See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/320843762

Tank-inflorescence in innocentii (): Three- dimensional model and development

Article in Botanical Journal of the Linnean Society · October 2017 DOI: 10.1093/botlinnean/box059

CITATIONS READS 11 321

7 authors, including:

Fernanda Nogueira Sofia Kuhn Universidade Federal do Oeste do Pará Universidade Federal do Rio Grande do Sul

11 PUBLICATIONS 55 CITATIONS 7 PUBLICATIONS 46 CITATIONS

SEE PROFILE SEE PROFILE

Felipe Luis Palombini Gabriel H. Rua Universidade Federal do Rio Grande do Sul Universidad de Buenos Aires

38 PUBLICATIONS 181 CITATIONS 53 PUBLICATIONS 596 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Conservação e caracterização dos recursos genéticos de espécies de gramíneas nativas do Brasil View project

Study of Partial Volume Effect on X-ray Microtomography's images View project

All content following this page was uploaded by Fernanda Nogueira on 15 November 2017.

The user has requested enhancement of the downloaded file. Botanical Journal of the Linnean Society, 2017, 185, 413–424. With 5 figures.

Tank-inflorescence in Nidularium innocentii (Bromeliaceae): three-dimensional model and development

FERNANDA M. NOGUEIRA1, SOFIA A. KUHN1, FELIPE L. PALOMBINI2, GABRIEL H. RUA3, AVACIR C. ANDRELLO4, CARLOS ROBERTO APPOLONI4 and JORGE E. A. MARIATH1*

1Laboratory of Anatomy LAVeg, Institute of Biosciences, Department of Botany, Federal University of Rio Grande do Sul (UFRGS), Av. Bento Gonçalves, 9500, Porto Alegre, RS, Brazil 2Laboratory of Design and Material Selection LdSM, Federal University of Rio Grande do Sul (UFRGS), Av. Osvaldo Aranha 99/604, 90033-190, Porto Alegre, RS, Brazil 3Universidad de Buenos Aires, Facultad de Agronomía, Cátedra de Botánica Sistemática, Avenida San Martín 4453, C1417DSE, Buenos Aires, Argentina 4Applied Nuclear Physics Laboratory, Londrina State University (UEL), Rod. Celso Garcia Cid, Km 380, Londrina, PR, Brazil

Received 8 March 2017; revised 30 May 2017; accepted for publication 11 August 2017

In Nidularium, inflorescence branches are subtended by large floral bracts, in which water accumulates. The branch- ing pattern is obscured because their internodes remain short, hampering their interpretation. This study focuses on the development of the inflorescence in N. innocentii, combining different approaches in order to understand its architecture and summarize it in a three-dimensional (3D) model. We also present the interpretation of tank- inflorescence development, recognizing the processes that have taken place in the evolution of this structure in this group. The inflorescence was typified based on Troll’s and Weberling’s systems. Development was studied using light microscopy and X-ray microcomputed tomography. The system is polytelic; the main axis ends in the main flores- cence and bears lateral paraclades with coflorescences. Each lateral branch develops in the axil of a bract, which is large and displays alternate arrangement. No prophylls were observed in the system. In the 3D reconstruction, the volume of the model was calculated. The volume of the empty region is c. 2.4 times higher than the plant material. Tank-inflorescence development seems to have occurred by the combination of three processes: bract disposition and its overgrowth; failure in internode elongation; and paraclade flattening. The tank-inflorescence evolved in a few groups of core bromelioids, and may be associated with floral protection.

ADDITIONAL KEYWORDS: – nidularioids – inflorescence architecture - X-ray microcomputed tomography (μCT).

INTRODUCTION structures is frequently insufficient for a proper under- standing of inflorescence morphology, resulting in a An inflorescence is usually defined as a shoot system in problem that requires more sophisticated analytical which flowers develop (Troll, 1964). A wide diversity of tools. Comparative typological analysis has proved to inflorescence patterns is exhibited by the angiosperms be a powerful approach for generating primary homol- (Weberling, 1965) as a result of different combinations ogy hypotheses. In this sense, descriptions based on of developmental events. Simple observation of mature Troll’s (1964, 1969) and Weberling’s (1989) systems, which consider the branch systems and the position of each structural unit within the inflorescence as a whole (Weberling, 1965), are accurate and reliable *Corresponding author. E-mail: [email protected]

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 413 414 F. M. NOGUEIRA ET AL. for examining, describing and naming morphologi- system and the subtending bracts and calculation of cal features of inflorescences and drawing homology the retained water volume, leading to an accurate mor- statements that can drive future evolutionary devel- phological and functional interpretation in virtually opmental research (Rua, 2003; Tortosa, Rua & Bartoli, any plane (Staedler, Masson & Schönenberger, 2013). 2004; Acosta et al., 2009; Stützel & Trovó, 2013). The nidular inflorescence shared by some nidulari- In Bromeliaceae subfamily Bromelioideae, the so- oids brings up some questions, especially when and called ‘nidularioid genera’ (Silvestro, Zizka & Schulte, why this particular morphology has evolved. Is it pos- 2014; Evans et al., 2015; Heller et al., 2015; Santos- sible that this structure evolved in tankless ancestors Silva et al., 2017) comprise an assemblage of genera acting as a water-impounding tank? Could water cap- that mostly occupy humid forest habitats, typical ture in floral bracts make an important contribution of the Brazilian Atlantic Forest. They share a set of to the water balance of floral development? This study morphological traits, which includes a particular focuses on the development of the inflorescence in inflorescence morphology (Leme, 1997, 1998, 2000; N. innocentii, combining different approaches in order Santos-Silva et al., 2017). Otherwise, the relationship to understand the architecture and summarize it in a among these genera is poorly understood, and the 3D model. We also discuss the interpretation of tank- inflorescence arrangement probably does not reflect inflorescence development, describing the processes a synapomorphy for this group (Santos-Silva et al., that may have occurred in the evolution of this struc- 2017). Nidularium Lem., one of the nidularioid gen- ture and its biological consequences. era, is widely distributed in the Neotropics, from the Brazilian states of Bahia to Rio Grande do Sul, occur- ring not only in the Atlantic Forest but also in con- MATERIAL AND METHODS trasting biomes such as caatinga (Leme, 2000; Flora do Brasil, 2020). Nidularium can be distinguished Botanical material from related genera [e.g. Canistropsis (Mez) Leme, Specimens of N. innocentii were collected in the field Wittrockia Lindm. and Edmundoa Leme] especially and maintained in the Living Collection of Plant with regard to inflorescence architecture, as it has Anatomy Laboratory (LAVeg) at the Universidade large floral bracts in which water accumulates. This Federal do Rio Grande do Sul (UFRGS), Porto Alegre, strategy might be associated with floral protection Brazil. Forty-nine individuals were analysed at dif- (Leme, 2000). Water-storage bract systems in inflo- ferent developmental stages: 28 at anthesis and 21 in rescences resemble the water-impounding foliage of early stages of development, 18 of which were studied Bromeliaceae and can accumulate up to 100 mL in with light microscopy and three with µCT. A voucher N. innocentii Lem. (Leme, 2000). specimen was deposited at ICN (ICN 190905). The inflorescence of Nidularium consists of a com- plex branched system, with large floral bracts, referred to as ‘nidular inflorescences’ by some authors (Benzing, Descriptive background 2000; Leme, 2000). The inflorescence has short inter- Inflorescences were analysed and described using a nodes, causing difficulty in its interpretation. Studies typology-based comparative approach (Troll, 1964, in the genus have not followed any typological criteria 1969; Weberling, 1965, 1989), to allow further compari- in describing inflorescences and have applied terms sons with other genera of Bromeliaceae in a coherent such as compound, cyathium or corymb, preventing framework. comparative morphological studies (Mez, 1934–1935; Smith & Downs, 1979; Leme, 2000). The enhancement of non-invasive methods, such as Preparation for light microscopy X-ray microcomputed tomography (μCT), is providing Inflorescences in early stages of development were new interpretations of structural and morphological collected and dissected. Inflorescence fragments were plant organization (Stuppy et al., 2003; Dhondt et al., fixed in 1% glutaraldehyde and 4% formaldehyde in 2010; Oskolski et al., 2015; Palombini et al., 2016). 0.1 M sodium phosphate buffer, pH 7.2 (McDowell & Through this method, thousands of two-dimensional Trump, 1976), and washed in 0.1 M sodium phosphate (2D), high-resolution radiographic projections are gen- buffer, pH 7.2. Material was dehydrated in an ethanol erated from a single sample and are later combined series (10−100%) and embedded in hydroxyethylmeth- and reconstructed in three dimensions by a computer acrylate (Gerrits & Smid, 1983). Thin sections (4 µm algorithm (Stock, 2009). The three-dimensional (3D) thick) were mounted on a glass blade and stained model can be segmented into multiple structures with 0.05% toluidine blue O, pH 4.4 (Feder & O’Brien, and thus enable the visualization of the arrangement 1968). Photomicrographs were acquired under a bright of inflorescence components, such as the branching field using a Leica DMR (Leica Microsystem GmbH,

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 TANK-INFLORESCENCE IN NIDULARIUM 415

Wetzlar, Germany) microscope with AxioVision LE binary masks for each structure, which were then (Carl Zeiss Meditec AG, Oberkochen, Germany). exported as surfaces (STL extension). Two main masks referring to the solid sample tissue and the void model of inflorescence, corresponding to the empty tank X-ray microcomputed tomography region volumes, were also segmented and quantified. Inflorescences in early stages of development were The corresponding area of both masks for each slice collected and fixed in 1% glutaraldehyde and 4% for- was assessed for determining the progressive and total maldehyde in 0.1 M sodium phosphate buffer, pH 7.2, amount of tank region compared with the solid sample for 24 h (McDowell & Trump, 1976) and washed in material, based on the relative volume method for μCT 0.1 M sodium phosphate buffer, pH 7.2. After wash- characterization presented by Palombini et al. (2016). ing, the material was immersed in 1% phosphotung- The method performs the sum of the values of the 2D stic acid/50% Formalin-Aceto-Alcohol for 28 days with relative area multiplied by the voxel size of the analy- solution substitution every 7 days to increase the con- sis, i.e. the thickness of the slice for each segmented trast (modified from Hayat, 2000). Material was thor- section, resulting in the relative volume. oughly washed in 0.1 M sodium phosphate buffer, pH 7.2. Samples were dehydrated in an acetone series and dried using the critical-point method (Gersterberger & RESULTS Leins, 1978), with BAL-TEC, CPD 030 equipment. Material was mounted in a plastic sample holder Synflorescence development takes place in the centre for the μCT analyses via a SkyScan 1172 (Bruker- of the rosette (Fig. 1A) from the main shoot apical mer- microCT, Kontich, Belgium). Two tomographic acquisi- istem (SAM) and several axillary SAMs subtended by tions were performed and further combined to scan the bracts. The central SAM originates a main florescence sample fully. Source voltage and current were 74 kV (MF), whereas c. 14 paraclades (Pc) bearing coflores- and 133 μA, respectively, and no filter was applied. cences (Fig. 1B, C) develop in the axils of large bracts Exposure time was 1700 ms, resulting in 1840 slices arranged in a spiral–alternate phyllotaxis (Fig. 1A) with 1.9908-μm voxel size. along an enrichment zone (EZ) c. 25 mm long at anthe- Open-source FIJI/ImageJ software (Schindelin sis (Fig. 1B, C). A simple 3D model, constructed using et al., 2012) was used to perform the 3D reconstruc- exclusively the angles between paraclades obtained tion, virtual sectioning, segmentation of structures from µCT data, captured this pattern. Such angles and volume calculation from the scanning data. μCT measure 142.37 ± 13.75° (Fig. 1D, E). stack images were first treated with brightness and Both MF and paraclades are composed of a varia- contrast adjustment to enhance different regions of the ble number of lateral flowers spirally arranged along sample. However, this operation also increased noise, an indeterminate main axis. The paracladial zone and the Sigma plug-in (Gonzalez & Woods, 2008) noise shows a basipetal development, whereas the flower reduction filter was applied as a result. The Sigma inception along the MF occurs in an acropetal spiral plug-in considers the grey-levels of neighbouring pix- (Figs 1C, 2A–C), and this pattern is reproduced along els to identify and attenuate noise regions while pre- each paraclade serving structure edges. Six iterations were performed . During the earliest developmental stages, the using the following parameters: pixel radius varying main florescence possesses few developed flowers in from 0.8 to 1.0; applying pixels within 1.5 sigmas; and the axils of the basalmost subtending leaves (sl), fol- minimum pixel fraction varying from 0.30 to 0.75. lowed by many axillary vegetative buds (Fig. 2A–C). The Sigma filter smooths the noise of images within Each paraclade has a variable number of flowers, a specific radius value, when performing the mean the distal branches possessing many developed calculation of neighbouring pixels. However, the filter flowers and one or two distal axillary vegetative only includes the adjacent pixels for which grey values buds (Figs 2D, E, G, 3A). In some distal branches do not vary from those of the current pixel in a given the entire system can develop flowers (Fig. 2F). The range. This range is given by the standard deviation outermost paraclades, proximal and distal, do not of neighbouring pixels within the used sigma number. develop second-order axes (Fig. 2D–G). In proximal If the number of pixels in the range is lower than the branches second-order axes (Pc’) mostly develop. minimum pixel fraction value, an average of all adja- Each secondary branch is also subtended by a sub- cent pixels is considered (Gonzalez & Woods, 2008). tending leaf (sl’) (Fig. 3B–E), and this system can After contrast and noise reduction adjustments, μCT either develop flowers (Fig. 3B, C, E) or maintain stack images were segmented into different structures meristem activity in an apical vegetative meristem to be individually analysed. The segmentation editor (Fig. 3D). Flower primordia (fp) have flattened api- plug-in (Schindelin et al., 2012) was used to create ces (Fig. 3C, G) and they can thus be distinguished

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 416 F. M. NOGUEIRA ET AL.

Figure 1. Plant and inflorescence of Nidularium innocentii. A, mature individual with tank-inflorescence. The arrows indi- cate bracts, with alternate arrangement. B, dissected inflorescence showing the removed paraclades from the enrichment zone (EZ) and the main florescence (MF). The arrows indicate the insertion points of paraclades in the EZ. C, schematic model of branched system, showing the flower opening in each paraclade. D, 3D schematic model highlighting the angle between each paraclade. (EZ) enrichment zone; (MF) main florescence. E, top view of 3D schematic model of inflorescence architecture. Scale bar = 5 cm (A); 2 cm (B). from vegetative meristems (vm), which have acute congested. In a basipetal sequence, the branches apices (Fig. 3C, F). Through the whole branch- become progressively flatter in an axial plane, and ing system the internodes remain short (Fig. 3H change from circular to elliptical in outline in trans- a), so that the synflorescence appears extremely versal section (Fig. 3H b). Following these changes,

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 TANK-INFLORESCENCE IN NIDULARIUM 417

Figure 2. Inflorescence of Nidularium innocentii A, transverse section of the main florescence (MF); the black circle indi- cates the apical vegetative meristem, and the arrows indicate subtending leaves (sl). B, polar view of μCT 3D reconstruc- tion of the MF, showing flowers and some vegetative buds. C, axial view, detail of the MF. D and E, axial view of μCT 3D reconstruction showing the paraclade-bearing distal branches; arrows indicate vegetative buds. F and G, transverse section of distal branches. In these paraclades the entire system can develop flowers (F), or maintain the meristem activity in an apical bud (G). Scale bar = 1 mm (A–G). the bracts become larger and cover the flowering bracts (Fig. 4E). The calculated volume of the model branches (Fig. 3H c). was 0.0621 mL, corresponding to the whole solid tis- In the 3D reconstruction, the inflorescence elements sue. The volume of the empty region, representing the were segmented in different perspectives: (1) the tank space, was 0.15 mL, i.e. c. 2.4 times greater than entire system; (2) bracts subtending paraclades that that of the solid tissue. bear flowers (Fig. 4A, B, E); (c) a void model of inflo- The area of solid tissue was constant over the sections, rescence corresponding to the empty space of the inflo- while the tank space varied depending on the height of rescence (Fig. 4C); (d) the inflorescence with the void the sample, producing a small area at the base and show- model (Fig. 4D); and (e) the distribution of paraclades ing a progressive increase of this area towards the top of in the whole system, allowing the visualization of the the sample (Fig. 5A). The relative area varied throughout entire branch system owing to transparency of the the sample, since the percentage of plant material was

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 418 F. M. NOGUEIRA ET AL.

Figure 3. μCT 3D reconstruction and light microscopy in inflorescences of Nidularium innocentii. A, lateral view of par- aclade-bearing distal branches; arrows indicate buds. B, lateral view of proximal branches, with buds (arrows). C–E, lon- gitudinal (C, E) and transverse (D) sections across a proximal branch, bearing second-order axes (Pc’). F and G, detail of longitudinal section of second-order axes, showing the vegetative meristem (vm) (F) and flower primordium (fp) (G). H, illustration summarizing the tank-inflorescence development: [a] internode shortage; [b] branches flattening in the axial plane, and shift from circular to elliptical outline in transversal section; [c] bract overgrowth, becoming larger and covering the flowering branches. (Pc) paraclade; (sl) subtending leaf; (b) bud. Scale bar = 1 mm (A–E); 500 µm (F, G).

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 TANK-INFLORESCENCE IN NIDULARIUM 419

Figure 4. μCT 3D reconstruction of inflorescences in Nidularium innocentii. A, lateral view of the entire system; bracts with axillary paraclades that bear flowers (in red). B, longitudinal section of the inflorescence. C, longitudinal view of the void model, corresponding to the empty space of inflorescence. D, inflorescence with the void model. E, paraclades perspec- tive in the system, allowing the visualization of all branch systems owing to transparency of the bracts. Scale bar = 1 mm. higher in the basal region, leaving a small tank space, growth (Mez, 1934–1935; Smith & Till, 1998). Lateral whereas towards the apex the relative area comprised a inflorescences occur less frequently (Smith & Downs, large percentage of empty space (tank) (Fig. 5B). 1974, 1979). In Bromelioideae, despite most genera having terminal inflorescences, both forms of develop- ment can arise (Mez, 1934–1935; Foster, 1945; Smith & Downs 1979; Smith & Till, 1998). In the context of DISCUSSION Troll’s typology (Troll, 1964, 1969), inflorescences of Inflorescences of Bromeliaceae are mostly terminal Nidularium constitute polytelic systems, since the and commonly pseudolateral because of sympodial main axis ends in a main florescence, i.e. a racemose

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 420 F. M. NOGUEIRA ET AL.

Figure 5. Absolute (mm2) and relative (%) area of the tank and the plant material spaces across the inflorescence of Nidularium innocentii.

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 TANK-INFLORESCENCE IN NIDULARIUM 421 system without a terminal flower. The branching been questioned by several authors (Jacques-Félix, system below the main florescence consists of para- 1988; Choob & Mavrodiev, 2001; Choob, 2002). When clades of first and second orders which replicate the present, the adaxial prophyll of monocots always main axis architecture, each ending in a coflorescence indicates the presence of a branch, but prophylls (Weberling, 1989; Rua, 1999). may eventually be lacking or obsolete (Jacques-Félix, During the development of polytelic systems, the 1988). Since the palea of Poaceae, long considered a main florescence is usually dominant (Weberling, prophyll, is currently regarded as an outer perianth 1989), with basipetal development of the paracladial whorl (Kellogg, 2000; Reinheimer & Kellogg, 2009), zone occurring, as described in N. innocentii. According the lack of prophylls in the floral axes seems to be a to Smith & Till (1998), an acropetal flowering sequence general feature among (Kircher, Kukkonen & occurs recurrently in Bromeliaceae. In fact, acropetal Müller-Doblies, 1983). Conversely, prophylls are usu- development occurs at the level of the florescences ally present in inflorescence branches (Reutemann (‘racemization’ sensu Sell, 1976, 1980), whereas the et al., 2012). In N. innocentii prophylls were never synflorescence (compound inflorescence) follows a observed in floral axes or in inflorescence branches, basipetal development. Such bi-directional develop- suggesting a general inhibition of prophyll primor- ment has been described elsewhere for Arabidopsis dia. In Vriesea, Costa et al. (2014) reported prophyll thaliana (L.) Heynh. (Hempel & Feldman, 1994) and occurrence ‘on the branch peduncle facing the main seems to be a widely generalized pattern among pol- axis’ of inflorescences, i.e. on the base of paraclades, ytelic synflorescences of angiosperms (Rua GH, our suggesting that the arrest of inflorescence prophylls in unpubl. data). axes other than the ultimate floral axes is not univer- There are few studies on inflorescences of sal among bromelioids. The question of whether such Bromeliaceae (Sideris & Krauss, 1938; Foster, 1945; a feature constitutes a synapomorphy of Nidularium Okimoto, 1948; Souza, Wanderley & Alves, 2008; Costa, and eventually of all nidularioid genera remains unan- Gomes-da-Silva & Wanderley, 2014). The inflorescence swered and deserves further research. in Aechmea Ruiz & Pav. was described as compound with spikes as basic units (Souza et al., 2008), rather than racemes as in N. innocentii. In Vriesea Lindl., a The tank-inflorescence – hypothesis on the typological approach was applied that showed simple evolution and role of water accumulation inflorescences (i.e. florescences) to be racemes, and Although the term ‘nidular inflorescence’ is common compound inflorescences (i.e. synflorescences) are het- in the literature in reference to the inflorescences erothetic double or triple racemes (Costa et al., 2014). of the nidularioid genera (Leme, 1997, 1998, 2000; Therefore, inflorescence patterns of Aechmea and Benzing, 2000), here we propose to call the inflores- Vriesea seem to be entirely similar to that observed cences of N. innocentii ‘tank-inflorescences’, consider- in N. innocenti, and internode length seems to control ing the morphology and function of such structures. the final morphology of the inflorescence. Indeed, it Inflorescences of Bromeliaceae can be recognized by is known that differential internode elongation has a their fleshy bracts, which are usually more colourful dramatic influence on the final aspect of inflorescences than the petals (Smith & Till, 1998; Benzing, 2000). that otherwise can follow identical branching patterns These bracts, bearing flowering branches in their axils, (Endress, 2010), and well-known inflorescence types seem to be slightly different in non-associated gen- such as racemes, spikes, umbels and capitula (i.e. basic era such as Aechmea, Billbergia Thunb., Canistrum florescences of polytelic systems) differ only because E.Morren, Guzmania Ruiz & Pav. and Vriesea. They of allometric differences in internode elongation are usually related to pollinator attraction, but are (Rua, 1999). also associated with the protection of floral primordia It is a general feature among angiosperms that (Gardner, 1986; Benzing, 2000). The tank-inflorescence each branch or flower is formed in the axil of a leaf. seems to be a syndrome associated with the simultane- These leaves are called subtending leaves or phero- ous occurrence of several processes, since the extraor- phylls (Briggs & Johnson, 1979; Endress, 2010). The dinary bract development to conform the tank is first leaves in a branch (usually one adaxial leaf in accompanied by non-elongation of internodes and the monocots and two lateral leaves in other angiosperms) dorsiventral flattening of paraclades. correspond to prophylls (Briggs & Johnson, 1979; As described above, the flowering branches of Weberling, 1989). Prophylls in monocots are typically Nidularium show a transversally flattened outline. adaxial and frequently two-keeled (Jacques-Félix, Both polystichous and distichous floral arrangements 1988). They have elsewhere been interpreted as dou- occur in bromeliad species (Smith & Downs, 1977, 1979; ble organs composed of two phyllomes (Turpin, 1819; Costa et al., 2014); in Nidularium the flower arrange- Rüter, 1918), although such an interpretation has ment along inflorescence branches is polystichous,

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 422 F. M. NOGUEIRA ET AL. although they may appear distichous because of the clade, including N. innocentii, referring the develop- flattened arrangement. ment of this feature to hypoxic conditions related to In Bromeliaceae, the water-impounding foliage water accumulation. makes the plant independent of the need to capture The use of typological criteria in inflorescence water through the roots and this enables species to analysis not only allows a correct interpretation, grow as epiphytes (Pittendrigh, 1948; Givnish et al., but also enables definition of morphological char- 2014; Silvestro et al., 2014). Water-impounding foli- acters that can be applied in comparative studies age has appeared at least three times in the evolu- with a phylogenetic perspective, and is important to tion of Bromeliaceae, in subfamilies Brocchinioideae, understand functional steps of the evolution of struc- Bromelioideae and Tillandsioideae. In bromelioids and tures. Tank-inflorescence development seems to have tillandsioids, this feature is associated with high rates occurred by a combination of many evolutionary pro- of diversification and appears to be linked to epiphyt- cesses, bract overgrowth, non-elongation of the inter- ism (Givnish et al., 2014). The reconstructed evolu- node and paraclade flattening. In Bromelioideae, the tion of the water-impounding tank in Bromelioideae tank-inflorescence has evolved in a few groups of shows that the common ancestor of the subfamily had core bromelioids, and may be associated with floral a tankless habit, a condition shared with the early protection. diverging lineages (Schulte, Barfuss & Zizka, 2009; Silvestro et al., 2014). On the other hand, a tank-form- ing habit involving vegetative structure appeared in ACKNOWLEDGEMENTS the core bromelioids and is shared by a large group in the subfamily (Silvestro et al., 2014). In other words, We thank the Laboratory of Plant Anatomy at the water-impounding foliage is present in a large group Federal University of Rio Grande do Sul (UFRGS), and in Bromelioideae (Silvestro et al., 2014), but only a few the Applied Nuclear Physics Laboratory at Londrina genera in core bromelioids share tank-inflorescences. State University (UEL) for technical support. This N. innocentii comprises terrestrial and epiphytic work was supported by the National Council for individuals, widely distributed in the Atlantic Forest Scientific and Technological Development (CNPq) and (Smith & Downs, 1979; Leme, 2000; Flora do Brasil, the Coordination of Improvement of Higher Education 2020). Its tank-inflorescence space volume was estab- Personnel (CAPES). lished here to be c. 2.4 times greater than that of plant tissues, indicating appropriate space to water reserve. The well-developed tank in Nidularium possibly rep- resents a differential significance for water provi- REFERENCES sion. Moreover, the tank-inflorescence of N. innocentii Acosta JM, Perreta M, Amsler A, Vegetti AC. 2009. The may play a protective role during floral development. flowering unit in the synflorescences of Amaranthaceae. Although not formally tested, it could be observed that Botanical Review 75: 365–376. in which the tank was kept dry during devel- Benzing DH. 2000. Bromeliaceae: profile of an adaptive radia- opment showed injuries in the final developmental tion. Cambridge: Cambridge University Press. stages of flowers (our unpubl. data). In Neoregelia Briggs BC, Johnson LAS. 1979. Evolution in the Myrtaceae L.B.Sm., maintenance during the development of sub- – evidence from inflorescence structure. Proceedings of the merged ovary and fruits may be involved in protection Linnean Society of New South Wales 102: 157–272. of these structures from herbivores (Benzing, 2000). Choob VV. 2002. On the problem of homologization of pro- Santos-Silva et al. (2017) discussed some aspects of phylls and cotyledons. Wulfenia 9: 73–76. the nidularioid inflorescence, concluding that this Choob VV, Mavrodiev EV. 2001. Morphological characters morphology probably evolved multiple times across of the leaf series in the Commelinaceae family with special the evolution of the tank-forming members of core emphasis on the number of prophylls and their homology in monocots. Botanichesky Zhurnal 86: 1–11. Bromelioideae. The water-impounding inflorescence of Costa AF, Gomes-da-Silva J, Wanderley MGL. 2014. Nidularium may represent a strategy for inflorescence Vriesea (Bromeliaceae, Tillandsioideae): taxonomic history, protection in humid forest environments, sprouting and morphology of the Brazilian lineage. The Journal of the out as a specialized life form in medium and low forest Torrey Botanical Society 141: 338–352. layers, as also reported for related nidularioid genera Dhondt S, Vanhaeren H, Van Loo D, Cnudde V, Inzé D. as well as for unrelated genera such as Guzmania, a 2010. Plant structure visualization by high-resolution X-ray member of Tillandsioideae (Leme, 2000). Additionally, computed tomography. Trends in Plant Science 15: 419–422. Nogueira et al. (2015) described the development of Endress PK. 2010. Disentangling confusions in inflorescence ovary aerenchyma in some species of the nidularioid morphology: patterns and diversity of reproductive shoot

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 TANK-INFLORESCENCE IN NIDULARIUM 423

ramification in angiosperms. Journal of Systematics and Morphology, Anatomy and Systematics. Acta Botanica Evolution 48: 225–239. Neerlandica 32: 370–371. Evans T, Jabaily RS, Faria APG, Sousa LOF, Wendt Leme EMC. 1997. Bromélias da Mata Atlantica: Canistrum. T, Brown GK. 2015. Phylogenetic relationships in Rio de Janeiro: Salamandra. Bromeliaceae subfamily Bromelioideae based on chloroplast Leme EMC. 1998. Bromélias da Mata Atlantica: Canistropsis. DNA sequence data. Systematic Botany 40: 116–128. Rio de Janeiro: Salamandra. Feder N, O’Brien TP. 1968. Plant microtechnique: some Leme EMC. 2000. Bromélias da Mata Atlântica: Nidularium. principles and new methods. American Journal of Botany, Rio de Janeiro: Sextante. 123–142. McDowell E, Trump BR. 1976. Histological fixatives for diag- Flora do Brasil. 2020 (in construction). Jardim Botânico do nostic light and electron microscopy. Archives of Pathology Rio de Janeiro. Available at: http://floradobrasil.jbrj.gov. and Laboratory of Medicine 100: 405–414. br/reflora/floradobrasil/FB6229. Accessed: 13 October Mez C. 1934–1935. Bromeliaceae. In: Engler A, ed. Das 2016. Pflanzenreich, Regni Vegetabilis Conspectus. Leipzig: Wilhem Foster MB. 1945. Lateral inflorescences in the Bromeliaceae. Engelmann. The National Horticultural Magazine 1: 14–22. Nogueira FM, Fagundes NF, Kuhn SA, Fregonezi JN, Gardner CS. 1986. Inferences about pollination in Tillandsia Mariath JEA. 2015. Ovary and ovule anatomy in the nidu- (Bromeliaceae). Selbyana 1: 76–87. larioid complex and its taxonomic utility (Bromelioideae: Gersterberger P, Leins P. 1978. Rasterelek­ Bromeliaceae). Botanical Journal of the Linnean Society 177: tronenmikroskopische Untersuchungen an Blütenknospen 66–77. von Physalis philadelphica (Solanaceae) Anwendung Okimoto MC. 1948. Anatomy and histology of the pineapple einer neuen Präparationsmethode. Berichte der Deutschen inflorescence and fruit. Botanical Gazette 110: 217–231. Botanischen Gesellschaft 91: 381–387. Oskolski A, Balthazar M, Staedler YM, Shipunov AB. Gerrits PO, Smid L. 1983. A new, less toxic polymerization 2015. Inflorescence and floral morphology of Haptanthus system for the embedding of soft tissues in glycol meth- hazlettii (Buxaceae, Buxales). Botanical Journal of the acrylate and subsequent preparing of serial sections. Journal Linnean Society 179: 190–200. of Microscopy 132: 81–85. Palombini FL, Kindlein W, Oliveira BF, Mariath JEA. Givnish TJ, Barfuss MH, Van Ee B, Riina R, Schulte K, 2016. Bionics and design: 3D microstructural characteri- Horres R, Gonsiska PA, Jabaily RS, Crayn DM, Smith zation and numerical analysis of bamboo based on X-ray JA, Winter K, Brown GK, Evans TM, Holst BK, Luther microtomography. Materials Characterization 120: 357–368. H, Till W, Zizka G, Berry PE, Sytsma KJ. 2014. Adaptive Pittendrigh CS. 1948. The bromeliad–Anopheles–malaria radiation, correlated and contingent evolution, and net spe- complex in Trinidad. I-the bromeliad flora. Evolution 1: 58–89. cies diversification in Bromeliaceae. Molecular Phylogenetics Reinheimer R, Kellogg EA. 2009. Evolution of AGL6-like and Evolution 71: 55–78. MADS box genes in grasses (Poaceae): ovule expression Gonzalez RC, Woods RE. 2008. Digital image processing, 3rd is ancient and palea expression is new. The Plant Cell 21: edn. Upper Saddle River: Pearson Prentice Hall. 2591–2605. Guarise NJ, Vegetti AC. 2008. Processes responsible of the Reutemann A, Lucero L, Guarise N, Vegetti AC. 2012. structural diversity of the Cyperaceae synflorescence: hypo- Structure of the Cyperaceae inflorescence. The Botanical thetical evolutionary trends. Flora 203: 640–647. Review 78: 184–204. Hayat M. 2000. Principles and techniques of electron micros- Rua GH. 1999. Inflorescencias, bases teóricas para su analisis. copy. Cambridge: Cambridge University Press. Buenos Aires: Sociedad Argentina de Botánica. Heller S, Leme EMC, Schulte K, Benko-Iseppon AM, Rua GH. 2003. Centothecoid grasses and the evolution of pan- Zizka G. 2015. Elucidating relationships in the Aechmea icoid spikelets. Plant Systematics and Evolution 240: 83–89. alliance: AFLP analysis of Portea and the Gravisia complex. Rüter E. 1918. Über Vorblattbildung bei Monokotylen. Flora Systematic Botany 40: 716–725. 110: 193–261. Hempel FD, Feldman LJ. 1994. Bi-directional inflorescence Santos-Silva F, Venda AKL, Hallbritter HM, Leme EMC, development in Arabidopsis thaliana: acropetal initiation of Mantovani A, Campostrini-Forzza R. 2017. Nested in flowers and basipetal initiation of paraclades. Planta 192: chaos: insights on the relations of the ‘nidularioid complex’ 276–286. and the evolutionary history of Neoregelia (Bromelioideae- Jacques-Félix, H. 1988. Les Liliopsida (ex Monocotyledones) Bromeliaceae). Brittonia 69: 133–147. n’ont pas de cotylédon. II. La préfeuille de la plantule: ses Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, rapport avec celles des axes feuillés. Bulletin du Muséum Longair M, Pietzsch T, Preibisch S, Rueden C, Saalfeld National d’Histoire Naturelle, 4e sér., 10, Section B, S, Schmid B, Tinevez JY, White DJ, Hartenstein V, Adansonia 3: 275–333. Eliceiri K, Tomancak P, Cardona A. 2012. Fiji: an Kellogg EA. 2000. The grasses: a case study in macroevolu- open-source platform for biological-image analysis. Nature tion. Annual Review of Ecology and Systematics 31: 217–38. Methods 9: 676–682. Kircher P, Kukkonen I, Müller-Doblies D. 1983. Spikelet Schulte K, Barfuss MH, Zizka G. 2009. Phylogeny of structure of graminoid monocots. 7th Symposium, Bromelioideae (Bromeliaceae) inferred from nuclear and

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424 424 F. M. NOGUEIRA ET AL.

plastid DNA loci reveals the evolution of the tank habit within Chevaliera (Bromeliaceae – Bromelioideae). Botanical the subfamily. Molecular Phylogenetics and Evolution 51: Journal of the Linnean Society 158: 584–592. 327–339. Staedler YM, Masson D, Schönenberger J. 2013. Plant tis- Sell Y. 1976. Tendances évolutives parmi les complexes inflo- sues in 3D via X-ray tomography: simple contrasting meth- rescentiels. Revue Générale de Botanique 83: 247–267. ods allow high resolution imaging. PloS One 8: e75295. Sell Y. 1980. Physiological and phylogenetic significance of the Stock SR. 2009. Microcomputed tomography: methodology direction of flowering in inflorescence complexes. Flora 169: and applications. Boca Raton: CRC Press. 282–294. Stuppy WH, Maisano JA, Colbert MW, Rudall PJ, Rowe Sideris CP, Krauss BH. 1938. Growth phenomena of pineap- TB. 2003. Three-dimensional analysis of plant structure ple fruits. Growth 2: 181–196. using high-resolution X-ray computed tomography. Trends Silvestro D, Zizka G, Schulte K. 2014. Disentangling in Plant Science 8: 2–6. the effects of key innovations on the diversification of Stützel T, Trovó M. 2013. Inflorescences in Eriocaulaceae: Bromelioideae (bromeliaceae). Evolution 68: 163–175. taxonomic relevance and practical implications. Annals of Smith LB, Downs RJ. 1974. Pitcairnioideae (Bromeliaceae). Botany 112: 1505–1522. In: Flora Neotropica Monograph. Vol. 14. New York: Hafner Tortosa R, Rua GH, Bartoli A. 2004. A typological analysis Press, 1–660. of the inflorescences of the genus Nassauvia (Asteraceae). Smith LB, Downs RJ. 1979. Bromeliaceae (Bromelioideae). Flora 199: 42–46. In: Flora Neotropica. Vol. 14. New York: New York Botanical Troll W. 1964. Die Infloreszenzen. Typologie und Stellung im Garden, 661–1492. Aufaubau des Vegetionskörpers, Bd. 1. Jena: Fisher. Smith LB, Downs RJ. 1977. Tillandsioideae (Bromeliaceae). Troll W. 1969. Die Infloreszenzen. Typologie und Stellung im In: Flora Neotropica. Vol. 14. New York: Hafner Press, Aufaubau des Vegetionskörpers, Bd. 2, 1. Stuttgart: Fisher. 1493–2141. Turpin PJF. 1819. Mémoire sur l’inflorescence des Graminées Smith LB, Till W. 1998. Bromeliaceae. In: Kubitzki K, ed. et des Cypérées. Paris: A. Belin. The families and genera of vascular plants IV. New York: Weberling F. 1965. Typology of inflorescences. Journal of the Springer. Linnean Society of London, Botany 59: 215–221. Souza GM, Wanderley MGL, Alves M. 2008. Inflorescence Weberling F. 1989. Morphology of flowers and inflorescences. architecture in Brazilian species of Aechmea subgenus Cambridge: Cambridge University Press.

© 2017 The Linnean Society of London, Botanical Journal of the Linnean Society, 2017, 185, 413–424

View publication stats