Abstract
Background and Aims
Acer, one of the largest genera of Sapindaceae, is well known for its diverse inflorescence and flower morphologies. Structural diversity of maples makes this genus a wonderful model to explore morphogenetic interaction between inflorescence and flower construction.
Methods
We investigated the development and structure of inflorescences and flowers using scanning electron microscopy in ten species of Acer from nine (of 17) sections. Observations are interpreted in the framework of molecular phylogenies.
Key Results
Inflorescences are represented by panicles, thyrsoids, botryoids and sciadioids. The evolutionary transition from thyrsoids or panicles to racemose inflorescences resulted in the loss of floral prophylls, another sequence of floral organ initiation and often the acquisition of a new flower groundplan. The main structural transformations are changes in the floral merism and the number of stamens caused by changes in the shape and size of the floral meristem. The order of sepal initiation, the position of the median sepal and flower orientation are largely related to mechanical pressures within the floral bud associated with variations in the inflorescence construction. The gynoecium position is also variable. In species with thyrsoids or panicles, the carpel position depends on the first sepal position or on the flower position within the inflorescence. In species with racemose inflorescences, the gynoecium is always inserted in the transverse plane.
Conclusions
The evolutionary transition from branched to racemose inflorescences was followed by changes in the geometry of the floral meristem and ultimately resulted in considerable changes in flower construction and development. The presence of floral prophylls and their developmental dynamics are labile among species of Acer. Inflorescence construction represents not only an important taxonomic marker but also a key feature that probably defines the diversity of floral construction and development in Acer and Sapindaceae as a whole.
Keywords: Inflorescence, flower, flower development, flower groundplan, morphogenesis, ontogenetic pattern, organ initiation sequence
INTRODUCTION
The order Sapindales is currently circumscribed to include nine families that comprise ∼2 % of the world angiosperm diversity (Joyce et al., 2023). Some representatives, including maples, are of great ecological and economic importance (Abrams, 1998; Perkins et al., 2009; Acevedo-Rodríguez et al., 2010). Despite being small, the flowers of Sapindaceae, the second largest family of the order after Rutaceae, show remarkable (but relatively poorly known) morphological diversity (partly reviewed by El Ottra et al., 2022). Currently, only a limited part of this diversity has been studied developmentally (Sokoloff et al., 2024).
Although family relationships within Sapindales are disputable, the generic and species relationships have been set, which makes this family a wonderful model for understanding the evolution of morphological characters of flowers and inflorescences. Among 144 genera of Sapindaceae, the genus Acer, with 155 accepted species, represents a well-defined and unusual group of predominantly temperate trees and shrubs (De Jong, 2004; Acevedo-Rodríguez et al., 2010; Areces-Berazain et al., 2021; Buerki et al., 2021). In the past, the morphological peculiarities of Acer allowed segregation of a separate family, Aceraceae (Radlkofer, 1933; De Jong, 1976; Muller and Leenhouts, 1976). Despite differences in gynoecium construction and flower symmetry, the flowers of maple species have many features in common with typical Sapindaceae, including a characteristic androecium of eight stamens (Acevedo-Rodríguez et al., 2010). Species of Acer demonstrate the same diversity in inflorescence and pollination biology that is found within Sapindaceae and even Sapindales as a whole. Thus, the genus-focused investigation of patterns of flower construction and development in Acer can serve as a case study to understand the principles of evolution of reproductive organs in this group of eudicots and to reveal ontogenetic mechanisms underlying the morphological transformations of the flowers.
Most maples share the same floral groundplan: five free sepals, five free petals, eight stamens and two carpels in a syncarpous gynoecium (De Jong, 1976; Acevedo-Rodríguez et al., 2010). This pattern is considered a plesiomorphic condition in the genus Acer and the family Sapindaceae. Various transformations of the typical flower structure occurred independently in different sections (De Jong, 1976). These transformations include the changes in floral merism and the number of some floral organs. Monoecy is the most common and ancestral condition for maples. Dioecy is observed in five sections, and the transition to this reproductive system occurred at least three times (Renner et al., 2007; Areces-Berazain et al., 2021). In some cases, the change in reproductive system is accompanied by a transition to wind pollination and, as a consequence, to a tendency to petal reduction (De Jong, 1976; Prenner, 2004).
It is possible tentatively to recognize four derived floral groundplans in Acer; three are associated with reductive tendencies and one with an increased number of organs (De Jong, 1976). The first type, with four (or five) sepals and petals and four to six stamens, is in section Indivisa. The second type, with isomerous perianth and androecium, i.e. with five (or six) sepals, petals and stamens, is in the sections Rubra and Pubescentia. The third type, tetramerous, is characteristic of sections Arguta and Negundo. The last type, with five (or six) sepals and petals and ten (or 12) stamens, occurs in sections Trifoliata and Macrophylla. This description of floral diversity in maples is very superficial. All these variants were described using only the organ numbers (De Jong, 1976) but not the organ positions within the flower; in other words, the floral formulas are known but not the floral diagrams. There are only few exceptions (Eichler, 1878; Ronse De Craene, 2022; Zavialov and Remizowa, 2024).
The homologies of derived floral patterns become even more obscure owing to an insufficient understanding of flower construction in typical flowers with eight stamens. Zavialov and Remizowa (2025) reviewed existing theories and proposed new insights on the origin and homology of the eight-staminate androecium of Sapindaceae based on developmental evidence. However, other questions regarding floral organ distribution remain. For example, there is no consensus in the literature about the position of the carpels. Eichler (1878) noted that the gynoecium is located differently in different species of the genus Acer. In some species, the carpels are located obliquely in the plane of the second sepal, whereas they are transverse in other species. Buchenau (1861) also emphasized this feature, although he did not trace any clear relationship with sepal positions.
The situation with inflorescences is similar. De Jong (1976) classified maple inflorescences into five main types, but he did not provide a structural interpretation of these types. The first two types are represented by compound branched inflorescences with cincinni (the first type) or without them (the second type). The third type is a raceme, the fourth type is an umbel in species of the section Rubra, and the fifth type is a compound raceme in staminate inflorescences of Acer negundo. De Jong (1976) put forward some suggestions about directions of inflorescence evolution. He believed that the first type is ancestral and that the second is the most widespread and is derived from the first one. The remaining three types were considered to be derived from the second type.
In addition to the specific questions on the evolution of individual traits, we can also ask a more global question: how do the flowers and inflorescences interact to attain the existing diversity of structural and developmental patterns? What characters can be combined and, vice versa, what combinations are prohibited? What are the morphological drivers of evolutionary transformations in flowers and inflorescences? Is it possible to identify morphological patterns and correlations in their development and, consequently, evolution? The genus Acer can become an interesting model for exploring morphological transformations of the flower groundplan and morphogenetic changes accompanying such transitions.
MATERIALS AND METHODS
This study provides data on the structure and development of inflorescences and flowers in ten species: Acer tataricum subsp. ginnala (Maxim.) Wesm. (section Ginnala), Acer tegmentosum (Maxim.) Maxim. (section Macrantha), Acer platanoides L. (section Platanoidea), Acer spicatum Lam. (section Spicata), Acer nikoense (Miq.) Maxim. (section Trifoliata), Acer barbinerve Maxim. (section Arguta), Acer saccharinum L. (section Rubra), Acer rubrum L. (section Rubra), Acer carpinifolium Siebold & Zucc. (section Indivisa) and Acer pseudosieboldianum (Pax) Komarov (section Palmata).
The plant material (generative buds and inflorescences) was collected for 3 years in several parks and botanical gardens: the Botanical Garden of Moscow State University, the Tsitsin Main Botanical Garden of the Russian Academy of Sciences, and the Saint Petersburg Botanical Garden. More detailed information, including vouchers, is given in Supplementary Data Table S1.
All materials were fixed in 70 % ethanol. For scanning electron microscopy, generative buds at different developmental stages were dissected under an Olympus SZX7 stereomicroscope (Olympus, Tokyo, Japan). Flowers were dehydrated through absolute acetone, critical-point dried using a Hitachi HCP2 critical-point drier (Hitachi, Tokyo, Japan), then coated with gold and palladium using an Eiko IB-3 ion-coater (Eiko, Tokyo, Japan). Observations were made using a CAMSCAN S2 scanning electron microscope (Camscan, Cambridge, UK), JSM-6380LA (JEOL, Tokyo, Japan) and Quattro S (Thermo Fischer Scientific, Waltham, MA, USA) at Moscow University.
To analyse the diversity of flower structure in A. platanoides, we examined 50 flowers and produced flower diagrams of each designated structural type, which were then combined into a generalized diagram using a graphical editor. Each diagram was placed in a separate layer, the transparency of which was proportional to the frequency of its occurrence. All illustrations were prepared in Krita v.5.1.5.
In this paper, we followed the terminology introduced by Endress (2010a). We describe a raceme with a terminal flower as a botryoid, an umbel with a terminal flower as a sciadioid, and a thyrse with a terminal flower as a thyrsoid. A panicle is an inflorescence that does not correspond strictly to the racemose or cymose branching patterns. Each branch terminates in a flower. Moving up the main branch, branching becomes successively less rich, encompassing fewer branching orders (Endress, 2010a). We described the subtending phyllome of the flower as a flower-subtending bract, also known as a pherophyll, and the two first phyllomes on a peduncle as floral prophylls (or bracteoles). For a description of stamen development and arrangement, we used the following characters: appearance from individual primordium or appearance from common stamen–petal primordium.
RESULTS
Inflorescence architecture
Species with branched inflorescences (panicles and thyrsoids)
In A. tataricum subsp. ginnala (section Ginnala), the inflorescence is a panicle or, rarely, a thyrsoid (Fig. 1A, K). In case of thyrsoid, there are phyllomes on the main axis and only two prophylls on the axes of the second and higher orders. In case of a panicle, numerous phyllomes are located on the main axis, and several phyllomes are inserted on the proximal axes of the second order (Figs 2 and 3). There are only prophylls on the distal axes of the second order and axes of higher orders. Branching occurs in the axil of one prophyll of a pair or the axils of both prophylls, forming monochasia and dichasia. In the upper part of the inflorescence, single flowers extend from the main axis, and phyllotaxis changes from decussate to spiral from the base to the top of the inflorescence (Figs 1A, K and 2A, B).
Fig. 1.
Diversity of inflorescences in maples. (A) Panicle of Acer tataricum subsp. ginnala. (B) Thyrsoid of Acer spicatum. (C) Thyrsoid of Acer platanoides. (D) Variants of thyrsoid of Acer pseudosieboldianum. (E) Three-flowered botryoid. (F) Five-flowered botryoid. (G) Seven-flowered botryoid. (H) Multiflowered botryoid. (I) Three-flowered sciadioid. (J) Five-flowered sciadioid. (K) Inflorescence tip of A. tataricum subsp. ginnala. (L) Middle stage of thyrsoid development in A. spicatum, spiral arrangement of future cymes, only first-order flowers are visible in each cyme. Purple and green symbols indicate neighbouring right parastichies; circles and squares indicate neighbouring left parastichies. (M) panicle of A. platanoides. (N) Pistillate three-flowered botryoid of Acer barbinerve. (O) Acer pseudosieboldianum; different cymes are painted in different colours, and numbers indicate the branching order of each flower. (P) Staminate seven-flowered botryoid of A. barbinerve. (Q) Pistillate five-flowered botryoid of Acer carpinifolium with sterile bracts under terminal flower (asterisks). (R) Multiflowered botryoid of Acer tegmentosum. (S) Three-flowered botryoid of Acer nikoense with leaves (lf). (T) Three-flowered sciadioid of Acer rubrum. (U) Five-flowered sciadioid of Acer saccharinum. Abbreviation: tfl, terminal flower. Scale bars: 100 µm in L, O, P–R; 300 µm in M, N, S–U; 500 µm in K.
Fig. 2.
Early stages of inflorescence and flower development in Acer tataricum subsp. ginnala and Acer pseudosieboldianum. (A–E) Acer tataricum subsp. ginnala. (A) Initiation of second-order axes. (B) Initiation of third-order axes. (C) Appearance of flower primordia. (D) Initiation of floral prophylls. (E) Initiation of sepals. (F–H) Acer pseudosieboldianum. (F, G) Early stages of cyme development, sepal initiation in fl1. (H) Petal and stamen initiation. Abbreviations: B, subtending bract of second-order axes; b, subtending bract of third-order axes; br, flower-subtending bract; fl, lateral flower (numbers indicate branching and initiation order); I, stamens from individual primordia; IA, inflorescence apex; p, petals; s, sepals (numbers indicate initiation sequence); α, first floral prophyll; β, second floral prophyll; *, stamen–petal common primordia. Scale bars: 50 µm in F–H; 100 µm in A–E.
Fig. 3.
Early stages of inflorescence and flower development in Acer platanoides and Acer spicatum. (A–G) Acer platanoides. (A) Flower initiation. (B) Initiation of first floral prophyll. (C) Initiation of second floral prophyll. (D) Appearance of next order flower in axil of the first prophyll. (E) Flower with a single floral prophyll and appearance of a flower of the next order in its axil. (F) Transition from dichasial branching pattern to monochasial. More developed flowers are in dichasium (right lateral flower removed); primordia of the next-order flowers are in monochasium. (G) Late stage of inflorescence development; all flowers are tightly packed on surface of spherical inflorescence. (H–N) Acer spicatum. (H) Early stage of inflorescence development, initiation of lateral cymes; only first-order flowers are initiated in each cyme. (I) Inflorescence apex, top view. (J) Primordium of first flower in a cyme. (K) Expansion of flower primordium. (L) Initiation of floral prophylls. (M) Appearance of next-order flower in floral prophyll axil; floral prophyll not visible. (N) Initiation of sepals, expansion of flower-subtending bract. Abbreviations: br, flower- subtending bract; fl, lateral flower (numbers indicate branching and initiation order); IA, inflorescence apex; lf, leaf; s, sepals (numbers indicate initiation sequence); α, first floral prophyll; β, second floral prophyll. Scale bars: 30 µm in E, J–L, N; 50 µm in A–C, M; 100 µm in D, H, I; 300 µm in F, G.
Acer platanoides (section Platanoidea) is similar to the previous species in branching pattern, but the shape is corymbose (Fig. 1C, M). Also, the difference is that in the case of a thyrsoid, the lateral inflorescence elements are dichasia that branch monochasially in their distal parts.
The inflorescence in A. spicatum (section Spicata) is a thyrsoid (Fig. 1B, L). However, unlike the species described above, numerous phyllomes (50 or even more) and second-order axes in their axils are spirally arranged on the main axis (Fig. 1L). There are only two prophylls on the second-order axes, and further branching occurs in the axil of only one prophyll. Thus, the thyrsoid is composed of monochasia.
In A. pseudosieboldianum (section Palmata), the inflorescence is a thyrsoid without visible phyllomes (Fig. 1D, O). From the main axis, two or four monochasia of four to five flowers extend in pairs. However, cymes are often reduced to a single flower (Fig. 1D, right).
Species with racemose inflorescences
The inflorescence structure in A. carpinifolium (section Indivisa), A. barbinerve (section Arguta) and A. nikoense (section Trifoliata) is uniform. The inflorescence is a botryoid (Fig. 1E–H). There are only flower-subtending bracts and no floral prophylls (Figs 4 and 5). Phyllotaxis is decussate. Sometimes, sterile empty phyllomes are developed under the terminal flower in A. carpinifolium (Fig. 1Q). Differences between these species are in the number of flowers: A. carpinifolium, five or seven flowers (Fig. 1F, G, Q); A. barbinerve, three to nine flowers (Fig. 1E–G, N, P); and A. nikoense, three or, rarely, five flowers (Fig. 1E, F, S).
Fig. 4.
Early stages of inflorescence and flower development in Acer barbinerve and Acer carpinifolium. (A–J) Acer barbinerve. (A) Initiation of bracts and flowers in their axils. (B) Transition from apical inflorescence meristem to terminal flower. (C) Initiation of transversal sepals. (D) Initiation of median sepals and adaxial petals (arrows). (E) Initiation of petals and stamens in terminal flower. (F) Flower with late(?) median abaxial sepal, stamen and petal primordia and gynoecium primordium. (G) Appearance of median adaxial sepal (arrow). (H) Unequal growth rates of median and transversal stamens; gynoecium primordium appears in the centre (staminate flower). (I) Stamens become equal (staminate flower). (J) Mature pistillate flower. (K–M) Acer carpinifolium. (K) Initiation of transversal sepals (arrows). (L) Initiation of petals and stamens. (M) Division of common primordia, formation of gynoecium primordium. Note difference in stamen sizes: the largest one is transversal; abaxial stamens are delayed in development. Abbreviations: br, flower-subtending bract; fl, lateral flower; I, stamens from individual primordia; II, stamens from common primordia; lf, leaf; ms, median sepal; p, petal; s, sepal; ts, transversal sepal; +, median stamens/staminodes; −, transversal stamens/staminodes; *, stamen–petal common primordia. Scale bars: 30 µm in D, E, K, L; 100 µm in A–C, F–I, M; 200 µm in J.
Fig. 5.
Early stages of flower and inflorescence development in Acer saccharinum and Acer nikoense. (A–G) Acer saccharinum. (A) Initiation of flowers; numerous trichomes cover the inflorescence axis. (B) Primordium of a lateral flower; arrow indicates possible cryptic bract. (C) Initiation of transversal sepals. (D, E) Initiation of median adaxial sepal (arrow); beginning of calyx tube formation. (F, G) Initiation of stamens. (H, I) Acer nikoense. (H) Flower initiation. (I) Initiation of floral organs. Abbreviations: br, flower-subtending bract; fl, lateral flower; I, stamens from individual primordia; ms, median sepal; s, sepal; tbs, transverse abaxial sepal; tfl, terminal flower. Scale bars: 30 µm in A–H; 100 µm in I.
In A. tegmentosum (section Macrantha), the inflorescence is a multi-flowered botryoid (Fig. 1H, R) with spirally arranged flowers.
Members of section Rubra (A. saccharinum and A. rubrum) possess sciadioids (Fig. 1I, J). The short inflorescence axis is covered with numerous trichomes (Fig. 1T, U). There are no flower-subtending bracts and no flower prophylls. The number of flowers is three or five; phyllotaxis is decussate.
Inflorescence development
Species with panicles and thyrsoids
In A. tataricum subsp. ginnala, we studied the earliest stages of inflorescence development. Initially, on the inflorescence meristem, bracts arise acropetally, and the primordia of second-order axes are formed in their axils (Fig. 2A). Phyllotaxis is initially decussate but almost immediately becomes spiral (Fig. 2A). Subsequently, the inflorescence meristem elongates, continuing to segregate new phyllomes. At the same time, the basal second-order axes start to produce decussate phyllomes (Fig. 2B). After the initiation of all phyllomes, the apical meristem of any axis becomes a terminal flower (Fig. 1K). The formation of second-order axes and their branching take place inside the bud. However, the main part of inflorescence development occurs outside the bud in the spring: the appearance of the third- and fourth-order axes and flower development.
Most of the inflorescence development of other species occurs in the bud. Acer spicatum differs from A. tataricum subsp. ginnala by a higher number of bracts and primordia of the second-order axes, located immediately in a spiral (Figs 1L and 3H, I). The inflorescence apex retains its activity for a long time (Fig. 3I), and only shortly before the bud opening it produces a terminal flower. In A. platanoides and A. pseudosieboldianum, the terminal flower of the main axis is formed early (Fig. 1O). As a result, they produce fewer second-order axes, but higher-order axes are branched more extensively. In A. platanoides, most new axes are formed on the side facing outwards (towards the bud scales). Flowers on axes of different orders are located side by side on the surface of the spherical inflorescence (Figs 1M and 3F, G).
In A. tataricum subsp. ginnala, A. spicatum and A. platanoides, the flowers are initiated in the axil of their subtending bracts as small, transversely elongated or oval-shaped primordia (Figs 2C and 3A, J, K). Next, two prophylls are initiated successively (A. platanoides; Fig. 3B, C) or almost simultaneously (A. tataricum subsp. ginnala and A. spicatum; Figs 2D and 3L, M). After the separation of prophylls, the flower meristem becomes round, and the calyx commences to appear (Figs 2E and 3D, N).
Next, species differ in the dynamics of the flower-subtending bract and growth of floral prophylls. In A. tataricum subsp. ginnala, by the time of the corolla and androecium initiation, the flower-subtending bract and prophylls are significantly expanded and cover the flower (Fig. 6B). The sepals grow fast after their initiation. During gynoecium development, they cover all other organs. In A. platanoides, both the bract and the prophylls do not increase in size after their appearance, hence they are soon hidden behind the developing flowers (Fig. 7A, C, G, H). In A. spicatum, the prophylls do not expand, and the flower-subtending bracts grow and cover the flowers during the middle stages of development (Fig. 7K, L).
Fig. 6.
Floral diagrams and young flowers in some species. (A–C) Acer tataricum subsp. ginnala. (A) Three-flowered dichasium. (B) Lateral flower on middle developmental stage; gynoecium is in the plane of the first sepal (s1). (C) Sepal initiation in terminal flower follows phyllotaxis of bracts. (D–F) Acer pseudosieboldianum. (D) Lateral branch of inflorescence, side and top views. (E, F) Examples of cymes; asterisk indicates flower with rare position of gynoecium. (G) Acer barbinerve, staminate flower. (H) Acer barbinerve, pistillate flower. (I, J) Acer carpinifolium. (K) Acer Acer tegmentosum. (L–N) Acer saccharinum. (O) Acer nikoense, terminal flower. (P) Acer nikoense, lateral flower. (Q–T) Acer rubrum. Abbreviations: s, sepals (numbers initiation indicate order); α, first floral prophyll; β, second floral prophyll. Light green, bracts and floral prophylls; dark green, sepals; pink, petals; orange, stamens from common primordia; yellow, stamens from individual primordia; violet, carpels. Scale bars: 100 µm.
Fig. 7.
Variability of floral organ positions in Acer platanoides and Acer spicatum. (A–H) Acer platanoides. (A, B) Flower with two floral prohylls bearing two well-developed flowers of the next order. (C, D) Flower with two floral prohylls and one well-developed flower of the next order. (E) Frequency of gynoecium positions in mature flowers with median adaxial sepal. (F) Frequency of gynoecium positions in mature flowers with median abaxial sepal. (G, H) Flowers without well-developed flowers of the next order; median sepal (s3) is abaxial in G and adaxial in H; gynoecium is located in the plane of the first sepal in both variants. (I–L) Acer spicatum. (I) Sepal initiation; median sepal (s2) is adaxial. (J) Flower with median abaxial sepal. (K, L) Flowers with differently inserted gynoecia. Abbreviations: br, flower-subtending bract; fl, lateral flower; s, sepals (numbers indicate initiation order). The colours and symbols are identical to Fig. 1. Light blue indicates flowers in the axils of floral prophylls. Scale bars: 100 µm.
In the ebracteate inflorescences of A. pseudosieboldianum, the appearance of a flower on each new-order axis is associated with the division of the meristem into two parts. One part (larger and located closer to the flower of the previous order) becomes a new flower, and the second, smaller part remains meristematic and continues to divide (Fig. 2F, G). The newly formed flower primordium is squeezed between the two older flowers of the previous orders and the meristem and becomes rectangular under their pressure (Fig. 2F).
Species with racemose inflorescences
Early development of racemose inflorescences is similar in most species, as exemplified by botryoids of A. carpinifolium and A. barbinerve. Initially, small bracts are initiated in an acropetal order on the cone-shaped inflorescence meristem (Fig. 4A, B). They are arranged decussately (Fig. 1N, P, Q). Floral primordia, transversally elongated, appear in their axils (Fig. 4B, K). After the lateral flowers are initiated, the apical meristem of the inflorescence forms a terminal flower. The terminal flower is usually larger and develops ahead of the lateral flowers closest to it (Fig. 4B).
In A. nikoense, the inflorescence meristem is spherical, and its larger central part becomes the terminal flower (Fig. 5H). The rest gives rise to two flower-subtending bracts. In their axils, small primordia of lateral flowers appear.
In A. saccharinum, the inflorescence meristem is flat-spherical (Fig. 5A). From the earliest stages of development, numerous trichomes cover the developing inflorescence (Fig. 5A). Flower-subtending bracts are absent. Flowers appear as narrow, transversely elongated primordia. Then, the floral primordia become oval (Fig. 5B). However, on the abaxial side of the floral primordium, an almost indistinguishable groove can be seen, separating the main mass of the meristem from its base (Fig. 5B, arrowhead). This groove is a possible boundary between the rudimentary flower-subtending bract and the flower.
The remaining species differ in the degree of expression of the bracts. In A. barbinerve, small flower-subtending bracts that support the flower only from below (Fig. 4B, C) are hidden behind the growing flowers. In A. carpinifolium, flower-subtending bracts look like small, weakly expressed protrusions of the inflorescence axis (Figs 1Q and 4K). They do not expand, although they remain visible owing to the elongation of the peduncles during flower development. In A. nikoense, flowers are completely covered by the flower-subtending bract (Fig. 5H). However, before sepal initiation, the flower primordium expands, protruding from the bract axil and leaving the bract below the flower (Fig. 5I).
Flower structure and development
Species with branched inflorescences (panicles and thyrsoids)
Acer tataricum subsp. ginnala, A. pseudosieboldianum, A. platanoides and A. spicatum have typical flowers with five sepals, five petals and eight stamens; three of them arise from common primordia with petals, and five arise from individual primordia (Figs 6A, D and 7A, C, K, L). The organ positions and their stability differ between these species.
In A. tataricum subsp. ginnala and A. pseudosieboldianum, the positions of the floral organs are strictly determined (Fig. 6A, D). Acer tataricum subsp. ginnala have transverse-adaxial prophylls. The first prophyll is larger, and the next-order axis appears predominantly in its axil (Fig. 6A, B). There are no prophylls in A. pseudosieboldianum (section Palmata); branching occurs transversally only on one side of the flower to produce a monochasium. The side (left or right) changes with each subsequent branching order, forming a ‘zig-zag’ cincinnus (Fig. 6D–F).
In both A. tataricum subsp. ginnala and A. pseudosieboldianum, the first sepal is located in a transverse-abaxial position opposite the next-order flower. The second sepal is located in a median adaxial position (Fig. 6A, B, D). The larger floral prophyll containing the next-order flower in its axil can be situated either on the left or on the right side of the flower; consequently, the first sepal is either right or left. Both enantiomorphs occur in the same inflorescence. In a terminal flower, the first sepal occupies a position opposite the uppermost phyllome on the axis (Fig. 6C).
The initiation of floral organs in A. pseudosieboldianum has not been described previously; therefore, we will describe it in more detail. The sepals are initiated in a spiral: the first sepal is transversal-abaxial, the second median is adaxial, etc. The fourth sepal is delayed in development owing to pressure from the older flower and is usually not visible (Fig. 2G, H). Long trichomes develop on the inflorescence axis, filling the space between the flowers (Fig. 2F, G). All petals and stamens are initiated almost simultaneously. Five stamens are initiated from individual primordia, and two stamens on the sides of the fifth sepal are initiated from common petal–stamen primordia (Fig. 2H). We observed this stage of development in flowers of the third and fourth orders, which often have only seven stamens; the stamen between the first and fourth sepals is lost. Flowers of the second order have eight stamens, three of which are most likely to arise from common primordia with petals, as in most other eight-staminate maples. Flowers of higher orders often have six stamens (Fig. 6E, F).
In A. platanoides and A. spicatum, neighbouring flowers (located nearby but not necessarily on the same axis) are in close contact with each other, forming a ‘puzzle’ (Figs 3G and 7I–L). The prophylls occupy transverse-abaxial positions (Figs 3L and 6A–D). In A. spicatum, the flower of the next order is formed only in the axil of one prophyll (Fig. 3M). The first sepal is formed in a transverse-abaxial position opposite this prophyll (Fig. 3N), and the second in a median adaxial position, as in A. tataricum subsp. ginnala and A. pseudosieboldianum, but the position of the median sepal can occasionally be abaxial in older flowers (Fig. 7I, J). In A. platanoides, if the next-order flower develops in the axil of only one prophyll, the first sepal is initiated in a transverse-abaxial position opposite this prophyll, and consequently, the third sepal is median abaxial (Fig. 7C, D). If flowers are initiated in the axils of both prophylls, the first sepal is transverse-adaxial, opposite the more developed flower primordium of the next order, and the second sepal is median abaxial (Fig. 7A, B). If the next-order flowers are not formed in the prophyll axils (at least at sepal appearance), the spiral of sepal initiation can begin in different positions (Fig. 7G, H). In one case, the first sepal was transverse-abaxial and the third sepal median abaxial (Fig. 7G). In another case, the first sepal was transverse-adaxial and the third sepal median adaxial (Fig. 7H).
The gynoecium is usually in the plane of the first (or less often the second) sepal in A. tataricum subsp. ginnala, A. platanoides and A. spicatum (Figs 6A, B and 7A–D, G, H). However, the carpels can be shifted slightly closer to the transversal or the median plane. In A. tataricum subsp. ginnala, calyx position is stable, hence the position of carpels is also stable. All variants of gynoecium position can be observed in A. platanoides, with unstable sepal positions: from strictly median (Fig. 7A) to strictly transversal (Fig. 7G) with all possible intermediate states. When the median sepal is adaxial (usually the third in the developmental spiral), the carpels are more often located obliquely or in the median plane (Fig. 7E). When the median sepal is abaxial (usually the second or the third), the carpels are often located in the transversal plane (Fig. 7F).
In A. pseudosieboldianum, the gynoecia of all flowers of a cincinnus are oriented in the same direction (Fig. 6D–F). Morphologically, the gynoecium of the first flower in the cyme is transversal, the gynoecium of the second flower is median and that of the third flower is again transversal, etc. In the terminal flower, the gynoecium parallels the gynoecium of the closest second-order flowers.
Species with racemose inflorescences
In all species, transversal sepals appear first (Figs 4C, K and 5C, I). In A. barbinerve and A. carpinifolium, flowers are tetramerous, with four sepals and four petals; two sepals are transversal and two are median (Fig. 6G–J). The median abaxial sepal is initiated second, and the median adaxial is the last sepal (Fig. 4D, F, G, L). Transversal sepals are larger than median ones; this difference remains during early and middle stages of flower development. In pentamerous A. nikoense, after initiation of transverse-abaxial sepals, median abaxial and one transverse-adaxial sepals arise. The second transverse-adaxial sepal is delayed (Fig. 5I). In apetalous A. saccharinum, the small adaxial sepal appears after the transversal sepals (Fig. 5D). At the same time, the formation of the calyx tube begins owing to zonal growth under the free sepal tips. The remaining two or three sepal primordia emerge already on the calyx tube. The boundaries between the sepals become poorly visible (Fig. 5E).
The subsequent stages of flower development vary significantly; therefore, each species is considered separately.
In A. barbinerve, before the sepals are fully formed, the petals and stamens appear (Fig. 4D–G). Adaxial petals initiate slightly earlier than abaxial ones (Fig. 4D). In the centre of the receptacle after stamen initiation, the meristematic area of the future gynoecium is oval and elongated in the transversal plane and becomes separated by the surrounding stamens (Fig. 4E–G). The median stamen primordia are longer and narrower, and the transversal ones are smaller and more rounded (Fig. 4G). Then, the pistillate and staminate flowers differ in development.
In pistillate flowers, the gynoecium begins to develop. Staminodes hardly increase in size. The ovary expands, occupying all the space available in the flower, hence the staminodes become fused to the gynoecium base (Fig. 4J). The median staminodes look like protrusions at the base of the ovary opposite the border between the carpels, and the transversal ones become almost indistinguishable. In several flowers, we found an atypical calyx. Such flowers possessed two lower and two upper sepals (Fig. 1N, left flower).
In staminate flowers, the gynoecium does not develop, forming only a solid pistillodium with two apices (Fig. 4H). Stamen primordia enlarge with different developmental rates: the median stamens are larger than the transversal ones (Fig. 4H). However, this difference gradually decreases (Fig. 4I).
In tetramerous A. carpinifolium, there are six or seven stamens (90 % of examined flowers had six stamens and 10 % had seven). Petals and stamens appear simultaneously. Two petals are initiated from common petal–stamen primordia, and the remaining petals and stamens are initiated from single primordia (Fig. 4L, M). Stamens arising from single primordia are considerably larger. The largest ones are located opposite the carpel backs (Fig. 4M). The difference in stamen size disappears at later stages. If there are six stamens, usually two of them are formed from common primordia and are located approximately opposite the petals, and the remaining four stamens are approximately opposite the sepals (Figs 4L and 6I, J). If there are seven stamens, two stamens are formed from common primordia, five from single primordia, and two of them are located opposite the same median abaxial or adaxial sepal (Fig. 4M). In two cases, we observed a terminal flower of A. carpinifolium with eight stamens.
In apetalous A. saccharinum and A. rubrum (members of section Rubra), flowers are pentamerous or hexamerous (Fig. 6L–N, Q–T). In pentamerous flowers, the median sepal is adaxial or, rarely, abaxial (Fig. 6L, Q, R). In hexamerous flowers (Fig. 6M, N, S, T), there are two median sepals in adaxial and abaxial positions or sepals in the median plane are absent. In A. saccharinum, the stamen primordia arise from single primordia. The number of stamens is the same as the number of sepals. They are usually located opposite the sepals (Fig. 5F), or some stamens may alternate with them (Fig. 5G). Petals are not initiated in this species. There are no data on the development of the flower of A. rubrum, but they appear to be similar except for the presence of petals and the absence of a calyx tube.
Acer tegmentosum (section Macrantha) has a typical flower groundplan, with five sepals, five petals and eight stamens, as in species with branched inflorescences. But unlike these species, four stamens arise from common primordia and four arise from individual primordia (Fig. 6K). The median sepal is always in an adaxial position. Flower development of A. tegmentosum was mostly described in a separate article (Zavialov and Remizowa, 2025).
Acer nikoense (section Trifoliata) differs by an increased number of organs: the lateral flowers are pentamerous, less often hexamerous (Fig. 6P). Terminal flowers are hexamerous (Fig. 6O), less frequently heptamerous. In pentamerous flowers, there are ten stamens in two whorls; five arise from common primordia, and the other five arise from individual primordia (Fig. 6P). Flower development of A. nikoense was mostly described in a separate article (Zavialov and Remizowa, 2025).
The carpels are transversal in all species (Fig. 6G–T). In the terminal flower, the gynoecium is parallel to the gynoecia of the nearest lateral flowers or the gynoecium is opposite the first sepal (only in A. tegmentosum).
In all species, flower and inflorescence development happens inside the bud. The sepals and petals grow weakly, almost do not cover the reproductive parts of the flower, and increase in size only shortly before flowering.
DISCUSSION
Main developmental drivers of flower evolution
Ronse De Craene (2024) reviewed two main factors that provide structural evolution of flowers: heterochronic shifts and mechanical forces. These factors usually co-occur in the same flower and represent flower development as a process in time and in some spatial conditions or, in another sense, as a reflection of programmed pathways of morphogenesis exposed to environmental conditions (Ronse De Craene, 2018). Current research on flower development and evolution is now aimed, among other things, at identifying these drivers and specific mechanisms of influence (e.g. Leins and Erbar, 1997; Remizowa et al., 2013; Ronse De Craene, 2016; Bull–Hereñu et al., 2022). However, there is a lack of clear understanding of the genetic basis of the enormous structural flower diversity. We know a lot about the genetic regulation of floral organ identity, flower symmetry and flower initiation; some mathematical models of flower development were proposed (reviewed by Kramer, 2018; also see Barrio et al., 2010; Abad et al., 2017). Nevertheless, we still have a poor understanding of how the number of organs is regulated genetically, and therefore, we tend to explain such changes by spatial–mechanical factors (Ronse De Craene, 2016; Bull–Hereñu et al., 2022). How is the size of the floral meristem related to the floral development and structure? What is the genetic basis for different patterns of organ arrangement? What are the molecular mechanisms underlying different types of sequential heterochrony?
Apparently, the use of a few model species [mostly Arabidopsis thaliana (L.) Heynh. and Antirrhinum majus L.] is insufficient to solve these issues. Studies on closely related and simultaneously morphologically diverse species can provide more information about fine aspects of genetic regulation. The genus Acer is one of the potential model taxa. Below, we discuss how maples can help to shed light on the role of genetic regulation in the merism, organ arrangement, positional stability, meristem size, etc.
Floral groundplans in the genus Acer
We examined flower structure and development in five sections with typical flowers (Platanoidea, Ginnala, Spicata, Macrantha and Palmata) and four sections (Arguta, Indivisa, Trifoliata and Rubra) that represent the main derivative types. Typical flowers are surprisingly diverse in developmental patterns and in the positions of organs. Meanwhile, the number of mechanisms that change the number of floral organs is relatively low. The first mechanism is the change in flower merism through the loss or addition of a flower sector (Ronse De Craene and Smets, 1994; Endress, 1995; Ronse De Craene, 2016). This situation can be seen best in A. carpinifolium. In this species, a flower sector with a sepal, a petal and two stamens, one of which is formed from a common petal–stamen primordium and the other from an individual primordium, has disappeared (Fig. 6I, J). In terminal flowers of A. carpinifolium, there may be a secondary regaining of an additional seventh antesepalous stamen or a complete return to eight stamens. In this case, the additional stamens are probably formed by dédoublement (Ronse De Craene and Smets, 1993), i.e. there are two stamen primordia in certain positions instead of one (two adaxial stamens in Fig. 4M). In species of the section Rubra and in A. negundo, merism can vary, increasing to hexamery or decreasing to trimery (the last only in A. negundo; Zavialov and Remizowa, 2024). The reason for such variation is probably the unstable size of the flower meristem. The mechanism of this change is probably similar to A. carpinifolium, but with another set of organs within a sector: one sepal, one antesepalous stamen and one petal (Fig. 6L–N, Q–T).
In developmental patterns with common primordia, another possible pathway of decrease in the number of organs is the loss of organs arising from common primordia (Ronse De Craene et al., 1993; Ronse De Craene, 2024). In eudicots, these organs are petals and stamens. Based on the flower development, we can suggest that loss of stamens arising from common primordia occurred independently in the sections Rubra, Arguta and Negundo and possibly in the yet unstudied section Pubescentia (De Jong, 1976). In sections Arguta and Negundo (we studied A. negundo and A. barbinerve), the loss of stamens arising from common primordia is superimposed on the loss of a flower sector (Fig. 8).
Fig. 8.
Phylogenetic tree according to Areces-Berazain et al. (2021), with mapped characters of inflorescence and flower developmental morphology. (A) Inflorescence type. (B) Changes in structure of corolla and androecium. (C) Gynoecium position. (D) Number of sepals and position of median sepal. (E) Sexual systems. Characters are based on the present study, Zavialov and Remizowa (2024), Zavialov and Remizowa (2025) and De Jong (1976). Question marks indicate data that are not obtained by direct observations, but by extrapolation owing to the similarity of the flower formula according to De Jong (1976). Petal loss is not shown because it occurred at the species level in three different sections. Data on sexual systems are based on Renner et al. (2007). Monoecy refers to all variants of the sexual system where there are bisexual individuals.
Common petal–stamen primordia is an important feature of morphogenesis that can distinguish even closely related species (Ronse De Craene et al., 2000; Cao et al., 2018; Ronse De Craene, 2024). The current predominating view is that common primordia appear owing to the late initiation of petals and fusion of their primordia with stamen primordia (Ronse De Craene et al., 1993; Ronse De Craene, 2024). The ontogenetic formation of common primordia can be considered a signal of petal reduction in a phylogenetic context (Ronse De Craene et al., 1993; Ronse De Craene, 2024). In our opinion, the only satisfactory explanation for the formation of common primordia is a general decrease in the floral meristem size (for more details, see Zavialov and Remizowa, 2025). Common primordia provide an economy of meristematic tissue. A decrease in the floral meristem size is a factor of reduction. Common primordia, in this case, act as a compensatory mechanism, changing morphogenesis but retaining the definitive structure as in A. tegmentosum (Fig. 6K). However, reduction of flower organs still occurs when the floral meristem decreases beyond a critical level, as in A. saccharinum, A. barbinerve and A. negundo (Zavialov and Remizowa, 2025). The critical level is a hypothetical value of the floral meristem size below which organ reduction occurs. Above this value, common primordia can stabilize the flower structure under a decrease in the meristem size. This value is a theoretical abstraction, because not only the size of the meristem but also its shape influences organ loss (Zavialov and Remizowa, 2025).
The petal loss in maples is most likely to be associated with the transition to wind pollination. Generally, it is difficult to accept that the reduction of petals in the genus Acer is related to the presence of common primordia. The loss of petals is characteristic of only three species (A. negundo, A. saccharinum and A. saccharum). In two of them, it was preceded in evolution by the reduction of stamens arising from common primordia (De Jong, 1976; Zavialov and Remizowa, 2024). The antepetalous stamens are the first to be lost in the evolutionary lineages leading to apetaly, and thus there are no common primordia. Therefore, the loss of petals cannot be connected with common primordia. The most straightforward hypothesis is that the petal initiation is shifted to progressively later ontogenetic stages until their complete loss. However, this idea is highly speculative, because species without common primordia but with petals in these evolutionary lineages (A. rubrum in the section Rubra and A. barbinerve in the Arguta–Negundo clade), which are transitional forms in this hypothetical series, do not show a pronounced delay in petal initiation. Also, there is no clear evidence of petal reduction in adult flowers of these species.
Interestingly, the loss of petals is accompanied by the appearance of a distinct calyx tube. Perhaps these morphogenetic transformations are related. The gene PETAL LOSS (PTL) in Arabidopsis thaliana is responsible not only for the normal development of petals but also for sepal boundaries (Lampugnani et al., 2012; Bull–Hereñu et al., 2022). In mutants, petals are usually trumpet shaped, thread like or stamenoid and occupy more space on the meristem, whereas sepals are fused to form a noticeable calyx tube. The same mechanism can be proposed for maples. However, the calyx tube is initiated differently in different maples. In A. negundo, it arises owing to zonal growth under free sepal teeth (Zavialov and Remizowa, 2024), whereas in A. saccharinum the calyx tube is first to be initiated, followed by the formation of free sepal tips.
When stamens are lost in maples, in most cases these are the stamens that arise from common primordia with petals. This pattern cannot be reconciled with any of the opinions described above. We believe that common primordia do not contribute to the reduction of either the petals or the stamens, which arise from the common primordia. The loss of stamens that arise from common primordia in A. barbinerve and A. rubrum is not necessarily a result of stamen delay. The stamens here could be lost owing to saltational change in the flower geometry and size (at least there are no transitional stages with those stamens initiated but with ceased growth). In our view, the complete elimination of either the stamen or petal part from the common primordium is to some extent problematic; these organs are intimately linked via common primordia and should appear in early development. The reduction of these stamens is possibly related to the fact that these stamens do not form a complete whorl; two stamens were already lost in the ancestor of Sapindaceae (Acevedo-Rodríguez et al., 2010), and the remaining stamens of incomplete whorl are located in geometrically disadvantageous positions, competing for space with the petal primordia. The petals are important for the overall appearance of the flower and thus for insect pollination, hence they are more stable under reduction caused by a decrease in meristem size.
Stamen reduction in maples is thought to be associated with gynoecium pressure (Eichler, 1878; Ronse De Craene, 2022). This is not the case for the two initially absent stamens, as discussed in detail in a separate study (Zavialov and Remizowa, 2025). Only one stamen that arises from a common primordium is located opposite the carpel and experiences its pressure. Such stamens do not show visible signs of suppression. The other two stamens initiated from common primordia do not experience any special pressure, hence their reduction is unlikely to be caused by the gynoecium. Also, the stamens can be pushed apart during development under the influence of the gynoecium, avoiding the excess pressure.
Unlike other species, A. nikoense exhibits an increase in the number of floral organs: the androecium consists of ten stamens (Fig. 6P). The species with obdiplostemony show another stable state, where the perianth and androecium whorls are isomerous and the flower buds are more spacious, such that the stamens compete less with each other. This flower construction can be regarded either as a reversal to the ancestral diplostemony or as a secondary regain of two stamens. The androecium of eight stamens is most likely to be a synapomorphy of Sapindaceae (Ronse De Craene et al., 2000; Zhang et al., 2022). This androecium evolved from a two-whorled androecium of ten stamens. Therefore, the appearance of a ten-staminate androecium can be considered a reversion to the ancestral state. This is not the only possible explanation, because the appearance of two additional stamens might be a secondary phenomenon. It is unlikely that additional stamens arise by duplication, but it is difficult to confirm or deny the possibility of their appearance de novo. In some cases, besides the addition of missing stamens there is a secondary doubling of some stamens, more often in antipetalous positions (see fig. 8E, F in the paper by Zavialov and Remizowa, 2025) and sometimes an increase in the flower merism. An increase in the floral meristem could explain such a change (Ronse De Craene and Smets, 1994; Ronse De Craene, 2016; Moyroud and Glover, 2017), but direct measurements should confirm this.
Inflorescence diversity and transition from panicle/thyrsoid to botryoid
Unlike flowers, which sometimes differ in structure in closely related species, inflorescences of a certain architecture are usually characteristic of genera, subfamilies, families or even orders (Prusinkiewicz et al., 2007; Endress, 2010a). The genus Acer is one of the few exceptions. The transition from panicle/thyrsoid to racemose inflorescences occurred five times independently among maples (Fig. 8). The transitions between the panicle and the thyrsoid are easy, because the difference between these inflorescences is in the number of phyllomes on the lateral axes and, as a consequence, the flowers in their axils (Endress, 2010a; Remizowa et al., 2013).
The transition to raceme is more complex and is associated with heterochrony. In racemose inflorescences, branching in the axils of the floral prophylls (bracteoles) can be arrested without losing the bracteoles, or the floral prophylls can be reduced together with the ability of branching in their axils. In the latter case, the reverse transition to the thyrse becomes difficult (Remizowa et al., 2013). Axes of the second order change their identity from indeterminate inflorescence meristem to determinate floral meristem (Prusinkiewicz et al., 2007; Coen and Prusinkiewicz, 2024). It can be described via the duration of expression of the genes TFL1 and LFY (Prusinkiewicz et al., 2007). Although existing models are based on these two genes, subtle changes in regulatory molecular networks are likely to mediate transformations of inflorescence types (Lemmon et al., 2016). Genetic regulation of branching has been studied for all inflorescence types but only for distantly related taxa (Souer et al., 1998; Conti and Bradley, 2007; Lippman et al., 2008; Cheng et al., 2018). Maples might become a unique model for studying such changes because different species fill almost the entire morphospace of possible inflorescence structural types (Prusinkiewicz et al., 2007). In both thyrsoids and botryoids, there are variants with different numbers of flowers and length of the main axis (Fig. 1).
The genus Acer is also unique because, in most cases, changes in inflorescence structure are accompanied by changes in flower structure (Fig. 9). The loss of prophylls is accompanied by a flattening of the floral meristem in maples, which could hypothetically lead to an increase in floral merism owing to the increase in the perimeter of the ellipse compared with the circle. However, in reality, an increase in merism is observed only in A. nikoense. The changes mostly result in reductions in the number of flower organs in other studied species (De Jong, 1976). A possible cause of organ loss is a decrease in flower meristem (Ronse De Craene, 2016; Moyroud and Glover, 2017), as confirmed by measurements at the stage of petal and stamen initiation (Zavialov and Remizowa, 2025). But why does a decrease occur? The first explanation is a reduction of inflorescence meristem, which was confirmed by the data on the size and foliage of generative shoots (Kostina, 2007). Such reduction can be a part of the mechanism providing a transition to a racemose inflorescence. Another explanation is the presence and duration of prophyll initiation (Fig. 9), which can be an important stage, because prophyll initiation might influence the floral meristem size. The change in the duration of this stage and its presence is an example of heterochrony. More detailed measurements at different developmental stages are required to test these hypotheses. Acer tegmentosum is the only example where a change in flower development accompanies a reduction of the floral meristem but not a significant change in flower structure. This is probably attributable to the significantly smaller size of sepal primordia, followed by changes in the initiation of stamens and petals.
Fig. 9.
Main patterns of flower development in maples. (A) Generalized pattern of flower development in typical flowers of Acer platanoides, A. pseudosieboldianum, A. spicatum and A. tataricum subsp. ginnala. Note that variability in structure and development is not shown here. (B–E) Developmental patterns in: (B) Acer carpinifolium, (C) A. barbinerve, (D) A. nikoence and (E) A. saccharinum. Developmental stages: I, flower primordium; II, initiation of prophylls; III, sepal initiation; and IV, initiation of corolla and androecium.
Position of median sepal
Sepals are initiated in a spiral sequence in most eudicots (Payer, 1857). Therefore, the position of the median sepal is determined by the position of the first sepal in the initiation spiral (Endress, 2010b; Bukhari et al., 2017). The order of sepal initiation is dependent on various factors. Among them are the pressure and spatial arrangement of surrounding structures (Remizowa et al., 2013; Bull-Hereñu et al., 2022; Ronse De Craene, 2024). In a typical case, these are the flower-subtending bract and the floral prophylls (bracteoles). In monocots, the position and initiation sequence of tepals are directly determined by the position of a single prophyll (Kirchoff, 1983; Remizowa et al., 2013). In eudicots, there are two opposite variants (with median adaxial or median abaxial sepal), but they do not depend on the presence of prophylls (Bukhari et al., 2017).
The role of prophylls in flower development is well documented in Fabaceae by Tucker (1996, 2002a, b, 2003). Tucker distinguished two types of flower development. The ‘omega-type’ pattern is characterized by large floral prophylls and a ‘vertical’ narrow floral meristem hidden between them. Here, the flower possesses only a single sepal and a single petal, and floral development is unidirectional. In the ‘circular type’, the rounded floral meristem is surrounded by small bracteoles, and five sepals arise in a spiral sequence (Tucker, 2002a, b). It can be seen clearly that the first initiated sepal is usually located in the median abaxial position at equal distances with two prophylls, which are located in transverse-adaxial positions. But in Astragalus compactus Lam., the position of the flower within the inflorescence has a greater influence on the initiation spiral than the prophylls do. However, it does not change the position of the median sepal (Naghiloo et al., 2012). Suppression or loss of prophylls in Faboideae also does not affect sepal positions but can influence the order of sepal initiation (Prenner, 2004; Ronse De Craene, 2024). In general, the position of perianth elements in zygomorphic flowers of Faboidae is highly stable and not determined by prophylls, but prophylls, if present, can influence the order of organ initiation and the number of perianth parts.
The position of the median sepal can vary in pentamerous actinomorphic flowers of Rosaceae, but not much attention has been paid to this fact (Evans and Dickinson, 1999a, b, 2005). However, there is no connection between the position of the median sepal and the presence/degree of development of prophylls (Evans and Dickinson, 1999b).
A sequential initiation of sepals in a 2/5 spiral was reported for all studied species of Sapindaceae (Ronse De Craene et al., 2000; Cao et al., 2006, 2017, 2018; Cao and Xia, 2009; Zhang et al., 2022). In most studies, the position of sepals relative to the flower-subtending bract is not precisely indicated. However, it seems that the median sepal is usually adaxial. In Eurycorymbus cavaleriei (H.Lév.) Rehder & Hand.-Mazz., the first sepal is initiated in the median abaxial position and then becomes pushed slightly transversely. The floral diagram of this species shows the median sepal in the adaxial position (Cao et al., 2017).
Maple flowers are always actinomorphic. Therefore, floral symmetry has no great influence on organ arrangement. Can the impact of external pressures explain the position of the perianth elements in maples? Probably only in part.
In A. tataricum subsp. ginnala, the floral prophylls are inserted closer to the adaxial side, both prophylls are well developed, and their axils contain the next-order flowers (Fig. 6A). As a consequence, the first sepal appears on the abaxial side (Fig. 6B), but is not equidistant from the prophylls (as in Fabaceae), being inserted slightly towards the smaller prophyll.
In A. pseudosieboldianum, the flower primordium is compressed on three sides, and therefore, the place of appearance of the first sepal is also clearly determined (Fig. 2F, G). But why does the first sepal occupy a transverse-abaxial position and not strictly a median abaxial one? Although flowers with a median abaxial sepal do occur, they are rare in this species. Probably, in a calyx with median abaxial sepal, the greater number of sepals are under strong pressure from the surrounding structures or older flowers. When the first sepal occupies an abaxial-transverse position, only the fourth sepal is under significant pressure and is delayed in development. If the first sepal occupies the median abaxial position, the fourth and the second sepals are shifted more abaxially. As a result, two sepals (fourth and second) face the neighbouring flower, and both experience some pressure.
In A. platanoides, the sepal positions are difficult to interpret. The flower-subtending bracts and the prophylls do not enlarge after their appearance and do not directly influence flower development. However, the next-order flowers may form in the axils of one or both prophylls. The prophylls and, consequently, the flowers in their axils are shifted to the abaxial side. If only one flower of the next order is formed, the first sepal is initiated on the opposite side in a transverse-abaxial position, but this position is more strongly shifted to the adaxial side, compared with A. tataricum subsp. ginnala, owing to the more abaxial position of the prophylls (Fig. 7C, D).
In the case of flowers appearing in the axils of both prophylls, the first sepal is initiated in a transverse-adaxial position, even more towards the adaxial side (Fig. 7A, B). It cannot take the median adaxial position, probably because of the neighbouring flower hanging over it. However, this interpretation is only part of the truth. In reality, the order of sepal initiation is influenced not only by its ‘own’ next-order flowers but also by all neighbouring flowers, because they are all densely placed on the surface of a spherical inflorescence and are in close contact with each other throughout development. At the stage of sepal initiation, we observed only the calyx orientation with the median abaxial odd sepal. It is possible that after initiation, the flower can rotate on a twisting pedicel and the sepals change position owing to the pressure of neighbouring flowers. In addition to this explanation, two more can be proposed. Probably, there are more variants of sepal initiation than we have found, and what we see in older flowers indicates this. However, we observed many flowers at the stage of sepal initiation, and it would be strange not to find this variation. Or it is not the flower rotation, but the ‘shift’ of sepals. The latter hypothesis does not seem to be true, because the carpels are always located in the plane of the first sepal. At the stage of gynoecium emergence, we did not see flowers with a median adaxial sepal (Fig. 7A, C, G).
A similar development was observed in A. spicatum. Prophylls do not expand after initiation. The flower of the next order arises only in the axil of one prophyll. Therefore, the first sepal is initiated on the opposite side in the transverse-abaxial position (Fig. 7I). But later, we rarely observed flowers with a median abaxial sepal (Fig. 7J). This shift apparently occurs owing to the pressure of neighbouring flowers.
We can conclude that, at least in two species, the close contact between the developing flowers destabilizes sepal positions and, thus, the entire flower. In other words, the surrounding structures shape the floral meristem, and this shape underlines the sequence of at least sepal initiation (see ‘imprinted shape’ in the paper by Endress, 2008). Acer platanoides and A. spicatum are characterized by extensively branched inflorescences (panicles/thyrsoids) and a complex flower shape at the stage of sepal initiation (Fig. 1B, C, M, L). As a result, each flower experiences a unique set of mechanical pressures. A complex mosaic of flowers is formed, where both the position of the first sepal and the position of the median sepal may be different.
However, in most cases, the floral structure in terms of organ positions is stable in dense inflorescences. Apparently, floral meristems of stable flowers should have a simple shape (rounded or of a certain outline) and a set number of surrounding structures in certain positions when ready to produce organs (Harris, 1991; Endress, 2008). Owing to such spatial constraints, similar flower constructions are often present in plants that are not closely related, for example, in genera Maundia and Aponogeton (Sokoloff et al., 2013), Potamogeton and Carludovica (Harling et al., 1998; Nunes et al., 2012). However, as we pointed out above, external pressure does not solely determine the stabilization or destabilization of the flower structure.
Loss of prophylls, in most cases, changes the order of sepal initiation (Fig. 9) and sometimes flower orientation, especially in monocots (Remizowa et al., 2013). Transversal sepals are the first to appear in the positions of lost floral prophylls in eudicots (Evans and Dickinson, 1999a; Prenner, 2004; Remizowa et al., 2013; Ronse De Craene, 2022). This trend is also observed in maples, except for A. tegmentosum with small sepal primordia (Fig. 6K; Zavialov and Remizowa, 2025).
In maples without floral prophylls, the median sepal can be either adaxial (A. tegmentosum; Fig. 6K) or abaxial (A. nikoense; Fig. 6P). The different configurations of the calyx are probably inherited from the closest related species with prophylls.
In the section Rubra, flower-subtending bracts and bracteoles are lost (Fig. 9E). The only evidence of a flower-subtending bract is a ridge at the base of the floral primordium (Fig. 5B). This is evidence of a so-called cryptic flower-subtending bract; a similar developmental feature occurs in Arabidopsis thaliana and in most members of the Brassicaceae (Erbar and Leins, 1997; Long and Barton, 2000). Representatives of this section are characterized by variability in the position of the median sepal. This can be explained not by multidirectional pressures but rather by the absence of any obvious mechanical forces prescribing the sepal positions, hence all possible variants are realized (Fig. 6L–N, Q–T). Similar correlations were observed in staminate inflorescences of A. negundo (Zavialov and Remizowa, 2024). Small partial inflorescences, which are under greater pressure from surrounding structures, contain tetramerous flowers that are more typical for this species. Large partial inflorescences contain a large proportion of flowers of a different merism (trimerous, pentamerous, etc.).
Position of carpels
Variability in the position of the gynoecium and its dependence on the position of other organs is not common (Douglas and Tucker, 1996; Bachelier and Endress, 2009; Karpunina et al., 2017; Sokoloff et al., 2018; Wei and Ronse De Craene, 2020). The variability in carpel position is often overlooked at the species level. An obvious condition for possible variability in the gynoecium position is that the gynoecium merism is not equal to the merism of all other flower whorls.
Species of the family Fabaceae possess a highly stable position of monocarpellate gynoecium. But in actinomorphic flowers with unstable merism in species of the genus Gleditsia L., the position of the carpel can vary (Tucker, 1991). It is believed that the loss of stability in this case is associated with the reduction of inflorescence bracts (Tucker, 1991), although there are other legumes without bracts, but with a conservative floral structure, for example, some representatives of the genus Trifolium (Retallack and Willison, 1990).
In subfamily Grevilleoideae (Proteaceae), carpel position can also vary (Douglas and Tucker, 1996). It has been shown that the carpel position or, more precisely, the position of the cleft, is determined by the flower proportions and the proportions of the meristematic apex that gives rise to the gynoecium (Douglas and Tucker, 1996). However, what causes these developmental differences remains unclear.
In species with panicles or thyrsoids, the position of carpels can be associated with the position of the first sepal (in A. platanoides, A. spicatum and A. tataricum subsp. ginnala) or with the position of the flower within the inflorescence (in A. pseudosieboldianum). Those correlations provide additional arguments against the ideas of bidirectional prepatterning of the floral meristem and the autonomy of gynoecium development (Choob and Penin, 2004; Sokoloff et al., 2017). Probably, strict determination of the gynoecium position is more characteristic for zygomorphic flowers, whereas in many actinomorphic flowers, more detailed studies would reveal variability in the gynoecium position. The position of the gynoecium would perhaps be more definite if the maple ancestor had zygomorphic flowers or actinomorphic ones recently inherited from zygomorphic ones.
The gynoecium in all studied maples with racemose inflorescences is inserted in the transverse plane (Figs 6G–T and 9B–E). The position of the carpels is related to the arrangement of the floral organs, which, in turn, is associated with the shape of the floral meristem determined by the inflorescence structure. Most studied species are characterized by few-flowered racemose inflorescences with predominantly decussate phyllotaxis. Already, Eichler (1878) believed that phyllotaxis in the inflorescence can play an important role in stabilizing the position of the gynoecium.
In racemose inflorescences, the floral meristem is transversally elongated (irrespective of prophyll reduction; Fig. 9B–E). This configuration is often retained in the floral meristem after the initiation of the perianth and androecium, which facilitates the orientation of the carpels in the transversal plane. It can be assumed that carpels appear where there is least pressure and where stamens are lost. However, in A. barbinerve and A. carpinifolium, the carpels are opposite the transverse stamens, meaning that they compete with them for space and experience mutual pressure (Fig. 4). In A. saccharinum and A. rubrum, the androecium configuration can vary, but at least one stamen is located opposite the carpel (Figs 1T, U and 6L–N, Q–T).
Floral and inflorescence evolution in a phylogenetic context
High-quality phylogenetic data were obtained for the genus Acer in recent years (Li et al., 2019; Areces-Berazain et al., 2021). The new data did not significantly impact the taxonomy of maples, generally agreeing with the morphology-based taxonomy. The reason is probably the concentration of morphological differences at the section level. A unique combination of generative and vegetative features characterizes each maple section (De Jong, 1976, 2004). For the same reason, combining sections based on morphology into larger taxonomic groups is impossible. In our study, which covers slightly more than half of the sections, no features important for the intersectional taxonomic level were found. Only the sections Arguta and Negundo clearly stand out as morphologically distinct groups with some synapomorphies, such as racemose inflorescences, dioecy, tetramerous flowers and the lack of common petal–stamen primordia (Fig. 8). However, until recently these sections did not come together (De Jong, 1976; Renner et al., 2007). Most traits previously known or described in the present study are highly homoplastic (Fig. 8). Interestingly, some features are combined only in a certain way. Although it is not a strict rule, derived floral structural types are combined predominantly with racemose inflorescences. The transversal gynoecium, tetramerous calyx and, surprisingly, dioecy are found only in species with racemose inflorescences (Fig. 8). The first two traits, as discussed above, are related to spatial and, possibly, dimensional developmental constraints (Fig. 9). The connection between the reproductive system and inflorescence type is more intriguing, especially because of the diversity of reproductive systems in maples (De Jong, 1976; Renner et al., 2007). More detailed studies are needed to clarify this structure–function correlation. The median sepal position in pentamerous flowers is highly unstable among maples, probably owing to connection with very fine developmental characteristics of inflorescences. Increasing the number of species studied would not help much in tracing evolutionary lines. Closely related maples often differ in flowering phenology and some quantitative characteristics of the inflorescence (the number of branches of different orders, the number of flowers, etc.) and, consequently, also in the position of the median sepal (De Jong, 1976).
Despite the large number of homoplasies, the data obtained generally allow us to describe the most probable ancestral state of the generative organs for maples. The inflorescence was a branched panicle or thyrsoid; it is possible that both types occurred together. In general, this coincides with the ideas of De Jong (1976), although he emphasized that the thyrsoid was the ancestral inflorescence (the first type, according to his classification), and the panicle originated from it (the second type, according to his classification). The flower had a typical structure: five sepals, five petals, eight stamens and two carpels. The ancestral position of the median sepal, as discussed above, cannot be determined, but in comparison with other Sapindaceae, it should be adaxial. The androecium most probably consisted of five stamens developed from individual primordia and three stamens initiated from common petal–stamen primordia (Fig. 9A).
Thus, we do not identify any taxonomically important trait, but we discover the important predictive properties of the ‘inflorescence-based’ traits. The transformations of the flower structure, the stability of the flower organ position, and the reproductive system are associated precisely with the transformations of the inflorescence structure. This observation is crucial for the entire family Sapindaceae. Species of Sapindaceae are extremely diverse and numerous, but many of them are endemic or grow in places that are hard to reach (Radlkofer, 1933; Acevedo-Rodríguez et al., 2010). Therefore, in the near future, it would be impossible to characterize the flower structure and development of most members in detail. We propose testing the correlations of structure and development found in the genus Acer in other genera that are accessible to researchers. If the patterns turn out to be sufficiently universal, we can use relatively meager taxonomic descriptions to predict important structural features that are usually overlooked (see also Endress, 2008). Our research shows that the number of carpels, which is given such great importance as a distinctive feature of maples in comparison to other Sapindaceae (Ronse De Craene, 2022; Zhang et al., 2022), does not play such a significant role in flower development. In maples, the number of carpels varies from one to three, and sometimes more than three (De Jong, 1976). Such variability does not affect the number of stamens and their position. The same can be said about Dipteronia (Zhang et al., 2022). In contrast, the number of carpels is reduced to two in some other Sapindaceae; for example, in Delavaya (Cao and Xia, 2009).
Conclusion
Structural diversity of flowers and inflorescences in the genus Acer is established mainly by heterochronic shifts in the initiation sequence and developmental rates accompanied by changes in the meristem size. Mechanical forces and spatial constraints are also important for the interaction between inflorescences and flowers. These factors determine the organ positions and the order of their initiation, starting in the calyx and followed by other organs. For now, it is difficult to say precisely which developmental features are attributable to spatial constraints and which are attributable to internal factors. Regardless of the mechanisms of their influence, inflorescences play a crucial role in the evolution of flowers by providing a basis for structural transformations of flowers via changes in inflorescence architecture. Even a slight change in size and arrangement of inflorescence components (axes and phyllomes) has the potential for subsequent changes in floral development and morphology. Such a connection allows recognition of precisely tuned patterns that are associated mainly with developmental conditions set by non-floral organs. These patterns are highly homoplastic in terms of phylogeny but, on the contrary, highly conservative, because they are present repeatedly in species with similar inflorescence architecture. These observations on morphogenetic correlations in inflorescences of Acer will contribute to a better understanding of the evolution of floral traits in Sapindaceae and eudicots in general.
Supplementary Material
ACKNOWLEDGEMENTS
We are grateful to G. A. Boyko for permissions to collect specimens from the Botanical Garden of Moscow State University, to G. A. Firsov for providing specimens from the Saint Petersburg Botanical Garden, and to T. E. Kramina, D. D. Sokoloff, A.D. Lisitsina and A. O. Astashkin for help in collecting material. Scanning electron microscopy studies were carried out at the Shared Research Facility ‘Electron microscopy in life sciences’ at Moscow State University (Unique Equipment ‘Three-dimensional electron microscopy and spectroscopy’). We also thank two anonymous reviewers and Dr Alexandru Tomescu for helpful comments and suggestions.
Contributor Information
Alexander E Zavialov, Department of Higher Plants, M.V. Lomonosov Moscow State University, Moscow 119234, Russia.
Margarita V Remizowa, Department of Higher Plants, M.V. Lomonosov Moscow State University, Moscow 119234, Russia.
SUPPLEMENTARY DATA
Supplementary data are available online at Annals of Botany online and consist of the following. Table S1: list of vouchers of studied species.
FUNDING
The study was conducted under the state assignment of Lomonosov Moscow State University.
AUTHOR CONTRIBUTIONS
Conceptualization: A.E.Z., M.V.R.; Material preparation, data collection and analysis: A.E.Z, M.V.R.; Writing—original draft preparation: A.E.Z.; Writing—review and editing: M.V.R.
DATA AVAILABILITY
All data supporting this study are available at the original published sources listed in this article and in the Supplementary Data.
REFERENCES
- Abad U, Sassi M, Traas J. 2017. Flower development: from morphodynamics to morphomechanics. Philosophical Transactions of the Royal Society B: Biological Sciences 372: 20150545. doi: 10.1098/rstb.2015.0545 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Abrams MD. 1998. The red maple paradox. BioScience 48: 355–364. doi: 10.2307/1313374 [DOI] [Google Scholar]
- Acevedo-Rodríguez P, Van Welzen PC, Adema F, Van Der Ham RWJM. 2010. Sapindaceae. In: Kubitzki K. ed. Flowering plants. Eudicots: Sapindales, Cucurbitales, Myrtaceae. Berlin: Springer, 357–407. [Google Scholar]
- Areces-Berazain F, Hinsinger DD, Strijk JS. 2021. Genome-wide supermatrix analyses of maples (Acer, Sapindaceae) reveal recurring inter-continental migration, mass extinction, and rapid lineage divergence. Genomics 113: 681–692. doi: 10.1016/j.ygeno.2021.01.014 [DOI] [PubMed] [Google Scholar]
- Bachelier JB, Endress PK. 2009. Comparative floral morphology and anatomy of Anacardiaceae and Burseraceae (Sapindales), with a special focus on gynoecium structure and evolution. Botanical Journal of the Linnean Society 159: 499–571. doi: 10.1111/j.1095-8339.2009.00959.x [DOI] [Google Scholar]
- Barrio RÁ, Hernández-Machado A, Varea C, Romero-Arias JR, Álvarez-Buylla E. 2010. Flower development as an interplay between dynamical physical fields and genetic networks. PLoS One 5: e13523. doi: 10.1371/journal.pone.0013523 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Buchenau F. 1861. Morphologische Bemerkungen über einige Acerineen. Botanische Zeitung 19: 265–269. https://www.biodiversitylibrary.org/item/104711#page/279/mode/thumb [Google Scholar]
- Buerki S, Callmander MW, Acevedo-Rodriguez P, et al. 2021. An updated infra-familial classification of Sapindaceae based on targeted enrichment data. American Journal of Botany 108: 1234–1251. doi: 10.1002/ajb2.1693 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bukhari G, Zhang J, Stevens PF, Zhang W. 2017. Evolution of the process underlying floral zygomorphy development in pentapetalous angiosperms. American Journal of Botany 104: 1846–1856. doi: 10.3732/ajb.1700229 [DOI] [PubMed] [Google Scholar]
- Bull-Hereñu K, Dos Santos P, Toni JFG, et al. 2022. Mechanical forces in floral development. Plants 11: 661. doi: 10.3390/plants11050661 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cao L-M, Liu J, Lin Q, Ronse De Craene LP. 2018. The floral organogenesis of Koelreuteria bipinnata and its variety K. bipinnata var. integrifolia (Sapindaceae): evidence of floral constraints on the evolution of monosymmetry. Plant Systematics and Evolution 304: 923–935. doi: 10.1007/s00606-018-1519-y [DOI] [Google Scholar]
- Cao L-M, Ronse De Craene LP, Wang Z-X, Wang Y-H. 2017. The floral organogenesis of Eurycorymbus cavaleriei (Sapindaceae) and its systematic implications. Phytotaxa 297: 234. doi: 10.11646/phytotaxa.297.3.2 [DOI] [Google Scholar]
- Cao L-M, Xia N. 2009. Floral organogenesis of Delavaya toxocarpa (Sapindaceae; Sapindales). Journal of Systematics and Evolution 47: 237–244. doi: 10.1111/j.1759-6831.2009.00023.x [DOI] [Google Scholar]
- Cao L-M, Xia N-H, Deng Y-F. 2006. Floral organogenesis of Handeliodendron bodinieri (Sapindaceae) and its systematic implications. Acta Phytotaxonomica Sinica 44: 393. doi: 10.1360/aps050126 [DOI] [Google Scholar]
- Cheng X, Li G, Tang Y, Wen J. 2018. Dissection of genetic regulation of compound inflorescence development in Medicago truncatula. Development 145: dev158766. doi: 10.1242/dev.158766 [DOI] [PubMed] [Google Scholar]
- Choob VV, Penin AA. 2004. Structure of flower in Arabidopsis thaliana: spatial pattern formation. Russian Journal of Developmental Biology 35: 224–227. doi: 10.1023/B:RUDO.0000036712.64312.e8 [DOI] [Google Scholar]
- Coen E, Prusinkiewicz P. 2024. Developmental timing in plants. Nature Communications 15: 2674. doi: 10.1038/s41467-024-46941-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Conti L, Bradley D. 2007. TERMINAL FLOWER1 is a mobile signal controlling Arabidopsis architecture. The Plant Cell 19: 767–778. doi: 10.1105/tpc.106.049767 [DOI] [PMC free article] [PubMed] [Google Scholar]
- De Jong PC. 1976. Flowering and sex expression in Acer L.: a biosystematic study. Wageningen: H. Veenman & Zonen B.V. [Google Scholar]
- De Jong PC. 2004. World maple diversity. In: Wiegrefe SJ, Angus H, Otis, Gregory DP. eds. International maple symposium. UK: Westonbirt Arboretum and Royal Agricultural College, 2–11. [Google Scholar]
- Douglas AW, Tucker SC. 1996. The developmental basis of diverse carpel orientations in Grevilleoideae (Proteaceae). International Journal of Plant Sciences 157: 373–397. doi: 10.1086/297355 [DOI] [Google Scholar]
- Eichler AW. 1878. Blüthendiagramme. Leipzig: Verlag von Wilhelm Engelmann. [Google Scholar]
- El Ottra JHL, De Albuquerque Melo-de-Pinna GF, Demarco D, Pirani JR, Ronse De Craene LP. 2022. Gynoecium structure in Sapindales and a case study of Trichilia pallens (Meliaceae). Journal of Plant Research 135: 157–190. doi: 10.1007/s10265-022-01375-y [DOI] [PubMed] [Google Scholar]
- Endress PK. 1995. Major evolutionary traits of monocot flowers. In: Rudall P. ed. Monocotyledons: systematics and evolution. Kew, UK: Royal Botanic Gardens, 43–79. [Google Scholar]
- Endress PK. 2008. The whole and the parts: relationships between floral architecture and floral organ shape, and their repercussions on the interpretation of fragmentary floral fossils. Annals of the Missouri Botanical Garden 95: 101–120. doi: 10.3417/2006190 [DOI] [Google Scholar]
- Endress PK. 2010a. Disentangling confusions in inflorescence morphology: patterns and diversity of reproductive shoot ramification in angiosperms. Journal of Systematics and Evolution 48: 225–239. doi: 10.1111/j.1759-6831.2010.00087.x [DOI] [Google Scholar]
- Endress PK. 2010b. Flower structure and trends of evolution in eudicots and their major subclades. Annals of the Missouri Botanical Garden 97: 541–583. doi: 10.3417/2009139 [DOI] [Google Scholar]
- Erbar C, Leins P. 1997. Different patterns of floral development in whorled flowers, exemplified by Apiaceae and Brassicaceae. International Journal of Plant Sciences 158: S49–S64. doi: 10.1086/297506 [DOI] [Google Scholar]
- Evans RC, Dickinson TA. 1999a. Floral ontogeny and morphology in subfamily Amygdaloideae T. & G. (Rosaceae). International Journal of Plant Sciences 160: 955–979. doi: 10.1086/314175 [DOI] [PubMed] [Google Scholar]
- Evans RC, Dickinson TA. 1999b. Floral ontogeny and morphology in subfamily Spiraeoideae Endl. (Rosaceae). International Journal of Plant Sciences 160: 981–1012. doi: 10.1086/314176 [DOI] [PubMed] [Google Scholar]
- Evans RC, Dickinson TA. 2005. Floral ontogeny and morphology in Gillenia (“Spiraeoideae”) and subfamily Maloideae C. Weber (Rosaceae). International Journal of Plant Sciences 166: 427–447. doi: 10.1086/428631 [DOI] [Google Scholar]
- Harling G, Wilder GJ, Eriksson R. 1998. Cyclanthaceae. In: Kubitzki K. ed. Flowering plants monocotyledons. Berlin, Heidelberg: Springer Berlin Heidelberg, 202–215. [Google Scholar]
- Harris E. 1991. Comparative inflorescence and floral development in the Asteraceae (Volumes I and II). PhD Thesis, The Louisiana State University and Agricultural and Mechanical College, USA.
- Joyce EM, Appelhans MS, Buerki S, et al. 2023. Phylogenomic analyses of Sapindales support new family relationships, rapid Mid-Cretaceous Hothouse diversification, and heterogeneous histories of gene duplication. Frontiers in Plant Science 14: 1063174. doi: 10.3389/fpls.2023.1063174 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Karpunina PV, Oskolski AA, Nuraliev MS. 2017. Patterns of carpel arrangement in gynoecia of Araliaceae: evidence from Polyscias. In: Nuraliev M, Oskolski A, Tilney P. eds. IX Apiales symposium. Abstract book. Guangzhou: The Gold Coast Marina Club, 15–16. [Google Scholar]
- Kirchoff BK. 1983. Floral organogenesis in five genera of the Marantaceae and in Canna (Cannaceae). American Journal of Botany 70: 508–523. doi: 10.1002/j.1537-2197.1983.tb07878.x [DOI] [Google Scholar]
- Kostina MV. 2007. On the significance of generative shoot characters in the systematics of Acer L. Byulleten’ Moskovskogo Obshchestva Ispytatelei Prirody Otdel Biologicheskii 112: 16–25. https://moip-bio.msu.ru/wp-content/uploads/moip_2007_112_5.pdf [Google Scholar]
- Kramer EM. 2018. Floral architecture: regulation and diversity of floral shape and pattern. Annual Plant Reviews online 121–148. doi: 10.1002/9781119312994.apr0167 [DOI] [Google Scholar]
- Lampugnani ER, Kilinc A, Smyth DR. 2012. PETAL LOSS is a boundary gene that inhibits growth between developing sepals in Arabidopsis thaliana. The Plant Journal 71: 724–735. doi: 10.1111/j.1365-313X.2012.05023.x [DOI] [PubMed] [Google Scholar]
- Leins P, Erbar C. 1997. Floral developmental studies: some old and new questions. International Journal of Plant Sciences 158: S3–S12. doi: 10.1086/297504 [DOI] [Google Scholar]
- Lemmon ZH, Park SJ, Jiang K, Van Eck J, Schatz MC, Lippman ZB. 2016. The evolution of inflorescence diversity in the nightshades and heterochrony during meristem maturation. Genome Research 26: 1676–1686. doi: 10.1101/gr.207837.116 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li J, Stukel M, Bussies P, et al. 2019. Maple phylogeny and biogeography inferred from phylogenomic data. Journal of Systematics and Evolution 57: 594–606. doi: 10.1111/jse.12535 [DOI] [Google Scholar]
- Lippman ZB, Cohen O, Alvarez JP, et al. 2008. The making of a compound inflorescence in tomato and related nightshades. PLoS Biology 6: e288. doi: 10.1371/journal.pbio.0060288 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Long J, Barton MK. 2000. Initiation of axillary and floral meristems in Arabidopsis. Developmental Biology 218: 341–353. doi: 10.1006/dbio.1999.9572 [DOI] [PubMed] [Google Scholar]
- Moyroud E, Glover BJ. 2017. The evolution of diverse floral morphologies. Current Biology 27: R941–R951. doi: 10.1016/j.cub.2017.06.053 [DOI] [PubMed] [Google Scholar]
- Muller J, Leenhouts PW. 1976. A general survey of pollen types in Sapindaceae in relation to taxonomy. In: Ferguson IK, Muller J. eds. The evolutionary significance of the exine. London: Academic Press for the Linnean Society of London, 407–445. [Google Scholar]
- Naghiloo S, Dadpour MR, Movafeghi A. 2012. Floral ontogeny in Astragalus compactus (Leguminosae: Papilionoideae: Galegeae): variable occurrence of bracteoles and variable patterns of sepal initiation. Planta 235: 793–805. doi: 10.1007/s00425-011-1538-1 [DOI] [PubMed] [Google Scholar]
- Nunes ELP, Lima MCD, Moço MCDC, Coan AI. 2012. Floral development in Potamogeton (Potamogetonaceae, Alismatales) with emphasis on gynoecial features. Aquatic Botany 100: 56–61. doi: 10.1016/j.aquabot.2012.02.006 [DOI] [Google Scholar]
- Payer J-B. 1857. Traité de organogénie comparée de la fleur. Paris: V. Masson. [Google Scholar]
- Perkins TD, Van Den Berg AK. 2009. Maple syrup—Production, composition, chemistry, and sensory characteristics. Advances in Food and Nutrition Research 56: 101–143. doi: 10.1016/S1043-4526(08)00604-9 [DOI] [PubMed] [Google Scholar]
- Prenner G. 2004. New aspects in floral development of Papilionoideae: initiated but suppressed bracteoles and variable initiation of sepals. Annals of Botany 93: 537–545. doi: 10.1093/aob/mch076 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Prusinkiewicz P, Erasmus Y, Lane B, Harder LD, Coen E. 2007. Evolution and development of inflorescence architectures. Science 316: 1452–1456. doi: 10.1126/science.1140429 [DOI] [PubMed] [Google Scholar]
- Radlkofer L. 1933. Sapindaceae. In: Engler A. ed. Das Pflanzenreich IV, 165 (Heft 98a-h). Leipzig: Verlag von Wilhelm Engelmann, 961–1114. [Google Scholar]
- Remizowa MV, Rudall PJ, Choob VV, Sokoloff DD. 2013. Racemose inflorescences of monocots: structural and morphogenetic interaction at the flower/inflorescence level. Annals of Botany 112: 1553–1566. doi: 10.1093/aob/mcs246 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Renner SS, Beenken L, Grimm GW, Kocyan A, Ricklefs RE. 2007. The evolution of dioecy, heterodichogamy, and labile sex expression in Acer. Evolution 61: 2701–2719. doi: 10.1111/j.1558-5646.2007.00221.x [DOI] [PubMed] [Google Scholar]
- Retallack B, Willison JHM. 1990. Floral development in Ottawa and florex red clover Trifolium pratense (Papilionoideae: Leguminosae). American Journal of Botany 77: 1222–1230. doi: 10.1002/j.1537-2197.1990.tb13621.x [DOI] [Google Scholar]
- Ronse De Craene LP. 2016. Meristic changes in flowering plants: how flowers play with numbers. Flora 221: 22–37. doi: 10.1016/j.flora.2015.08.005 [DOI] [Google Scholar]
- Ronse De Craene LP. 2018. Understanding the role of floral development in the evolution of angiosperm flowers: clarifications from a historical and physico-dynamic perspective. Journal of Plant Research 131: 367–393. doi: 10.1007/s10265-018-1021-1 [DOI] [PubMed] [Google Scholar]
- Ronse De Craene LP. 2022. Floral diagrams: an aid to understanding flower morphology and evolution, 2nd edn. New York: Cambridge University Press. [Google Scholar]
- Ronse De Craene LP. 2024. The interaction between heterochrony and mechanical forces as main driver of floral evolution. Journal of Plant Research 137: 697–717. doi: 10.1007/s10265-024-01526-3 [DOI] [PubMed] [Google Scholar]
- Ronse De Craene LP, Clinckemaillie D, Smets E. 1993. Stamen-petal complexes in Magnoliatae. Bulletin du Jardin botanique national de Belgique/Bulletin van de National Plantentuin van België 62: 97. doi: 10.2307/3668268 [DOI] [Google Scholar]
- Ronse De Craene LP, Smets EF. 1993. Dédoublement revisited: towards a renewed interpretation of the androecium of the Magnoliophytina. Botanical Journal of the Linnean Society 113: 103–124. doi: 10.1111/j.1095-8339.1993.tb00333.x [DOI] [Google Scholar]
- Ronse De Craene LP, Smets EF. 1994. Merosity in flowers: definition, origin, and taxonomic significance. Plant Systematics and Evolution 191: 83–104. doi: 10.1007/BF00985344 [DOI] [Google Scholar]
- Ronse De Craene LP, Smets E, Clinckemaillie D. 2000. Floral ontogeny and anatomy in Koelreuteria with special emphasis on monosymmetry and septal cavities. Plant Systematics and Evolution 223: 91–107. doi: 10.1007/BF00985329 [DOI] [Google Scholar]
- Sokoloff DD, Karpunina PV, Nuraliev MS, Oskolski AA. 2018. Flower structure and development in Melanophylla (Torricelliaceae: Apiales): lability in direction of corolla contortion and orientation of pseudomonomerous gynoecium in a campanulid eudicot. Botanical Journal of the Linnean Society 187: 247–271. doi: 10.1093/botlinnean/boy020 [DOI] [Google Scholar]
- Sokoloff DD, Nuraliev MS, Oskolski AA, Remizowa MV. 2017. Gynoecium evolution in angiosperms: monomery, pseudomonomery, and mixomery. Moscow University Biological Sciences Bulletin 72: 97–108. doi: 10.3103/S0096392517030105 [DOI] [Google Scholar]
- Sokoloff DD, Remizowa M, Ronse De Craene LP, Bachelier JB, Jabbour F. 2024. Jean-Baptiste Payer (1818–1860), prolific botanist and enlightened French citizen: his legacy in flower development and a need for a new, global compendium. Botany Letters 171: 433–455. doi: 10.1080/23818107.2024.2373433 [DOI] [Google Scholar]
- Sokoloff DD, Von Mering S, Jacobs SWL, Remizowa MV. 2013. Morphology of Maundia supports its isolated phylogenetic position in the early-divergent monocot order Alismatales: Maundia and evolution of Alismatales. Botanical Journal of the Linnean Society 173: 12–45. doi: 10.1111/boj.12068 [DOI] [Google Scholar]
- Souer E, Krol AVD, Kloos D, et al. 1998. Genetic control of branching pattern and floral identity during Petunia inflorescence development. Development 125: 733–742. doi: 10.1242/dev.125.4.733 [DOI] [PubMed] [Google Scholar]
- Tucker SC. 1991. Helical floral organogenesis in Gleditsia, a primitive Caesalpinioid legume. American Journal of Botany 78: 1130–1149. doi: 10.1002/j.1537-2197.1991.tb14520.x [DOI] [Google Scholar]
- Tucker SC. 1996. Trends in evolution of floral ontogeny in Cassia sensu stricto, Senna, and Chamaecrista (Leguminosae: Caesalpinioideae: Cassieae: Cassiinae); a study in convergence. American Journal of Botany 83: 687–711. doi: 10.1002/j.1537-2197.1996.tb12758.x [DOI] [Google Scholar]
- Tucker SC. 2002a. Comparative floral ontogeny in Detarieae (Leguminosae: Caesalpinioideae). 1. Radially symmetrical taxa lacking organ suppression. American Journal of Botany 89: 875–887. doi: 10.3732/ajb.89.6.875 [DOI] [PubMed] [Google Scholar]
- Tucker SC. 2002b. Comparative floral ontogeny in Detarieae (Leguminosae: Caesalpinioideae). 2. Zygomorphic taxa with petal and stamen suppression. American Journal of Botany 89: 888–907. doi: 10.3732/ajb.89.6.888 [DOI] [PubMed] [Google Scholar]
- Tucker SC. 2003. Floral development in legumes. Plant Physiology 131: 911–926. doi: 10.1104/pp.102.017459 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wei L, Ronse De Craene LP. 2020. Hofmeister’s rule’s paradox: explaining the changeable carpel position in Caryophyllaceae. International Journal of Plant Sciences 181: 911–925. doi: 10.1086/710493 [DOI] [Google Scholar]
- Zavialov AE, Remizowa MV. 2024. Unstable patterns of flower structure and development in the almost dioecious species Acer negundo (Sapindaceae, Sapindales). Botanical Journal of the Linnean Society 204: 267–281. doi: 10.1093/botlinnean/boad048 [DOI] [Google Scholar]
- Zavialov AE, Remizowa MV. 2025. Androecium homologies in eight-staminate maples: a developmental study. Journal of Plant Research 1–22. doi: 10.1007/s10265-025-01641-9 [DOI] [PubMed] [Google Scholar]
- Zhang Y, Cao L-M, Ronse De Craene LP. 2022. Inflorescence and floral development of Dipteronia sinensis (Sapindaceae): demonstration of developmental constraints in the shaping of the flower. Flora 291: 152069. doi: 10.1016/j.flora.2022.152069 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data supporting this study are available at the original published sources listed in this article and in the Supplementary Data.









