Siqueiranthus Leme, Zizka, E.H. Souza & Paule, 2022

Leme, Elton M. C., Zizka, Georg, Souza, Everton Hilo De, Paule, Juraj, De Carvalho, Jordano D. T., Mariath, Jorge E. A., Halbritter, Heidemarie & Ribeiro, Otávio B. C., 2022, New genera and a new species in the “ Cryptanthoid Complex ” (Bromeliaceae: Bromelioideae) based on the morphology of recently discovered species, seed anatomy, and improvements in molecular phylogeny, Phytotaxa 544 (2), pp. 128-170 : 149-159

Siqueiranthus Leme, Zizka, E.H. Souza & Paule , gen. nov.

Diagnosis:—This new genus differs from Cryptanthus by its sepals shortly connate at the base for 1/5–1/3 of their length (vs. connate for 1/2–2/3 of their length), petals free or subfree (vs. basally connate for 1/7–1/3 of their length), appendaged (vs. unappendaged but sometimes bearing well-developed lateral callosities), erect except for the suberect apex and forming a prevailingly tubular corolla (vs. arcuate-recurved at anthesis and forming an open, fan blade-like corolla), stamens unequal in length (vs. equal), anthers basifixed or subbasifixed (vs. dorsifixed), pollen with reticulum showing slightly flat muri and broader lumina (vs. muri typically thick with narrower lumina), sulcus covered with exine elements partially forming a net (vs. sulcus completely covered with exine elements forming a net), stigma simple-erect with tendency to simple-patent (vs. conduplicate patent), epigynous tube well-developed (vs. usually inconspicuous or lacking, rarely distinct), fruits longer and narrower [18–25 × 7–8 (–9.2) mm vs. 12–20 × 9–12 mm], seeds smaller [2–3.5 (–4) × 1.5–3 mm vs. 3.5–5 × 2.5–4 mm], with small and underdeveloped embryo (vs. differentiated fully developed embryo).

Type:— Cryptanthus cinereus D.M.C. Ferreira & Louzada, Systematic Botany 45: 460. 2020

[≡ Siqueiranthus cinereus (D.M.C. Ferreira & Louzada) Leme, Zizka, E.H. Souza & Paule ].

Distribution and habitat:—In the “Cryptanthoid complex” the geographical distribution proved to be important to strengthen taxa circumscription and clarify relationships, since monophyletic clades that reveal joint ecological preferences and/or sympatric ranges can in many cases be clearly correlated with morphological characters ( Leme et al. 2017b). Some of the obtained monophyletic clades often consist of species with similar ranges (and possibly origin), at least at the scale of co-occurrence within a given geographical area, e.g., a specific mountain chain.

The known population of Siqueiranthus in the county of Ibateguara, Alagoas state, northeastern Brazil, is situated at the confluence of the ranges of Cryptanthus , which prevails along coastal areas in Atlantic Forest related habitats, and Orthophytum with a more inland geographical distribution, more often associated with open, rocky outcrop environments, in Northeastern Brazil, north of São Francisco river (fig. 10). However, no sympatric occurrence was observed within any species of these three genera, despite the current strong habitat fragmentation and biological impoverishment observed at the type locality of Siqueiranthus just reveal a pale scenario of the local diversity in the present days.

The reported confluent occurrence in theAtlantic Forest domain is clearly in accordance with the close morphological relationship of Siqueiranthus and Cryptanthus ‒ it is part of the earliest branching “Cryptanthoid clade” and sister to Cryptanthus itself (fig. 1), and also in some aspect it shares with Orthophytum a given number of vegetative and floral characters (see topic “distinctive characters” below). On one hand, there is no correspondence between the range of Siqueiranthus and the geographic pattern of any other Cryptanthoid genus or subgenus (see distribution maps in Leme et al. 2017b, figs. 1–4 ). On the other hand, it was already noted that the distribution of Cryptanthus species seldomly overlaps with that of the remaining closely related Cryptanthoid genera ( Leme et al. 2017b).

Siqueiranthus is microendemic, known only from a small population at the type locality, in the county of Ibateguara. This place belongs to the microregion of “Serrana dos Quilombos”, east of Alagoas state, known to have one of the lowest temperatures in northeastern Brazil. The area belongs to a biogeographic unit, along with areas in the states of Pernambuco, Paraíba, and even Rio Grande do Norte, Ceará, and Piauí, known to be an important South American center of endemism, a “hotspot’s hotspot”, the so called “ Pernambuco Endemism Center”, or simply “Centro Pernambuco ” ( Tabarelli & Santos 2004, Silva & Casteleti 2005). The county of Ibateguara is one of the sites of the “Centro Pernambuco ” that have been tagged as epicenter of imminent massive species extinction ( Ricketts et al. 2005, Leme & Siqueira-Filho 2007). For instance, the endemic Aechmea marginalis Leme & J.A. Siqueira (2001: 149) , also found originally in an Atlantic Forest fragment in Ibateguara two decades ago, was never recollected despite of several attempts to locate it ( Leme & Siqueira-Filho 2006), which gives a real perspective of what will happen to Siqueiranthus in a very near future in a clear ongoing scenario of mass extinction, as reported by Pontes et al. (2016).

Etymology:—This new genus honors the biologist Prof. Dr. José Alves Siqueira Filho, of the Universidade Federal do Vale do São Francisco – UNIVASF, in Pernambuco, Northeastern Brazil, founder and Curator of the Herbarium HVASF, founder and executive director of the Centro de Referência para a Recuperação em Áreas Degradadas – CRAD, for Caatinga biome. Besides his expertise in taxonomy of Bromeliaceae , reproductive biology, and conservation of the Atlantic Forest and Caatinga biomes, and its flora and fauna, Dr. Siqueira has greatly contributed to the knowledge on the biodiversity of the states of Pernambuco, Alagoas, and neighboring regions in northeastern Brazil, and on the countless critically endangered species by means of his books, scientific articles, and lectures.

Distinctive characters:— Siqueiranthus has a general aspect that recalls a member of Orthophytum due to its habit, leaf conformation, texture, trichome coverage, and corolla structure. However, its andromonoecy, pollen, fruits and seeds clearly demonstrate its close relationship with Cryptanthus , justifying the original position of the taxon within this genus proposed by Ferreira & Louzada (2020). This close relationship is also highly supported by molecular data obtained here and reflected by its sister position to the whole clade of Cryptanthus . On the other hand, misinterpretation/overlooking of some important morphological characteristics indicates the need for a reassessment of the combination of its diagnostic features and species delimitation in comparison to Cryptanthus .

Ferreira & Louzada (2020) stated in the prolologue that C. cinereus propagates by axillary shoots and rhizomes 5–99 cm long, and that the propagation type is not a consistent character because it may vary within the same species, citing the synonym C. burle-marxii Leme ‒ originally conceived as a long-stolonifeous species ‒, in relation to C. zonatus ( Visiani, 1847: 4) Visiani (1854: 9) ‒ typically propagating by axillary shoots ‒ as previously proposed by Ferreira (2016) and also in Ferreira et al. (2021). These authors might have misinterpreted the long and thick stem of C. cinereus as rhizomes because the stem in this species may lay on the floor with age, mainly the proximal older portion, and generates some axillary shoots. Undoubtely, C. cinereus does not produce any rhizome or stolons, which raises questions about the denied importance of propagation type in species delimitation in this group.

Concerning flower morphology, it was not noted that prevailingly tubular corolla of C. cinereus is not observed in any Cryptanthus species, but characterizes other Cryptanthoid genera, like Lapanthus and Orthophytum . For instance, the thick petal appendages of C. cinereus , which are scutelliform with tendency to cupuliform, was interpreted by Ferreira & Louzada (2020) as conspicuous callosities, as well as its simple-erect (with tendency to simple-patent) stigma type ‒ a kind of stigma not present in Cryptanthus ‒ was considered to be conduplicate-patent (typical of Cryptanthus and Lapanthus ), minimizing the importance of useful morphological differences of this unusual taxon from all other Cryptanthoid members.

The highly supported sister position of Siqueiranthus to Cryptanthus in the phylogeny together with the number of distinctive morphological and anatomical characters form the basis for our decision to recognize it as a distinct new monotypic genus. A full description of the species is presented below based on living specimens that flowered at the type locality.

In the “Cryptanthoid complex” the presence of andromonoecy in Siqueiranthus is only shared with Cryptanthus , which immediately points towards their close parentage. However, the new genus differs from Cryptanthus by its distinctly caulescent habit (vs. usually stemless or when caulescent stem to 60 cm tall). The caulescent habit, however, is also observed in Hoplocryptanthus , Krenakanthus , Orthocryptanthus , Orthophytum subg. Capixabanthus , and Rokautskyia , despite Siqueiranthus has the longest (to 160 cm long) and thickest (to 2.8 cm in diameter) stem ever documented in the group (figs. 12 E, F). Besides, its sepals are shortly connate at the base for 1/5–1/3 of their length (fig. 13 E) when compared to the sepals of Cryptanthus , which are connate for 1/2–2/3 of their length. On the other hand, petals are free or subfree (fig. 13 F) as in Krenakanthus , Lapanthus , Orthocryptanthus , Orthophytum subgg. Orthophytum , Capixabanthus, and Clavanthus , and Sincoraea . In contrast, petals in Cryptanthus are basally connate for 1/7–1/3 of their length (fig. 13 I).

Petals of Siqueiranthus are erect except for the suberect apex at anthesis (figs. 13 C, D) and form a prevailingly tubular corolla [type II, according to Leme et al. (2017b)], like the corolla architecture of Lapanthus and Orthophytum (figs. 13 J, K), while Cryptanthus has an open, fan blade-like corolla of type I (cf. Leme et al. 2017b) formed by arcuate-recurved petals (figs. 13 G, H). It is important to highlight that petal appendages are scutelliform in this new genus (figs. 14 A–C), as in Orthophytum subg. Clavanthus , with a tendency to cupuliform as also observed in Krenakanthus (fig. 14 H) and Orthophytum subg. Capixabanthus . In contrast, Cryptanthus has unappendaged petals, sometimes with well developed lateral callosities (figs. 14 D, E).

Filaments in this new genus are unequal in length (vs. equal in Cryptanthus ), similar to Clavanthus , Krenakanthus , Orthocryptanthus , and Orthophytum . Anthers have a narrowly subtriangular shape before and at anthesis, with a pronounced bilobed base and an obtuse and inconspicuously apiculate apex (figs. 16 A–D). Size varies greatly before anthesis (to 5 mm long) and afterwards (2.5–3 mm long) due to dehydration during dehiscence, as usually observed in all members of the “Cryptanthoid complex” (see fig. 16).

Anthers in Siqueiranthus are considered basifixed or at least subbasifixed because filaments are fixed right at the base of the connective which distinctly separates the two thecae for 3/4 of their length (figs. 16 A, C). Despite being similar, the anthers of Cryptanthus are dorsifixed (figs. 16 F, I) or at least dorsifixed near the base but not basifixed (figs. 16 K, L) like in other members of the “Cryptanthoid complex”. Dehiscence is adaxial similarly to Cryptanthus (anthers not laterally compressed at anthesis), contrasting with the prevailing lateral dehiscence of the anthers of Orthophytum that become laterally compressed at anthesis (figs. 16 N, P).

In general, the morphology of the anthers has been neglected in the taxonomy of Bromeliaceae due to the lack of a detailed characterization in species/genera descriptions. However, in a pioneering study, Gardner (1986) associated filament and anther characteristics as the basis for the recognition of five morphological groups in Tillandsia . Later on, Leme et al. (2017a) used anther dehiscence (introrse or latrorse) and the connective exposure during dehiscence to distinguish species of the genera Waltillia Leme et al. (2017a: 29) and Vriesea Lindley (1843: 11) , with the anther dehiscence introrse, from Alcantarea (É. Morren ex Mez) Harms (1929b: 802) , with anther dehiscence latrorse. In a morphoanatomical study of some species of Dyckia Schultes & Schultes f. (1830: 1194), Carvalho et al. (2016) and Carvalho et al. (2017) highlighted the importance of anther characteristics in species delimitation, analysing, e.g. connective structure and thickening, anther dehiscence, and stamen morphometry. Recently, Büneker (2021) provided detailed illustration and descriptive morphological data of anthers of a complex of Dyckia species, reinforcing the utility of shape, dimension, apex structure, level of development, and influence of the connective in anther dehiscence in circumscribing species and group/complex of species.

In Bromelioideae, Leme et al. (2021) adopted length-width ratio of the anthers and the point of filament fixation to distinguish species in Hylaeaicum (Ule ex Mez, 1934: 40) Leme et al. (2021: 45), as well as documented an unusual prolongation structure at the distal end in one species. However, the use of anther characteristics in taxonomy in species/genera delimitatons greatly depends on the accumulation of detailed data, which is still rudimental, and remains challenging.

Pollen in Siqueiranthus is smaller when comparared to Cryptanthus (ca. 55 µm vs. 60–70 μm), bearing a reticulum with slightly flat muri and broad lumina (vs. muri typically thick with narrower lumina), sulcus covered with exine elements partially forming a net (vs. sulcus completely covered with exine elements forming a net), somewhat resembling pollen of Krenakanthus and Orthophytum subg. Capixabanthus (see fig. 12).

The key-importance of stigma morphology in generic delimitations and the recognition of monophyletic lineages was widely explored by Barfuss et al. (2016) and Leme et al. (2017b), but the existence of several intermediate structures and subtypes makes the use of stigma morphology in Bromeliaceae taxonomy challenging, requiring constant reevaluation mainly when the optimal condition for study is achieved. The precise interpretation of stigma morphology requires caution and some experience because the structures change during development ( Varadarajan & Brown 1988): compaction, lobe position, interwining, twisting and spiral folding and length, as well as lobe and margin curvature and presence of papillae ( Brown & Gilmartin 1989). Thus, determination of stigma type should ideally be based on flowers immediately before anthesis ( Brown & Gilmartin 1989), as was done here to identify the stigma type of Siqueiranthus as simple-erect with a tendency to simple-patent (figs. 15 D–F), differing from the conduplicate-patent stigma of Cryptanthus (figs. 15 A–C), which is also observed in Lapanthus . Simple-erect and simple-patent types were reported by Leme et al. (2017b) for Forzzaea , Orthocryptanthus (figs. 15 G–I), and Sincoraea .

The epigynous tube of Siqueiranthus is amazingly well-developed (fig. 18 A), varying from 5 to 8 mm long, while members of the “Cryptanthoid complex” usually lack an epigynous tube or present an inconspicuous one, rarely a distinct epigynous tube is found. However, the reexamination of some Cryptanthus species revealed a larger epigynous tube than previously reported by Leme et al. (2017b), but no longer than 3 mm (figs. 18 B–D).

Fruit size, shape, color, fragrance, and calyx length, and persistency proved to be important in species and genera circumscriptions and delimitations in Bromelioideae ( Leme 1997, 1998, 2000, Leme et al. 2017b, 2021), and the amazing fruit diversity in this subfamily was for the first time documented in detail by Leme et al. (2021). The large fruits of Siqueiranthus reinforce its close relationship with Cryptanthus , but a closer analysis reveals it has longer and narrower fruits (figs. 18 E, F) than Cryptanthus [18–25 × 7–8 (–9.2) mm vs. 12–20 × 9–12 mm (figs. 18 G–J)], despite some overlapping measures (value in parenthesis indicated in the protologue but not observed in the specimens we studied). In fruit, the sepals are persistent or the distal portion decays in Siqueiranthus (figs 18 E, F), similarly to Cryptanthus (figs 18 G–J), with remnants 6–7 times shorter than the fruit length. Differences in relation to Cryptanthus in this aspect are subtle, because their sepals always decay and the remnants are not more than 2–4 times shorter than the fruit length.

The first attempt to systematically use seed morphology on a large scale in Bromelioideae to reinforce genera cincumscriptions was done by Leme et al. (2017b, 2021). They adopted size, shape, number of seeds per fruit, coverage of an aril-like, whitish-translucent jelly substance, and presence of chalazal/micropylar monocaudate or bicaudate appendages as distinctive characters. In Siqueiranthus , seeds are few (10–20 per fruit) and considered large [2-3.5 (-4) x 1.5-3 mm] compared to the seeds of other genera of the “Cryptanthoid complex”, except Cryptanthus for which even larger seeds are reported (3.5–5 × 2.5–4 mm) (see fig. 13). The morphology and anatomy, however, revealed striking differences of the seeds of Siqueiranthus in relation to Cryptanthus from one side, as well as demonstrated important similarities with the seeds of Orthophytum from the other, like its underdeveloped embryo. The observed differences are fully addressed in the item discussion above.

A full redescription of Siqueiranthus cinereus , here provided as a descriptio generico-specifica [Shenzhen Code, art. 38.5; Turland et al. (2018)], is presented below.

Species:— Siqueiranthus is monotypic.

Figures

FIGURE 1. Maximum cladecredibilitytree of apost-burninBayesian analysis (15 million generations) based onnuclear (agt1, ETS, phyC), plastid (matK, trnH-psbA, ycf1 pos. 1113-2103, ycf1 pos. 4492-5440) as well as ecological and morphological data. Above the branches, Bayesian posterior probabilities (PP) and maximum-likelihood bootstrap support (BS) are shown (PP/BS). The scale bar below the tree shows the branch length for 0.005 substitutions per nucleotide position.

FIGURE 2. Maximum clade credibility tree of a post-burnin Bayesian analysis (15 million generations), based on nuclear and plastid data only. Bayesian posterior probabilities (PP) and maximum-likelihood bootstrap support (BS) are shown (PP/BS). The scale bar below the tree shows the branch length for 0.004 substitutions per nucleotide position.

FIGURE 3. Seed morphology. Seeds in lateral view (A, C–E, G–J, M, O, Q–T) and raphe view (B, F, H, L, N, P). A–B. Cryptanthus aff. beuckeri (Leme 3891). C. C. colnagoi (Leme 9154). D. C. pickelii (Leme 9665). E–F. C. bahianus (Leme 9616). G-I. Siqueiranthus cinereus (Leme 9785). G and I show two morphological variations for this species. J. Krenakanthus roseolilacinus (Leme 8922). K–L. Orthocryptanthus vasconcelosianus (Leme 8673). M–N. Orthophytum maracasense (Leme 9617). O–P. O. compactum (Leme 5604). Q. O. leprosum (Leme 7897). R. O. pseudovagans (Leme 6821). S. O. vagans (Leme 505). T. O. saxicola (Leme 6058). c = chalazal region or inconspicuous chalazal appendage. m = micropylar region. r = raphe. Bars = 0.5 mm.

FIGURE 4. Embryo morphology. A–L. Isolated embryo in lateral view of Cryptanthus, Orthophytum, and Siqueiranthus. A. C. aff. beuckeri (Leme 3891). B. C. aff. burle-marxii (Leme 9487). C. C. pickelii (Leme 9665). D. C. colnagoi (Leme 9154). E. C. bahianus (Leme 9616). F. S. cinereus (Leme 9785). G. O. saxicola (Leme 6058). H. O. maracasense (Leme 9617). I. O. compactum (Leme 5604). J. O. pseudovagans (Leme 6821). K. O. leprosum (Leme 7897). L. O. vagans (Leme 505). cn = cotyledonary node; cot = cotyledon; h-r = hypocotyl-radicle axis. s = suspensor. Bars = 250 μm.