How to lose a whorl: the evolutionary and developmental biology of apetaly
Department of Botany and Beaty Biodiversity Museum, University of British Columbia, Vancouver, Canada V6T 1Z4
University College Dublin, Ireland
Correspondence: quentin.cronk@ubc.caAbstract
The general plan of the core eudicot flower involves two perianth whorls, a calyx, and a corolla. However, numerous eudicots, in multiple lineages, have only one perianth whorl, generally assumed to be the calyx, resulting in apetaly. Historically, these plants were placed in the polyphyletic taxonomic group ‘Monochlamydeae’ and, unsure about the nature of the single perianth whorl, the 19th century botanist de Candolle coined the word ‘tepal’ to indicate this uncertainty. This review surveys the molecular mechanisms of apetaly, its evolutionary drivers, and its possible role as a gateway to the unisexual flower through the B-switch hypothesis. The pollination biology of the petal (and petaloidy) is considered. Many apetalous plants have evolved petaloidy of the remaining (calyx) whorl. It is suggested that more attention be paid to the molecular dissection of the three-whorl floral ground plan. Fortunately, excellent genomic resources to aid further research are now available for species with a unipartite perianth, such as Beta, Cannabis, Fagopyrum, Portulaca, Spinacia, and Urtica.
Teaser
In this review I look at how and why so many eudicot lineages jettisoned the petal, often to reinvent it via petaloidy of the remaining perianth whorl.
When a flower has two floral envelopes, the outer one is always a calyx; there can therefore be no doubt in this regard. But if it has only a single envelope, what is its nature? The question is generally resolved by assuming that this simple perianth is a calyx, even though it sometimes acquires a delicacy and coloration that make it resemble a corolla (Duchartre, 1867).
Introduction
The problem
Members of the eudicot clade (which includes most dicot plant species) generally have a differentiated perianth of calyx and corolla. This is particularly true of the Pentapetalae subclade that has many variations, reductions, and elaborations on the fundamental floral structure of K5 C5 A5 G5 (i.e. four floral whorls each of five members). In this notation K = calyx, C = corolla, A = stamens, and G = carpels (other conventions of the standard floral formula are omitted here). The Pentapetalae include Arabidopsis K4 C4 A2 + 4 G2 and Antirrhinum K5 C5 A4 G2. One very common variation on the basic pattern is the repeated reduction of two perianth whorls to one, to become, for example, P5 A5 G1 (which is the floral notation of the Marvel of Peru, Mirabilis jalapa, among other plants). Here, P is used to denote the single perianth category.
This raises three types of questions. (i) Nature: when the number of perianth whorls is reduced from two to one, what is the nature of the remaining whorl—calyx or corolla? Where there is a biseriate perianth, the calyx (outer whorl) and corolla (inner whorl) may be defined positionally, but with only one whorl, there is no positional reference, and we must fall back on phenotypic and gene expression definitions (discussed below). (ii) Development: how has a whorl been lost in molecular and developmental terms? As this has occurred multiple times in evolution, has the same underlying mechanism been used, or is every case different? (iii) Evolution: why has a perianth whorl been lost repeatedly? What is the selective advantage to those clades that have a single perianth whorl? The biseriate perianth is maintained throughout most of the core eudicots, presumably because it is important to separate the functions of protection in the bud (calyx) from attraction to pollinators at anthesis (corolla). In that case, what might be the fitness advantage of jettisoning a whorl?
In this review, I will attempt to answer these questions from an evo-devo perspective, as far as is possible given the present fragmentary state of knowledge.
Historical background
Linnaeus’s Systema Sexualis (sexual system) was an artificial classification based almost entirely on the inner two whorls of the flower, namely the number of fertile floral organs (stamens and pistils). In the tentative steps towards a natural system, the perianth played a greater role. So de Jussieu (1789) had a group, ‘Apetalae’, for those plants that, in his opinion, had a calyx but no corolla (or no perianth of any kind). Augustin Pyramus de Candolle took on this idea in his Regnum Vegetabile (de Candolle, 1818) but called the group the ‘Monochlamydeae’ (i.e. one covering). This change reflected the fact that de Candolle was not so sure of the morphological nature of the remaining perianth. ‘Monochlamydeae’ makes no claims in that regard, whereas ‘Apetalae’ does.
John Lindley popularized the natural system in Britain in his book ‘An Introduction to the Natural System of Botany’ (Lindley, 1830), coining the English word ‘monochlamydeous’. He writes: ‘If it [the calyx] is unaccompanied by the corolla, plants are said to be Monochlamydeous’. Similarly, those plants which had both a differentiated calyx and corolla were termed dichlamydeous. The history of the monochlamydeous groupings in early botany is set out in Table 1.
| Author (name of class) | Families (ordo) in class |
|---|---|
|
de Jussieu (1789)
(Apetalae) | Stamina epigyna: Aristolochiae; Stamina perigyna: Elaeagni, Thymelaeae, Proteae, Lauri, Polygoneae, Atriplices; Stamina hypogyna: Amaranthi, Plantagines, Nyctagines, Plumbagines |
|
de Candolle (1818)
(Monochlamydeae) | Phytolaccaceae, Salsolaceae, Basellaceae, Amarantaceae, Nyctaginaceae, Polygonaceae, Lauraceae, Myristicaceae, Proteaceae, Penaeaceae, Geissolomaceae, Thymelaeaceae, Elaeagnaceae, Grubbiaceae, Santalaceae, Hernandiaceae, Begoniaceae, Datiscaceae, Papayaceae, Aristolochiaceae, Nepenthaceae, Stackhousiaceae, Euphorbiaceae, Daphniphyllaceae, Buxaceae, Batidaceae, Empetraceae, Cannabineae, Ulmaceae, Moraceae, Artocarpeae, Urticaceae, Piperaceae, Chloranthaceae, Garryaceae, Cupuliferae, Corylaceae, Juglandeae, Myricaceae, Platanaceae, Betulaceae, Salicineae, Casuarineae |
|
Bentham and Hooker (1880)
(Monochlamydeae) | Curvembryeae: Nyctagineae, Illecebraceae, Amarantaceae, Chenopodiaceae, Phytolaccaceae, Batideae, Polygonaceae; Multiovulatae Aquaticae: Podostemaceae; Multiovulatae Terrestres: Nepenthaceae, Cytinaceae, Aristolochiaceae; Micrembryeae: Piperaceae, Chloranthaceae, Myristiceae, Monimiaceae; Daphnales: Laurineae, Proteaceae, Thymelaeaceae, Penaeaceae, Elaeagnaceae; Achlamydosporeae: Loranthaceae, Santalaceae, Balanophoreae; Unisexuales: Euphorbiaceae, Balanopseae, Urticaceae, Platanaceae, Leitnerieae, Juglandeae, Myricaceae, Casuarineae, Cupuliferae; Anomalae: Salicineae, Lacistemaceae, Empetraceae, Ceratophylleae |
de Candolle (1818) had the idea that the ‘single covering’ often represented the merging of the double covering into one. In his original definition (de Candolle, 1818), he contrasted the dichlamydeous ‘perigonio duplici’ (double perianth) with the monochlamydeous flower, which was: ‘Simplici, petalis nempe aut nullis aut cum calice coalitis’ (‘simple, namely with either no petals, or petals and calyx merged [coalitis]’). In support of the merger of calyx and corolla, that de Candolle called an ‘opinion’ given as a ‘simple hypothesis’, he noted: ‘that this envelope is often green on the outside and coloured internally, that it always has stomata on the outer surface and none on the inner, we shall perhaps be inclined to conclude, that this perigone is formed of a calyx lined, thus to speak, with a petaloid expansion of the torus’ (de Candolle, 1840). This raised the question of the sometimes dubious nature of the single perianth whorl. De Candolle himself solved this by coining a new term: ‘It is well to retain for these doubtful cases of a single envelope a particular name … Following the analogy of the terms sepals and petals, I propose … the name of Tepals’ (de Candolle, 1840). In subsequent usage, this definition was expanded to include a perianth of two identical whorls, as in tulips (Tulipa spp.)
De Candolle’s natural system was adopted, in modified form, by Bentham and Hooker for their Genera Plantarum. This was an encyclopaedic survey of plant diversity that, although written entirely in Latin, had a considerable and lasting influence on botany, particularly in English-speaking countries. They defined their Monochlamydeae (Bentham and Hooker, 1880) by the following characters: ‘Perianth simple, with lobes or segments usually in 1–2 series, similar to each other, often resembling a calyx, sometimes tiny or completely absent. Otherwise similar to Polypetalous plants.’ Bentham and Hooker were well aware that this was a highly artificial group, but were forced to take a pragmatic approach. They wrote: ‘The class Monochlamydeae does not form a natural or well-defined group: many orders are minimally related to each other and often come closer to one or another among the class Polypetalae. Many expert botanists have attempted to distribute them among the Polypetalae. However, they have been only modestly successful, and no system has truly bettered that proposed by Candolle.’ Bentham and Hooker therefore adopted and updated de Candolle’s system (which was by then very out of date). Now, some 150 years later, that we have a reliable outline phylogeny of angiosperms, Bentham and Hooker’s supposition that these plants are widely scattered over many disparate groups is amply confirmed (Table 2).
| Bentham and Hooker family | Modern family | Modern order | Typical floral organ numbers |
|---|---|---|---|
| Batideae | Bataceae* | Brassicales | P0/4 A4 G2 |
| Nyctagineae | Nyctaginaceae* | Caryophyllales | P5 A5 G1 |
| Illecebraceae | Caryophyllaceae (part) | Caryophyllales | P5 A2–10 G2 |
| Amarantaceae | Amaranthaceae (part)* | Caryophyllales | P3–5 A3–5 G1 |
| Chenopodiaceae | Amaranthaceae (part)* | Caryophyllales | P3–4 A2–3 G2 |
| Phytolaccaceae | Phytolaccaceae | Caryophyllales | P5 A5 G5 |
| Polygonaceae | Polygonaceae* | Caryophyllales | P2–6 A6–9 G1 |
| Nepenthaceae | Nepenthaceae** | Caryophyllales | P3–4 A8–9 | P3–4 G3–4 |
| Empetraceae | Ericaceae (part*) | Ericales | P4–5 A3–4 | P4–5 G3–5 (Corema) |
| Euphorbiaceae | Euphorbiaceae**, Phyllanthaceae**, and other segregates | Euphorbiales | P0 A1 | P0 G3 (Euphorbia), P4–6 A2–5 | P5–6 G3 (Phyllanthus) |
| Juglandeae | Juglandaceae* | Fagales | P0–6 A3–50 G3 |
| Myricaceae | Myricaceae** | Fagales | P0 A2–15 | P0 G2 |
| Casuarineae | Casuarinaceae** | Fagales | P1–2 A1 | P0 G1 |
| Cupuliferae | Fagaceae** and Betulaceae** | Fagales | P0–9 A5–10 | P–9 G2 |
| Podostemaceae | Podostemaceae | Malpighiales | P2 A2 G2 |
| Balanopseae | Balanopaceae* | Malpighiales | P0 A1–10 | P0 G2 |
| Salicineae | Salicaceae (part**) | Malpighiales | P0/5 A4–5 | P0/5 G2 |
| Lacistemaceae | Lacistemataceae | Malpighiales | P4 A1 G2–3 |
| Cytinaceae | Cytinaceae** | Malvales | P4 A∞ | P4 G4–8 |
| Thymelaeaceae | Thymelaeaceae | Malvales | P4 A8 G2 |
| Penaeaceae | Penaeaceae | Myrtales | P4 A4 G4 |
| Proteaceae | Proteaceae | Proteales | P4 A4 G1 |
| Platanaceae | Platanaceae** | Proteales | P3–4 + 3–4 A3–4 | P3–4 + 3–4 G5–8 |
| Elaeagnaceae | Elaeagnaceae* | Rosales | P2–4 A4 G2 |
| Urticaceae | Cannabaceae**, Moraceae**, Ulmaceae*, Urticaceae** | Rosales (urticalean Rosales) | P4 A4 | P4 G1–2 |
| Loranthaceae | Loranthaceae* and Santalaceae (part**) | Santalales | P4–6 A4–6 | P4–6 G1–3 |
| Santalaceae | Santalaceae*, Schoepfiaceae*, and other segregates | Santalales | P4–5 A4–5 G2 |
| Balanophoreae | Balanophoraceae** | Santalales | P3–4 A3–4 | P0–2 G1–2 |
| Leitnerieae | Simaroubaceae (part**) | Sapindales | P0–4 A10–12 | P4 G1 |
Among the monochlamydeous plants, the perianth whorl that is lost is overwhelmingly considered to be whorl 2, the petals. For this reason, this review uses the more familiar and accessible term ‘apetalous’ instead of ‘monochlamydeous’, and mostly they mean much the same. However, where the monochlamydeous perianth is so petaloid that ‘apetalous’ does not seem quite suitable, where two whorls exist but with no differentiation, or where there is genuine concern about the whorl identity, then ‘apetalous’ may be ill chosen, and so ‘tepalous’ or ‘monochlamydeous’ is preferred. For instance, buckwheat (Fagopyrum: Fig. 1) has five perianth members derived from sepals, but they are highly petaloid and have the function of attracting insects (buckwheat honey is well known). In non-specialist literature, they are often called petals. Here, de Candolle’s term tepal can be used advisedly rather than sepal.
The de Candolle hypothesis that a single perianth whorl could be formed out of an evolutionary fusion of sepal and petal was concerned with the admixture of physical entities. However, since the development of the concept of organ identity, we can see this as potentially a merger of sepal and petal identities in a single organ. The de Candolle hypothesis was revisited by Saunders (1933). Looking at the vascular traces to the perianth members of monochlamydeous plants, Saunders was able to discern two types, the first with vascular traces characteristic of single organs (sepals) and tending to be green. The other type was one with additional side traces indicating (she thought) that the petals had been incorporated into the sepal, being split between neighbouring sepals. Each perianth member can therefore be represented by the formula 0.5–1–0.5, where the 1 indicates the sepal, and 0.5 half the neighbouring petal on either side. This interpretation was never widely accepted. She interpreted the monochlamydeous perianth of Saraca (Leguminosae: Fig. 1) in this way, as did Hartog (1888), but this was later discounted (Rao and Sirdeshmukh, 1954). Saraca is correctly regarded as having four sepals (even though the sepals are highly coloured). Again, with the development of the concept of organ identity, the problem of petaloidy of sepals is largely removed, without having to regard them as ontogenetic fusion products.
The apetalous plants of the old Monochlamydeae are those of families in which the unipartite perianth is the norm. Of course, it was well known that there were very many anomalous apetalous plants in groups normally with a bipartite perianth. The abundance of other characters associating these plants with their true families meant that no one was misled. Examples of recently derived apetaly within petaloid lineages include both early diverging (non-core) eudicots such as Ranunculaceae [e.g. Thalictrum (Fig. 1), Beesia, and Enemion] and Papaveraceae (e.g. Bocconia and Macleaya), and core eudicots such as Rosaceae (e.g. Alchemilla and Sanguisorba), Primulaceae (Lysimachia maritima), Leguminosae [e.g. Ceratonia, Talbotiella, Swartzia apetala (Fig. 1), and Saraca (Fig. 1)], Brassicaceae (Rorippa in part, e.g. Rorippa sessiliflora; Lepidium in part, e.g. Lepidium ruderale, petals lacking or rudimentary). These anomalies are significant as they are all potentially informative case studies of the mechanisms of petal loss.
Nature: some terms and homology considerations
Homology is well known to be a complicated matter (Rieppel, 2013) and will not be addressed in detail here. At a very basic level, for organs to be homologous, there must be some sort of similarity, in either position, form, or development. This similarity may be obscured by subsequent evolutionary change, but (at some deep level at least) it must be there for a conjecture of homology to be made (primary homology assessment). Next, this similarity must result from evolutionary descent. Operationally, the similarity of topological relations (position and connections) is often regarded as a powerful, often the most powerful, indicator, over and above molecular and developmental similarity. This works well when we attribute positional relations to the calyx (outer whorl) and corolla (inner whorl).
However, when there is only one whorl, then the phenotype (or gene expression underlying the phenotype) is used instead (i.e. petaloidy and petal identity or sepaloidy and sepal identity). It is certainly possible to define sepals and petals by gene expression, but if we wish to use gene expression to ‘explain’ organ categories, we introduce a level of circularity if we also use it to ‘define’ organ categories.
The problem is expressed by Ronse De Craene as follows: ‘The fact that B genes are sometimes expressed in sepals is not an indication of petal nature, but of petaloidy’ [quoted from Ronse De Craene (2010), p. 50; see also Ronse De Craene (2007)]. However, this begs the question of whether a distinction between ‘petal nature’ and ‘petaloidy’ is really secure. The common action of B-class genes is also a similarity derived from common descent. The antennapedia mutant of Drosophila is commonly regarded as having legs on the head, not ped-oidy of the antennae, as it ‘causes the antennae to be replaced by legs’ (Frischer et al., 1986).
The evolution of perianth of differing homologies is particularly wondrous in the Caryophyllales (Brockington et al., 2009). This order is ancestrally monochlamydeous with a sepal-derived perianth. However, during the course of its evolution, numerous lineages have become dichlamydeous by recruiting an extra whorl from bracts or stamens, and it is often a challenge to homologize different perianth parts. An example of the problems can be found in the Spring Beauties (Claytonia: Montiaceae). Claytonia has a pseudocalyx of two bracts (which perform the function of a calyx), five ordinary-looking petals, and five stamens. One obvious conclusion is that the ‘petals’ could be derived from the monochlamydeous perianth (of sepal origin). However, dos Santos and co-authors find temporal and positional evidence for an androecial origin of the petals. They are initiated after the stamens and at the base of the stamens, and—on that evidence—the petals were interpreted as androecial petaloids (dos Santos et al., 2012). Under this interpretation, Claytonia, with its attractive pink or white ‘petals’, is actually without any true perianth. Floras such as the ‘Flora of North America’ consistently refer to these organs as petals, and it should be borne in mind that they might develop their phenotypes (indistinguishable from true petals) using molecular developmental mechanisms that are homologous to those employed by true petals. An alternative possibility remains that rather than the recruitment of novel petaloids, a petaloid calyx whorl has been delayed in development to become suggestive of staminal origin rather than being actually of staminal origin. In some members of the Nyctaginaceae, such as M. jalapa, there is also a pseudocalyx of bracts and a petaloid perianth. However, in this case, there is no doubt that the perianth whorl is equivalent to the calyx (Ronse De Craene and Brockington, 2013). The Claytonia example is illuminating of the difficulties of deriving unequivocal evolutionary scenarios of organ origins from development alone. In such cases, it will be advantageous to analyse developmental evolution within the Chamberlinian paradigm of multiple working hypotheses (Chamberlin, 1931; Elliott and Brook, 2007).
It is also necessary to consider the process by which organs can be lost in evolution and, in particular, the distinction between suppression and deletion. An organ series (such as the petals) may be suppressed (i.e. removed or reduced by a process). This implies that genes are expressed that suppress development, or that genes necessary for the completion of development are not expressed. Very often, suppression is not complete but leaves vestigial traces of the suppressed organs. This is because the selective pressure to remove an organ becomes vanishingly small when the organs are minute. Alternatively, organs may be deleted (i.e. never formed, so there is nothing to suppress). Deletion may arise as a corollary of another process, such as a change of identity genes to remove a given identity or the acceleration of development so that particular primordia never arise, so they do not have to be suppressed. Characteristically, deleted organs have no trace of development (even early development), and no vestigial organs form. These two development types were first clearly expounded and well exemplified by the work of Tucker in Leguminosae. Primordia may be absent, as in the petals of Ateleia (Tucker, 1990), or primordia may form, but the organs may be suppressed at some point after initiation (Tucker, 1992).
Finally, in the molecular developmental literature, it is common to find ‘organ whorl’ used as a near synonym of organ identity, as in ‘four whorls of the flower: sepals, petals, stamens, carpels’. This obscures the fact that these organ identities (especially the stamens) may exist as multiple whorls and may not even be whorled but rather spiralled. In this context, it is better and more precise to use the term ‘organ series’ instead.
Development: the molecular developmental control of apetaly
In monochlamydeous plants it is the petal whorl that has been lost, not the sepals (at least in the overwhelming majority of cases). This raises the question of whether there is some feature of petals that makes this whorl most likely to be lost, and the answer is undoubtedly yes. Petal primordia often delay development, reaching a mature size only late in floral development (Ronse De Craene, 2008, 2018). By contrast, the sepal and stamen whorls develop rapidly, competing for available space with the arrested petal primordia and potentially overwhelming them. It is easy to see how increasing retardation of petal primordium growth could lead to increasing reduction and even loss. The genetic mechanism behind the widespread phenomenon of petal primordium retardation is not known.
Genes capable of suppressing and facilitating petal initiation and growth
In natural variants, population variation for petal number (including apetaly) may be under polygenic control. So in Cardamine hirsuta, petal number variation (from zero to four) is a polygenic trait (Pieper et al., 2016). This is a caution that looking for genes of large effect on petal initiation using mutant screens may be misleading. Lepidium shows a reduction series of petals (Bowman and Smyth, 1998), also apparently under polygenic control. Petal reduction in Lepidium is associated with allopolyploidy (Lee et al., 2002) and loss of the two lateral stamens. The association with allopolyploidy implies that the subfunctionalization of duplicated genes may be involved. Capsella bursa-pastoris has a natural mutant that parallels the variation in Lepidium (Klepikova et al., 2021). These Lepidium-like (lel) plants show earlier calyx initiation and delayed petal initiation compared with the wild type. The lel phenotype in C. bursa-pastoris is controlled by two independent loci, one dominant (gain of function) and one recessive (loss of function).
Nevertheless, despite the polygenic nature of many apetaly phenotypes, there are important single genes affecting petal initiation and growth that mutant screens have discovered and that might be involved in the evolution of this trait. PETAL LOSS (PTL) is a trihelix family gene with major effects in whorl 2 (petals) (Griffith et al., 1999). As well as promoting petals, PTL also inhibits the growth of the sepal whorl, so loss-of-function mutants show sepal fusion as well as petal reduction. It also activates the petal development gene RABBIT EARS (RBE) (Seiji et al., 2004). In this context, it is worth mentioning the extremely interesting, but as yet molecularly uncharacterized, gene lodiculeless spikelet(t) [ld(t)] from rice (Maeng et al., 2006), which cleanly deletes lodicules from rice flowers (lodicules being the grass homologues of petals, i.e. the inner perianth members).
Local auxin maxima are known to control the formation of organ primordia (Lampugnani et al., 2013), and therefore any repatterning of auxin dynamics in the floral meristem could be involved in organ loss. For instance, in the auxin transport mutant PINOID (PID), floral organ initiation is severely disrupted (Bennett et al., 1995).
Dominant gain-of-function mutations in organ-suppressing genes could also potentially cause apetaly. Floral meristems require meristem control genes to keep the size, and organ numbers, of flowers in check. The pluripetala (plp) loss-of-function mutants have larger meristems and increased floral organ numbers (Running et al., 2004). Overexpression of PLURIPETALA is therefore expected to have a suppressive effect. In a genetic analysis of the apetalous Brassica napus line AP01, PLURIPETALA was found to be the major negative quantitative trait transcript associated with apetaly, suggesting that PLP specifically negatively regulates petal development in this line (Yu et al., 2018).
MADS-box genes
At first sight, the floral homeotic ABCDE genes are unpromising vehicles to explain the evolution of apetaly, as these genes are primarily identity genes, and apetaly in monochlamydeous lineages tends to be by whorl 2 suppression rather than altered identity of whorl 2. However, as identity genes they can, and do, cause apetaly in mutants by changing the identity of petal primordia to sepals or stamens. The mutation of the stamenoid petals of C. bursa-pastoris involves a mutation in the AGAMOUS regulatory sequence (Nutt et al., 2006; Ziermann et al., 2009; Hameister et al., 2013), increasing the number of stamens from six to 10 at the expense of petals. Similarly, petals can be transformed into extra sepals (or at least reduced and sepaloid petals) in B. napus by whorl-specific down-regulation of a B-class gene, as a result of expressing a PISTILLATA RNAi construct under an A-class (APETALA1) promoter (Byzova et al., 2004).
In the Ranunculaceae (Duan et al., 2020), it appears that a combination of the inactivation of the petal-specific AP3-3 and an expansion of the AGAMOUS1 (AG1) domain is involved in the evolution of apetaly, with a consequent increase in stamen number (but as stamen numbers in Ranunculaceae are highly variable, it is difficult to be sure). Core eudicot monochlamydeous lineages generally show no signs of such homeotic transformations. Instead, the petal primordia never form, or form but do not develop. It should be borne in mind, though, that floral identity genes have roles outside identity. For instance, AGAMOUS has a floral meristem determinacy function and represses meristem activity. When expressed in whorl 2 (the petal whorl) under the control of the APETALA3 promoter (Jack et al., 1997), the petal whorl fails to develop (or at most weakly develops as one or two organs that are converted into stamens). It should be noted too that there are examples of loss of petaloidy of sepals in connection with regulatory changes in BCE genes (Soza et al., 2016; Martínez-Gómez et al., 2021).
The gene LEAFY (LFY) is a floral meristem gene that can directly activate the transcription of all floral identity genes (A, B, C, and E), achieving domain specificity by acting in concert with domain-specific partners, for instance B genes with UFO/LFY, and C genes with WUSCHEL/LFY (Bowman and Moyroud, 2024). In a fascinating and important study, LFY genes (with their cis-regulatory regions) from various Brassicaceae were used to complement the lfy-6 mutant of Arabidopsis, which does not have petals or stamens (Yoon and Baum, 2004). Idahoa scapigera LFY1 (there are two, but only one was tested) complemented the mutant to the extent of forming stamens, but no petals (or a reduced number) were formed. Conversely, Leavenworthia crassa LFY complemented the mutant better, forming wild-type flowers, except in three lines, which produced extra petals in the petal whorl. The results are suggestive that evolutionary changes at the LFY locus could potentially alter petal number, including producing apetaly. The second LFY homologue from Idahoa was not tested, but it may be that there is subfunctionalization, and LFY2 is a petal specialist and LFY1 is a stamen specialist. The important point is that heterologous complementation did not cause a shift in organ identity but rather the complete suppression, or proliferation, of the petal whorl.
Evolution: the evolutionary ecology of petaly and apetaly—some considerations
In the core eudicots, the bipartite perianth is interpreted as ancestral. Why then should one perianth whorl be lost repeatedly? Many apetalous eudicot lineages have been very successful, such as the Fagales and urticoid Rosales (Fig. 2). Many of these have converted to particular pollination types that render petals either unnecessary or a positive hindrance (i.e. wind pollination, or generalist pollen-feeding insects such as flies and beetles). There are several evolutionary factors to be considered: (i) the function of the corolla whorl; (ii) conversion to wind pollination, which appears to be the main and most obvious driver to apetaly; (iii) conversion to self-pollination or agamospermy; (iv) conversion to alternative insect pollinators; and (v) transference of function of attractiveness to other organs, primarily bracts, sepals, or stamens.
There are a multitude of different functions of the petal in floral biology, but attraction is obviously a major one. Stellaria is a genus with both petaliferous (petal-bearing) and apetalous species, and Lovell (1901) artificially removed the petals of Stellaria media with the following result: ‘I found on trial that a flower [of Stellaria media] could be distinctly seen at a distance of twenty-five feet; but after removing the petals it was visible only about four feet.’ The petals of S. media are rather modest (certainly not large or showy), but the 6-fold increase in visibility to a vertebrate visual system is indicative that there is likely to be a major impact on insect visitation, although that remains to be tested.
Depetalling studies
There is a very long history in the field of pollination biology of experimental flower mutilation (including petal removal). The results are surprisingly inconsistent. Sometimes depetalling (i.e. the experimental removal or reduction of petals) seems to make no difference to the pollinator visitation rate, or even increases it. In other experiments, the expected reduction in pollinators, or seed set, is noted, as when the petals of Rosa acicularis were experimentally shortened (Clements and Long, 1923). The unmutilated flower over 1 h received 42 visits from six species of bees, whereas the flower with shortened petals received 13 visits from four species of bees during the same amount of time. Although unreplicated, the observation is definitely suggestive, but both visitation rates are likely to allow adequate pollination. A more detailed study of petal reduction in Hibiscus moscheutos (Kudoh and Whigham, 1998) reported little difference in reproductive success after 50% petal reduction (although there were some marked differences on 100% reduction). Interestingly, a non-pollinating pollen-feeding beetle (that was also a larval seed predator) preferred flowers with unreduced petals, apparently as petalled flowers provided a better living space. On these results, it is hard to see why flowers with smaller petals have not been selected for.
Extensive experiments on poppies (Papaver) were performed by Emil Giltay of Wageningen and Félix Plateau of Brussels in the early 1900s, who disagreed (somewhat vigorously) about the interpretation. Results from depetalled Papaver rhoeas (Giltay, 1904) showed a decrease in seed set: ‘For 215 flowers from which the corolla was removed, these [seed samples] weighed 10.770 grams … for an equal number of normal flowers, however, they weighed 25.230 grams … As a control, artificial pollination was later carried out on 28 flowers from which the corolla had been removed. These produced almost as many seeds as the normal ones, namely 0.115 grams per fruit, which shows that the lower quantity in the decorollated flowers is not a result of the injury that occurred.’ As P. rhoeas is obligately self-incompatible, Giltay, perfectly reasonably, attributed decreased seed set in depetalled flowers to a lack of insects attracted to cross-pollinate. Working with Papaver orientale, Plateau (1902) agreed that there was a decrease in seed production in the depetalled plants, but disagreed that this had anything to do with attraction. Plateau measured insect visitation and found that there were twice as many visitors to the depetalled flowers, perhaps because the androecium is very conspicuous and also because the androecium is the direct source of the pollen reward for bees. He therefore attributed the decrease to a change in the way the insects worked the flowers, writing: ‘the hymenopteran that goes to a decorollated flower does not land on the stigmas. Without the support provided by the petals, it flies directly to the male organs and immediately hangs on one of the stamens, which, then weighed down by the insect, assumes a vertically descending position, the anther hanging lower than the ovary’.
This raises the point that attraction is only one of the functions of the corolla: it is also part of the floral biomechanics and plays a part in controlling the behaviour of the insect within the flower. A key term here is synorganization (Endress, 2016), defined as ‘the provision of a novel or more efficient function by different plant organs working in concert’ (Cronk and Borges, 2024). The greater the synorganization of the flower, the greater the functional change in the pollination mechanism on petal loss. It is important to distinguish between the role of the corolla in attraction to pollinators and its role in the effectiveness of pollination. Thus, while the naked androecium was more attractive than the corolla in Plateau’s poppy, the corolla’s role in determining bee behaviour at the flower was important in optimizing pollination. Of course, it is possible that Giltay and Plateau were both right. Under conditions of high pollinator density, when pollinators do not have to be attracted from a large distance, the conspicuous androecium is an adequate and effective short-distance signal for bees. However, under circumstances of pollinator scarcity, when bees must be attracted from long distances, a colourful (red) corolla may then be more important than the androecium in attracting pollinators (but it should be recalled that insects have no red receptor in their visual system). This is consistent with a study on altitudinal variation in Trollius ranunculoides (Ye et al., 2011). This plant has showy petaloid sepals and small, relatively inconspicuous nectarial petals. At low elevations where pollinators are abundant, relatively more resources are partitioned to petals (i.e. nectar reward), whereas at high elevations where pollinators are scarce, resource allocation is increased to the showy sepals. Bees have good learning abilities and can use many visual cues in foraging, not just petals, although conspicuous petals may be important initially. Balfour and Ratnieks (2023) removed the ‘petals’ (ray corolla lobes) from the showy capitula of Rudbeckia hirta and Helenium autumnale (Asteraceae) and found that experienced bee pollinators showed no detectable preference for the intact capitula. However, naive bees (first-time visitors) preferentially visited the intact capitula. The authors conclude that the main function of the showy ray florets is to attract these naïve first-timers.
As pollinator visitation rates and pollinator behaviour are hard to measure, but seed set is relatively straightforward to measure, the latter is often used as a proxy measure for some aspect of pollination, as in Giltay’s study. This alone is not able to disentangle various possibilities, as Plateau pointed out. In an interesting study of Halimium atriplicifolium (Teixido, 2014), it was found that depetalled flowers set more seeds but, among petalled flowers, those with larger petals tended to set more seeds. These results were attributed to a trade-off between the indirect costs of maintaining petals against the increased attractiveness of large petals. However, the possibility of greater pollinator visitation to the depetalled flowers (as in P. orientale) was not addressed. Direct observations of pollinator behaviour are vital. For instance, in the fungus-gnat-pollinated Mitella pauciflora (Katsuhara et al., 2017), fruit set and pollen dispatch both significantly decreased with petal removal. However, this was not due to a failure to attract pollinators, as petal removal did not change visitation frequency or duration but instead decreased pollinator landings. In this case, the petals are not functioning for attraction but as landing pads. The diversity of petal function is very great. As another example, Corbett et al. (1992) show that petal removal in the arctic poppy Papaver radicatum reduces ovary temperature and seed production, highlighting the role of petals in increasing irradiance and heat capture at the ovary, so affecting reproductive success.
Mutants lacking petals
It is possible to study the fitness and pollination of apetalous crop varieties or naturally occurring mutants in the wild. Such studies can be very informative, but the pollination and floral biology of these systems are rarely studied rigorously, although there are some exceptions. Particular interest attaches to the ‘stamenoid petals’ (Spe) mutation of C. bursa-pastoris, a mutation in the AGAMOUS regulatory sequence (Nutt et al., 2006; Ziermann et al., 2009; Hameister et al., 2013). This mutant successfully persists in the wild and has been suggested as a possible incipient species, ‘Capsella apetala’ (Hameister and Neuffer, 2017) or a ‘hopeful monster’ caught in the act (Hintz et al., 2006). The persistence of Spe is partly aided by the high degree of selfing in C. bursa-pastoris, so insect attraction by petals is less important.
Nevertheless, a small amount of outcrossing can be highly significant to the long-term fitness of primarily inbreeding species, and the petaloid wild type does attract twice as many potential pollinators (Ziermann et al., 2009). The flower visitors include various Hymenoptera, Diptera, and Coleoptera, with bees (Apidae) and hoverflies (Syrphidae) particularly common. In one study, 43 visitors were collected from the wild type, but only half as many (22) from the Spe mutant. The type of pollinator apparently most responsive to petals was bees, 21 to six, versus the total insect visitation of 43 to 22 (Hameister and Neuffer, 2017). Interestingly, pleiotropic effects may be most important in the persistence of Spe plants in mixed populations, as Spe tends to flower later than the wild type, pushing it into a different phenological niche (Neuffer et al., 2020).
A separate apetalous mutation of C.bursa-pastoris (the Lepidium-like, lel, mutant) has also been observed (Klepikova et al., 2021), but pollinator visitation has not been reported. This phenotype is controlled by two, as yet unidentified, independent loci, and is particularly interesting as it appears to mirror the phenotype of the apetalous species of Lepidium.
A number of apetalous mutants are known in the genus Brassica, for example in turnip (Brassica rapa) (Ramanujam, 1940). Apetalous rapeseed (B. napus) is an agriculturally important variant, possibly resulting from a pathway that leads to direct suppression of petals by the gene PLURIPETALA (PLP) (Yu et al., 2016, 2018). Despite the lack of petals, it has high seed yield and the agriculturally important trait of high klendusity to Sclerotinia sclerotiorum (a disease partly dispersed by falling petals). Pierre et al. (1996) studied the pollination of petaliferous and apetalous B. napus. They found honeybee density to be higher on the apetalous plants and, interestingly, honeybees showed foraging constancy, tending to keep to one type only, restricting cross-breeding between petalled and apetalous types. An important role of the petals is to determine pollinator movements on the flower. Honeybees on petalled flowers use the platform provided by the petal limbs, thus contacting the anthers and stigma. In contrast, on apetalous flowers, nectar was often foraged from between the sepals, leading to less contact with the reproductive organs.
Transference of function: taking over of visual attractiveness by bracts, sepals, or showy stamens
Many Ranunculaceae are apetalous (e.g. Caltha, Thalictrum, and the tribe Anemoneae, such as Clematis and Anemone). Here, the attractive functions normally held by petals are taken over by showy sepals. Thalictrum (meadow rues) are a good example. Ancestrally, they are insect pollinated with showy white (Fig. 1) or purple sepals (Martínez-Gómez et al., 2023). Subsequently, the petaloidy (colour, size, and conical epidermal cells) of the sepals was lost on transition to wind pollination. Here, apetaly preceded wind pollination, so wind pollination was not a driver of apetaly. Subsequently, insect pollination re-evolved in some species, mainly associated with showy stamens with the filaments coloured, flattened, or both (Di Stilio et al., 2009). This parallel evolution of apetaly in Ranunculaceae has also been studied at the molecular level in other groups (Duan et al., 2020).
A similar ‘transference of function’ (Corner, 1958) from petals to sepals has occurred in the legume genus Saraca, the species of which are apetalous but have brightly coloured sepals. In Saraca asoca, the tubular calyx holds nectar and appears to be an adaptation to butterfly pollination, although it is also attractive to the Giant Asian Honeybee (Apis dorsata) (Smitha and Thondaiman, 2016). Some apetalous legumes have retained the ancestral bee pollination by converting the attractive function to stamens, a direct signal as the reward in such flowers is pollen only. Swartzia apetala (Moço and Pinheiro, 1999) is one such example. The visual display is produced by 80 short stamens arranged in a brush (the species is heterantherous, also having two long stamens). The staminal brush is a sufficient cue to attract a variety of tropical bees in the genera Centris and Xylocopa to visit for pollen collection by vibration, making it a rare case of buzz pollination in which the anthers are not poricidal but longitudinal in dehiscence. The absence of petals might increase the efficiency of the vibrational collection of pollen by reducing damping by corolla.
In some examples, the transference of function of petaloidy is to bracts. An example is Chrysosplenium alternifolium (golden saxifrage). Here, the apetalous flowers have inconspicuous calyces but conspicuous yellow floral bracts (Ronse De Craene et al., 1998). These bracts, together with the exposed nectar, are sufficient to attract a wide variety of small insects as flower visitors (from Coleoptera, Diptera, and Hymenoptera). One of the most spectacular examples of petaloid bracts is the handkerchief tree, Davidia involucrata (Sun et al., 2008; Vekemans et al., 2012), the aggregated flowers of which lack any perianth but have large, insect-attracting, white, UV-absorbing bracts. These bracts, besides attracting insects, also function as rain umbrellas.
This raises the problem of why, when petals have evolved for petal function, have petal functions been transferred so frequently to other organs? In this connection, we might note that the transfer of the attractive function to stamens in plants where pollen is the reward is what, in signalling theory, is called an honest signal (Kooi et al., 2023). An honest signal is one that provides reliable information about the quality of the sender and, in the case of the androecium, it signals both the quantity and state of the pollen reward. In the case of the sepal, petaloidy can potentially be conferred readily by partial homeosis if a single perianth whorl is all that a floral meristem can sustain through available space. With bracts, the reinvention of petaloidy through bracts, involving flowers with reduced perianth, is common in pseudanthial inflorescences (Baczynski and Classen-Bockhoff, 2023). It is, indeed, the only possibility when flowers are perianthless, as in Davidia involucrata and Euphorbia pulcherrima.
Apetaly associated with a transition of pollination mode
Many groups of plants have transitioned to wind pollination from ancestral biotic pollination, and many of these are apetalous (Culley et al., 2002). The Fagales (Juglandaceae, Myricaceae, Betulaceae, Casuarinaceae, and Fagaceae) are apetalous and largely wind pollinated. The same is true of most of the Rosales (Fig. 2), in which the apetalous urticoid Rosales (Ulmaceae, Cannabaceae, Urticaceae, and Moraceae; Fig. 3) are mainly wind pollinated (Ficus being a major exception). Small or absent petals are part of the wind pollination syndrome (Friedman and Barrett, 2009). In abiotic pollination, attraction and guidance of insects are not necessary, so the primary function of petals is removed. In addition, a large perianth may be aerodynamically unfavourable to passive pollen dispersal and capture. However, it remains to be determined how often apetaly precedes wind pollination rather than vice versa. In some cases where we have phylogenetic evidence, apetaly actually precedes wind pollination, and wind pollination evolves from an apetalous insect-pollinated ancestor, as in Thalictrum (Martínez-Gómez et al., 2023). In such cases, apetaly facilitates wind pollination, not the other way round.
As self-pollination does not require the attraction of insects, the selfing syndrome sometimes involves apetaly. For instance, the evolution of apetaly has been ‘rampant’ in the genus Stellaria, and in that group it is associated with selfing (Sharples et al., 2021). The evolution of small flowers with reduced petals (for instance in small ephemerals) will have the effect of tending to reduce herkogamy and thereby increase selfing. If selfing becomes dominant, loss-of-function mutations in the petal development pathway will not be selected against, and petals may disappear through drift. The violets (Viola ssp.) are well known for their obligately selfing and apetalous cleistogamous flowers, produced late in the season after their earlier, large-petalled chasmogamous flowers. Sometimes, a whole series of flowers showing progressive petal loss can be found. Viola canadensis is slow to produce cleistogamous flowers in the summer, first producing intermediates. The transition from petaliferous to apetalous flowers is therefore a rather gradual one: ‘the petals become gradually smaller and finally disappear completely, one or several at a time’ (Nieuwland and Kaczmarek, 1914).
In petaliferous, insect-pollinated plants, apetaly may be associated with a change of pollinator. Whereas bees, moths, and butterflies are strongly attracted by visual signals and are usually associated with petaliferous plants, beetles, ants, and flies are often attracted by other cues. A remarkable case of apetaly-driven reinforcement is provided by Ranunculus eschscholtzii (Fisher et al., 1979) which exists in two forms, apetalous (ant pollinated and rare) and petaliferous (syrphid pollinated and common). Where it grows together with the more abundant petaliferous R. suksdorfii, it is apetalous. Any petaliferous individuals will be overwhelmed by syrphid-mediated pollen gene flow from R. suksdorfii, resulting in sterile progeny (the syrphids ignore the apetalous plants). This is an elegant case of reproductive character displacement (RCD)—trait evolution selected to reduce negative reproductive interactions between species (Pfennig and Pfennig, 2009).
The B-switch hypothesis of the unisexual flower and its relationship to apetaly
Examination of Table 2 reveals that there are large numbers of plants that are monochlamydeous but also have unisexual flowers (monoecious and dioecious). The statistical strength of the association would be hard to test, but might be illuminating. The association, if genuine, might be through independent but correlated adaptation to wind pollination, as many wind-pollinated lineages (Fagales, for instance) are monoecious or dioecious, and (as discussed above) wind pollination has no need of an attractive corolla, which may hinder wind pollination. However, there is also an idea that simplifying floral morphology by losing a perianth whorl might make it developmentally easier to switch other perianth whorls (stamens and carpels) on and off, as required for unisexual flowers. This suggests that a unipartite perianth might be a gateway to the unisexual flower. An example of a possible mechanism is the ‘B-switch hypothesis’.
The B-switch hypothesis is the suggestion that the natural overexpression of B-class MADS-box gene expression leads to male flowers, but when B-class gene expression is switched off, flowers are female (Fig. 4). Control of sex therefore works directly by toggling these B-class genes (PI and AP3) on and off (Cronk and Müller, 2020). B-class MADS-box genes are important identity genes for second and third whorl organs (petals and stamens). Their role in stamen identity makes them candidates for involvement at some level in floral sex. However, their additional role in petal identity would have an additional impact on the petal whorl, unless, of course, the plant was naturally apetalous.
Just finding differences in B-class gene expression between male and female flowers is of no use as this may be purely a collateral effect; without stamens or petals, B-class genes have no organs to be expressed in. However, there is emerging evidence that, at least in some cases, it is indeed causal, and regulation of B-class genes directly determines floral sex (the B-switch hypothesis).
Studies on the apetalous plants meadow rue (Thalictrum dioicum) and spinach (Spinacia oleracea) played a crucial role in developing the B-switch hypothesis. Spinach is a dioecious species where male and female flowers develop unisexually from inception, not by organ abortion. This means that male and female flowers differ from the beginning in terms of in which organs they initiate. The spinach homologues of APETAL3 (SpAP3) and PISTILLATA (SpPI) are expressed in early stages throughout the male floral meristem, but they were not detectable in female flowers (Pfent et al., 2005). These authors observed that B-class gene expression precedes the development of stamens in males, while SpPI is not expressed in female flowers at any stage. SpAP3 is expressed at only low levels in female flowers. It therefore appeared that the regulation of B-class genes might have a role in the development of sexual dimorphism and dioecy in spinach. A later study (Sather et al., 2010) conducted a functional analysis of SpPI, SpAP3, and a C-class gene, SpAG. The authors demonstrated that the suppression of B-class gene expression in male spinach plants resulted in the development of female flowers, which was not simply a result of homeosis, because the number and whorl location of the organs also changed in addition to organ identity. This result provided key evidence that the regulation of B-class genes is a major control point in sex determination in spinach. The study suggested that the feminizing mutation in spinach involves the suppression of B-class gene expression. The authors also proposed that the masculinizing mutation regulates the termination of the flower in the third whorl and that this mutation occurred in the spinach B-class genes or just downstream.
At the same time, research on T. dioicum (Di Stilio et al., 2005), also a dioecious species with unisexual flowers from inception, likewise supported a role for B-class genes. It was observed that some B-class gene paralogues in Thalictrum are largely male-specific, while others are not expressed at all in females or express male and female-specific alleles. This suggested a homeotic mechanism for sex determination via differential regulation of B- and C-class floral organ identity genes. Morphology and cell-level differences in male and female perianth were also considered, with female perianth appearing smaller and more sepaloid than male perianth. Further studies (LaRue et al., 2013) used virus-induced gene silencing (VIGS) of the B-class gene PISTILLATA (ThdPI) in T. dioicum. It was found that silencing ThdPI in male plants caused stamen primordia to develop into carpels, resulting in a male-to-female flower conversion. In a hermaphroditic Thalictrum species (T. thalictroides), silencing of ThtPI resulted in stamen primordia developing into supernumerary carpels, resulting in a hermaphroditic to female flower conversion. These results constituted telling experimental evidence for the B-switch hypothesis and that B-class genes have a direct role in sexual system transitions. These experiments indicated that the evolution of female plants could result from a loss-of-function mutation in a B-class gene. They proposed a two-step evolutionary model for the transition from hermaphroditic to unisexual plants in Thalictrum involving two independent mutations at a B-class gene locus.
Cronk and Müller (2020) suggested a B-switch hypothesis for poplar and aspen trees (Populus), linking it to the concept of a ‘default sex.’ Populus, like Spinacia and Thalictrum, has apetalous flowers and organ suppression from inception with a reduced disc-like single perianth whorl (calyx). The authors suggested that, in poplar, the default sex is male with B-class gene (PI/AP3) expression blocking carpel formation. A feminizing factor then acts as a ‘master regulator’, converting male flowers to female by blocking the expression of PI and AP3, inhibiting stamens and instead allowing carpels to form (Fig. 4). This is consistent with expression studies showing a strongly male-biased expression of AP3 and PI in poplar (Cronk et al., 2020). Müller et al. (2020) experimentally confirmed the gene popARR17 as the feminizing factor in Populus. The CRISPR/Cas9 arr17 knockout in female poplars resulted in the conversion of plants from female to male, with the development of functional male flowers. This allowed a comparison of gene expression differences in an identical genetic background (Leite et al., 2018). This study suggested that popARR17 represses the expression of UFO, which in turn is necessary for the activation of the B-class genes needed for male floral organ development, pointing to popARR17 triggering female development by repressing the UFO–PI cascade.
Further examples are discussed in another review (Zhang et al., 2022). The B-switch mechanism has also been put forward (Yang et al., 2019) for the florally very different species, persimmon (Diospyros), which has a bipartite perianth and organ suppression by abortion. Here, the feminizing gene MeGI has been suggested to negatively regulate the B-class gene PI through a putative regulatory cascade, MeGI–SVP–PI, with MeGI directly controlling the expression of SVP, which is an upstream repressor of B- and C-class genes. This suggests a somewhat different developmental process as the negative regulation of B-class genes does not seem to affect the inner perianth whorl (although interestingly there are some minor morphological differences between male and female inner perianth). In Diospyros, like other plants with unisexual flowers and a bipartite perianth, there is a problem that controlling sex through B genes will have a major effect on the inner perianth.
However, one can see that the absence of an inner perianth whorl (apetaly) conveniently solves that problem, as there are no petals to be affected when the B-class genes developmentally function as a switch to control the presence or absence of stamens. In solving this problem so conveniently, apetaly can be seen as a facilitator and gateway to unisexual flowers by the B-switch mechanism.
Conclusions and future work
It should be pointed out that many discussions of the ‘floral ground plan’ are actually discussions of a four-whorl ground plan, regardless of the fact that entire large clades of eudicots, the monochlamydeous families, have a three-whorl ground plan. Work on the three-whorl ground plan is consequently a Cinderella discipline in need of further study. It would be unwise to consider the three-whorl ground plan as a minor variant of the four-whorl type, as many of the lineages have been stably three-whorl for 80 million years or more (such as the ‘urticoid Rosales’ and the Fagales). They are therefore likely to have acquired what is in essence a distinctly different organization. The urticoid Rosales (Ulmaceae, Cannabaceae, Urticaceae, and Moraceae: Fig. 2) are an interesting example as they have never reverted to a bipartite perianth and have been highly successful.
This is a good time to start, as there are excellent genome resources for many monochlamydeous plants. Some are trees (e.g. Morus and Populus) which are challenging to work on because of their long life cycle, but many are short life cycle herbaceous species and crop plants, and among these can be mentioned Fagopyrum (buckwheat: Polygonaceae), Cannabis (Cannabaceae), Urtica (nettle: Urticaceae), Portulaca (Portulacaceae), and many from the Amaranthaceae (Amaranthus: amaranth, Beta: beet, Chenopodium: quinoa, Spinacia: spinach). Furthermore, in the Ranunculales (which include numerous monochlamydeous examples) there is a concerted sequencing effort (RanOmics group et al., 2024).
It may be argued that such plants still have a longer life cycle than Arabidopsis and lack the ease of transformation of Arabidopsis. The transformation problem may or may not be solved in the future, but it is worth noting on the subject of life cycle that Oxybasis (Chenopodium) rubra can be brought to flower in 21 d (Cumming, 1959), on account of it being phenomenally sensitive to florally inductive photoperiod, even at the cotyledon stage, permitting 100 seedlings to be flowered in a Petri dish. The seed yield of such micro plants is minimal, but if larger quantities of seed are required, a normally grown plant will produce ∼70 000 seeds (Williams, 1969).
Acknowledgements
I thank Niels Müller for many fruitful discussions of dioecy.
Funding
I thank the Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grants program for the award RGPIN-2019-04041.