Synonymies and typification of the Rhamnaceae of Brazil

In the course of preparing the taxonomic revision of the Rhamnaceae of Brazil for the Catalogue of Plants and Fungi of Brazil, several problems related to typification were detected. Taxonomic observations and nomenclatural notes are here reported based on the analysis of type material, as well as classic and recent collections, in 63 herbaria in Brazil and abroad, together with analyses of virtual herbaria. Types were sought, the majority of which were located and studied. Eleven synonymies, 37 lectotypes and one neotype are designated and defined for seven genera in the family.


Introduction
The family Rhamnaceae is distributed in tropical, subtropical and temperate areas.In Brazil, it occurs from north to south in forests as well as in open vegetation.The family is characterized by woody, herbaceous and scandent species with flowers with cucullate, convolute or conchiform petals, antipetalous stamens and a nectariferous disc internally enclosing the floral receptacle.The occurrence of 47 species within 14 genera, representing seven of the tribes in the family, have been reported for the Brazilian flora (Lima 2010).
The Rhamnaceae species occurring in Brazil were studied by Reissek (1861), who recognized 47 species in 12 genera.Subsequent studies undertook taxonomic treatments that were more comprehensive for some of these genera, such as Condalia (Johnston, 1962), Colubrina (Johnston, 1971), Scutia (Johnston, 1974), Rhamnus (Johnston & Johnston, 1978), Discaria (Tortosa, 1983) and Colletia (Tortosa, 1989).Lima (2000) produced and presented a taxonomic revision of the Rhamnaceae of Brazil, encompassing the diversity and occurrence of the family in different plant habitats throughout the country.That work and the studies developed thereafter involved analyses of initial descriptions (protologues) and type material of Rhamnaceae species, which led to the detection of nomenclatural problems, mainly regarding typification and synonymy, which are presented here.

Materials and methods
The nomenclatural revision presented here was based on studies of herbarium collections, of personal field collections throughout Brazil and of the specialized literature.The Rhamnaceae of Brazil were studied in the collections of the following herbaria: ALCB, B, BHCB, BM, BR, C, CEN, CEPEC, CESJ, CH, COL, CVRD, EAC, EAN, ESA, ENCB, FCAB, FLOR, G, GUA, HB, HRB, HRCB, HUEFS, IAN, IBGE, ICN, INPA, IPA, JPB, K, L, LIL, M, MAC, MBM, MBML, MG, MEXU, NY, OXF, P, PACA, PEUFR, R, RB, RBR, RFA, S, SI, SP, SPF, SPSF, TEX, U, UB, UEC, UFP, UPS, US, VEN, VIC, VT and W; which acronyms are in accordance with the Index Herbariorum (Thiers continuously updated).Studies were complemented with field collections by the authors, collections in virtual herbaria, as well as the original descriptions of the species.Exsiccata of the type material examined are indicated with an exclamation mark.Exsiccata of the type material indicated with †, deposited in B was destroyed.
Among the Rhamnaceae species occurring in Brazil, lectotypes were selected for those described by Reissek (1861).Except in specific cases, the selected lectotype was the specimen from W, where Reissek worked and produced the chapter on Rhamnaceae published in Flora Brasiliensis (Reissek, 1861).

Colletia exserta
= Colletia insidiosa Reissek, in Martius. Fl. bras. 11(1): 100.1861.Type: Brazil Meridional, Sellow 904 (holotype B †, lectotype BR! designated here).Syn.nov.Tortosa (1989) placed Colletia exserta in synonymy with C. paradoxa (Sprengel) Escalante, explaining that the species proposed was a morphological variant corresponding to the juvenile stage of C. paradoxa.Analysis of the type material of both species, together with observations of more recent collections available in Brazilian and Argentine herbaria, as well as from field collections, enabled the recognition of the two species.Colletia exserta is distinguished from C. paradoxa mainly by presenting narrow, delicate, curved ascending branches and tubular flowers, whereas in C. paradoxa the branches are broad, thick, straight and patent, even in young plants, and the flowers are urceolate.The lectotypification of C. exserta is required because the holotype in B was destroyed.
The proposal to place Colletia insidiosa in synonymy with C. exserta is supported by the analysis of the type material, which was collected by Sellow for both taxa, there being no significant differences between the two.As in the case of C. exserta, the holotype of C. insidiosa (in B) was destroyed, making its lectotypification necessary.Reissek,in Martius. Fl. bras. 11(1): 98. 1861.Type: Brazil, Bahia, Serra d' Açuruá, Blanchet 2832 (lectotype M! designated here; isolectotypes BM!, BR!, F!, K 2ex!, P 3ex!).

Crumenaria choretroides
The description of this species was based on the following syntypes: Brazil, Minas Gerais, Serro Frio, Jun, Martius s.n.(M!); Brazil, Goiás, Jul 1846, Gardner 3422 (BM!, BR!, G!, K!, OXF!, NY!, W!); and Brazil, Goiás, rio São Francisco, Pohl s.n.(W!, NY!).The material of Pohl s.n. is complete and clearly shows the characteristics of the species.In addition, it is deposited in the herbarium where the author of the species worked and therefore for these reasons it is selected as the lectotype.The collection of Martius 2479 in M is composed of two sheets, which were examined by Johnston in 1968.He labeled one as the holotype but made no annotation on the other.Here, the sheet considered by Johnston as the holotype was chosen as the lectotype.

Crumenaria decumbens
Crumenaria diffusa was known only from the holotype.No differences in relation to C. decumbens were found, particularly regarding the leaves and flowers, and we therefore propose placing C. diffusa in synonymy with C. decumbens.Analysis of the types of C. steyermarkii and of C. decumbens revealed the similarity between these taxa, both in vegetative as well as in reproductive characters, and it was impossible to draw any distinction between the two.Therefore, we propose that C. steyermarkii also be placed in synonymy with C. decumbens.This concept widens the Acta bot.bras.28(3): 376-381. 2014.area of distribution of C. decumbens, previously restricted to the southeast and northeast of Brazil, mainly in areas of caatinga (shrublands), and the species now presents disjunct distribution in Guatemala and Honduras.Reissek, in Endlicher. Nov. stirp. dec. Mus. Vind. 4: 28-29. 1839.Type: Brazil, Minas Gerais, rio São Francisco, Pohl s.n.(lectotype W! designated here; isolectotypes BR! [2 exs.],C!, G!, OXF!, W! [2 exs.]).Reissek (1839) described Crumenaria erecta based on the collection Pohl s.n.from Minas Gerais.The specimen deposited in W, where the author of the species carried out his studies, is composed of three sheets.The same collection is also found deposited in BR, C, G and OXF.The material in W is selected as the lectotype because it best fits the description of the species.Subsequently, when treating C. erecta for Flora Brasiliensis, Reissek (1861) also cited the material Lhotsky s.n., without indicating its provenance.The holotype of Crumenaria glaziovii deposited in B was destroyed, necessitating its lectotypification.The specimen found in R is complete and is perfectly diagnostic of the taxon.Reissek,in Martius. Fl. bras. 11(1): 113.1861.Type: Brazil, Paraná, campo, 1828, Sellow 4768 (holotype B †, lectotype US! designated here).

Crumenaria polygaloides
The collection of Sellow 4768 deposited in B was destroyed, requiring that Crumenaria polygaloides be lectotypified.Therefore, the duplicate found in US is designated as the lectotype.

Gouania acreana
The holotype of Gouania acreana deposited in B was destroyed.Therefore, one of the isotypes was used as the lectotype.Reissek (1861) did not cite the herbarium where the holotype of Gouania colurnifolia was deposited when he described the species.Therefore, the designation of a lec-totype is required.The material deposited in W, where the author of the taxon worked, was selected.Reissek in Martius. Fl. bras. 11(1): 107, tab. 26, fig. 4. 1861.Type: Brazil, Amazonas, "in silvis japurensibus", Martius s.n.(lectotype M! designated here; isolectotypes M!, W!).

Gouania cornifolia
When Reissek (1861) described Gouania cornifolia, he did not cite the herbarium where the holotype was deposited.The lectotype selected here is the collection of M and contains all of the information found in the protologue.Examination of the type material of Gouania corylifolia and G. petiolaris showed that there were no differences between the two taxa, both of which had been collected in the Rio de Janeiro state.The main difference was the smaller leaves of G. petiolaris.In the many collections of G. corylifolia examined, we observed different leaf sizes, all within the range of variation of the two taxa.Taking into account the law of priority, G. petiolaris is placed in synonymy with G. corylifolia.

Gouania hypochroa
Reissek (1861) described Gouania hypochroa without mentioning the herbarium where the holotype was deposited.Of the type material of this species, only one sheet was found in W. That sheet had been examined in 1969 by Johnston, who considered it as the holotype and recorded this information on the specimen.Here, that material is regarded as the lectotype.Reissek,in Martius. Fl. bras. 11(1): 109.1861.Type: Brazil, Minas Gerais, Martius s.n.(lectotype M! designated here; isolectotypes BR!, W!).

Gouania inornata
The specimen of Gouania inornata deposited in M was chosen as the lectotype because it is complete and because part of the collection of Martius for Flora Brasiliensis is deposited in that herbarium, where the author worked for many years after returning from Brazil.Although the specimen in W (the isolectotype) bears illustrations and handwritten annotations made by Reissek, it is composed of leaf fragments, flowers and dehiscent fruits.
Acta bot.bras.28(3): 376-381.2014.Reissek,in Martius. Fl. bras. 11 The specimen of Gouania latifolia deposited in W was examined in 1969 by Johnston, who considered it as the holotype, recording this information on the sheet itself, as well as regarding the material deposited in M as an isotype.

Gouania latifolia
Here, those specimens are recognized as the lectotype and isolectotype, respectively.
The specimen of Gouania chrysophylla examined in M was selected as the lectotype of the species.Gouania mollis was described by Reissek (1861) based on the following syntypes: Brazil meridional, Sellow s.n.; Brazil, Minas Gerais, Lhotsky s.n.; and Brazil, Minas Gerais, Ackermann s.n.The material of Sellow in B was destroyed.Of the Lhotsky material, only photographs were found (one in F and other one G), and the Ackermann collection was not located.Therefore selection of a neotype is necessary.The Martius material deposited in W was chosen as the neotype because it contained handwritten annotations and illustrations made on the sheet itself by Reissek, author of the species.According to the diagnosis of G. latifolia, the leaves are large, the laminas measuring 8.3-11.0× 6.8-9.6 cm and the petiole being 1.3-2.0cm in length, whereas in G. chrysophylla the leaves are smaller (laminas 6.8-8.3 × 4.0 cm; petiole, 1.3-1.4cm).Analysis of several specimens of G. chrysophylla, as well as type material, showed that the laminas can be 4.6-11.0× 3.5-8 cm and the petioles can be 0.5-2.5 cm in length.Therefore, G. chrysophylla is here placed in synonymy under G. latifolia.
Gouania latifolia and G. mollis have been treated as distinct species, and the two names are often both accepted as valid in floristic lists.Analysis of the original descriptions, along with studies of classic and recent collections, revealed that the features referred to as characteristic of Gouania latifolia overlap with those considered characteristic of G. mollis, suggesting that the description of G. mollis was based on juvenile material of G. latifolia, whose branches display only buds.Therefore, G. mollis is regarded as a synonym of G. latifolia in this work.Reissek,in Martius. Fl. bras. 11(1): 110.1861.Type: Brazil, Pará, Martius s.n.(lectotype W! designated here; isolectotype M! [2 exs.]).

Gouania trichodonta
Three sheets of Gouania trichodonta were found and examined in W. The one containing the illustrations and annotations made by the author of the species was selected as the lectotype.Hook. & Arn., in Hook. Bot. Misc. 3: 174. 1833.Type: Uruguay, Tweedie s.n.(lectotype K! designated here; isolectotypes K!, W!).
Although a search was made in different herbaria, only three sheets of the Tweedie s.n.collection of Gouania ulmifolia were found: two duplicates in K and one in the herbarium in W. One of the specimens in K is selected as the lectotype, because that is where the author of the species worked.
Gouania urticifolia was described based on the following syntypes: Martius s.n.(M!), from Minas Gerais; Sellow s.n.(G!, P!, W!) and Silva Manso 347 (BR!), from Cuiabá, Mato Grosso.The specimen of Martius s.n. in M was selected as the lectotype, because it is the most complete.

Gouania velutina
When Reissek (1861) described Gouania velutina, he did not mention the herbarium in which the holotype was deposited.The sheet in W is selected as the lectotype, because it best displays the characters contained in the description and is at the institute where the author of the species worked.Reissek,in Martius. Fl. bras. 11(1): 104.1861.Type: Brazil, Minas Gerais, Barra do rio das Velhas, Pohl s.n.(lectotype W! designated here; isolectotype W!).
Schomburgk 711 (BM!, G!, K!, W!), both collected in Guyana; and Friedrichsthat 646 (BR!, W!), collected in Nicaragua.One of the specimens of Pohl s.n.deposited in W is chosen as the lectotype, because it best corresponds to the protologue, contains original illustrations on the sheet itself and is at the herbarium where the author of the species worked.(holotype B †, lectotype P! designated here).Syn.nov.

Rhamnidium
Rhamnidium elaeocarpum was described based on the following syntypes: Pohl s.n.(no precise location); and Riedel s.n.(P!), Manso s.n. and Lhotzky s.n., all three collected in Cuiabá, Mato Grosso, the last two collections not seen.The Pohl material deposited in W is selected as the lectotype, because the author of the species worked at that herbarium.
Rhamnidium cognatum was a species known only from the original description and the holotype in B, which was destroyed.A duplicate of the holotype, found in P, was chosen here as the lectotype of the species.
When Reissek (1861) described Rhamnidium cognatum, based on sterile material, he regarded the species as close to R. elaeocarpum.A comparative study of vegetative characters of the types of both taxa indicated that they are in fact similar.There is an overlap between the two taxa in terms of characters such as length, width, number of veins and indumentum of the leaf lamina, as well as length of the petiole.Therefore, R. cognatum is here placed in synonymy with R. elaeocarpum.Reissek,in Martius. Fl. bras. 11(1): 95.1861.Type: Brazil, Rio de Janeiro, Lagoa Feia, Sep 1835, Luschnath s.n.(lectotype BR! designated here; isolectotype BR!).

Rhamnidium molle
When Reissek (1861) described Rhamnidium molle, he did not cite the herbarium where the holotype was deposited.Therefore, the lectotype selected here is the specimen from W, where the author of the taxon worked.
7. Ziziphus Mill.Triana & Planch. Ann. Sci. Nat. 5(16) Ziziphus cinnamomum was described based on the syntypes: Linden 1548 (P, BM, F, G, K, W), from Colombia; and Funk 794 (K!), from Valencia, Venezuela.Therefore, designation of a lectotype is necessary.After analysis of type material, the duplicate of the Linden collection in P, where Triana worked, was selected as the lectotype.That material had also been examined by Johnston, who regarded it as the holotype, in 1961.

Ziziphus cinnamomum
Analysis of the type material and more recent collections of Ziziphus cinnamomum and of Z. itacaiunensis revealed the close similarity of these species, mainly in the floral characters, the flowers being monochlamydeous, with a tricarpellate ovary and a velutinous indumentum, but also in the shape and size of the leaves.Because there are no morphological differences that would support maintaining a distinction between the two taxa, we propose placing Z. itacaiunensis (the more recently described species) in synonymy with Z. cinnamomum.Reissek,in Martius. Fl. bras. 11 Reissek (1861) described Ziziphus cotinifolia but did not mention where the holotype was deposited.Therefore, the specimen in W, where the author of the species worked, is here selected as the holotype.

Ziziphus cotinifolia
When Mansfeld (1927) described Ziziphus pseudojoazeiro, it is possible that he chose that epithet because he considered the species to be similar to Z. joazeiro Mart.Analysis of several specimens of Ziziphus cotinifolia collected in the semi-arid region of the Northeast, together with the study of type specimens of both species, revealed that the type of Z. pseudojoazeiro are young stems of Z. cotinifolia at the beginning of flowering.It is therefore acceptable to place Z. pseudojoazeiro in synonymy with Z. cotinifolia.
There is no information in the protologue of Ziziphus pseudojoazeiro as to where the holotype would be.Of the collection Luetzelburg 12396, only one duplicate was found, which was examined by Johnston in 1969, who designated it as the lectotype on that sheet; this view is being followed here.