|
|
||||||||
|
What's this? |
Anatomy and Morphology |
Institute of Botany, Academia Sinica, Taipei, Taiwan 115, Republic of China; Institute of Systematic Botany, The New York Botanical Garden, Bronx, NewYork, New York 10458-5126 USA
Received for publication June 25, 2006. Accepted for publication March 2, 2007.
ABSTRACT
The subfamily Lecythidoideae of Lecythidaceae (Brazil nut family) is a dominant group in neotropical forests, especially those of Amazonia. New World members of the family have large showy flowers that are either polysymmetric or monosymmetric. In this study, floral organogenesis of all 10 neotropical genera was examined using SEM. Our observations of floral development are put into the context of a molecular phylogeny based on sequences of the ndhF and trnL-F genes (Am. J. Bot. 94: 289301). Floral evolution of the subfamily is explained as having undergone four different levels of complexity in regard to floral symmetry. The basal most genera, Grias and Gustavia, have polysymmetric flowers. At level two, represented only by Couroupita, monosymmetry is established through the expression of abaxial dominance and the development of an androecial hood; at this level, abaxial dominance impacts the perianth and androecium, but not the gynoecium. At the third level, monosymmetry is developed in groups of Couratari and Cariniana domestica; but, in the Allantoma/Cariniana decandra lineage, a reversal back to polysymmetric flowers, resulting from a gradual weakening of abaxial dominance, and the loss of the hood has occurred. Finally, in level four, including Bertholletia, Corythophora, Eschweilera, and Lecythis, monosymmetry is so strongly expressed that the gynoecium is also influenced by abaxial dominance. In this group, the hood is complicated in both structure and function, and the floral axis is changed from straight to slightly inclined. This study demonstrates that the development of floral abaxial dominance is the proximate cause of monosymmetry in the Lecythidoideae. We suggest that monosymmetric flowers are more efficiently pollinated, and therefore the bees and bats that pollinate the monosymmetric flowers in this group are ultimately responsible for the monosymmetry.
Key Words: floral evolution floral organogenesis floral symmetry Lecythidaceae Lecythidoideae monosymmetry pollination polysymmetry
The Brazil nut family (Lecythidaceae), consisting of small to very large trees, is composed of the following subfamilies: Planchonioideae, Foetidioideae, and Lecythidoideae (Prance and Mori, 2004
). In the New World, this family is represented by 10 genera and approximately 230 species belonging to the subfamily Lecythidoideae (Prance and Mori, 1979
, 2004
; Mori and Prance, 1990
). The monophyly of the Lecythidoideae is strongly supported by our recent DNA sequence data (Mori et al., 2007
), a base chromosome number of x = 17 in contrast to x = 13 in the Planchonioideae (Kowal et al., 1977
; Kowal, 1989
), and the normally oriented cortical bundles in the stem (phloem outside, xylem inside) in contrast to having inversely oriented cortical bundles (xylem outside, phloem inside) in the other two subfamilies (Prance and Mori, 1979
).
Lecythidoideae are an important element of the flora of Central and South America between 19° N and 25° S latitudes (Mori and Prance, 1990
). They usually rank as one of the most important tree families of primary forests in this region (Black et al., 1950
; Cain et al., 1956
; Gentry, 1982
; Mori and Boom, 1987
; Mori and Lepsch-Cunha, 1995
). In central Amazonian Brazil, 38 different species of the Brazil nut family are found in a 100 hectare plot with 1114 species and 45149 individuals per hectare (Mori et al., 2001
). The high diversity and density of Lecythidoideae in Latin American lowland forests more than likely correlates with the evolution of monosymmetric flowers pollinated by bees and bats (Mori and Boeke, 1987
) and efficient seed dispersal by animals, wind, and water (Tsou and Mori, 2002
).
Flowers of Lecythidoideae are showy, often aromatic (Knudsen and Mori, 1996
), and pollinated by various groups of bees, infrequently by bats (three species of Lecythis), and perhaps by beetles (species of Grias) (Prance, 1976
; Mori et al., 1980a
; Mori and Boeke, 1987
; Knudsen and Mori, 1996
). Floral diversity of the Lecythidoideae, to a large extent expressed in the androecium (Knuth, 1939
, fig. 1; Prance and Mori, 1979
, figs. 813; Mori and Prance, 1990
, fig. 31), has captured the imagination of all botanists who have monographed or studied the family (Berg, 1858
; Knuth, 1939
; Miers, 1874
; Mori and Prance, 1990
; Prance and Mori, 1979
; Tsou, 1994
). The major variations include floral orientation with the polysymmetric species usually facing upward and the monosymmetric species facing horizontally or downward, differences in floral symmetry with three genera polysymmetric and seven monosymmetric, flower size ranging from 1 to 10 cm or more in diameter, differences in sepal number (from four to six) and calyx fusion (either free or completely fused in the bud), petal number ranging from four to eight but the majority of species with six, stamen number ranging from 10 to over 1000, and last, but most importantly, the variation in the structure of androecial hood of species with monosymmetric flowers. The hood, which is an extension that initiates on the abaxial side of the androecial primordium, arches over the receptacle to become the most prominent part of the flower. This kind of androecium is not found in any other angiosperm in the world. The form of the androecial hood provides important characters for generic circumscriptions (Prance and Mori, 1979
; Mori and Prance, 1990
; Prance, 1990
) and serves as a guide into the flower, a landing platform, and provides a food source (fodder pollen or nectar) for pollinators (Mori and Boeke, 1987
).
The pollinators of monosymmetric flowers, mostly bees and less frequently bats, are attracted to the flowers by their large size, monosymmetry, conspicuous colors, and aromas. In many species, including Bertholletia excelsa and many species of Eschweilera and Lecythis, the androecial hood is yellow throughout or yellow only at the distal end of the hood where pollinators enter the flower to gather nectar. Moreover, bees recognize and presumably are attracted by monosymmetric flowers, but more primitive insects do not recognize the monosymmetry (Giurfa et al., 1996
). The hood also serves as a landing platform for pollinators as well as limits entry into the flower to pollinators with enough strength to force their way into flowers with hoods that are tightly pressed against the summit of the ovary. Flat hoods with an opening between the hood and the summit of the ovary, such as in Couroupita guianensis and some species Lecythis, allow pollinators as well as nonpollinators to enter the flowers. Some species with flat hoods, for example, all species of Corythophora, also exclude nonpollinating insects because their hoods are thickened and pressed against the summit of the ovary. In general, the more complicated the hood the more difficult it becomes for a pollinator to enter the flower. The most complex hoods are found in species of Eschweilera and Couratari (Mori and Prance, 1990
, fig. 31; Prance, 1990
); in these genera pollination is limited to bees strong enough to force their way into the flower, but also with tongues long enough to reach the nectar at the end of the coil. The hood of Bertholletia excelsa (Prance and Mori, 1979
, fig. 9) and some species of Lecythis (see L. holcogyne in fig. 31 in Mori and Prance, 1990
) is not fully coiled, and in these species the petals are tightly pressed against the hood such that a great deal of force is required for entering the flower (Mori and Prance, 1990
, fig. 45).
There are three types of pollinator reward in the Lecythidoideae. In the polysymmetric-flowered Grias and Gustavia, the pollen is homomorphic and the fertile pollen also serves as the pollination reward. The second type is called fodder pollen, i.e., sterile pollen that is mostly found in the hood, but sometimes occurs in the stamens of the staminal ring on the abaxial side of the flower (e.g., Corythophora labriculata). Finally, the third type of reward is nectar, produced only in monosymmetric-flowered species by staminodes at the apex of the inwardly coiled hood. The specialized pollinator rewards of neotropical Lecythidaceae, fodder pollen and nectar, are only produced by species with monosymmetric flowers (Mori and Boeke, 1987
).
Hirmer (1918)
was the first to point out that stamen initiation of the Lecythidaceae was centrifugal. Thompson (1921
, 1927
) proposed that the development of the hood was mainly due to cellular gigantism and that there was dimorphic pollen in Couroupita guianensis. Centrifugal stamen initiation (Leins, 1972
; Endress, 1994
, 1999
) and dimorphic pollen (Jacques, 1965
; Mori et al., 1980b
) have been confirmed by others. Endress (1994
, 1999
) reported an interesting change in the floral symmetry of C. guianensis in which the flowers are first monosymmetric, then polysymmetric, and finally monosymmetric at different stages of development. Mori and Boeke (1987)
studied the pollination biology of Lecythidaceae in French Guiana and observed that the principal pollinators of species that have monosymmetric, nectar-producing flowers and complex androecial hoods are euglossine bees and noted that the distribution of euglossines and that of species of Lecythidaceae with complex hoods and nectar production nearly overlap. They proposed that long-tongued euglossine bees and neotropical Lecythidaceae with complex flowers co-evolved. This does not, however, rule out pollination of neotropical Lecythidaceae by other pollinators such as species of Xylocopa (a xylocopine) and Epicharis (an anthophorid) bees (Nelson et al., 1985
; Gamboa-Gaitán, 1997
).
Mori and Boeke (1987)
further speculated that floral evolution in neotropical Lecythidaceae went from polysymmetric species that produce only fertile pollen (species of Gustavia and Grias), to monosymmetric species with both fertile and fodder pollen (e.g., all species of Corythophora, Couroupita guianensis, and Lecythis sect. Pisonis), and finally to species with complex androecial hoods in which nectar is produced (e.g., Couratari and Eschweilera).
For the past 3 yr, we and Arne Anderberg have been working on a molecular phylogeny of Lecythidoideae based on sequences from the ndhF (Anderberg Lab) and trnL-F (Tsou Lab) genes (Mori et al., 2007
). Concurrent with this study, we also studied floral organogenesis of neotropical Lecythidaceae using the SEM to trace the early floral development. The Lecythidoideae were the major emphasis of this study, and all 10 genera of the subfamily were studied.
Based on a molecular study (Mori et al., 2007
), a general evolutionary scheme within the monophyletic subfamily Lecythidoideae is suggested (Appendix S1, see Supplemental Data accompanying online version of this article). The 10 lecythidoid genera are arranged in four major levels of advancement. In level one Grias and Gustavia form the basal most clade; level two includes only Couroupita, which is sister to the remaining seven genera; level three encompasses Allantoma, Cariniana, and Couratari, which form parallel clades; and level four, the most advanced, includes Bertholletia, Corythophora, Eschweilera, and Lecythis. The last two genera are not monophyletic as species of Eschweilera and Lecythis are scattered in two and four clades, respectively. According to the molecular phylogenetic tree (Mori et al., 2007
), floral symmetry of the Lecythidoideae evolved from polysymmetric to monosymmetric flowers once. However, the molecular phylogeny suggests a reversal back to polysymmetric flowers in the Allantoma/Cariniana decandra lineage, which had not been proposed until the recent study of Allantoma and Cariniana by Huang (2005)
based on morphological and anatomical features.
In this paper, we report on floral organogenesis for all 10 genera of Lecythidoideae and for Barringtonia of the Planchonioideae and focus on the evolution of monosymmetric flowers.
MATERIALS AND METHODS
Samples from 13 species of all 10 genera of the Lecythidoideae and Barringtonia racemosa of the Planchonioideae were collected in the field or from cultivated plants in botanical gardens. The species studied are Allantoma lineata (Mart. ex O. Berg) Miers (M. A. de Freitas 521, tree no. 135, Amazonas, Brazil, INPA), Barringtonia racemosa (L.) Spreng. (Chih-Hua Tsou 1911, Taiwan, HAST), Bertholletia excelsa Bonpl. (Bruce Nelson s.n., Amazonas, Brazil), Cariniana decandra Ducke (Alexandre Adalardo de Oliveira 294, tree no. 1785, Brazil, Amazonas, NY), Cariniana domestica (Mart.) Miers (Ghillean Tolmie Prance 8834, Brazil, Rhondônia, NY), Cariniana micrantha Ducke (Scott Alan Mori 20191, Brazil, Amazonas, NY), Corythophora amapaensis Pires ex S. A. Mori & Prance (Scott Alan Mori 18550, French Guiana, NY), Couratari sandwithii Prance (Lisa Campbell 419, Venezuela, Amazonas, NY), Couroupita guianensis Aublet (Scott Alan Mori 25755, Costa Rica, San José, cultivated, photograph on a sheet at NY), Eschweilera micrantha (O. Berg) Miers (Ghillean Tolmie Prance 15424, Brazil, Amazonas, NY), Eschweilera rankiniae S. A. Mori (M. A. de Freitas et al. 241, Brazil, Amazonas, NY), Grias peruviana Miers (Juan Revilla s.n., Peru), Gustavia macarenensis Philipson subsp. paucisperma S. A. Mori (Michael Nee & Scott Alan Mori 4159, Venezuela, Mérida, MO, NY VEN), Lecythis pisonis Cambess. (Chih-Hua Tsou 1549, Singapore Bot. Gard., cultivated, HAST). Flowers of representatives of 10 of these 11 genera are shown in Figs. 112.
|
In the floral formula given for each sample, Sattler's (1976)
floral formulae are adopted with a slight modification. The symbols "*," "
," and "·|·" represent polysymmetry, disymmetry, and monosymmetry, respectively. The letters S, P, A, and G designate sepals, petals, androecium, and gynoecium in that order. The symbol
after A means more than 10 stamens in the androecium. The abbreviations Aats and Aatp refer to stamens opposite the sepals and petals, respectively. The symbol
above G means that the gynoecium is inferior during the early stages of development that we observed. A parenthesis "( )" is used to enclose those organs with congenital fusion, and the symbol < > designates postgenital fusion. In the sequence of organ initiation, the symbol " , " is used to separate organs that initiate consecutively; and the symbol " - " is used if they appear synchronously or nearly so. For example, the formula "S1,2,3,4, P14" indicates that the four sepals initiate consecutively, then the four petals initiate synchronously. In the third line, symmetries at calyx initiation, corolla initiation, and maturity are given in that order because floral symmetry may change during development. Occasionally, the symmetry of a specific stage may vary and then the less common type is given in parentheses "( )".
RESULTS
Floral organogenesis in general
Flowers of the 11 genera studied develop from various types of indeterminate inflorescences. Flowers at anthesis are either polysymmetric (Figs. 13, 6, 7) or monosymmetric (Figs. 4, 5, 812). Individual flower primordia are always subtended by an abaxial bract and preceded by two lateral bracteoles. Most of the time, these three appendages were removed during dissection of the buds.
Sepals are initiated first. Four to six sepals are initiated individually and asynchronously in all the species except that in Grias peruviana sepals are completely congenitally fused, and a zonal calyx is initiated. A high degree of congenital fusion of sepals leaving short sepal apices is developed in Allantoma lineata, Barringtonia racemosa, Bertholletia excelsa, Cariniana decandra, and Gustavia macarenensis. Then, the four to eight petals are initiated individually and asynchronously from one or two whorls. At early stages of development, the floral ground plan is either monosymmetric or polysymmetric depending on the degree of abaxial dominance. In Barringtonia, Grias, and Gustavia, the influence of abaxial dominance is not evident, and growth on the abaxial and adaxial sides is about equal. In these genera, the floral apex is rounded, elliptic, or tetragonal in polar view at sepal initiation. Whereas in the remaining seven genera with abaxial dominance, the floral apex is obovate in polar view at sepal initiation, and the abaxial sepal and the two abaxial petals initiate first in their respective whorl. After the corolla initiation, the receptacle differentiates into well-defined androecial and gynoecial regions. In most genera, the receptacle is rather flat at the apex; but in members characterized by strong abaxial dominance (Bertholletia, Corythophora, Eschweilera, and Lecythis), the receptacle has a shallow or deep transverse depression in the middle at the apex, but with the abaxial side or both the abaxial and adaxial sides slightly raised, thus the receptacle apex becomes boat-like or chair-like. Henceforth, we call this feature the "transverse depression."
At stamen initiation, the androecium first appears as an androecial ring meristem. The outer circumference of the ring meristem is influenced by the number and arrangement of petals, which cause the ring meristem to become tetragonal, pentagonal, or hexagonal in outline. Polysymmetric members usually have a ring meristem of uniform width. In contrast, in monosymmetric species the abaxial side of the ring is broader. The androecial ring meristem in all species gives rise to many or numerous stamen primordia on the upper surface; in addition, the ring meristem further proliferates into the hood from the abaxial side in all monosymmetric-flowered species and produces a distinct lip from the adaxial side of the meristem in some monosymmetric genera (Bertholletia, Corythophora, Couratari, Eschweilera, and Lecythis).
Stamen initiation is centrifugal in all species except in Cariniana decandra in which there are only 10 stamens initiated from one or two irregular whorls and in Allantoma and Couratari in which the sequence was not observed. The androecial hood (Fig. 13) is the most prominent structure of the flower in monosymmetric species. It proliferates from the abaxial side of the ring meristem by the end of stamen initiation in Couroupita or somewhat earlier in most other monosymmetric genera. The young hood then extends in the available space between the receptacle and corolla. In Couratari, Bertholletia, Corythophora, Eschweilera, and Lecythis, more or less the entire abaxial half of the androecial margin is involved in the growth of the hood. During the early stages of development, the hood may produce stamen primordia in the monosymmetric-flowered species of Cariniana or staminode primordia in Bertholletia, Corythophora, Couratari, Couroupita, Eschweilera, and Lecythis. The subsequent staminodes developing from these primordia may be (1) without anthers, (2) have anthers with fodder pollen, or (3) produce nectar. With time, the hood becomes highly diversified, but we did not follow development beyond its early stages.
|
Floral organogenesis of individual species
In the following descriptions for individual species, the floral formula, sequence of organ initiation, and floral symmetry at three stages of development are documented first, then a description of early developmental stages is described. Later developmental features, for example, the various types of androecial hoods and the development of inferior vs. half-inferior ovaries are not described.
Most micrographs were taken from a polar view of an inflorescence or a floral apex, with their abaxialadaxial axis in an updown position. Some micrographs were taken from an apical/lateral view to show the relative positions of the abaxial, medial, and adaxial sides; or some micrographs show longitudinal profiles of a floral bud.
The order of the following descriptions starts with Barringtonia (Planchonioideae), which is followed by species of Lecythidoideae. Among the latter species, the order follows the four levels of the molecular phylogeny as described in the introduction.
PlanchonioideaeBarringtonia racemosa
)
(24)
in centrifugal sequence The floral apex is elliptical-oblong in polar view at sepal initiation, with the abaxial and adaxial sepals initiated first (F1, F2 in Fig. 14), then followed by the two lateral ones (F3 in Fig. 14). The four sepals are congenitally fused with only the apices remaining free (Fig. 15). Then the four petals are initiated synchronously (not shown) and alternate to the sepals (Fig. 16). At the beginning of stamen initiation, the androecial ring meristem has an obovate-rhombic circumference with a round abaxial side and an angular adaxial side (Figs. 17, 18). Stamens initiate starting from the inner side of the ring meristem in a centrifugal direction (Figs. 17, 18), and soon four to five irregular series of stamen primordia appear and cover the entire androecial ring meristem surface (Figs. 18, 19). Later on, the obovate-rhombic young androecium is spherical and polysymmetric. The stamen primordia of the middle rows grow faster and soon become the tallest ones; meanwhile, the body of the ring meristem is slightly thickened (Fig. 20). Filaments are congenitally fused basally, which becomes more evident at later stages. The young gynoecium is composed of three or four carpel primordia, and the style elongates rapidly (Figs. 1820). During floral development, the receptacle surface is perpendicular to the floral longitudinal axis (Figs. 19, 20).
|
)
(45)
in centrifugal direction The floral apex appears nearly square in polar view after the two bracteoles have initiated (Fig. 21). The sepal primordia are so thoroughly congenitally fused that the young calyx appears as a thick ring without evidence of sepal lobes at the apex (Figs. 2123). The four distinct petals are initiated from abaxial-lateral and adaxial-lateral positions. The initiation sequence was not followed. They are of similar sizes as they overlap and arranged polysymmetrically (Fig. 24). The androecial ring meristem is four-sided at stamen initiation, the two adaxial sides are wider, and stamens initiate first on the adaxial side (Fig. 25). About four irregular whorls of stamens are produced in a centrifugal sequence, but stamens of the outermost whorl soon become the longest and the innermost the shortest; soon the ring meristem is fully covered by stamen primordia (Figs. 2628). Fusion of filaments at the base takes place at later stages. There are four or five carpel primordia in the gynoecium (Figs. 25, 26). The style remains short in the early stages illustrated as well as at anthesis. During floral development, the receptacle surface is perpendicular to the floral longitudinal axis (Fig. 28).

(48)
in centrifugal direction
), *, *
|
The floral apex is circular to elliptical in polar view at sepal initiation. The abaxial sepal initiates first, followed by the adaxial one, and then the lateral pair (Figs. 29, 30). A calyx ring is formed shortly due to the congenital fusion of sepals. The eight petals initiate in two successive whorlsthe four in positions alternate to the sepals initiate first and synchronously (F3 in Fig. 29), and then another four initiate in between the first four more or less synchronously (Fig. 31). These eight petal primordia later merge as a single whorl as the receptacle expands. Occasionally, buds with seven petals in a single corolla whorl are found (Fig. 32). Nevertheless, in a small percentage of the inflorescences there are only six petal primordia in each floral bud; the floral apex is oblong at sepal initiation. During sepal initiation, the abaxial sepal initiates first (F1 in Fig. 36), then the adaxial sepal, one lateral sepal, and the other lateral sepal follow successively (F2 in Fig. 36). During petal initiation, the two abaxial petals initiate first (Fig. 37). The floral symmetry is monosymmetric at these stages (Fig. 37). In normal cases, when stamens begin to initiate, the androecial ring meristem usually has a multisided outer circumference and a circular inner circumference (Fig. 32). Numerous stamens are initiated in about seven or eight irregular whorls following a centrifugal sequence and finally cover the entire surface of the ring meristem (Fig. 33). When the stamen initiation is about complete, these numerous primordia reach similar sizes, and the ring meristem beneath the stamens thickens slightly (Fig. 34). Later on, stamens of the outer whorls exceed the inner ones in height (Fig. 35). Usually six carpel primordia are found in the gynoecium (Figs. 32, 33), and the style remains short in the early stages illustrated as well as at anthesis. During floral development, the receptacle plane is perpendicular to the longitudinal axis.

(67)
in centrifugal direction
), ·|· The floral apex is obovate in polar view at sepal initiation (F1, F2 in Fig. 38). The abaxial sepal initiates first, followed by the adaxial one, the abaxial-lateral pair, and lastly the adaxial-lateral pair (F1, F2, F3 in Fig. 38). The six petal primordia are alternate with the sepals, the abaxial pair initiate first, then the adaxial pair, and then the lateral pair (Fig. 39). Stamen initiation and carpel initiation begin about the same time; by then, the androecial ring meristem is hexagonal in outer circumference, and the two abaxial sides are longer than the other sides (Fig. 40). Stamen initiation proceeds centrifugally, and nine to 11 irregular whorls of stamen primordia are soon developed; these primordia display a weak gradient in decreasing height outward (Figs. 41, 42). Shortly after this, the abaxial side of the ring meristem starts to proliferate to form a roof-shaped ridge with an angle of ca. 120° (Fig. 42); stamens are produced on the proliferation in a centrifugal sequence (Figs. 42, 43). Due to the restricted space beneath the petals, the androecial proliferation curves and extends along the receptacle surface. A flat hood with numerous enlarged stamens and a bare ligule are thus formed (Fig. 4). The body of the ring meristem thickens to form a cushion-like structure below the ring stamens, which is called the staminal ring. No proliferation takes place on the adaxial side of the ring meristem. During floral development, the receptacle surface is perpendicular to the longitudinal axis.

(3)
in centrifugal sequence
(*), ·|·
|

(3)
in centrifugal sequence The inflorescence and flower primordia are small. Mature flowers are c. 0.8 cm in diameter. The young bract has a pubescent outer epidermis (Fig. 49). The abaxial sepal initiates first (F1 in Fig. 49), then the abaxial-lateral pair, then the adaxial one, and lastly the adaxial pair (not shown). The six petals are initiated more or less synchronously and alternate with the sepals (Fig. 50). During stamen initiation, the androecial ring meristem is hexagonal in circumference, with the abaxial region wider (Fig. 51). Stamens initiate from the inner side of the ring meristem in a centrifugal direction (Fig. 51). Approximately three irregular whorls and ca. 36 stamens are initiated and cover the entire surface of the androecial ring meristem (Fig. 52). A hood is derived from the abaxial rim of the ring meristem, but no proliferation takes place on the adaxial side (Fig. 52). Stamens on the ring meristem, the ligule, and the hood form a continuum without free space in between. During floral development, the receptacle plane is perpendicular to the longitudinal axis.
(3)
|
(4) The floral apex is pentagonal with the abaxial side slightly wider or nearly circular in polar view at sepal initiation (F1, F2 in Fig. 62). The five sepals initiate in a quincuncial sequence, with the abaxial one initiating first (F1, F2 in Fig. 62). Or the abaxial sepal and one adaxial sepal initiate first and second, followed irregularly by the remaining three. The five young sepals soon fuse together to form a synsepalous calyx. The sequence of petal initiation was not traced. The five petal primordia are in positions alternate with the sepals and of similar sizes during early development; they are polysymmetric at this stage (Fig. 63). Around 30 stamens are initiated and arranged in two whorls (Fig. 64), but the initiation pattern is not observed. Later on, a thick staminal tube develops at the base of the androecium (Fig. 65), and the upper part of each filament bends inward with the anther facing downward (Figs. 66, 67). No proliferation develops from the androecial periphery (Fig. 7). There are four carpels constituting the gynoecium, and the style is short (Fig. 66). During floral development, the receptacle plane is perpendicular to the longitudinal axis.

(3)
(in two whorls), sequence unknown
|

(4)
in centrifugal sequence
|

(34)
in centrifugal sequence
|

(2)
in centrifugal sequence
|

(4)
in centrifugal direction
|
-shaped ridge (Figs. 108, 109) and on the adaxial side to give rise to a short staminal lip (Fig. 111). The young hood soon encounters the stout style, and the style is slightly inclined (Fig. 110). The hood then develops into a more or less flat structure covering the remaining androecium (Figs. 112, 114, 115) and generates two types of staminodes on the inner surfaceone type without anthers distally and another with anthers containing fodder pollen proximally (Figs. 111, 112). On the other hand, the ring stamens are well developed on the adaxial side, whereas the staminal lip becomes indistinct (Fig. 114). The gynoecium is composed of four carpels (Fig. 107). The style is short but wide in the beginning (Figs. 110, 113), then becomes wider in the middle part (Fig. 111), which eventually develops into a circular expansion around the style that we call the stylar collar (Fig. 112). The abaxial dominance is so strongly expressed that not only the gynoecium, but also the whole floral bud is inclined toward the adaxial side (Figs. 112, 113). DISCUSSION
Floral organogenesis in all 10 genera of Lecythidoideae was followed using SEM. The only previous study using this methodology for the Lecythidoideae was on Couroupita guianensis (Endress, 1994
). The great floral diversity displayed at maturity by species of neotropical Lecythidaceae is, to a large extent, initiated in the early developmental stages of the flower. There are a few highly conserved features, which are uniformly expressed throughout the 10 genera, e.g., sepal primordia are free at initiation (except Grias) and initiated from one whorl, petal primordia are free and initiated from one (or two) whorl(s), a common androecial ring meristem is developed before initiation of individual stamens, and the gynoecium is composed of congenitally fused carpels.
On the other hand, many characters are variable within the subfamily. The majority of the characters observed during early floral development play a role in the display of floral symmetry found at anthesis in different species. Physiological factors exert an impact on the type of symmetry, such as the gain or loss of abaxial dominance, the strength of abaxial dominance, the initiation sequence of perianth parts, and the timing of hood initiation. Structural members also contribute to a specific symmetry, such as the shape of floral apex, the positions of perianth parts, the presence of an abaxial hood and an adaxial staminal lip, and the orientation of the style and floral axis. When the floral organogenesis of the 10 lecythidoid genera is compared with a phylogenetic scheme, it is evident that the direction of floral evolution is from polysymmetry to the establishment and continued strengthening of monosymmetry. Such is the major floral evolutionary trend within the Lecythidoideae. In the basal most Lecythidoideae, Grias and Gustavia, flowers are polysymmetric. The next stage in advancement is found in the monosymmetric flowers of Couroupita in which monosymmetry is established in the perianth and androecium, but polysymmetry may appear shortly before hood initiation. In Couratari, representing the next evolutionary level, the monosymmetry of the perianth and androecium is slightly strengthened, the hood initiates earlier, and an adaxial staminal lip develops. Finally, at the most advanced level, represented by species of Bertholletia, Corythophora, Eschweilera, and Lecythis, monosymmetry is established in the gynoecium as well as in the androecium, and both an abaxial hood and an adaxial staminal lip are present in these genera. An exception to this trend has occurred in the Allantoma/Cariniana decandra clade in which there is a reversal back to polysymmetry. The acquisition of monosymmetry in the Lecythidoideae is first expressed in the perianth, then the androecium, and finally in the gynoecium. Lecythidaceae with monosymmetric flowers are only found in the neotropics where the majority of the species are of this floral type.
The floral monosymmetry of the Lecythidoideae results mainly from two factors: the expression of abaxial dominance and the prolonged meristematic activity of the androecial ring meristem. The phenomenon of abaxial dominance can be influenced by a strong acropetal developmental gradient within the main shoot (inflorescence) (Endress, 1999
) or by the interactions of the TCP1/CYC-like gene and other related genes that are known to express at the adaxial domain of the floral meristem or the adaxial floral organs and lead to monosymmetry with a single dorsiventral axis (Coen, 1996
; Costa et al., 2005
; Hileman et al., 2003
). In the Lecythidoideae, the earliest sign indicating the presence of abaxial dominance is an obovate floral apex with the abaxial side broader than the adaxial side at calyx initiation and the abaxial sepal initiating first. The androecial ring meristem of the monosymmetric-flowered Lecythidoideae is remarkable because of the different meristematic activities expressed in different regions.
Androecial ring meristems have been reported in at least 29 families in the angiosperms (Tucker, 2003
). In Swartzieae (Leguminosae), the ring meristem functions in the proliferation in stamen number (up to 150 stamens per flower) (Mansano et al., 2002
; Tucker, 2003
). In Lecythidoideae, most species have hundreds of ring stamens (Prance and Mori, 1979
; Mori and Prance, 1990
), which is also associated with a well-developed ring meristem (Figs. 34, 42, 82, 90, 103). In addition to producing ordinary stamens on the surface, the ring meristem of Lecythidoideae also proliferates from its abaxial side to give rise to a hood as in all monosymmetric genera and from its adaxial side to give rise to a staminal lip in the most advanced monosymmetric genera. The hood of monosymmetric-flowered species of Lecythidoideae is unique in the angiosperms. In addition to abaxial dominance, cellular gigantism in the hood (Thompson, 1921
, 1927
) becomes evident during later developmental stages and is important in shaping the hood.
In the following, we interpret floral organogenesis of the Lecythidoideae from an evolutionary perspective by following the four stages of evolutionary advancement suggested in a recent molecular phylogenetic study (Mori et al., 2007
).
Level 1. Polysymmetric form, the basal form (Grias and Gustavia)
Abaxial dominance is not present, and the floral apex shows an elliptic, tetragonal, or circular shape at calyx initiation. The perianth is basically tetramerous in the species studied. Floral symmetry is polysymmetric throughout the developmental stages, with the exception that disymmetry rarely appears at calyx initiation in Gustavia. The receptacle surface is rather flat during stamen initiation, and the androecial ring meristem does not proliferate from the periphery. The receptacle axis is perpendicular to the floral axis, which is straight and upright. These features are also common to members of the Planchonioideae, the Old World outgroup of the Lecythidoideae.
Level 2. Establishment of the monosymmetric form (Couroupita)
Abaxial dominance is present but does not have any impact on the gynoecium at this level. In Couroupita, the perianth is hexamerous. As a result of abaxial dominance, the abaxial sepal and the abaxial petal pair initiate first during calyx and corolla initiation and the floral apex and the androecial ring meristem are broader on the abaxial side. Androecial hood initiation takes place when the initiation of fertile stamens is complete, but no adaxial staminal lip develops at the adaxial side. The receptacle is perpendicular to the floral axis, which is straight. Floral monosymmetry is present from the time of calyx initiation; nevertheless, there is a short period during stamen initiation that the floral structure is nearly disymmetric as seen in this study (Fig. 41) or polysymmetric as shown by Endress (1999)
. In either case, the symmetry changes back to monosymmetry when the hood starts to develop.
Among the existing genera of Lecythidoideae, Couroupita is the first genus in the molecular phylogeny to have monosymmetric flowers (Mori et al., 2007
). We suggest that monosymmetric flowers of Couroupita may have arisen from Gustavia-like ancestors. Some flowers of Gustavia macarenensis subsp. paucisperma have only six petals instead of eight as is normal for the species. In this Gustavia, abaxial dominance is weakly expressed and floral symmetry becomes monosymmetric during certain early stages of development (Figs. 36, 37). These features may well represent the first signs of abaxial dominance in ancestors that led to present-day species of Couroupita. The possible evolutionary route from ancestral species of Gustavia to Couroupita has been proposed based on morphological and embryological characters by Tsou (1994
, p. 92).
Level 3. Monosymmetric form (Cariniana monosymmetric group and Couratari) and reversed polysymmetric form (Cariniana polysymmetric group and Allantoma)
At level 3, three parallel clades are present. In the Couratari lineage, monosymmetric characteristics of Couroupita, such as abaxial dominance are expressed in the perianth and androecium, but not the gynoecium; the hexamerous perianth and an androecial hood are present. In addition, two advanced features associated with monosymmetry have evolved, i.e., the staminal lip developed from the adaxial periphery of the androecial ring meristem and the convoluted hood, which is a synapomorphy for species of Couratari. In the two other lineages, floral size has decreased to 6 to 15 mm in diameter, and the strength of abaxial dominance is weakened. In the Cariniana domestica group, even though the hexamerous perianth number is retained, the abaxial sepal initiates first, and a hood is developed; the monosymmetry is weakened and the floral apex at the stage of corolla initiation is nearly a disymmetric or polysymmetric hexagon and not monosymmetric. In the Allantoma/Cariniana decandra clade, the perianth is pentamerous, the monosymmetry appears only at calyx initiation after which flowers become polysymmetric until anthesis, and no hood develops. The decreased perianth merism and the reduction of the androecial ring meristem are probably influenced by the reduction in floral size. Because there may be a mechanism that couples CYC expression with organ position (Coen, 1996
), the change in the perianth from the hexamerous to the pentamerous configuration may alter the domain of expression for the CYC gene, thus weakening the abaxial dominance. As a result of the reduction, the number of stamens is decreased to only 10 to 30, and these are placed in only one or two whorls. Other specialized features of this group are the thickening of the filaments, especially of the outer stamens, and the high degree of congenital fusion of filaments during later development. A reversal from monosymmetric to polysymmetric flowers in this clade is suggested by both molecular (Mori et al., 2007
) and morphological (Huang, 2005
) data.
We view the floral polysymmetry of Grias and Gustavia as primary within the Lecythidoideae, and the floral organogenesis in these two genera is fairly similar to that of Barringtonia of the Planchonioideae. On the other hand, the floral polysymmetry of the Allantoma/Cariniana decandra group is secondary, and its floral organogenesis is different from that of Grias, Gustavia, or Barringtonia in that the former group has a pentamerous perianth, weak expression of abaxial dominance during perianth initiation, small floral size, and stamen number fewer than 30; whereas the latter has tetramerous perianth, no expression of abaxial dominance, medium to large flowers, and many to numerous stamens. It is plausible that reduction in floral size, a decrease in abaxial dominance, loss of the hood, and return to polysymmetry started with a Couroupita-like ancestor that evolved into species similar to the Cariniana domestica group and ended with species of today's Allantoma/Cariniana decandra group. Other support for this scenario comes from fruit and seed morphology. The primitive fruit type is indehiscent as found in members of level 1 (Grias, Gustavia) and level 2 (Couroupita) and the Planchonioideae, whereas members of level 3 (Cariniana and Allantoma included) and level 4 possess a cylindrical, woody dehiscent fruit type. In addition, the unilaterally winged seed type of the Allantoma/Cariniana decandra group is the same as that of the monosymmetric Cariniana-domestica group (Tsou and Mori, 2002
). An exception is found in the water-dispersed Allantoma lineata, which has only a vestigial seed wing.
Level 4. Advanced monosymmetric form (Clade 6, Bertholletia, Corythophora, Eschweilera, and Lecythis
Four genera and about 150 species are represented in level 4; this represents 73% of the known species of Lecythidoideae. Although the flowers of these four genera are diverse, especially in the morphology of the androecial hood, the diversified androecial hood of these four genera shows no essential difference in early floral development. And even though Bertholletia has a fused calyx that splits into two lobes at anthesis, there is an underlying similarity in the formation of the calyx. At calyx initiation, six sepals are initiated in Bertholletia as in the other three genera of this level. Thus, the presence of a distinctive two-lobed calyx is not that different from the six-lobed calyx common to all other monosymmetric-flowered Lecythidaceae.
Species in level 4 express the strongest monosymmetry throughout development and possess additional advanced traits not found at the other levels, e.g., the abaxial dominance is also expressed in the gynoecium; the hood initiates earlier (i.e., during stamen initiation instead of after stamen initiation); the receptacle has a transverse depression and a raised abaxial regiontherefore, the receptacle surface is oblique to the floral axis from early stages; and the young style bends toward the adaxial side in the same direction as the extension of the hood, thus the floral axis is changed from straight as in levels 1, 2, and 3 to oblique. In addition, the evolution of gynoecial monosymmetry takes place in two phases in level 4. In phase one, only the style becomes inclined to the adaxial side and assumes a monosymmetric form, but the ovary remains straight as in Bertholletia (Figs. 84, 85). In phase two, however, the entire gynoecium is inclined from early development as is found in Corythophora (Figs. 90, 91), Eschweilera (Fig. 99), and Lecythis (Figs. 110112). The evolution of monosymmetry is most likely caused by an increase in abaxial dominance.
Significance of the evolution of floral symmetry in the Lecythidoideae
We have demonstrated that there is a trend from polysymmetric to monosymmetric flowers in the New World Lecythidaceae and that the proximate cause of the evolution of monosymmetry from polysymmetry is the evolution of abaxial dominance and the prolonged meristematic activity of the androecial ring meristem. Cellular gigantism occurs at later stages of development (Thompson, 1921
, 1927
).
The polysymmetric flowers of Gustavia are open to any animal seeking pollen as a reward, but they are most efficiently pollinated by bees that buzz the anthers (Mori and Boeke, 1987
). A synapomorphy uniting all species of Gustavia is the presence of poricidal anthers, a feature associated with buzz pollination. The polysymmetric flowers of Grias have been suggested to be pollinated by beetles because their aromas are consistent with beetle pollination (Knudsen and Mori, 1996
), and a few beetles have been seen in the flowers. The pollination of the species of the polysymmetric-flowered Allantoma/Cariniana decandra group is completely unknown, but we suggest they will also prove to be beetle pollinated because their carnose flowers are morphologically similar to those of Grias. In all of the polysymmetric-flowered genera, the symmetry of the flower does not direct the pollinator into the flower from a predictable direction.
In contrast, the monosymmetric-flowered species of all other genera are directed into the flower from a predictable direction. Because the hood of the flower is open on the adaxial side, the bees and bats that have been recorded pollinating these species land on the hood and crawl into the flower from the adaxial side. Moreover, there is often a spot of yellow that serves as a guide to the adaxial side of the flower where the bees enter. In some species, the hood is at least partially open (e.g., Lecythis idatimon and L. pisonis), but in most monosymmetric-flowered species the hood is pressed against the summit of the ovary (e.g., Bertholletia excelsa, all species of Corythophora, Couratari, Eschweilera, and some species of Lecythis). The hood, thus, allows entry into the flower to the bees and bats strong enough to open it. Upon entering the flower, the pollinators are positioned in such a way that their head and back press against the staminal ring while they are collecting either pollen or nectar from the hood. In this case, the hood pollen is fodder pollen, which can usually be distinguished from fertile pollen by its yellow versus white color. The most specialized flowers are those of Eschweilera and Couratari in which there is a nectar chamber at the apex of the inwardly coiled hood. The most efficient pollinators are euglossine bees because they are both robust enough to enter the flower and have a long enough proboscis to reach the nectar (Nelson et al., 1985
; Mori and Boeke, 1987
; Prance, 1976
). A special adaptation of Couratari is the external flap that covers the outside of the hood and forms a false chamber. We have noted that trigonid bees often penetrate the hood of other genera to gather nectar by burrowing through the hood, but this is seldom seen in species of Couratari because drilling through the external flap leads to a false chamber with no nectar.
Evolutionary adaptation to pollinators, we hypothesize, is the ultimate cause of monosymmetry in the Lecythidoideae. It is well known that floral monosymmetry allows for precise placement of pollen and, thus, enhances the success of pollination (Endress, 2001
); the Lecythidoideae are another example of this phenomenon. It is because of the expression of abaxial dominance and the evolution of an androecial hood that monosymmetric species of Lecythidaceae have radiated from their polysymmetric ancestors. The resulting monosymmetric flowers of Lecythidoideae have contributed to the ecological success of species with this flower type, especially in Amazonian lowland forests where Lecythidaceae are among the most dominant plant families, reaching up to as many as 24 species and 149 individuals in a single hectare (Mori et al., 2001
).
The restriction of monosymmetric-flowered Lecythidaceae to the neotropics is likely the result of selection for monosymmetry by bees, especially euglossines, which also only occur in the New World. The evolution of monosymmetry is, in turn, caused by abaxial dominance more strongly expressed in monosymmetric than in polysymmetric flowers. Thus, the amazing range of flower types in neotropical Lecythidaceae is a response to selection for more efficiently pollinated flowers caused by the evolution of a wide variety of androecial hoods, a feature found in no other plant family in the world.
FOOTNOTES
1 The authors thank L. Campbell, M. A. de Freitas, R. Kew, S. Lee, M. Nee, B. Nelson, A. A. de Oliveira, G. T. Prance, and J. Revilla for helping obtain collections. They also thank T.-H. Chen, C. A. Gracie, and A. Henderson for providing photographs used in Figs. 112, B. Angell for drawing Fig. 13, and Y.-L. Fu for helping with darkroom work. Dr. S. Tucker's and one anonymous reviewer's careful reading and very detailed comments on the manuscript are highly appreciated. The authors are grateful to the National Science Council, Republic of China, for grants NSC 92-2621-B-001003 and NSC 93-2621-B-001002, which supported work on this project at the Academia Sinica. ![]()
4 Author for correspondence (chtsou{at}gate.sinica.edu.tw
) ![]()
LITERATURE CITED
Berg O.. 1858. Myrtaceae, tribus II. Barringtoniaceae and trib. III. Lecythideae. In C. F. P. von Martius [ed.], Flora Brasiliensis 14: 469-526 pls 59-74.
Black G. A. Dobzhansky T. Pavan C.. 1950. Some attempts to estimate species diversity and population density of trees in Amazonian forests. Botanical Gazette 111: 413-425.
Cain S. A. De Castro G. M. O. Pires J. M. Da Silva N. T.. 1956. Application of some phytosociological techniques to Brazilian rain forest. American Journal of Botany 43: 911-941.[CrossRef][Web of Science]
Coen E. S.. 1996. Floral symmetry. EMBO Journal 15: 6777-6788.[Web of Science][Medline]
Costa M. M. R. Fox S. Hanna A. I. Baxter C. Coen E.. 2005. Evolution of regulatory interactions controlling floral asymmetry. Development 132: 5093-5101.
Endress P. K.. 1994. Diversity and evolution of tropical flowers Cambridge University Press, Cambridge, UK.
Endress P. K.. 1999. Symmetry in flowers: diversity and evolution. International Journal of Plant Sciences 160: S3-S23.[CrossRef][Web of Science][Medline]
Endress P. K.. 2001. Evolution of floral symmetry. Current Opinion in Plant Biology 4: 86-91.[CrossRef][Web of Science][Medline]
Gamboa-Gaitán M. Á.. 1997. Biología reproductiva de Eschweilera bogotensis (Lecythidaceae), en la Cordillera Occidental de Colombia. Caldasia 19: 479-485.
Gentry A. H.. 1982. Patterns of neotropical plant species diversity. Evolutionary Biology 15: 1-84.[Web of Science]
Giurfa M. Eichmann B. Menzel R.. 1996. Symmetry perception in an insect. Nature 382: 458-461.[Medline]
Hileman L. C. Kramer E. M. Baum D. A.. 2003. Differential regulation of symmetry genes and the evolution of floral morphologies. Proceedings of the National Academy of Sciences 100: 12814-12819.
Hirmer M.. 1918. Beiträge zur Morphologie der polyandrischen Bluten. Flora 110: 140-192.
Huang Y. Y.. 2005. The phylogeny of Cariniana (Lecythidaceae) based on morphological and anatomical data Master's thesis, H. Herbert Lehman College, The City University of New York and The New York Botanical Garden, Bronx, New York, USA.
Jacques F.. 1965. Morphologie du pollen et des ovules de Couroupita guianensis Aubl. (Lecythidacees). Pollen et Spores 7: 175-180.[Medline]
Knudsen J. T. Mori S. A.. 1996. Floral scents and pollination in neotropical Lecythidaceae. Biotropica 28: 42-60.[CrossRef][Web of Science]
Knuth R.. 1939. Lecythidaceae. In Engler [ed.], Pflanzenreich IV, vol. 219a 83-146.
Kowal R.. 1989. Chromosome numbers of Asteranthos and the putatively related Lecythidaceae. Brittonia 42: 131-135.
Kowal R. Mori S. A. Kallunki J. A.. 1977. Chromosome numbers of Panamanian Lecythidaceae and their use in subfamilial classification. Brittonia 29: 399-410.[CrossRef][Web of Science]
Leins P. Von.. 1972. Das zentrifugale Androeceum von Couroupita guianensis (Lecythidaceae). Beiträge zur Biologie der Pflanzen 48: 313-319.
Mansano V. De F. Tucker S. C. Tozzi A. M. G. De A.. 2002. Floral ontogeny of Lecointea, Zollernia, Exostyles, and Harleyodendron (Leguminosae: Papilionoideae: Swartzieae s.l). American Journal of Botany 89: 1553-1569.
Miers J.. 1874. On the Lecythidaceae. Transactions of the Linnean Society, London 30: 157-318.
Mori S. A. Becker P. Kincaid D.. 2001. Lecythidaceae of a central Amazonian forest. Implications for conservation. In R. O. Bierregaard, J. C. Gascon, T. E. Lovejoy, and R. Mesquita [eds.], Lessons from Amazonia. Yale University Press, New Haven, Connecticut, USA.
Mori S. A. Boeke J. D.. 1987. Chapter XII. Pollination. Memoirs of the New York Botanical Garden 44: 137-155.
Mori S. A. Boom B. M.. 1987. Chapter II. The forest. Memoirs of the New York Botanical Garden 44: 9-29.
Mori S. A. Lepsch-Cunha N.. 1995. The Lecythidaceae of a central Amazonian moist forest. Memoirs of the New York Botanical Garden 75: 1-55.
Mori S. A. Mattos Silva L. A. Dos Santos T. S.. 1980a. Observações sobre a fenologia e biologia floral de Lecythis pisonis Cambess. (Lecythidaceae). Anais da sociedade de botânica do Brasil 30: 109-116.
Mori S. A. Orchard J. E. Prance G. T.. 1980b. Intrafloral pollen differentiation in the New World Lecythidaceae, subfamily Lecythidoideae. Science 209: 400-403.
Mori S. A. Prance G. T.. 1990. Lecythidaceae. Part II. The zygomorphic-flowered New World genera (Couroupita, Corythophora, Bertholletia, Couratari, Eschweilera, Lecythis). Flora Neotropica Monograph 21: 1-376.
Mori S. A. Tsou C. H. Anderberg A. A. Wu C. C. Cronholm B.. 2007. Evolution of Lecythidaceae with an emphasis on the circumscription of neotropical genera. Information from combined ndhF and trnL-F sequence data. American Journal of Botany 94: 289-301.
Murphy J. A.. 1978. Non-coating techniques to render biological specimens conductive. Scanning Electron Microscopy 2: 175-186.
Nelson B. W. Absy M. L. Barbosa E. M. Prance G. T.. 1985. Observations on flower visitors to Bertholletia excelsa H. B. K. and Couratari tenuicarpa A. C. Sm. (Lecythidaceae). Acta Amazonica 15: 225-234.
Prance G. T.. 1976. The pollination and androphore structure of some Amazonian Lecythidaceae. Biotropica 8: 235-241.[CrossRef][Web of Science]
Prance G. T.. 1990. Couratari. Flora Neotropica 21: 119-158.
Prance G. T. Mori S. A.. 1979. The actinomorphic-flowered New World Lecythidaceae (Asteranthos, Gustavia, Grias, Allantoma & Cariniana). Flora Neotropica 21: 1-270.
Prance G. T. Mori S. A.. 2004. Lecythidaceae. In K. Kubitzki [ed.], The families and genera of vascular plants. VI. Flowering plants. Dicotyledons. Celastrales, Oxalidales, Rosales, Cornales, Ericales. Springer-Verlag, New York, New York, USA.
Sattler R.. 1976. Organogenesis of flowers. A photographic text-atlas University of Toronto Press, Buffalo, New York, USA.
Thompson J.. 1921. Studies in floral morphology. II. The staminal zygomorphy of Couroupita guianensis Aubl. Transactions of the Royal Society of Edinburgh 53: 1-15.
Thompson J.. 1927. A study in advancing gigantism, with staminal sterility, with special reference to the Lecythidaceae. Publications of the Hartley Botanical Laboratories 4: 5-44.
Tsou C. H.. 1994. The embryology, reproductive morphology, and systematics of Lecythidaceae. Memoirs of the New York Botanical Garden 71: 1-110.
Tsou C. H. Mori S. A.. 2002. Seed coat anatomy and its relationship to seed dispersal in subfamily Lecythidoideae of the Lecythidaceae (the Brazil nut family). Botanical Bulletin of Academia Sinica 43: 37-56.[Web of Science]
Tucker S. C.. 2003. Floral ontogeny in Swartzia (Leguminosae: Papilionoideae: Swartzieae): distribution and role of the ring meristem. American Journal of Botany 90: 1271-1292.
![]()
CiteULike
Complore
Connotea
Del.icio.us
Digg
Facebook
Reddit
Technorati
Twitter What's this?
This article has been cited by other articles:
![]() |
F. Jabbour, C. Damerval, and S. Nadot Evolutionary Trends in the Flowers of Asteridae: Is Polyandry an Alternative to Zygomorphy? Ann. Bot., August 1, 2008; 102(2): 153 - 165. [Abstract] [Full Text] [PDF] |
||||
![]() |
Compiled by, F. Tooke, T. Chiurugwi, and N. Battey Flowering Newsletter bibliography for 2007 J. Exp. Bot., July 18, 2008; (2008) ern109v1. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |