseshoe-shaped form, perhaps as a result
of inhibition caused by the close proximity of its subtending
bract because the primordium is concave on the side of the
bract. If this hypothesis is correct, the inhibitory activity of
the bract is almost constant despite the great variability in its
shape. Indeed, the inhibitory activity is present even when no
visible trace of a bract can be detected.
On the other hand, the extraordinary plasticity in bract
presence/absence and structure among scattered flowers in the
lower (bisexual) region is more difficult to explain. Cases of
stage. D, Primordium at the center of the spike apex that will produce an apparently abracteate female flower. This is delayed in development
compared with other flower primordia. One of the bracteate flowers has a very small bract. E, Detail of D.F, Last-formed flower primordium
(black arrowhead, upper right) is abracteate. Note also a developing functionally female flower with abortive stamen primordia (white
arrowheads). G, Last-formed flower primordium (white arrowhead) is abracteate. A young flower primordium on the other side of the spike apex
has a very small bract (black arrowhead). H, Detail of Gshowing female flower primordium subtended by very small bract. A large glandular hair
is present on the bract. I, Gynoecia of three female flowers showing ascidiate gynoecium development. The entire gynoecium represents a single
ascidiate carpel. J, Ascidiate gynoecium development. BR ¼flower-subtending bract (or its primordium) of normal size, br ¼reduced flower-
subtending bract primordium, fm ¼female flower primordium, G¼gynoecium primordium, ST ¼stamen primordium. Scale bars: A,B,
G¼100 mm; C,D,F¼60 mm; E,J¼30 mm; H¼20 mm; I¼50 mm.
935
REMIZOWA ET AL.—REPRODUCTIVE DEVELOPMENT IN PEPEROMIA FRASERI
Fig. 8 Peperomia fraseri, variation in spike apex development. Young spike apices with some flower-subtending bracts removed. A,B, Female
flower primordia subtended by bracts. In A, the right-hand bract (BR, black letters) subtends the flower primordium (fm, black letters) but is
inserted between two flower primordia. C, Single abracteate female flower primordium (arrowhead). D, Single abracteate young female flower
(arrowhead); this is delayed in development. E,F, Two abracteate female flower primordia (marked by black letters fm). Note also a reduced bract
in E.BR ¼flower-subtending bract (or its primordium) of normal size, br ¼reduced flower-subtending bract primordium, fm ¼female flower
primordium, G¼gynoecium primordium, ST ¼stamen primordium. Scale bars: A–C,E¼100 mm; D¼200 mm; F¼60 mm.
936
Fig. 9 Spike apex morphology typical of the genus Peperomia (Peperomia dolabriformis Kunth; plant cultivated in the Princess of Wales
Conservatory at the Royal Botanic Gardens, Kew). A, Young spike, lateral view. B, Spike slightly before anthesis, lateral view. Note peltate bracts
subtending bisexual flowers, each with two bisporangiate stamens and a gynoecium. C, Top view of young spike showing residual apex. D, Distal
portion of spike showing residual apex. Scale bars: A,C¼500 mm; B¼1 mm; D¼400 mm.
937
somatic instability within a single inflorescence are rare in
flowering plants and may be a result of epigenetic factors such
as variation in methylation resulting in gene silencing (Cubas
et al. 1999; Rudall and Bateman 2003). Furthermore, we can-
not exclude the possibility that this type of bract variation is
teratological. Sastrapradja (1968) reported the occurrence of
teratological inflorescences in other Peperomia species but pro-
vided no details. However, natural variation may be greater
than previously reported; e.g., Piper betel displays consider-
able variation between stamens in flowers of the same spike or
sometimes even in the same flower (Johnson 1910).
Bract and stamen loss in the upper (female) flowering re-
gion of the spike. With respect to floral gender distribution
along the spike, most Piperaceae that have been investigated
developmentally are uniform, although flowers are either
bisexual or unisexual in different spikes. For example, all
American species of Piper produce bisexual flowers, whereas
Old World species of Piper have unisexual flowers and are
usually dioecious (Yuncker 1958). An exception is P. betel,in
which the degree of development of stamens and carpels dif-
fers markedly in flowers of the same spike or sometimes in
different stamens of the same flower (Johnson 1910). Simi-
larly, among other basal angiosperms there are comparable
examples of floral polymorphism of oligomerous flowers
within a given inflorescence, as in the long spike of the basal
monocot Lilaea scilloides (Posluszny et al. 1986), though
Lilaea is not a close relative of Peperomia.Peperomia hispi-
dula displays a gender distribution similar to that of P. fra-
seri; most flowers are subtended by bracts, but bracts formed
at the top of the spike are either sterile (not subtending any
flower) or sometimes subtend female rather than bisexual
flowers (Johnson 1914). The presence of sterile bracts at the
top of the spike mirrors the condition in some Saururaceae
(Tucker 1985). Perhaps significantly, both P. hispidula and
P. fraseri possess relatively few flowers per spike in contrast
to many other peperomias.
Bract and stamen loss at the top of the spike seems to be
part of a general tendency to flower reduction toward the
spike apex in P. fraseri. Floral primordia become narrower
toward the apex. The lower primordia are distinctly horse-
shoe shaped, the upper primordia more elliptical, and the up-
permost primordia almost rounded. Narrowing of the flower
primordium implies reduction of its lateral flanks (i.e., the
positions of stamen initiation). Abortive stamen primordia
are initiated on flower primordia of intermediate shape, and
no stamens are initiated on elliptical and rounded primordia.
Thus, spatial constraints in the inflorescence control the size
and shape of the floral primordium. In turn, constraints of
size and shape in the floral primordium control floral sex by
suppression of the stamens in smaller primordia.
We propose the following model to explain this phenome-
non. The first event of flower organogenesis (FO
1
), which
precedes visible primordium formation, is the establishment
of a region (sector) on the spike apical meristem, which
Fig. 10 Diagrams showing different regions of a spike (not to
scale). A, Most Peperomia species. B,Peperomia fraseri.
Fig. 11 Diagrams showing range of flower types in Peperomia
fraseri often present on the same spike, together with intermediate
forms. Sequences of floral development are also shown; note differ-
ences in shape of floral primordia. A, Bracteate bisexual flower. B,
Abracteate bisexual flower. C, Bracteate functionally female flower
with abortive stamen primordia. D, Bracteate female flower with no
vestige of stamens. E, Abracteate female flower. fsb ¼flower-
subtending bract. Open shapes with thin lines indicate floral
primordia. Filled circles indicate spike axis.
938 INTERNATIONAL JOURNAL OF PLANT SCIENCES
ultimately will form the flower and its subtending bract.
Then, a significant widening of the spike apical meristem
takes place (FO
2
). Finally (FO
3
), formation of bract and
flower primordium occurs within the region identified during
FO
1
.AtFO
3
, because of the widening of the spike apical
meristem, the flower sector is much larger than at the time of
its establishment (FO
1
). In P. fraseri,FO
2
(meristem widen-
ing) is arrested toward the end of the meristem activity
(i.e., a nonterminal deletion in an ontogenetic transition se-
ries). Thus, widening occurs after the establishment of the
first flower regions but not after establishment of the last
flower regions. However, subsequent primordium formation
is not arrested when meristem widening is arrested, and the
primordia therefore occupy a smaller space than in normal
conditions. When the space allowed for flower formation is
slightly smaller, stamens are suppressed, and female bracteate
flowers are initiated. When the space is yet smaller, the bract
also is suppressed. However, it is difficult to understand why
in this case an abracteate flower is formed rather than a sterile
bract, as in P. hispidula (Johnson 1914).
Spike Apex
In contrast with most species of Peperomia and other Pipera-
ceae, P. f r a s e r i has no apical residuum on mature spikes. In most
Piperaceae, the distal portion of the apical meristem remains
uncovered by primordia at the time when cessation of its activity
takes place (Tucker 1980, 1982a, 1982b). The upper bracts,
although initiated, remain small and undifferentiated, and flow-
ers either are not initiated or do not fully develop in their axes
(fig. 9). By contrast, in P. f r a s e r i , the entire surface of the spike
apical meristem produces primordia (fig. 10), and significantly,
most of these primordia produce fully developed organs.
In some spikes of P. fraseri, an abracteate female flower is
initiated in a precisely terminal position on the spike apical
meristem (fig. 3D; fig. 7D,7E; fig. 8D). Thus, in these cases,
the partial inflorescence possesses an apparently terminal
flower. A terminal flower is highly atypical for Piperaceae but
rather common among perianth-bearing Piperales such as
Aristolochiaceae (Gonza
´lez 1999). However, the condition in
P. fraseri, although anomalous in Piperaceae, may not be ge-
netically similar to the ‘‘true’’ terminal flower of Aristolochia-
ceae, in which the inflorescence is a rhipidium (cyme).
Rather, it resembles the condition in some Saururaceae,
in which there is a cluster of sterile bracts, reduced flowers,
and abracteate stamens (Nozeran 1955; Rohweder and
Treu-Koene 1971; Tucker 1979, 1981). In P. fraseri, the exact
position of initiation of the abracteate flower(s) is variable;
sometimes a single female abracteate flower is initiated to
one side of the apex. Female flower(s) situated at the distal-
most region of the spike of P. fraseri are delayed in develop-
ment compared with surrounding flowers. In angiosperms
with a true terminal flower, it is normally accelerated in de-
velopment compared with the nearest lateral flowers (Weberling
1981, 1989). Thus, one explanation is that in P. fraseri, the
abracteate flower may sometimes be shifted to a terminal po-
sition. If so, this is a pseudoterminal flower rather than a
true terminal flower, and the spike is therefore homologous
with an indeterminate structure.
In an extensive review of distribution regularities of differ-
ent sexual types of flowers within compact angiosperm inflo-
rescences, Fedorova (2004) found that in simple racemose
inflorescences without a terminal flower, the lower flowers
are normally either (functionally) female or bisexual and the
upper flowers are either (functionally) male or bisexual. For
example, this pattern occurs in spadices of many Araceae
and inflorescences of Lilaea (Juncaginaceae). If an inflores-
cence has a terminal flower, it is either (functionally) female
or bisexual. Lower flowers may be (functionally) male or bi-
sexual. The cyathium of Euphorbia is an obvious example of
this developmental pattern.
However, if spikes of P. fraseri are interpreted as indeter-
minate, they do not fit Fedorova’s (2004) model because they
have bisexual flowers in the lower part and female flowers in
the upper part. This also applies to spikes of P. hispidula.
This may be because Fedorova’s (2004) flower type distribu-
tion applied to simple racemose inflorescences rather than
a raceme of spikes, as in P. fraseri. On the other hand, if we
accept the flowerlike nature of the mature spike apex in
P. fraseri, it fits better into Fedorova’s (2004) model. The en-
tire abracteate female flower of P. fraseri is represented by
a single gynoecium, which may be interpreted as a single as-
cidiate carpel. Thus, the entire spike apex resembles an apo-
carpous multicarpellate gynoecium, and in this respect the
spike apex of P. fraseri has a flowerlike nature. Flowers with
several free ascidiate carpels in a spiral arrangement do not
occur in Piperales, but they are present in many other basal
angiosperms (e.g., Endress 2001). Subtending bracts of upper
female flowers in P. fraseri may be compared with the peri-
anth of such a flower. In some spikes, we found upper bracts
somewhat displaced to one side or the other relative to the
subtended carpel (fig. 8A). This may be regarded as a shift
toward alternate attachment of the upper carpels and bracts,
making the spike apex appear even more flowerlike.
The terminal flowerlike structure in P. fraseri is quite dif-
ferent from typical flowers of Peperomia. However, this is
not an uncommon phenomenon in angiosperms. In species
that normally possess determinate inflorescences (cymes), the
morphology of the terminal flower often differs from that of
lateral and wild-type flowers. Similarly, terminal flowers of
atypical structure have often been observed in spontaneous
mutants of species that normally possess indeterminate inflo-
rescences (Rudall and Bateman 2003), so it is sometimes un-
clear whether this structure is a terminal flower or a terminal
pseudanthium. Indeed, Worsdell (1916) cited spontaneous
terminal peloric flower mutations in Digitalis as an example
of fasciation (i.e., teratologically fused flowers). Recent inves-
tigations on model organisms with normally indeterminate
inflorescences have produced artificially induced, heritable
terminal flower mutations, e.g., by suppression of the
CENTRORADIALIS (CEN) gene in Antirrhinum (Coen and
Nugent 1994; Bradley et al. 1996) or the orthologous
TERMINAL FLOWER1 (TFL1) gene in Arabidopsis (Bradley
et al. 1997; Ratcliffe et al. 1998; see also Ezhova and Penin
2001; Penin et al. 2005). This type of mutation may influ-
ence inflorescence structure in other ways. For example,
in the ‘‘delayed terminal flower’’ phenotype of some partial
revertants of an X-radiation-induced cen mutant of Antir-
rhinum majus, delayed formation of the terminal flower
resulted in a variable number of lateral flowers (Cremer et al.
2001).
939
REMIZOWA ET AL.—REPRODUCTIVE DEVELOPMENT IN PEPEROMIA FRASERI
Inflorescence Architecture
Peperomia fraseri differs from most other species of Peper-
omia, and from other Piperaceae and Saururaceae, in the
presence of numerous spikes arranged in a long raceme. This
inflorescence structure is unusual for Piperales but also oc-
curs in Peperomia peltigera, another Ecuadorean species (see
illustration in Sodiro 1900). However, similar inflorescence
construction with spikes arranged in a raceme occurs in some
unrelated angiosperm groups, e.g., in some species of Astrag-
alus (Fabaceae) (Barneby 1964).
The entire raceme of P. fraseri is similar in some respects
to the spicate inflorescences of most peperomias. In particu-
lar, it possesses spirally arranged spike-subtending bracts
with 8 þ13 contact parastichies. This is the most common
type of floral-subtending bract phyllotaxis in simple spikes of
Peperomia (Zhitkov 1977), in which there is a major change
in phyllotaxis between the vegetative part of the shoot and
the spike. The same is true for the raceme of spikes in P. fra-
seri, in which leaves below the raceme are opposite or verti-
cillate. The position of spike insertion is highly variable in
other Peperomia species with more orthodox inflorescence
structure; spikes can be either leaf-opposed or axillary, even
within the same species (Yuncker 1958).
Thus, the raceme of spikes in P. fraseri is similar to the simple
spikes of the majority of other perianthless Piperales. Interest-
ingly, the entire simple spike of perianthless Piperales has some
features in common with typical flowers, such as those of
perianth-bearing Piperales and other magnoliids. This similar-
ity is most prominent in Anemopsis and Houttuynia (Saurura-
ceae), in which the lower flower-subtending bracts in the spike
are petaloid. It is possible that such similarities between the
flower and simple spike on the one hand, and the simple spike
and double inflorescence on the other, are similar in nature.
We outline three evolutionary scenarios (A1–A3) to ex-
plain the similarity between spikes of Houttuynia and Ane-
mopsis and flowers of more orthodox magnoliids (fig. 12).
Scenarios A1 and A2 require that flower and inflorescence
morphology of extant perianth-bearing Piperales such as Sar-
uma (Aristolochiaceae) are similar to those of a shared ances-
tor of Piperales. Scenario A3 requires that the spikes in
perianthless Piperales are similar to the reproductive struc-
tures of ancestral Piperales. Thus, scenario A3 is the least
plausible based on phylogenetic optimization because flowers
of some extant perianth-bearing Piperales such as Saruma
closely resemble those of other basal angiosperms (Jaramillo
and Kramer 2004; Jaramillo et al. 2004).
Scenario A1. Spikes of perianthless Piperales evolved as
a result of strong reduction in flowers and their aggregation
into partial inflorescences. Showy inflorescence bracts evolved
in some Saururaceae, such as Anemopsis and Houttuynia
(Tucker and Douglas 1996).
Scenario A2. Partial loss of flower meristem identity oc-
curred in ancestors of perianthless Piperales so that instead
of producing a flower, the floral meristem started to produce
structures combining features of a flower and an inflores-
cence. This may explain why some perianthless Piperales
(Anemopsis and Houttuynia) possess flowerlike spikes.
Scenario A3. In the lineage of perianth-bearing Piperales,
spikes resembling those of extant perianthless Piperales were
transformed into flowers (scenario A3a). Another possibility
(scenario A3b) is that a common ancestor of Piperales had
no well-defined flowers but its reproductive structures were
similar in some respects to extant perianthless Piperales, es-
pecially Saururaceae. The latter scenario could indicate that
‘‘true’’ flowers evolved multiple times from a prefloral state,
as suggested by several authors (Meeuse 1975). However, the
hypothesis of multiple origins of true flowers is not supported
by data from flowers of extant basal angiosperms (Endress
2001) or early fossil flowers (Friis et al. 2000). The only early
angiosperm fossil that might be interpreted as a ‘‘prefloral’’
state is Archaefructus (Sun et al. 2002), but it is equally pos-
sible to interpret its reproductive structures as a reduced in-
florescence (Friis et al. 2003).
We propose three further related scenarios (B1–B3) to ex-
plain the similarity between the simple spikes of most perianth-
less Piperales and the raceme of spikes in P. fraseri. Scenarios
B1 and B2 both require that inflorescence morphology in
Saururaceae and most Piperaceae is closely similar to that of
a common ancestor of perianthless Piperales. Scenario B3 re-
quires that inflorescence morphology of the P. fraseri type is
primitive within perianthless Piperales. Thus, as with scenario
A3, scenario B3 is the least plausible based on phylogenetic op-
timization because inflorescence morphology of the P. fraseri
type is relatively unusual in angiosperms.
Scenario B1. The raceme of spikes in P. fraseri evolved as
a result of aggregation of simple spikes and reduction in
spike structure. This hypothesis is consistent with widely ac-
cepted views of inflorescence evolution in basal angiosperms
(scenario A1, above). It is also at least partially consistent
with the ‘‘pseudocycle concept’’ of evolution (Kusnetzova
1986, 1988, 1998), which hypothesized gradualistic, cyclic
evolution of similar morphological patterns.
Scenario B2. In ancestors of P. fraseri and its relatives,
a partial loss of flower meristem identity took place. Instead
of a flower, the floral meristem produced structures sharing
features of a flower and a partial inflorescence, as in Ane-
mopsis and Houttuynia, which would explain why the apex
of the ‘‘simple’’ spike of P. fraseri has some flowerlike fea-
tures. This scenario is consistent with (most likely sudden) re-
duction or loss of function or expression of floral meristem
identity genes, such as LEAFY (LFY) or its ortholog FLORI-
CAULA (FLO). For example, in leafy mutants of Arabidop-
sis, lower inflorescence nodes form secondary short shoots
rather than flowers (Ratcliffe et al. 1998), thus resembling
the condition seen in P. fraseri. In these mutants, a bract pri-
mordium subtending an axillary apex is formed instead of
a flower primordium. When the apex of the axillary meri-
stem has grown clear of the influence of the subtending bract,
the inflorescence architecture gene TFL1 is expressed in the
apical cells, thereby terminating the shoot. Interestingly, dif-
ferent floral identity genes (AP1 and CAL) become activated
at upper nodes in the main inflorescence in these mutants, re-
sulting in formation of carpellate flowers rather than shoots.
Similar mutants, both artificial and spontaneous, have been
described in Fabaceae and putatively ascribed to mutation of
a gene orthologous to LFY or FLO (Hirsch et al. 2002).
Scenario B3. Reduction in spike structure resulted in trans-
formation of entire spikes of the P. fraseri type into flower-
like structures. Further reduction resulted in transformation
940 INTERNATIONAL JOURNAL OF PLANT SCIENCES
of the entire raceme of spikes (as in P. fraseri) into a simple
raceme, spike, or spadix, as in most perianthless Piperales.
Conclusions
Current evidence is insufficient to demonstrate either evo-
lutionary scenarios B1 or B2. Ultimately, a synthesis of evi-
dence is required from paleontology, floral and inflorescence
morphology, and ontogeny and developmental genetics in
a phylogenetic context across a wide range of extant peri-
anthless Piperales. Whether a meristem develops as a flower
or a shoot could depend on simple heterochronic shifts in ex-
pression of genes responsible for inflorescence architecture
and/or floral identity (Ratcliffe et al. 1998). Furthermore,
spatial constraints in the developing spike are important for
floral construction. Peperomia fraseri and other atypical
Fig. 12 Summary of broader phylogenetic context of Piperaceae (as reviewed in Stevens 2005; see also Qiu et al. 2000), illustrating
evolutionary scenarios outlined in text.
941
REMIZOWA ET AL.—REPRODUCTIVE DEVELOPMENT IN PEPEROMIA FRASERI
members of perianthless Piperales, such as Peperomia hispi-
dula (Johnson 1914) and Piper betel (Johnson 1910), repre-
sent potentially useful models for further examination of
patterns of floral and inflorescence construction and bract
and stamen loss in Piperales and other basal angiosperms.
Acknowledgments
M. Remizowa and D. Sokoloff gratefully acknowledge
funding from the Systematics Research Fund and the Royal
Society, respectively, which enabled them to visit the Royal
Botanic Gardens, Kew, for a month in August 2004. This
work was also partially supported by a grant from the
Russian Foundation for Basic Research (03-04-48427) and
a President of Russia’s grant for support of leading scientific
schools (1898.2003.4). Thanks also to C. Prychid, N. A.
Zvonkova, and G. N. Davidovich for assistance with SEM
work. We are grateful to R. M. Bateman for critically read-
ing the manuscript and to M. McMahon and an anonymous
reviewer for helpful comments.
Literature Cited
APG (Angiosperm Phylogeny Group) 1998 An ordinal classifi-
cation for the families of flowering plants. Ann Mo Bot Gard 85:
531–553.
——— 2003 An update of the Angiosperm Phylogeny Group
classification for the orders and families of flowering plants: APG
II. Bot J Linn Soc 141:399–436.
Barneby RC 1964 Atlas of North American Astragalus. Mem NY
Bot Gard 13:1–1188.
Bradley D, R Carpenter, L Copsey, C Vincent, S Rothstein, E
Coen 1996 Control of inflorescence architecture in Antirrhinum.
Nature 379:791–797.
Bradley D, O Ratcliffe, C Vincent, R Carpenter, E Coen 1997
Inflorescence commitment and architecture in Arabidopsis. Science
275:80–83.
Buzgo M, DE Soltis, PS Soltis, H Ma, BA Hauser, J Leebens-Mack,
B Johansen 2005 Perianth development in the basal monocot
Triglochin maritima. Aliso (forthcoming).
Cheng-Yih W, K Kubitzki 1993 Saururaceae. Pages 586–587 in
K Kubitzki, JG Rohwer, V Bittrich, eds. The families and genera
of vascular plants. II. Magnoliid, Hamamelid and Caryophyllid
families. Springer, Berlin.
Coen ES, JM Nugent 1994 Evolution of flowers and inflorescences.
Development 120:107–116.
Cremer F, WE Lonnig, H Saedler, P Huijser 2001 The delayed
terminal flower phenotype is caused by a conditional mutation in
the CENTRORADIALIS gene of snapdragon. Plant Physiol 126:
1031–1041.
Cubas P, C Vincent, E Coen 1999 An epigenetic mutation responsible
for natural variation in floral symmetry. Nature 401:157–161.
Endress PK 2001 The flowers in extant basal angiosperms and
inferences on ancestral flowers. Int J Plant Sci 162:1111–1140.
Ezhova TA, AA Penin 2001 A new BRACTEA (BRA) gene control-
ling the formation of an indeterminate bractless inflorescence in
Arabidopsis thaliana. Russ J Genet 37:772–775.
Fedorova TA 2004 Compact inflorescences: regularities of the origin
and details of distribution of sexual types of flowers as a basis of
dioecism origin. Pages 110–115 in NP Savinykh, ed. Materials of
the tenth school of theoretical plant morphology: construction units
in plant morphology. Vyatka State University of Humanities, Kirov.
(In Russian.)
Friis EM, JA Doyle, PK Endress, Q Leng 2003 Archaefructus:
angiosperm precursor or specialized early angiosperm? Trends
Plant Sci 8:369–373.
Friis EM, KR Pedersen, PR Crane 2000 Reproductive structure and
organization of basal angiosperms from the Early Cretaceous
(Barremian or Aptian) of western Portugal. Int J Plant Sci
161(suppl):S69–S182.
Gonza
´lez F 1999 Inflorescence morphology and the systematics of
Aristolochiaceae. Syst Geog Pl 68:159–172.
Gonza
´lez F, PJ Rudall 2001 The questionable affinities of Lactoris:
evidence from branching pattern, inflorescence morphology, and
stipule development. Am J Bot 88:2143–2150.
Hirsch AM, RSN Krupp, Y Lin, SS Wang, W Yang, SC Tucker
2002 Inflorescence and flower development in wild-type and sid
mutant Melilotus alba, white sweetclover. Can J Bot 80:732–740.
Hufford L 1997 The roles of ontogenetic evolution in the origins of
floral homoplasies. Int J Pl Sci 158(suppl):S65–S80.
Igersheim A, PK Endress 1998 Gynoecium diversity and systematics
of the paleoherbs. Bot J Linn Soc 127:289–370.
Jaramillo MA, EM Kramer 2004 APETALA3 and PISTILLATA
homologs exhibit novel expression patterns in the unique perianth
of Aristolochia (Aristolochiaceae). Evol Dev 6:449–456.
Jaramillo MA, PS Manos, EA Zimmer 2004 Phylogenetic relation-
ships of the perianthless Piperales: reconstructing the evolution of
floral development. Int J Plant Sci 165:403–416.
Johnson DS 1910 Studies in the development of the Piperaceae. I.
The suppression and extension of sporogenous tissue in the flower
of Piper betel L. var. monoecium C.DC. J Exp Zool 9:715–749.
——— 1914 Studies in the development of the Piperaceae. II. The
structure and seed-development of Peperomia hispidula. Am J Bot
1:357–397.
Kusnetzova TV 1986 On a phenomenon of pseudocyclic similarity in
higher plants. J Gen Biol 47:218–233. (In Russian with English
summary.)
——— 1988 Angiosperm inflorescences and different types of their
structural organization. Flora 181:1–17.
——— 1998 Reduction in inflorescence, its essence and role in the
evolution of modul organisms. J Gen Biol 59:74–103. (In Russian
with English summary.)
Lei LG, HX Liang 1999 Variations in floral development in
Peperomia (Piperaceae) and their taxonomic implications. Bot J
Linn Soc 131:423–431.
Leinfellner W 1953 Die hypopeltaten Brakteen von Peperomia.
Oesterr Bot Z 100:601–615.
Liang HX 1994 On the systematic significance of floral organogen-
esis in Saururaceae. Acta Phytotax Sin 32:425–432.
Liang HX, SC Tucker 1989 Floral development in Gymnotheca
chinensis (Saururaceae). Am J Bot 76:806–819.
——— 1990 Comparative study of the floral vasculature in Saurur-
aceae. Am J Bot 77:607–623.
——— 1995 Floral ontogeny of Zippelia begoniaefolia and its
familial affinity: Saururaceae or Piperaceae? Am J Bot 82:681–689.
Meeuse ADJ 1975 Changing floral concepts: anthocorms, flowers,
and anthoids. Acta Bot Neerl 24:23–36.
Murty YS 1959 A contribution to the study of floral morphology of
some species of Peperomia. J Indian Bot Soc 38:120–139.
Nozeran R 1955 Contribution a l’e
´tude de quelques structures
florales. Ann Sci Nat Bot 16:1–224.
942 INTERNATIONAL JOURNAL OF PLANT SCIENCES
Penin A, V Choob, T Ezhova 2004 Basic principles of terminal flower
formation. Russ J Dev Biol 36:65–69.
Posluszny U, WA Charlton, DK Jain 1986 Morphology and de-
velopment of the reproductive shoots of Lilaea scilloides (Poir.)
Hauman (Alismatidae). Bot J Linn Soc 92:323–342.
Qiu YL, J Lee, F Bernasconi-Quadroni, DE Soltis, PS Soltis, M Zanis,
EA Zimmer, Z Chen, V Savolainen, MW Chase 2000 Phylogeny of
basal angiosperms: analyses of five genes from three genomes. Int J
Plant Sci 161(suppl):S3–S27.
Ratcliffe OJ, DJ Bradley, ES Coen 1998 Separation of shoot and
floral identity in Arabidopsis. Development 126:1109–1120.
Rohweder O, E Treu-Koene 1971 Bau und morphologische Bedeutung
der Inflorescenz von Houttuynia cordata (Saururaceae). Viertel-
jahrsschr Naturforsch Ges Zuer 116:195–212.
Rudall PJ, RM Bateman 2003 Evolutionary change in flowers and
inflorescences: evidence from naturally occurring terata. Trends
Plant Sci 8:76–82.
Sastrapradja S 1968 On the morphology of the flower in Peperomia
(Piperaceae) species. Ann Bogor 4:235–244.
Sodiro SJ 1900 Contribuciones al conocimiento de la flora ecuatori-
ana: Monografia. I. Piperaceas Ecuatoriensis. Escuela de Artes y
Oficios, Quito.
Stevens P 2005 Angiosperm Phylogeny Web site, Missouri Botanical
Garden. http://www.mobot.org/MOBOT/Research/APweb/welcome.
html.
Sun G, Q Ji, DL Dilcher, S Zheng, KC Nixon, X Wang 2002 Ar-
chaefructaceae, a new basal angiosperm family. Science 296:
899–904.
Takhtajan AL 1997 Diversity and classification of flowering plants.
Columbia University Press, New York.
Tucker SC 1975 Floral development in Saururus cernuus. 1. Floral
initiation and stamen development. Am J Bot 62:289–301.
——— 1979 Ontogeny of the inflorescence of Saururus cernuus
(Saururaceae). Am J Bot 66:227–236.
——— 1980 Inflorescence and flower development in the Piperaceae.
I. Peperomia. Am J Bot 67:686–702.
——— 1981 Inflorescence and floral development in Houttuynia
cordata (Saururaceae). Am J Bot 68:1017–1032.
——— 1982aInflorescence and flower development in the
Piperaceae. II. Inflorescence development of Piper. Am J Bot 69:
743–752.
——— 1982bInflorescence and flower development in the Piper-
aceae. III. Floral ontogeny of Piper. Am J Bot 69:1389–1401.
——— 1985 Initiation and development of inflorescence and
flower in Anemopsis californica (Saururaceae). Am J Bot 72:
20–31.
Tucker SC, AW Douglas 1996 Floral structure, development, and
relationships of paleoherbs: Saruma,Cabomba,Lactoris,and
selected Piperales. Pages 141–175 in DW Taylor, LJ Hickey, eds.
Flowering plant origin, evolution and phylogeny. Chapman & Hall,
New York.
Tucker SC, AW Douglas, HX Liang 1993 Utility of ontogenetic
characters in determining phylogenetic relationships of Saururaceae
and Piperaceae. Syst Bot 18:614–641.
Weberling F 1981 Morphologie der Blu
¨ten und Blu
¨tensta
¨nde. Ulmer,
Stuttgart.
——— 1989 Morphology of flowers and inflorescences. Cambridge
University Press, Cambridge.
Wood CE 1971 The Saururaceae in the southeastern United States.
J Arnold Arbor Harv Univ 52:479–485.
Worsdell WC 1916 The principles of plant teratology. Ray Society,
London.
Yuncker TG 1958 The Piperaceae, a family profile. Brittonia 10:1–7.
Zhitkov VS 1977 Forms of phyllotaxis in the genus Peperomia Ruiz
et Pav. and their morphogenesis. Bull Mosc Soc Nat, Biol Ser 82:
103–119. (In Russian with English summary.)
943
REMIZOWA ET AL.—REPRODUCTIVE DEVELOPMENT IN PEPEROMIA FRASERI
