Clarifying the nomenclatural history of Tovomitopsis, a Brazilian endemic genus of Clusiaceae

Abstract Tovomitopsis Planch. & Triana is a Brazilian Atlantic Forest endemic genus composed of two species: T.paniculata (Spreng.) Planch. & Triana and T.saldanhae Engl. An investigation was conducted to clarify the nomenclatural history of Tovomitopsis. We report the results of this investigation, provide an updated description of the genus, and propose lectotypes for T.paniculata and its synonyms: Tovomitafoliosa C.Presl and Tovomitapaniculata Cambess. We also propose lectotypes for T.saldanhae and for the new synonym Clusiaangustifolia Engl.


Introduction
Tovomitopsis Planch. & Triana is a Brazilian endemic genus currently composed of two species: T. paniculata (Spreng.) Planch. & Triana and T. saldanhae Engl. Both species occur in preserved remnants of Atlantic Forest in southeastern Brazil (Marinho 2021). Tovomitopsis was proposed in 1860 (Planchon and Triana 1860) as a replacement name for the illegitimate Bertolonia Spreng. (non Bertolonia Raddi 1820) and to accommodate presumably tetramerous flowered species. Together with Chrysochlamys Poepp. and Tovomita Aubl., these three genera were known by Planchon and Triana (1860: 225) as Les Tovomitées, being differentiated from each other especially by the arrangement of sepals on the floral bud: in Chrysochlamys and Tovomitopsis the outer sepals are smaller, exposing the inner sepals in bud, while in Tovomita the outer sepals are larger, covering the inner sepals and petals. Moreover, Planchon and Triana (1860) indicated that aril anatomy could be useful to differentiate, or at least, better circumscribe Les Tovomitées, but surprisingly this topic has not yet been further investigated.
Although some recent studies still indicate floral merosity as relevant to distinguish Tovomitopsis from Chrysochlamys (e.g. Hammel 2010), the latter includes species with four or five petals (e.g. Hammel 1999;Martínez y Pérez et al. 2015), and a clear morphological distinction between these two genera is yet missing. Taxonomic treatments and checklists carried out in Mexico (Martínez y Pérez et al. 2015), Central America (e.g. Hammel 1999), northern South America (Kearns 1998), where Chrysochlamys is distributed, and the Brazilian Atlantic Forest (Oliveira-Filho 2006) considered these two genera as congeneric.
Molecular phylogenetic evidence shows Tovomitopsis in a politomy with Dystovomita and the rest of Clusieae, and thus not very closely related to Chrysochlamys despite their gross morphological similarity (Marinho et al. 2019). Pollen morphology and aril anatomy (Planchon and Triana 1860;Hammel 1999;Stevens 2007;Marinho et al. 2019) have been suggested as promising to distinguish these two genera, but were so far gathered from only a few species of Chrysochlamys. The presence of resin glands in the anther dorsal region of Tovomitopsis could be a synapomorphy of the genus, and the absence of a pistillode in staminate flowers of Chrysochlamys could be also relevant to distinguish these genera (Bittrich and Marinho pers. com.).
Tovomitopsis consists of dioecious small trees or shrubs with prop roots and yellowish viscous exudate. The opposite leaves are petiolate, entire, chartaceous or coriaceous, with numerous closely arranged veins. The flowers have two pairs of sepals, the outer ones being smaller than the inner ones, and two pairs of whitish petals. Staminate flowers have yellow subclavate resiniferous stamens and a pistillode; pistillate flowers have staminodes similar to the stamens and a green-yellowish pistil with expanded stigmas. The fruits are green fleshy capsules that expose seeds with an orange vascularized aril when ripe (Bittrich 2003;Stevens 2007).
Although Tovomitopsis includes only two species, the genus has a long taxonomic history (see Hammel 1999), with several species floating among the three genera of Les Tovomitées (sensu Planchon and Triana 1860). The type species of the genus, Tovomitopsis paniculata, was described two hundred years ago, but a few nomenclatural issues remain to be addressed. Here, we clarify the nomenclatural history of Tovomitopsis, provide an updated description for the genus, and propose lectotypes for T. paniculata and its synonyms: Tovomita foliosa C.Presl and Tovomita paniculata Cambess. We also propose lectotypes for T. saldanhae and its new synonym, Clusia angustifolia Engl.

Material and methods
This study is based on the analysis of the protologues of Tovomitopsis names and some of its synonyms, on visits to historical collections in Europe (B, K, M, P, W) and the Americas (A, GH, NY, R, RB; herbaria acronyms according to Thiers 2021), and by analyzing specimens from virtual herbaria. Data on collectors and botanists were accessed in the Taxonomic Literature II website (Stafleu andCowan 1976-1988). All nomenclatural decisions follow the International Code of Nomenclature for algae, fungi, and plants (Turland et al. 2018).

Nomenclature and discussion
Tovomitopsis was proposed by Planchon and Triana (1860) as a replacement name for Bertolonia Spreng. [with just one species, B. paniculata, which was initially attributed to Chenopodiaceae (as "Chenopodieae")], a later homonym of Bertolonia Raddi (Melastomataceae). Along with the newly transferred T. paniculata they described five additional new species, all six of which they felt could be distinguished from Chrysochlamys and Tovomita. Bertolonia paniculata Spreng. was based on a pistillate specimen according to the illustration provided by the author (see Fig. 1, 1-4b). Sprengel (1821) did not mention either location, number or collector name for this collection.
In the "Flora Brasiliae Meridionalis", edited by Auguste de Saint-Hilaire et al., Jacques Cambessèdes (1828) used the same epithet "paniculata" when he published the new species Tovomita paniculata Cambess. This binomial is sometimes mistakenly interpreted as a new combination for Bertolonia paniculata Spreng. However, Cambessèdes is clearly indicated as the author of the Guttiferae monograph at the end of the treatment, and the † sign was used to indicate a new species throughout "Flora Brasiliae Meridionalis". The complete description of Tovomita paniculata that Cambessèdes (1828) provided included stamens, pollen grains and gynoecium, indicating that he studied both staminate and pistillate specimens. The protologue contains the following statement: "in sylvis primaevis propè vicum Aguassu, haud longè ab urbe Rio de Janeiro. Florebat Februario", and an illustration of a branch with many flowers, a feature found only in staminate specimens, while the illustration details show a pistillate flower and the detail of an ovary, as well as stamens with wellformed anthers (see Fig. 2). Cambessèdes (1828) did not indicate a type collection, but one specimen housed at P that was collected by Saint-Hilaire is probably the original material. The specimen P00093861 bears a label indicating the same location as in the protologue.
Presl (1834) proposed Tovomita foliosa C.Presl, as a new species and provided a detailed description and an illustration, but again without indication of the material he used. The author cited only "Habitat in Brasilia ad Rio de Janeiro". Although he described the species as a Tovomita, the pair of outer sepals not covering the inner sepals and other floral parts allow us to recognize a Tovomitopsis paniculata specimen in the  (1821) in "Neue Entdeckungen im ganzen Umfang der Pflanzenkunde II". illustration. In 1860, Planchon and Triana described Tovomitopsis and included, in addition to Tovomitopsis paniculata (Spreng.) Planch. & Triana (≡ Bertolonia paniculata Spreng.), five new species (currently placed in Chrysochlamys), and justified this decision based, in part, on the number of floral parts: while Tovomitopsis was circumscribed to include tetramerous flowers, Chrysochlamys retained the species with pentamerous flowers. In that work, the authors indicated two specimens for T. paniculata: "Brésil, Rio-de-Janeiro (Aug. de Saint-Hilaire; Sellow)".
In the "Flora brasiliensis", Engler (1888)  Febr. The author also described two new species in the genus: Tovomitopsis spruceana Engl., based on specimens from the Peruvian Amazon, and Tovomitopsis saldanhae Engl., based on specimens from Rio de Janeiro state.
The material used in the original description of Tovomitopsis paniculata was not mentioned or indicated by either Planchon and Triana or Engler, but the protologue and subsequent publications provided two important clues to find the type: i) both the description of the pistil and the illustration point to a pistillate specimen; ii) several points indicate that the type was collected in Rio de Janeiro state. For the locality of his Tovomita paniculata, Cambessèdes (1828) mentioned the village of Aguassu, which we believe to be Nova Iguaçu (before spelled as Iguassú), today a municipality in the metropolitan region of Rio de Janeiro city. Planchon and Triana (1860) cited Rio de Janeiro as the only locality of Tovomitopsis paniculata, and also all specimens cited by Engler (1888) came from the state of Rio de Janeiro. Both Planchon and Triana (1860) and Engler (1888) cited specimens collected by Friedrich Sellow and Auguste de Saint-Hilaire as the basis for their respective descriptions. Sellow collected in the state of Rio de Janeiro between 1814 and 1821 (Stafleu andCowan 1976-1988) and most of his specimens were deposited in B, but there are additional specimens originally from the Müller-Sprengel herbarium which were purchased by B in 1890 (Stafleu andCowan 1976-1988). We speculate that Kurt Sprengel, the German botanist who first described Tovomitopsis paniculata (as Bertolonia paniculata), likely studied one or more specimens collected by Sellow rather than the ones collected by Saint-Hilaire.
During a visit to European herbaria in 2016, a search at B was made for specimens assigned to the names Bertolonia paniculata, Tovomita paniculata, and Tovomitopsis paniculata, but none were found. Likewise, there are no Macbride negatives of such specimens in the Chicago Field Museum. Two relevant Sellow specimens were located in K (K001231050, image seen in Reflora Virtual Herbarium 2021), with the ♂ symbol indicating that it is a staminate specimen, and US (01882513) herbaria, both with floral buds; however, it is uncertain that these specimens are duplicates of a presumed type specimen housed in B. Therefore, we selected the illustration provided by Sprengel (1821) as the lectotype of Tovomitopsis paniculata, as this is the only unambiguous original material known to us. We also choose the illustration provided by Presl (1834) as the lectotype of Tovomita foliosa; and followed Planchon and Triana (1860) and Engler (1888) (Fig. 4) Type. lectotype (designated here), illustration in Sprengel (1821: Tab. I). (Fig. 1, 1-4b) Notes about Tovomitopsis saldanhae Tovomitopsis saldanhae was described by Engler (1888) based on specimens from the Serra dos Órgãos (Rio de Janeiro state) as "Habitat in Brasiliae provincia Rio de Janeiro, in Serra dos Orgâos ad Theresopolim: J. de Saldanha n. 6777, 6780, 6781, Glaziou n. 13576 in herb. Eichler". His description and the illustration point to the use of both staminate and pistillate specimens. In the same contribution, Engler also described Clusia angustifolia Engl. based on Saldanha 7335, which was collected in the same locality of T. saldanhae. The specimen (pistillate) clearly matches with specimens of T. saldanhae, especially the oblanceolate leaves with dark dots below, four petals and two series of resiniferous staminodes with subapical antherodes. As the specimen of Clusia angustifolia deposited in B was destroyed during World War II in 1943, we select a duplicate at R as the lectotype. Tovomitopsis saldanhae was later transferred to Chrysochlamys as C. saldanhae (Engl.) Oliveira-Filho, but Oliveira-Filho (2006) did not mention or indicate a type. Chrysochlamys saldanhae is now a synonym of T. saldanhae (Marinho 2021). Among the T. saldanhae syntypes, we chose A. Glaziou 13576 (P01901232, Fig. 5), which is the best preserved specimen housed at P, as the lectotype.    Note. Saldanha's collection numbers listed in a close numerical sequence may well belong to the same gathering (R. Forzza pers. com.). However, it is not possible for us to authenticate the preceding.
Regarding the other species of Tovomitopsis described by Engler (1888), Vesque (1893) transferred Tovomitopsis spruceana to Clusia (as C. trochiformis Vesque) and inadvertently proposed the lectotype by citing Spruce 4569 at the Boissier herbarium (now housed at G) as the type, but unfortunately this specimen was not found. Description. Small dioecious trees or shrubs with prop roots; axillary shoots with internodes regularly spaced from each other, grouped at the branch apex; exudates yellowish viscous on the branches and leaves. Leaves simple, opposite, decussate, petiolate; leaf blades chartaceous or coriaceous, light green in vivo, greenish to grayish in sicco, margin entire; venation simple brochidodromous, midvein prominent abaxially, flat adaxially; secondary veins slightly arcuate or straight, prominent abaxially, flat adaxially, forming angles between 40°-65° with the midvein; major secondary spacing generally regular; intersecondary veins parallel to major secondary veins, one per intercostals area; intramarginal secondary veins sometimes present. Inflorescences terminal, cymose, widely lax or congested, a single dichasium or a closed thyrse (the staminate more floriferous than the pistillate); bracteoles 2, triangular. Buds green, spheroid, apex rounded; sepals 2 pairs, green, decussate, base truncate, margin entire, apex rounded, outer pair smaller than inner pair and not enclosing the bud; petals 4, whitish, base truncate, margin entire, apex rounded. Staminate flowers with ca. 25 resiniferous stamens, filaments dorsiventrally compressed, yellow, sometimes the outer ones smaller than the inner ones, anthers lateral, yellow, thecae with longitudinal dehiscence, resiniferous glands present at the dorsal side of the anthers, pistillode inconspicuous. Pollen with general format in polar view subtriangular; isopolar, tricolporate, and with long ectocolpi; reticulate, and the semitectum not solid, but composed of twisted bacula. Pistillate flowers with staminodes similar to stamens; ovary 4-locular, 1 ovule per locule, styles 4, very short, distinct and persistent; stigmas 4, capitate, free from each other, persistent. Capsule pendant or straight on the branch, with 4 valves, epicarp smooth, green, mesocarp light red to purplish red. Seeds 1 per locule, each enclosed by a vascularized orange aril. Fig. 7.

Tovomitopsis and Chrysochlamys relationship
The phylogenetic relationship between Tovomitopsis and Chrysochlamys remained unknown until DNA sequence data became available for most genera of Clusieae. Taxonomic errors of attributing new Central American species to Tovomitopsis instead of Chrysochlamys (Maguire 1977) were likely due to characterizing the latter as having cauliflorous inflorescences, a condition only observed in Dystovomita (Engl.) D'Arcy among Clusieae. Moreover, Maguire (1977) almost certainly did not look at the type species of Chrysochlamys, i. e. C. multiflora Poepp., and it is doubtful that he analyzed species of Tovomitopsis from southeastern Brazil, including the type T. paniculata. A few years later, D'Arcy (1980) synonymized Chrysochlamys under Tovomitopsis, even though the latter is more recent (Hammel 1999).
Recent phylogenetic analyses of Clusieae demonstrated that Chrysochlamys is not sister to Tovomitopsis, but to Clusia (Gustafsson et al. 2007;Ruhfel et al. 2011;Marinho et al. 2019), and that floral resin evolved more than once in Clusieae and, possibly, even in Clusia (Gustafsson et al. 2007;Ruhfel et al. 2011). However, few species of Chrysochlamys were sampled in these studies, and phylogenetic relationships in this genus remain uncertain. Moreover, in two species of Chrysochlamys with resiniferous androecia (C. tenuifolia Cuatrec. and C. chrisharonii Vásquez & R. Rojas), male plants produce resin in a "capitulum" in the center of the flowers, where stamens are basally inserted (Hammel 1999;Vásquez Martínez and Rojas González 2009), while this structure is absent in flowers of Tovomitopsis. Further phylogenetic studies are needed to investigate if these Chrysochlamys species with the uncommon androecial morphology really belong to the genus if would be better placed elsewhere.

Final remarks
We present a brief contribution to the nomenclature of Tovomitopsis, a small endemic genus of Clusiaceae from the Brazilian Atlantic Forest. Except for the gross similarity with Chrysochlamys, the recognition of the genus is easy if associated with geographic distribution, as Chrysochlamys only occurs in Mexico, Central America and northern South America. Even so, further information on the geographic distribution and morphological limits between T. paniculata and T. saldanhae is needed. These two species are usually distinguished based on leaf size, shape and texture, but these features vary considerably both along an altitudinal gradient and from the coast to more inland sites. The presence and quantity of tiny blackish resinous glands on the abaxial surface of the leaves also should be considered and further investigated to distinguish species. An integrated taxonomic approach involving population genetics, geometric morphometrics of leaf outlines and classical taxonomy could shed light on species delimitations in Tovomitopsis.