Fracture of the dimorphic fruits of Aethionema arabicum (Brassicaceae)

Fruits exhibit highly diversified morphology, and are arguably one of the most highly specialised organs which evolved in higher plants. Fruits range in morphological, biomechanical and textural properties, often as adaptations for their respective dispersal strategy. While most plant species possess monomorphic (of a single type) fruits and seeds, here we focus on Aethionema arabicum (Brassicaceae). Its production of two distinct fruit (dehiscent and indehiscent) and seed types on the same individual plant provides a unique model system with which to study structural and functional aspects of dimorphism. Using comparative analyses of fruit fracture biomechanics, fracture surface morphology, and internal fruit anatomy, we reveal that the dimorphic fruits of Ae. arabicum exhibit clear material, morpho-anatomical, and adaptive properties underlying their fracture behaviour. A separation layer along the valve-replum boundary is present in dehiscent fruits, while indehiscent fruits have numerous fibres with spiral thickening, linking their winged valves at the adaxial surface. Our study evaluates the biomechanics underlying fruit opening mechanisms in a heteromorphic plant species. Elucidating dimorphic traits aids our understanding of adaptive biomechanical morphologies that function as a bet-hedging strategy in the context of seed and fruit dispersal within spatially- and temporally-stochastic environments. biomechanics Arrows indicate proposed directionality of DEH fruit rain-mediated seed dispersal (ombrohydrochory). of (“dehiscence zone”) and retaining a single seed within the pericarp during dispersal. Scale 75 X-ray tomographic microscopy. Morphology of the experimentally-fractured valves underlying the observed biomechanical differences between dehiscent (DEH) and indehiscent (IND) fruits of Aethionema arabicum . Macroscopic features of separated fruits indicate the presence of a replum and septum in DEH fruits (a), while a dysfunctional replum and lack of septum characterise IND fruits (f,g). SEM images of the abaxial (c,h) and adaxial (d,e,i,j) edges of fractured pericarps indicate the even structure of DEH pericarp fracture surface, in comparison with the uneven structure of IND pericarp fracture surface. Both abaxial and adaxial fracture surfaces of DEH pericarps are identical in morphology (c,d). The distinct thick-walled IND fruit endocarp layer (i,j) possesses numerous fibres with spiral thickening (j, arrowhead) across the adaxial fracture surface, where they were previously connected across the two halves of the pericarp. Scale bars = 1 mm (a,f), 100 µm (b,g) and 20 (c–e,


Introduction
Across the plant kingdom, fruits are highly diversified in their morphology, representing remarkable botanical architecture and reproductive ingenuity -from the giant pumpkins of Cucurbita maxima (Cucurbitaceae), to the microscopic fruits of Wolffia (Araceae) duckweeds which are no larger than 300 µm. The fruit is arguably one of the most highly specialised organs which evolved in higher plants, mediating the maturation and dispersal of seeds, and representing the end of the reproductive cycle in angiosperms (Ferrándiz et al. 1999;Linkies et al. 2010).
Fruits may range in biomechanical and textural properties from being fleshy and fibrous, to dry and papery. Dry fruits are broadly classified as either indehiscent, in which the pericarp (mature ovary wall) remains closed at maturity, or dehiscent, in which fruits split or open in some manner to release or expose the seed(s) (Spjut 1994). A classic example of the latter category, is in the cabbage family (Brassicaceae). Brassicaceae comprises species with great economic importance for food, fodder, industrial crops and ornamentals. It also includes important model plants such as Arabidopsis, Brassica, Lepidium and Boechera species (Heywood et al. 2007;Mummenhoff et al. 2008;Hohmann et al. 2015;Christenhusz et al. 2017). Typical fruit morphology within Brassicaceae consists of dehiscent, dry siliques ("capsules"), formed by a pistil composed of two or more carpels with persistent membranous placental tissue (septum). During seed dispersal, the pericarp valves detach from the replum along a separation layer, with varying dehiscence patterns (Fig. 1a) that exemplify evolutionarily labile morphologies.
The phenomenon of heterodiaspory is of particular importance for relatively short-lived species in spatio-temporally unpredictable environments, and may function as a bet-hedging survival strategy (Venable 1985), particularly in species distributed in desert, saline, and other frequently-disturbed habitats (Imbert 2002;.
In this study, we focus on the dimorphic species Aethionema arabicum (L.) Andrz. ex DC., a small, annual, herbaceous species belonging to the earliest diverging sister tribe (Aethionemeae) within the Brassicaceae (Franzke et al. 2011;Hohmann et al. 2015;Mohammadin et al. 2017). The genus Aethionema occurs mainly in the western Irano-Turanian region, an often-hypothesised cradle of the Brassicaceae (Hedge 1976;Al-Shehbaz et al. 2006;Beilstein et al. 2006;Mandáková et al. 2017). This divergence is thought to have occurred sometime during the Eocene (ca. 34-56 Ma) (Franzke et al. 2011;Hohmann et al. 2015; Mohammadin et al. 2017). Aethionema arabicum is characterised by two types of fruits and seeds produced on the same individual infructescence (Fig. 1b): dehiscent (DEH) fruits with 2-6 mucilaginous (M + ) seeds, and indehiscent (IND) fruits each containing a single nonmucilaginous (M -) seed (Lenser et al. 2016). Dehiscence of the DEH fruit morph causes local dispersal of M + seeds, which adhere to substrates via seed coat mucilage upon imbibition. M + seeds possess low dormancy and likely represent an anti-telechorous dispersal mechanism. In comparison, the more dormant IND fruit abscises from the mother plant in its entirety, and has the capacity to disperse longer distances (telechory) via wind and water (Arshad et al. 2019). The production of no intermediate morphs, together with a published genome sequence (Haudry et al. 2013), contributes to the suitability of Ae. arabicum as an excellent model system for fruit and seed dimorphism (Lenser et al. 2018;Mohammadin et al. 2018;

Fruit valve tensile testing and energy absorption relations
Data from Arshad et al. (2019) were re-analysed to obtain the energy absorption of fruit valve separation. Mature, dry fruits of Aethionema arabicum (Turkish accession ES1020, obtained from Eric Schranz, Wageningen University and Research Centre) (Wilhelmsson et al. 2019) were clamped on each side of the fruit wing, leaving a 2 mm gap between the jaws of a single-column tensile testing machine (Zwick Roell ZwickiLine Z0.5, Ulm, Germany) configured with a 200 N load cell. A constant speed for separation was set at 1 mm min -1 .
Force-displacement data were obtained using 30 replicates from 3 mature main branch infructescences. All fruits were freshly-harvested from plants grown under long-day conditions (16 h 20°C : 8 h 18°C, light : dark) in a glasshouse, and mechanically tested at room temperature (20ºC) and 31% relative humidity. The total area under the resultant forcedisplacement curve was calculated as the mechanical energy consumed by the pericarp in straining it to its fracture point. Using a digital camera (Canon EOS 5D Mark II, fitted with a EF 100 mm f/2.8 macro lens) and Fiji (Schindelin et al. 2012), the area representing the pericarp fracture zone was determined for 50 manually-separated replicates each of DEH and Polaron SEM Coating Unit E5100 (Bio-Rad Microscience Division, UK). Pericarp fracture surfaces were studied using SEM (Hitachi S-3000N, Japan) at an acceleration voltage of 20 kV, with images subsequently contrast adjusted in Adobe Photoshop CC.  (Stampanoni et al. 2006). Data were acquired using a 10× objective and a sCMOS camera (PCO.edge, PCO, Kelheim, Germany), with an exposure time of 80 ms at 12 keV (isotropic voxel dimensions = 0.65 µm). A total of 1501 projections were acquired equiangularly over 180°, post-processed and reconstructed using a Fourier-based algorithm (Marone and Stampanoni 2012). For verification, three replicates were examined for each fruit morph. Axial tomographic slice data derived from the scans were analysed and manipulated using Avizo ™ 9.5.0 (Thermo Scientific ™ , Visualization Science Group Inc., Burlington, MA) for Windows 10 Pro 64-bit, and contrast adjusted in Adobe Photoshop CC.

Synchrotron-based
Page 10 of 38

Distinct biomechanical events lead to dimorphic fruit failure
To investigate the biomaterial profiles underlying fruit opening mechanisms of the two distinct fruit morphs in Aethionema arabicum (DEH and IND), a uniaxial tensile test was performed. Such tests determine the resistance of the component against elongation, and thus enable the derivation of several key material properties and parameters of the tested material (Farquhar and Zhao 2006;Steinbrecher and Leubner-Metzger 2017). We observed two modes of fruit fracture and characterised the force-displacement curves associated with the distinct morphs (Fig. 2). Fruits from the DEH morph were typified by force-displacement curves exhibiting an initial elastic and plastic deformation, followed by a pre-failure event ( Fig. 2a). This fracturing event typically initiated at the valve-replum border adjacent to the fruit-pedicel junction, extending along the longitudinal axis of the replum. The crack wake was temporarily held together before a second elastic and plastic deformation phase, preceding complete fruit fracture after which no further change in force was detected. In contrast, IND fruits were characterised by a consistent loading phase, comprising nonuniform deformation of the pericarp prior to uniform and linear elastic deformation leading to fruit fracture (Fig. 2b).
The force-displacement curves for DEH fruits show the typical biomechanical response for loading of a benignly "ductile" and elastic material, which is able to be deformed in multiple stages without causing a complete fracture of the material. The DEH fruits are initially compliant and exhibit a degree of flexibility as the load is increased (Fig. 2a). The more brittle IND fruits, however, failed with less deformation when subjected to loading, with, on average, a 2.6-times higher force. Thus, the dissimilarities in the mechanism and ability of the dimorphic fruits to resist the extension of the initial crack are profound. The comparative biomechanical properties between fruit morphs were also associated with significant differences in the mean mechanical energy consumed by the pericarp in straining it to its failure point (t 58 = -9.704, P < 0.001, d = 2.5) (Fig. 2c). The energy taken up by each sample is represented by the total area underneath the force-displacement curve up to failure (Hourston et al. 2017;Steinbrecher and Leubner-Metzger 2017). The IND fruits (mean ± SE = 0.227 J mm -2 ± 0.02) had a ca. 12-fold increase compared with DEH fruits (0.019 ± 0.002).
Taken together, results show that fruit valve opening in Ae. arabicum has two clear biomaterial and mechanical energy profiles, associated unequivocally with the two morphs.

Comparative fracture surface morphology reveals distinct properties of the fruit endocarp
Since the fruits exhibit clear biomechanical failure patterns, pericarp fracture surface morphology was investigated to determine if it contributed to the observed fruit failure patterns. At the macroscopic level, experimentally-fractured valves revealed the replum and septum typical for dehiscent brassicaceous siliques in the DEH morph ( Fig. 3a), while experimentally-fractured valves in the IND morph (Fig. 3f) revealed a fruit with a dysfunctional replum and lacking a septum. Scanning electron microscopy (SEM) revealed structural differences at several hierarchical levels of organisation. During the fracturing process, the internal tissues of the two fruit morphs split in two distinct ways; the comparatively even structure of the DEH pericarp fracture surface contrasted with the uneven structure of the IND pericarp fracture surface, which often had protrusions at the valve edge ( Fig. 3g). In the DEH fruit, an exocarp layer of thick-walled cells, together with a thin-walled mesocarp and endocarp were visible at the fractured edge ( Fig. 3c-e). Both adaxial and abaxial surfaces were identical in morphology (Fig. 3c,d). Furthermore, the region of follicle splitting along the replum during dehiscence exhibits a concave surface (Fig. 3e). The IND Page 12 of 38 Botany fruit pericarp morphology, in contrast, exhibited a different structure (Fig. 3h-j). At the abaxial margin (Fig. 3h), the valve fracture surface appeared consistently "rough" in texture and comprised cell walls that had been mechanically torn. At the adaxial margin (Fig. 3i, j), the endocarp consisted of a very distinct thick-walled single cell layer, oriented at a perpendicular angle to the longitudinal axis of the fruit. Here, numerous fibres with spiral thickening (Fig. 3i, j) can be seen to run across the adaxial surface, where they were previously connected across the two halves of the pericarp. Thus, there are clear fracture surface morphologies, at various hierarchical levels, which underlie the two observed fruit fracturing behaviours.

Comparative internal anatomy confirms absence of a separation layer in IND fruits
To explore the internal anatomy of the dimorphic fruits, we conducted non-destructive investigations of the internal structure of mature fruits prior to the onset of ripening, with a particular focus at the region of fruit failure. Reconstructed digital sections (orthoslices) obtained by SRXTM revealed high resolution cell and tissue details (Fig. 4) without destruction of the sample or risk of artefacts associated with traditional histology (Betz et al. 2007;Smith et al. 2009). Differences, otherwise determined by tissue and cell wall composition, were highlighted by varying X-ray attenuation. Our schematic interpretation of fruit layers (Fig. 4c, d) indicates that cells of the exocarp, mesocarp, and endocarp layers were all readily distinguishable in the digital sections from both fruit morphs. Two distinct layers of the endocarp were observed; endocarp a (ena) comprised an inner epidermis of longitudinally-elongated, thin-walled cells, while a subepidermal endocarp b (enb) layer consisted of one to three layers of tightly packed, isodiametric cells. These observations correlate with the thick cell-walled endocarp layer of the IND fruit, as observed by SEM (Fig.   3j), which leads to the fibres with spiral thickening after fracturing.
The Ae. arabicum DEH morph is typical for many brassicaceous siliques, in that the margins of the two carpels and the parietal placentae, between which the septum is attached, form a replum. However, of particular significance is the separation between cells of the replum and endocarp layers (ena and enb) in the DEH fruit morph (Fig. 4a,c). This distinct tissue "separation layer" forms part of the "dehiscence zone", extending along the entire longitudinal axis of the pericarp at the valve-replum boundary. In contrast, tissue organisation within the IND morph pericarp is distinctly different (Fig. 4b,d). The mesocarp contributes to a distinct layer composed of large cells, with more densely-packed cells adjacent to the comparatively smaller, dysfunctional replum. The IND morph appears unilocular and only the remnants of a septum persist; the ovary thus contained a single ovule (Mseed; Fig. 4b,d).
Though there is a comparatively reduced replum, cells of the endocarp and the replum were not separated by a dehiscence zone, instead forming a continuous layer. Furthermore, while the enb layer becomes multi-layered proximal to the replum in the DEH morph, only a few cell-layers are present in the enb in the IND morph. Thus, the absence of a "separation layer" or "dehiscence zone", as well as septum absence, are major differences in the internal anatomy of the IND morph pericarp.

Distinct fracture biomechanics of dimorphic fruits
The integration of biomechanics and mechanobiology has been a significant methodological advancement to address questions in plant sciences, and has seen a renaissance over recent decades (Read and Stokes 2006;Moulia 2013). Our biomechanical evaluation of fruit opening mechanisms of a heteromorphic plant species links pericarp-specific properties to adaptive seed and fruit dispersal. In contrast to the single fruit fracture mechanism -and associated dispersal strategy -of monomorphic plants, the distinct biomechanical profiles for the Aethionema arabicum fruit morphs are correlated with their adaptations for different modes of dispersal (Arshad et al. 2019). Our comparisons of fruit fracture biomechanics, and the morpho-anatomical features which contribute to the observed patterns, show that the behaviours are unequivocally associated with the two fruit morphs. In natural conditions, it is hypothesised that recurring forces from raindrop impacts and/or wind on the DEH fruit most likely induce fatigue crack growth along the replum. DEH fruits exhibit a multi-staged biomechanical response with several ranges of linear behaviour during pericarp opening, while the IND pericarp provides a more brittle breaking behaviour and prolonged loading phase requiring a significantly higher opening energy (Fig. 2c). Torn out structures (IND fruit, Fig. 3g) indicate that friction between cell layers had to be overcome during the fruit fracture process, creating the "rough" fracture surface texture. As previously described by Beismann et al. (2000) shearing cell layers during the tearing process may contribute to the toughness of the material. The winged, dispersal-enhancing feature of the IND pericarp allows the seed to remain encased during dispersal (Arshad et al. 2019); however, as a rare adaptation for long-range dispersal in desert plants, pericarp "wings" may also serve as a protective measure against the adverse environment during germination and seedling establishment (Ellner and Shmida 1981).
That fruit opening in monomorphic Brassicaceae fruits is dependent on the positioning and formation of the valve margin and its dehiscence zone is well established (Spence et al. 1996;Avino et al. 2012), but little is known about differences in the opening mechanisms operating in heteromorphic fruits. We found that the IND fruit lacks the distinctive anatomical organisation present in typical dehiscent siliques, and that the cells of the endocarp layers form a continuous band around the replum, thus preventing fruit dehiscence. The IND fruit pericarp, therefore, not only confers enhanced dispersal ability and degree of dormancy (Arshad et al. 2019), but also a mechanism for remaining as a closed disseminule after dispersal. This may suggest that pericarp-mediated dormancy in the Ae. arabicum system may be partly physically-and physiologically-imposed on Mseeds. In Lepidium didymum (Brassicaceae) indehiscent fruits, Sperber et al. (2017) found that the thick, hard pericarp imposed a mechanical constraint on the germination of encased seeds by influencing water uptake patterns into seeds inside fruit valves, and that fungi induced selective weakening of pericarp tissue (at distinct predetermined zones), lowering its mechanical resistance to breakage. The mechanisms by which the Ae. arabicum IND pericarp may impose a mechanical constraint to full water uptake by the Mseed is little investigated. Ongoing ecophysiological, biomechanical, and molecular analyses on the influence of the pericarp tissue during and after IND fruit germination should shed light on its specific role.
The presented fruit fracture biomechanics prompts questions on the contrasting development and molecular regulation underpinning the morph-specific determination of Ae. arabicum fruits. Within the Brassicaceae, the evolutionary transition from dehiscent to indehiscent Page 16 of 38 Botany fruits has been investigated in Lepidium, where both dehiscent and indehiscent fruits are produced (Mummenhoff et al. 2008;Mühlhausen et al. 2013). Anatomical changes at the valve-replum border were connected with altered expression patterns of various genes orthologous to the known fruit developmental genes in Arabidopsis, including ALCATRAZ (ALC), INDEHISCENT (IND), SHATTERPROOF1 (SHP1) and SHATTERPROOF2 (SHP2) (Rajani and Sundaresan 2001;Liljegren et al. 2004;Ballester and Ferrandiz 2017).
Orthologues were shown to be expressed in the dehiscence zone-forming fruit valve margin of Lepidium campestre dehiscent fruits while, in the corresponding tissue of L. appelianum indehiscent fruits, expression patterns were down-regulated (Mühlhausen et al. 2013).
Indeed, in Ae. arabicum, expression analyses in mature fruits have previously revealed that the orthologue IND (which is involved in the differentiation of the separation layer and acts as the key regulator in controlling valve margin specification) was down-regulated in IND fruits compared with DEH fruits (Lenser et al. 2016). The genetic pathways operating in DEH and IND fruits therefore indicate an avenue for more detailed developmental and molecular time-course characterisations in Ae. arabicum, particularly with reference to valve margin-specific and dehiscence zone identity genes (Avino et al. 2012;Lenser et al. 2018).
Our microscopy approaches indicated fundamental differences in the structure and organisation of fruit valve layers. Combining SEM with non-invasive and non-destructive methods such as SRXTM provides new possibilities for the visualisation and analysis of the external and internal structure of fossil and extant plant material (Friis et al. 2014;Benedict et al. 2015). Such imaging solves problems associated with cutting or histological sectioning by minimising the introduction of artefacts (e.g. tears, gaps), and allows multiple planes of section through the same specimen to be acquired at high quality resolution (Smith et al. of valve tissue morphology was of particular significance at the adaxial fracture surface. As structural differences in the valves are associated with the mode of fracturing, the IND fruit tissue organisation could perhaps be interpreted as a remnant of a once active dehiscence apparatus (Fahn and Zohary 1955). Differential timing of anisotropic growth patterns, in turn coordinating the development of fruit growth and maturation leading to dispersal, may also influence the anatomical organisation and material properties observed between DEH and IND fruits. The dimorphic fruits in Ae. arabicum therefore provide an ideal system with which to model post-fertilisation gynoecium growth and shape formation, as has been demonstrated in the monomorphic Capsella rubella (Eldridge et al. 2016), to identify tissuespecific activities required to obtain the two distinct morphologies.

Hypothesis for fracturing biomechanics and dispersal in natural conditions
The semi-arid environment in which Ae. arabicum grows is characterised by highly variable rainfall in space and in time, and therefore presents challenging climatic and edaphic conditions for plant growth (Arshad et al. 2019). At the macroscale this may include sporadic rain events, with high precipitation rates over relatively small spatial scales, while at the microscale, topographic factors and soil surfaces are thought to contribute to the variability of water availability (Kigel 1995). The contrasting morpho-anatomical and biomechanical properties of Ae. arabicum fruits contribute to bet-hedging adaptations for successful dispersal in the scree and steppe habitats of Anatolia (Arshad et al. 2019). Prior to dehiscence, the DEH fruit pericarp dries as the fruit approaches maturity. Our working hypothesis is that the highly co-ordinated events causing tissue separation and endocarp lignification create spring-like tensions in the mature DEH fruit, the elastic energy from which is retained during the dry period until rain-induced (ombrohydrochory) impact events cause "pod-shattering" and M + seed dispersal to occur. This is consistent with the passive,

Page 18 of 38
Botany drying forces acting on microstructures that have been demonstrated during fruit and seed dispersal of other species (Elbaum and Abraham 2014). For example, the shedding of living twigs (in Salix spp. and Populus spp.) provides a reproductive mechanism (via twig dispersal and subsequent establishment in new habitats) that also relies on fracture mechanics. The relative roughness of the twig fracture surface in the genus Salix correlates with the classification of brittle and non-brittle species (Beismann et al. 2000).
Dehiscence upon wetting has generally been associated with plants adapted to arid environments (Gutterman 2002;Pufal et al. 2010), and the work of fracture is typically negatively correlated with moisture content (Farquhar and Zhao 2006). Tensions are created due to the differential drying of the parenchymatic and sclerenchymatic tissues of the pericarp, while degeneration of the middle lamellae of the separation layer cells (via cell wall degrading enzymes) forms a pre-determined breaking zone along the longitudinal axis. This, together with moisture-induced (hygrochastic) movements of fruit pedicels (Lenser et al. 2016), is thought to contribute to the distinct biomechanical fracture mechanism observed in DEH fruits. The mature IND fruit, in contrast, abscises from the mother plant in its entirety and thus has inherently different biotic and abiotic factors influencing its pericarp fracture biomechanics. Post-dispersal time-lapse data obtained during seedling establishment (not shown) suggest that IND fruit valve separation only occurs after completion of germination, as a result of radicle protrusion between two adjoining pericarp valves.

Conclusions
The species richness and divergent fruit shapes in the Brassicaceae provide an invaluable framework to address questions on seed and fruit dispersal. Here, we have shown that the dimorphic fruit fracture patterns in Aethionema arabicum are associated with distinct morpho-anatomical features influencing the deformation behaviours of the pericarp during opening. A distinct "separation layer" along the DEH fruit valve-replum boundary contributed to the multi-staged fracture events leading to failure. In contrast, IND fruits were shown to possess a distinct endocarp layer with spirally-thickened fibres linking its winged valves at the adaxial surface. This, and a lack of a dehiscence zone, mediate the more brittle   fruits in contrast show a sudden and complete failure preceded by a characteristic loading phase. Numbered panels above force-displacement curves and the corresponding photographs illustrate the process of DEH (1, 2, 3 in blue) and IND (1, 2 in red) fruit fracturing. Scale bars separation of fruit valves from fruits differ significantly (t 58 = -9.704, P < 0.001, d = 2.5) between DEH and IND morphs. N = 30. Error bars ± 1 standard error of the mean. Data are normalised relative to the mean area of the fracture zone (DEH: 6.94 ± 0.14 mm 2 , IND: 0.87 ± 0.03).  Morphology of the experimentally-fractured valves underlying the observed biomechanical differences between dehiscent (DEH) and indehiscent (IND) fruits of Aethionema arabicum. Macroscopic features of separated fruits indicate the presence of a replum and septum in DEH fruits (a), while a dysfunctional replum and lack of septum characterise IND fruits (f,g). SEM images of the abaxial (c,h) and adaxial (d,e,i,j) edges of fractured pericarps indicate the even structure of DEH pericarp fracture surface, in comparison with the uneven structure of IND pericarp fracture surface. Both abaxial and adaxial fracture surfaces of DEH pericarps are identical in morphology (c,d). The distinct thick-walled IND fruit endocarp layer (i,j) possesses numerous fibres with spiral thickening (j, arrowhead) across the adaxial fracture surface, where they were previously connected across the two halves of the pericarp. Scale bars = 1 mm (a,f), 100 µm (b,g) and 20 µm (c-e, h-j). SEM = scanning electron microscopy. M + = mucilaginous. M -= non-mucilaginous. ex = exocarp. ms = mesocarp. en = endocarp.
Comparative SRXTM results obtained from digital transverse sections of mature Aethionema arabicum dehiscent (DEH) (a) and indehiscent (IND) (b) fruits with schematic representations (c and d) at the valvereplum region. Inset fruits (not to scale) depict the region from which the slice is taken. The exocarp and outer epidermis (ex), two to three cell layers of mesocarp (ms), and two zones within the endocarp plus inner epidermis (ena and enb) can be distinguished in the fruit valves. A "separation layer" at the valve margin, extending in continuity around the replum when the fruit is further dried, allows DEH fruit valves to detach from the replum. Thus, the endocarp (enb), valve margins and regions of the mesocarp, together with cell wall degradation in the ena layer, contribute to "pod-shattering" biomechanics during DEH fruit dehiscence. Arrows indicate proposed directionality of DEH fruit drying tensions, which contribute to rainmediated seed dispersal (ombrohydrochory). The IND fruit, characterised by its absence of a separation layer ("dehiscence zone") and septum, does not undergo this highly co-ordinated process, instead retaining a single seed within the pericarp during dispersal. Scale bar = 75 µm. SRXTM = synchrotron radiation X-ray tomographic microscopy. M -= non-mucilaginous.