A function for the pleurogram in physically dormant seeds

Abstract

Background and Aims

Different structures have been shown to act as a water gap in seeds with physical dormancy (PY), and in Fabaceae they are commonly located in the hilar region. However, the function of the pleurogram, a structure in the extra-hilar region that is common in legume seeds, remains unknown. Our aims were to review the literature for occurrence of the pleurogram in Fabaceae, determine if the pleurogram can open, and compare the functional morpho-anatomy of water gaps in seeds of 11 Senna species.

Methods

Imbibition tests showed that all 11 species had PY. Structural features of the hilar and extra-hilar regions of the seeds were investigated using light and scanning electron microscopy, and dye-tracking was performed to trace the pathways of water through the seed coat.

Key Results

A pleurogram has been reported for 37 legume genera. Water gaps differed among Senna species, with lens, hilum, micropyle and pleurogram taking up water after PY was broken. In Senna alata seeds, only the pleurogram acted as a water gap, whereas in S. reniformis and S. silvestris water entered the seed through both the pleurogram and the hilar region. In the pleurogram of S. alata and S. reniformis, the palisade layer moved outward, exposing the hourglass cells, whereas in S. silvestris the palisade layer was broken.

Conclusions

The pleurogram acts as a water gap in some of the 11 Senna species examined, but it is non-functional in others. Opening the pleurogram occurs due to the formation of a linear slit in the palisade layer. The pleurogram is of functional significance by creating a wide opening, whereby water can reach the embryo and start germination. This is the first report of the pleurogram functioning as a water gap. Because this structure is shared by at least 37 genera, it also may be a water gap in many other legume species.

INTRODUCTION

Seeds with physical dormancy (PY, water-impermeable) have a specialized structure (the ‘water gap’) that opens during dormancy break, thereby allowing water to enter the seed (Baskin et al., 2000; Baskin, 2003). In seeds of Fabaceae, structures such as the hilum, lens and micropyle, or a combination thereof, function as the water gap, depending on the species (Gama-Arachchige et al., 2013; Geneve et al., 2018). Gama-Arachchige et al. (2011) referred to the region of a physically dormant seed where water enters as the ‘water-gap complex’, and they identified three types of water-gap complexes. Type-I has a linear opening obstructed by modified palisade cells, Type-II has an opening obstructed by a lid-like structure of palisade cells and Type-III has an opening occluded by a plug-like structure of sclerenchyma cells. These three types were further classified as ‘simple’ if one opening is formed and ‘compound’ if more than one is formed (Gama-Arachchige et al., 2013). However, in some species of Fabaceae the exact site where water enters the seed has not been identified, and Morrison et al. (1998) suggested that the water gap may not be located in the hilar region in seeds of some of them.

In seeds of some members of the Fabaceae subfamilies Caesalpinioideae and Mimosoideae [in the new subfamily classification Mimosoideae is within Caesalpinioideae (LPWG, 2017)], a visibly demarcated structure called the pleurogram is present in the lateral region of the seed coat (Corner, 1976; Werker, 1997) (see Fig. 1). The pleurogram is caused by the difference in cell height along the coat, which forms an accurate line demarcating this structure in seeds (Corner, 1951). The pleurogram is noticeable even in immature seeds. In addition to reports for these subfamilies of Fabaceae, Corner (1976) reported a pleurogram for seeds of Trichosanthes anguina (Cucurbitaceae), but the occurrence of a pleurogram in this family is not yet clear. In Fabaceae, a pleurogram is more common in seeds of Mimosoideae than in those of Caesalpinioideae (Gunn, 1981). It is externally visible on both sides of the seed and can be distinct from the rest of the seed coat (Corner, 1976; Werker, 1997). In some cases, more than one pleurogram occurs on each side of the seed, as detailed for Chamaecrista species (Caesalpinioideae) by De-Paula and Oliveira (2008). The morphology of the pleurogram varies in the Mimosoideae and Caesalpinioideae, and in some species of both subfamilies it covers most of the seed coat (Gunn, 1981).

Fig. 1.

The pleurogram in seeds of Senna species: (A) S. occidentalis, (B) S. spectabilis, (C) S. obtusifolia.

Gunn (1981) hypothesized that the pleurogram acts as a hygroscopic valve, similar to that of the hilum (Hyde, 1954), due to the presence of a fissure similar to the hilar groove. The opening and closing movements of the hilum that are driven by the humidity surrounding the seed allow water loss during seed maturation (Hyde, 1954). According to Gunn’s hypothesis, the pleurogram could also be involved in seed drying, a fundamental condition for acquisition of PY in seeds. This assumption is also made in other studies (Kelly et al., 1992; Morrison et al., 1998). However, there is no experimental evidence to support the ‘hygroscopic valve’ hypothesis for the pleurogram. In legumes, certain species have overgrown seeds (those whose coat is poorly differentiated) that lack a pleurogram (Corner, 1951; Maumont, 1993). These seeds are not water-impermeable (compared to impermeable pleurogrammic seeds), and Maumont (1993) considered them to be an adaptive response to wet environments in the tropics. Indeed, the association between the pleurogram and seed drying could be indicative of the ecological significance of this structure, but a detailed investigation is needed to prove it.

Furthermore, Rodrigues-Junior et al. (2014) confirmed that although there are multiple fissures in the pleurogram of Senna multijuga seeds they were not deep enough to allow water to penetrate the water-impermeable palisade layer of cells. Kelly et al. (1992) and Werker (1997) showed a wide fissure in the pleurogram region of seeds of the legumes Dichrostachys cinerea and Prosopis farcta (both Caesalpinioideae); however, they did not determine if the fissure was deep enough to allow water to enter the seed. The presence of fissures in the pleurogram of D. cinerea, P. farcta and S. multijuga seeds led us to propose the general hypothesis that the pleurogram, at least in some species, may serve as a water gap. In fact, if the two hypotheses (‘hygrocopic valve’ and ‘water gap’ hypotheses) described for pleurogram function are true, this structure is involved in both onset and release of PY. However, with the exception of S. multijuga, the function (or lack thereof) of the pleurogram has not been investigated.

Senna (Fabaceae, Caesalpinioideae, Cassiinae) is a large and diverse genus of approx. 350 species that occur in various habitats throughout the world. It is one of the largest genera in Fabaceae and includes herbs, shrubs and trees (Irwin and Barneby, 1982; Randell and Barlow, 1998; Marazzi and Sanderson, 2010). All Senna species whose dormancy has been investigated have PY (De Paula et al., 2012; Baskin and Baskin, 2014; Rodrigues-Junior et al., 2014; Erickson et al., 2016), and the vast majority of them produce seeds with a pleurogram (Irwin and Barneby, 1982). The overlap of seed morphology with the phylogeny of Senna (Marazzi and Sanderson, 2010) suggests that the evolution of a pleurogram occurred multiple times in this group, with the presence/absence of this structure occurring in distinct clades in Senna. From an evolutionary point of view, this remarkable structure was described as an ‘archaic feature’ in seeds of Senna by Irwin and Barneby (1982), being suppressed in some series within this genus. The pleurogram of Caesalpinioideae including that of Senna is classified as closed (Gunn, 1981), i.e. the line demarcating the seed coat is complete.

The lens is the only seed structure described as a water gap in Senna (De Paula et al., 2012; Rodrigues-Junior et al., 2014; Erickson et al., 2016). However, this genus has great seed diversity, varying widely in hilum and pleurogram morphology and in some cases with a wide pleurogram bordering the hilar region, as in Senna alata seeds. Thus, Senna is an appropriate genus with which to test our hypothesis that the pleurogram can function as a water gap in some species. Because information on occurrence of the pleurogram is widely scattered in the literature, we start by reviewing all published studies on the pleurogram. We then used seeds of 11 Senna species from Brazil as the study material to explore the function of the pleurogram in these species. Thus, we addressed the following questions: (1) What is currently known about the occurrence of the pleurogram in seeds of Fabaceae species? (2) Does the pleurogram function as a water gap in some species? (3) If so, does pleurogram structure differ between species in which it functions as a water gap and in those for which it does not? (4) Are there morpho-anatomical differences between the pleurogram and other sites of water entry in the hilar region?

MATERIALS AND METHODS

Review on occurrence of the pleurogram in Fabaceae

Literature data were retrieved from studies on legumes in which seed morphology was described in detail. We searched for reports of the pleurogram in legume species included in the conventional subfamilies Caesalpinioideae and Mimosoideae. The data for pleurogram occurrence were kept separate for subfamilies and genera in Fabaceae.

Plant species

Seeds of 11 Senna species were collected (Table 1) in various biomes in Brazil; nine of them are shrubs, and two are trees (see Supplementary Data Table S1). All species are native to Brazil, and Senna cana var. hypoleuca, S. reniformis and S. trachypus are narrow endemics (Souza and Bortoluzzi, 2015). These species produce seeds with distinct morphology among them, especially for the pleurogram (or even lacking it), which allows a broad investigation on the role of this structure in seeds. Seeds were placed in semi-permeable plastic bags and stored under laboratory conditions (25 ± 5 °C, 40-60% relative humidity) until used. Predated seeds were discarded before initiation of experiments.

Table 1.

PY-breaking treatment, features of the pleurogram, description of the hilar water gaps and water gap type in the 11 Senna species used in this study

Species	PY-breaking treatment*	Pleurogram			Water gap in hilar region	Water gap type†
		Features	Contrast	Functional
S. alata	HW or SA	Oblong, large, on the margins	Texture, colour	Yes	None	II-simple
S. cana	HW	Oblong, small, on lateral sides	Colour	No	Lens	I-simple
S. cernua	BW	Oblong, large, on the margins	Texture	No	Lens/hilum	I-compound
S. hirsuta	SA	Rounded, small, on lateral sides	Texture	No	Lens	I-simple
S. obtusifolia	BW	Oblong, large, on lateral sides	Texture	No	Lens	I-simple
S. occidentalis	HW	Oblong, large, on lateral sides	Texture	No	Lens	I-simple
S. pendula	HW	Absent	None	None	Lens	I-simple
S. reniformis	HW or SA	Oblong, large, on lateral sides	Texture	Yes	Lens	I/II-compound
S. silvestris	SA	Oblong, large, on lateral sides	Texture	Yes	Hilum/micropyle	I-compound
S. spectabilis	SA	Oblong, small, on lateral sides	Texture	No	Hilum	I-simple
S. trachypus	SA	Oblong, small, on the lateral sides	Texture	No	Hilum/micropyle	I-compound

* BW, boiling water (10 s); HW, hot water (80 °C/15 min); SA, sulphuric acid (30 min).

^† According to Gama-Arachchige et al. (2013).

Presence of physical dormancy

To determine if seeds had PY, 25 individual intact or manually scarified seeds of each species were placed in Petri dishes on germination paper moistened with distilled water and incubated at 25 °C (light/dark, 12/12 h). Light was provided by cool white fluorescent tubes (40 µmol m⁻² s⁻¹). Seeds were individually weighed at different intervals for 48 h with a Shimadzu AUX220 precision balance. Prior to each weighing, seeds were removed from Petri dishes and blotted dry, weighed and then returned to the moistened paper. Changes in seed mass were calculated.

Dormancy break

Preliminary tests were performed prior to this experiment (Supplementary Data Fig. S1), and treatments that increased germination to the highest percentage were used for each species (Table 1). The following three treatments were effective in breaking PY: immersion in hot water (80 °C) for 15 min; immersion in boiling water for 10 s; and immersion in sulphuric acid (98 %) for 30 min. For S. alata and S. reniformis seeds, both hot water and sulphuric acid were effective in breaking dormancy (Table 1 and Supplementary Data Fig. S1). There were four replicates of 25 seeds for each species. Treated and non-treated (control) seeds of each species were placed in Petri dishes on moistened germination paper and incubated at 25 °C (light/ dark, 12/12 h). Germinated seeds were scored at 3-d intervals for 30 d, and the criterion for germination was emergence of the radicle.

Structural changes during dormancy break

Dormant and non-dormant seeds [made water-permeable according to the best method for each species (see Table 1)] were mounted directly on stubs using double-sided carbon tape and sputter-coated with gold (10 nm) (Robards, 1978). Fifteen seeds per species were used. Samples were then observed with a scanning electron microscope (Quanta 200; FEI). The whole seed was analysed, and the disrupted structures were described.

Dye tracking

Seeds were made water-permeable and then immersed in 0.1 % methylene blue (modified from Johansen, 1940) for 15 and 30 min and 1, 3 and 6 h. Fifteen seeds were evaluated for each immersion period, after which they were rinsed with tap water, blotted dry and sectioned longitudinally and transversely to observe the presence of the dye and its route in the seed tissues. Seeds were observed under a stereomicroscope (Zeiss Stemi 2000-C) and pictures were taken with a digital camera (Canon A650 IS). This procedure was carried out to investigate if the intact seed structures that changed during dormancy break create a pathway for water entry into the seed.

Anatomy of the pleurogram and water-gap complexes

To investigate the structure of the pleurogram and compare it with hilar water gaps, 15 seeds of each species were made water-permeable and then fixed with a formalin/acetic acid/50 % ethanol solution for 48–h, dehydrated in a graded ethanol series and embedded in (2-hydroxyethyl)-methacrylate (Paiva et al., 2011). Then, seed material was sectioned (8 µm) longitudinally and transversally in the hilar and pleurogram regions using a microtome (Zeiss Hyrax M40), stained with 0.05 % toluidine blue, pH 4.7 (modified from O’Brien et al., 1964), and mounted in synthetic resin. Sections were observed with an optical microscope (Leica DM500) and photomicrographs were taken with a digital camera (Leica ICC50 HD).

Pleurogram of Senna alata seeds

Senna alata seeds have a wide and dark green pleurogram located close to the hilar region, due to an unusual compression of seeds during maturation. Thus, we investigated changes that occurred during dormancy break in this species. Twenty seeds were immersed in sulphuric acid (98 %) or hot water (80 °C) for 30 and 15 min, respectively. Then, morphological changes in the pleurogram were observed and seeds were placed in germination conditions (described above) to evaluate differences between dormant and non-dormant seeds. Images were taken using a digital camera (Canon A650 IS) coupled to a stereomicroscope (Zeiss Stemi 2000-C).

RESULTS

Review on occurrence of the pleurogram in Fabaceae

We found literature reports for the presence of a pleurogram in 37 genera of Fabaceae (see Supplementary Data Table S2). Eight of these genera are in subfamily Caesalpinioideae and 29 are in subfamily Mimosoideae.

Presence of physical dormancy

The mass of scarified seeds of the nine species tested increased 260–360 % following 48 h of imbibition, whereas that of intact (non-scarified) seeds did not increase (Supplementary Data Fig. S2, all P < 0.001). Thus, seeds of all tested species have PY.

Structural changes during dormancy break

Structural changes during dormancy break are summarized in Table 1. Dormant S. alata seeds had well-defined and wide pleurograms with several superficial fissures (Fig. 2A). After the hot-water treatment, the palisade layer in the pleurogram was uplifted, thereby exposing the hourglass cells below it. The palisade layer remained attached to the seed coat only at the margins of the pleurogram (Fig. 2B). When sulphuric acid was used to break PY in S. alata, the pleurogram had several cracks in it where the palisade layer was uplifted (Fig. 2C). For S. cana, S. cernua, S. hirsuta, S. obtusifolia and S. occidentalis, the pleurogram did not change during dormancy break. Senna pendula did not have a pleurogram. However, for these species, the lens was uplifted during dormancy break. This displacement in the lens created a gap beneath it, but the other structures in the hilar region remained unchanged (Figs 2D–M and 3A, B).

Fig. 2.

Structural changes during dormancy break in seeds of Senna. (A–C) Senna alata. (A) Pleurogram of dormant seeds. (B) Pleurogram of non-dormant (using hot water) seeds showing lifting of the palisade layer (arrowheads). (C) Pleurogram of non-dormant (using sulphuric acid) seeds with several cracks and the palisade layer lifted up (arrowheads). (D, E) Senna cana. (D) Hilar region of dormant seeds. (E) Hilar region of non-dormant seeds with the lens lifted up. (F, G) Senna cernua. (F) Hilar region of dormant seeds. (G) Hilar region of non-dormant seeds with slits formed around the lens (arrowheads). (H, I) Senna hirsuta. (H) Hilar region of dormant seeds. (I) Hilar region of non-dormant seeds with an elevated lens and stretched groove around the lens (arrowheads). (J, K) Senna obtusifolia. (J) Hilar region of dormant seeds. (K) Hilar region of non-dormant seeds with displacement of the lens creating spaces around it (arrowheads). (L, M) Senna occidentalis. (L) Hilar region of dormant seeds. (M) Hilar region of non-dormant seeds with the lens lifted up. hi, hilum; hg, hourglass cells; le, lens; mi, micropyle; pl, pleurogram.

Fig. 3.

Structural changes during dormancy break in seeds of Senna species. (A, B) Senna pendula. (A) Hilar region of dormant seeds. (B) Hilar region of non-dormant seeds with displacement of the lens creating slits around it (arrowheads). (C–E) Senna reniformis. (C) Hilar region of dormant seeds. (D) Hilar region of non-dormant seeds with the emerged lens. (E) Pleurogram of non-dormant seeds with the palisade layer lifted up (arrowheads). (F–H) Senna silvestris. (F) Hilar region of dormant seeds. (G) Hilar region of non-dormant seeds with disruption of micropyle creating a gap (arrowhead). (H) Pleurogram of non-dormant seeds with several cracks (arrowheads). (I, J) Senna spectabilis. (I) Hilar region of dormant seeds with a separation between hilum and lens (circle). (J) Hilar region of non-dormant seeds with a hole in the hilum (arrowhead) and absence of separation between hilum and lens (circle). (K–M) Senna trachypus. (K) Hilar region of dormant seeds. (L) Hilar region of non-dormant seeds with disruption in the entire hilar region. (M) Hilar region of non-dormant seeds with disruption of hilum and displacement of micropyle creating gaps (arrowheads). hi, hilum; hg, hourglass cells; le, lens; mi, micropyle; pl, pleurogram.

In the case of S. reniformis seeds, the lens was changed when hot water was used to break PY (Fig. 3C, D). However, in addition to the lens, the pleurogram was disrupted when dormancy was broken using sulphuric acid. The palisade layer moved outward in different parts in the pleurogram, exposing the hourglass cells (Fig. 3E). In S. silvestris seeds, the entire hilar region was corroded during dormancy break, but the micropyle was disrupted forming a gap (Fig. 3F, G). Also, several cracks were formed in the pleurogram in non-dormant S. silvestris seeds (Fig. 3H). In dormant S. spectabilis seeds, the hilum and lens were morphologically separated (Fig. 3I). However, after dormancy was broken the space between them disappeared. The lens remained unchanged in all samples of S. spectabilis, but the hilum and micropyle sometimes were corroded, with an evident hole in the hilum (Fig. 3J). The hilar region was corroded during dormancy break in S. trachypus seeds, and in some cases this region changed completely, or only the hilum and micropyle were damaged. Visible openings were present only on the hilum and micropyle (Fig. 3K–M). For the latter two species, the pleurogram did not change during dormancy break.

Dye tracking

The dye only penetrated the seed coat after breaking PY, but structures stained by methylene blue varied among the species (Fig. 4, Table 2 and Supplementary Data Fig. S3). The dye always entered the seed through a specific structure, and the palisade layer and hypodermal cells beneath it stained blue (Fig. 4 and Supplementary Data Fig. S3). The pleurogram stained blue in S. alata, S. reniformis and S. silvestris (Fig. 4), which had the pleurogram disrupted following PY break in the previous experiment (Figs 2 and 3).

Fig. 4.

Non-dormant seeds of Senna species stained with methylene blue to trace pathways of water into the seeds. (A, B) Senna alata. (A) Seed following immersion in hot water, with arrowheads indicating cells stained in the pleurogram. (B) Seed following immersion in sulphuric acid, with cells in the pleurogram region stained. (C–E) Senna reniformis. (C) Seed following immersion in hot water, with the lens stained. (D) Seed following immersion in sulphuric acid, with the lens stained. (E) Seed following immersion in sulphuric acid showing the dye penetrating through the pleurogram only. (F, G) Senna silvestris. (F) Seed with dye entering through the hilum/micropyle. (G) Seed with dye entering through the pleurogram. (H) Senna spectabilis seed with the hilum stained. hi, hilum; le, lens; mi, micropyle; pl, pleurogram.

Table 2.

Dye-tracking in seeds of Senna species after breaking PY.

Species	Methylene-blue-stained structure
	Hilum	Lens	Micropyle	Pleurogram
S. alata	–	–	–	+
S. cana	–	+	–	–
S. cernua	+	+	–	–
S. hirsuta	–	+	–	–
S. obtusifolia	–	+	–	–
S. occidentalis	–	+	–	–
S. pendula	–	+	–	N
S. reniformis	–	+	–	+
S. silvestris	+	–	+	+
S. spectabilis	+	–	–	–
S. trachypus	+	–	+	–

+, Palisade layer and hypodermis stained blue; –, tissues were not stained in that region; N, a pleurogram is not present.

Anatomy of the pleurogram and water-gap complexes

The pleurogram was a weak region in some of the Senna species investigated and had several disruptions in it after PY was broken (Fig. 5A). An opening across the palisade layer that reaches the hourglass cells in seeds of S. alata (Fig. 5B) and S. reniformis (Fig. 5C) indicates that the pleurogram acts as a water gap. The complete disruption in the palisade layer of the pleurogram formed a linear fissure that allowed contact of the inner tissues with the moist substrate (Fig. 5B, C). The linear fissure in the pleurogram was similar to the disruption in the palisade layer in the lens, as evidenced in the lens of S. obtusifolia (Fig. 5D, E). However, a decrease in palisade layer thickness in the lens slit was observed, unlike what occurred for the slits in the pleurogram region.

Fig. 5.

Transverse sections of the water gaps. (A) Pleurogram region showing several fissures in the palisade layer indicating the weak regions (arrows) in S. alata seeds. (B) Detail of an opening in the pleurogram of S. alata seeds. (C) Pleurogram of S. reniformis seeds showing an opening. (D) Lens of S. obtusifolia seed showing linear slits (arrows). (E) Detail of the water pathway in the lens (dotted circle) in S. obtusifolia. hg, hourglass cells; op, opening; pa, palisade layer; pc, parenchyma cells.

Pleurogram of Senna alata seeds

Dormant seeds of S. alata are slightly wrinkled and have two pronounced dark green pleurograms without cracks near the hilar region (Fig. 6A, B). During both dormancy-breaking treatments, the pleurogram was disrupted (Fig. 6C). After immersion in sulphuric acid, fissures were formed, and apparent tissue corrosion appeared in the pleurogram (Fig. 6D). The pleurogram also contained fissures after immersion in hot water, and some of the outer layers became detached (Fig. 6E). After 24 h, water-permeable seeds were completely imbibed and the radicle protruded (Fig. 6F, G). There were no visible changes in the hilar region during the dormancy-breaking treatments.

Fig. 6.

Morphological changes in Senna alata seeds during dormancy break and germination. (A, B) Dormant seed. (A) General view. (B) Detail of the intact pleurogram. (C–E) Seed following dormancy break. (C) General view. Arrow indicates disruption of the pleurogram. (D) Detail of the pleurogram after exposure to sulphuric acid. Arrow indicates the presence of fissures and asterisk the corrosion in some parts. (E) Pleurogram after immersion in hot water. Arrow indicates detachment of the palisade layer and asterisk exposure of inner tissues. (F, G) Germinated seed after 24 h of imbibition. (F) General view. (G) Details of the pleurogram covered by a mucilaginous layer. hr, hilar region; ml, mucilaginous layer; pl, pleurogram; ra, radicle.

DISCUSSION

Seeds of all 11 species of Senna included in this study have PY, which was broken by various treatments. We documented for the first time PY in seeds of nine species, while it previously was reported for two of the species, namely S. obtusifolia (Baskin et al., 1998) and S. silvestris (L. F. Daibes and A. T. Fidelis, Universidade Estadual Paulista Júlio Mesquita Filho, unpubl. res.). A diversity of water gaps was identified in our study, with the lens, hilum, micropyle and pleurogram opening during the dormancy breaking process, depending on the species. Seeds of S. alata, S. reniformis and S. silvestris had a pleurogram that acted as a water gap (Table 1). Thus, our hypothesis that the pleurogram acts as a water gap in some species was supported. This is the first demonstration that the pleurogram can function as a water gap. In S. alata seeds, the pleurogram was the only water gap; the lens was non-functional (did not open). For the other two species in which water entered the seed through the pleurogram (S. reniformis and S. silvestris), a hilar water gap was also present. In Fabaceae, the subfamilies Caesalpinioideae and Mimosoideae were recently revised and grouped according to their phylogenetic relationships, resulting in 148 recognized genera (see LPWG, 2017). A pleurogram was reported in 37 of the 148 genera, which highlights the wide distribution of this structure in legume seeds. More studies are needed to investigate the occurrence of a pleurogram in seeds of the other genera of Fabaceae, because in seeds of some species this structure is quite small, with no distinction in colour compared to the rest of the seed coat. For several species, there is no published information on seed morphology, and thus the distribution of the pleurogram in legumes may be extended to other genera in future investigations. Furthermore, it needs to be determined if the pleurogram is also a trait in the Cucurbitaceae and other families.

Hyde (1954) showed that the hilum in Fabaceae subfamily Faboideae acts as a hygroscopic valve, which opens when external humidity is low and closes when it is high. Furthermore, Gunn (1981) suggested that the pleurogram acts in the seed dehydration process in a similar way to that of the hilum. The function of the pleurogram as a hygroscopic valve needs to be investigated, but it cannot be explained by structural features of the pleurogram. Unlike the Faboideae hilum, the pleurogram does not have a tracheid bar, a specialized vascular structure most likely involved in the hilum movements (Hyde, 1954; Lersten, 1982), and the presence of multiple fissures in the pleurogram is different from the single groove in the hilum. On the other hand, the pleurogram can open during PY break and thus allow passage of water into the seeds. However, unlike the hilum in Faboideae the open pleurogram cannot be reclosed, because its palisade is completely disrupted during opening. This pathway for water entry is then permanently open, making impermeable seeds permeable. Thus, involvement of the pleurogram in the process of dormancy release has been proven in our study; however, its role in dormancy acquisition (i.e. in seed drying) remains to be determined.

Pleurogram morphology varied among species (Table 1) but these differences do not explain why the pleurogram functions as a water gap in some species but not in others. No function has been confirmed for the pleurogram other than it acting as a water gap, as reported here. In relation to water uptake by the seed, opening of the pleurogram is similar to that of the other seed structures. Breaks in the palisade cells were seen in seeds of all 11 Senna species, but they occurred in different seed structures, i.e. lens, hilum and pleurogram. Unlike breaks in the lens (Rodrigues-Junior et al., 2018), breaks in the pleurogram were not indicated by a clear decrease in palisade layer thickness.

Irwin and Barneby (1982) described the morphology of seeds of 146 Senna species from the New World, 76 % of which had a pleurogram. Based on a phylogenetic study of Senna by Marazzi and Sanderson (2010), there is evidence that the pleurogram has evolved multiple times during evolution of the genus. However, two of the three species reported here to have a functional pleurogram (S. silvestris and S. alata) are closely related, being included in the primitive clades in Senna (clades I and II) (Marazzi et al., 2006; Marazzi and Sanderson, 2010). The third species (S. reniformis) was not included in the phylogenetic analysis. Thus, a broad analysis should be done to determine if the functional pleurogram is an ancestral trait in seeds.

The lens is the most common water gap in seeds of Fabaceae species (Rolston, 1978; Karaki et al., 2012; Baskin and Baskin, 2014; Rodrigues-Junior et al., 2014), but the hilum and micropyle often are reported as secondary openings in legume seeds from tropical environments (De Paula et al., 2012; Delgado et al., 2015; Geisler et al., 2017). We found that the hilum and/or micropyle function(s) as water gaps in three Senna species (S. silvestris, S. spectabilis and S. trachypus) in which the lens was non-functional. Thus, there is a diversity of specialized structures (water gaps) that open in response to environmental cues, thereby controlling the timing of germination of seeds with PY.

The various kinds of water gaps found in seeds of Senna species can be fitted into the classification scheme of water gaps proposed by Gama-Arachchige et al. (2013). Most of the 11 Senna species have Type-I simple/compound, wherein the lens is the main functional water gap (Table 1). For S. alata and S. reniformis, the palisade cells are dislodged only in the pleurogram, creating a wide pathway for water entry into the seed. Most palisade cells in the pleurogram are detached from the seed coat after PY is broken, but a minor part of the pleurogram remains attached to the seed. This structure is analogous to Type-II in the classification scheme. In fact, the pleurogram also can act by forming a linear opening in the palisade cells, as in S. silvestris, which is analogous to Type-I.

Heretofore, water gaps in seeds with PY have been described as small channels that direct the water to the inner tissues of the seed (Gama-Arachchige et al., 2013). In contrast, our study shows that the pleurogram, for which no function previously has been reported, can open, thereby creating a wide path for entrance of water into seeds of some leguminous species. Because the pleurogram is a seed structure that occurs in at least 37 genera of Fabaceae, it may be functional in a large number of species. Thus, it may play a key role in the life cycle of plants in several species by controlling the timing of germination.

SUPPLEMENTARY INFORMATION

Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Table S1: collection and ecological data on species of Senna used in this study. Table S2: genera and subfamilies of Fabaceae that produce seeds with a pleurogram. Fig. S1: preliminary germination (mean ± s.d.) tests of Senna seeds following different dormancy-breaking treatments. Fig. S2: mass (mean ± s.d.) of intact and manually scarified seeds of nine Senna species incubated under moist conditions. Fig. S3: non-dormant seeds of Senna species stained with methylene blue to trace pathways of water into the seeds.