﻿Erythrina L. (Phaseoleae, Papilionoideae, Leguminosae) of Brazil: an updated nomenclatural treatment with notes on etymology and vernacular names

﻿Abstract Erythrina L. is a genus that comprises ca. 120 to 130 species distributed throughout the tropics and subtropics of the world. Linnaeus established the genus in Genera Plantarum (1737) and the first binomial name given to a Brazilian Erythrina was E.crista-galli L., described by himself in Mantissa Plantarum (1767). Vellozo proposed in Florae Fluminensis (1790–1881) the first treatment of the genus in Brazil, where he treated three species from the states of Rio de Janeiro and São Paulo. Martins and Tozzi proposed the most recent treatment in 2018, where the authors recognized 11 valid names and presented three new synonyms. Despite extensive efforts already made in the genus, previous works did not treat all names related to the valid ones for Brazilian Erythrina. The present work is the most comprehensive and up-to-date nomenclatural treatment for the genus in Brazil, covering all 84 related names found on digital nomenclatural databases. Here we analyze 64 protologues, update typification statuses, propose five new synonyms, 13 new lectotypes (11 first-step, two second-step) and one neotype, linking all protologues and type specimens with their corresponding available digital sources, and make additional notes on etymology and vernacular names.


Introduction
Erythrina L. (Phaseoleae, Papilionoideae, Leguminosae) is a genus that comprises ca.120 to 130 species (Du Puy et al. 2002;Schrire 2005) distributed throughout the tropics and subtropics of the world in a wide range of habitats.The species can be found in arid tropical deserts or lowland alluvial vegetations to montane forests beyond 3 000 m of altitude, varying from rhizomatous perennial subshrubs to trees measuring more than 40 m high (Neill 1988).
Linnaeus established the genus in Genera Plantarum (Linnaeus 1737), expanded later in Species Plantarum (Linnaeus 1753), where he described three species (E.herbacea L., E. corallodendrum L., E. piscipula L.) and two varieties (E.corallodendrum var.occidentalis L., E. corallodendrum var.orientalis L.) based on plants mentioned earlier by other authors before the establishment of Linnean binomials.The type species of the genus was designated by Walpers (1853) as E. herbacea.The first binomial name given to a Brazilian Erythrina was E. crista-galli L., described by Linnaeus in Mantissa Plantarum (Linnaeus 1767).
The first treatment of the genus in Brazil was proposed by Vellozo's Florae Fluminensis, which was done in 1790 but the text was only partially published in 1829, the plates in 1831, and the complete text in 1881 (Vellozo 1829(Vellozo , 1831(Vellozo , 1881;;Carauta 1969Carauta , 1973)).In this work, the author treated three species from the states of Rio de Janeiro and São Paulo: E. corallodendrum (which was considered to be a mention of E. corallodendrum L. rather than a new publication of this name), E. verna Vell., and E. mediterranea Vell.
The most recent treatment of Brazilian Erythrina was proposed by Martins (2014) in a PhD thesis.In 2015, the author published a taxonomic treatment of the genus in Flora e Funga do Brasil (Martins 2023).The nomenclatural treatment was published later in Martins and Tozzi (2018), where the authors also recognized 11 valid names (E. amazonica, E. crista-galli, E. falcata, E. fusca, E. mulungu, E. poeppigiana, E. similis, E. speciosa, E. ulei, E. velutina, E. verna), designated 12 lectotypes and one epitype, and proposed three new synonyms: the previously accepted E. dominguezii as E. mulungu (which had

Results and discussion
From the nomenclatural search in digital databases, 84 names published in a total of 64 protologues were found and analyzed (Suppl.material 1).The nomenclatural revision resulted in the maintenance of 57 synonyms and confirmed the 11 currently valid names for Brazilian Erythrina.Here we propose five new synonyms, designate 11 first-step and two second-step lectotypes, and one neotype.Seven names remained as doubtful synonyms for several reasons explained in the species' commentaries: E. crista-galli L. var.corallina N.F.Mattos, E. crista-galli L. var. laurifolia Tod., E. crista-galli L. var. speciosa Tod., E. laurifolia Tod. and E. speciosa Tod. as potential synonyms of E. crista-galli L.; and E. atrosanguinea Ridl.and E. picta Blanco as potential synonyms of E. fusca Lour.These two accepted species represent the Brazilian Erythrina with most names given throughout the years, from 1741 to 2010, being E. crista-galli one of the most cultivated Erythrina species around the world since the 18 th century, and E. fusca the most geographically widespread species in the genus (Krukoff and Barneby 1974).Moreover, five names remained unplaced (E.indica sensu R.Vig., E. mediterranea Vell., E. moelebei Vieill.ex Guillaumin & Beauvis., E. secundiflora Brot., E. velutina Jacq.) and eight were excluded (Corallodendron nervosum (DC.)Kuntze, E. adansonii hort.ex Colla, E. argentea Blume ex Miq., E. compacta W.Bull, E. compacta W.Bull ex K.Koch, E. corallodendrum Vell., E. fusca Lour.var.inermis Pulle, E. nervosa DC.) for several reasons explained in the corresponding section.

Nomenclatural treatment
Notes.There are no nomenclature issues with E. amazonica, as the name was validly published and its type specimen was correctly cited (Fig. 1).However, all databases and studies of the genus state that the protologue was published in 1939.Still, according to the journal's website, the publication date is October 1938.
Etymology.The specific epithet "amazonica" is derived from Latin, meaning "from the Amazon", and it was chosen due to the species being native to the Amazon Forest domain.
(1) = Erythrina graefferi hort.ex Tineo, Cat. Pl. Hort. Panorm.: 278. 1827   Notes.The first known name for the species was published in Linnaeus (1767), who described it from a specimen collected by Vandelli in Brazil.Yet, he did not mention any collection number nor cite any herbarium.Krukoff and Barneby (1974) accessed an exsiccata from the herbarium LINN consisting of a drawing with the inscription "Vandelli" and a single flower that undoubtedly represents the species, considering this to be the original collection.However, this citation cannot be considered a lectotypification of the name, as the authors did not mention it as the type collection.In Flora of the Lesser Antilles, Howard (1988) correctly designated the LINN collection as the type specimen (Fig. 2).Walpers (1851Walpers ( , not 1850) ) published Micropteryx mentioning E. crista-galli as a synonym of M. crista-galli, but the genus was later synonymized into Erythrina in Engler and Prantl (1894).Kuntze (1891) published Corallodendron mentioning E. crista-galli as a synonym of C. crista-galli, but the genus was also synonymized under Erythrina in Engler and Prantl (1894).
(1) Jacquin ( 1768) published E. laurifolia with an illustration of a specimen from a botanical garden in Rome that is morphologically identical to E. crista-galli.Lozano and Zapater (2010) correctly designated this illustration as the lectotype, and the name has been synonymized under E. crista-galli since Lamarck (1786).
(3) Bentham ( 1849) published E. fasciculata from Brazil, mentioning two collections that can be considered syntypes of the name, one from the state of Minas Gerais (Regnell I.73) and one from Goiás (Pohl 763).Krukoff and Barneby (1974) recognized those syntypes but did not designate a lectotype, which was then correctly designated by Lozano and Zapater (2010) as the collection from Regnell in herbarium K.Additional material: P (P02934612, photo of K000674145; P02934614, photo of W0027159).(4) Todaro ( 1861) published E. pulcherrima based on a cultivated specimen at the Orto Botanico di Palermo (Italy).The original protologue could not be found online, but later publications by himself (Todaro 1863(Todaro , 1876) also cited this name, including a well-made illustration in 1876 that even mentions the similarity with E. crista-galli.Martins and Tozzi (2018) correctly designated his illustration in 1876 as the lectotype, but wrongly considered this publication as the protologue of the name, which is the publication in 1861.
(5) Backer (1908) published the variety E. crista-galli var.hasskarlii with a short description from cultivated plants on Java Island (Indonesia), but did not assign any type specimen.Three exsiccatae of a plant cultivated in the "Hortus Bogoriensis" [Kebun Raya Bogor] (Java) were found in herbarium L, with the inscription "Erythrina crista-galli Linn.var.hasskarlii Backer", that could represent the original material.One of them was designated here as the presumed lectotype, and the name has been synonymized under E. crista-galli since Krukoff and Barneby (1974).( 6) Lombardo (1964) published the variety E. crista-galli var.leucochlora from Uruguay with a short description, but did not assign any type specimen.As there are no evident characters that support this variety besides the white flowers, which is a common mutation found in cultivated Erythrina species (Guedes-Oliveira et al. manuscript in preparation), the name has been synonymized under E. crista-galli since Krukoff and Barneby (1974), and Lombardo's illustration in the same publication was designated here as lectotype.(7) Lozano and Zapater (2010) published the variety E. crista-galli var.longiflora from Argentina with a full description, an illustration and mentioning the type specimens.However, as there are no evident characters that support this variety besides the size of leaves and flowers, which is a character with well-documented morphological plasticity in Erythrina species (Guedes-Oliveira et al. manuscript in preparation), the name was here synonymized under E. crista-galli.
Etymology.The specific epithet "crista-galli" is derived from Latin, meaning "the crest of the rooster", and it was presumably chosen after the usual association of the corolla with the shape and color of chicken combs.

Type. Unknown.
Notes.As the original protologue for E. speciosa (Todaro 1860) could not be found, it is not possible to determine if this name was validly published and thus illegitimate, as it was already validly published by Andrews (1807), or it can be considered as a nomen nudum.Moreover, the protologue is cited on digital databases as being published in 1860, but according to TL-2 it was published either in 1859 or 1862.Todaro (1862) later published the variety E. crista-galli var.speciosa based on his previous E. speciosa, but made no additional informative description that could help confirm the identity of the mentioned specimen.Notes.As the original protologue (Mattos 1977) could not be found, nor any collection made by Mattos was found in any herbaria, this variety could not be confirmed as a valid name or synonym.(Martius s.n.), one from Bahia (Wied-Neuwied s.n.), and two from Minas Gerais (Claussen s.n. and Saint-Hilaire s.n.), but without mentioning any herbaria.Krukoff and Barneby (1974) accessed the collection by Martius in herbarium M and selected it as the type specimen.Because there are three different exsiccatae in the mentioned herbarium, Martins and Tozzi (2018) designated one of them as a second-step lectotype (Fig. 3).Despite not mentioning the herbaria of the other syntypes from Bahia and Minas Gerais, some collections are cited here with what was found in digital herbaria and believed to be the original ones.Kuntze (1891) published Corallodendron mentioning E. falcata as a synonym of C. falcatum, but the genus was later synonymized under Erythrina in Engler and Prantl (1894).Additional material: BR (BR0000013473808, photo n.v.), F (neg. 6301, negative of M0213337), IAN (IAN001758, photo of F neg. 6301), MO (MO-1680376, photo of F neg. 6301), P (P02951457, photo of M0213337).

Erythrina falcata
(1) Colla ( 1834) published E. martii as a doubtful species from Brazil with a short description that resembles E. falcata, but without mentioning any collections.Krukoff (1938) already placed it as a doubtful synonym of E. falcata, but did not see the type specimen, which he believed could be at herbarium TO.Moraes et al. (2013), in a treatment of Brazilian plants distributed by Martius in 1827 and published by Colla in "Herbarium Pedemontanum" (Piedmont, Italy), confirmed the synonymy and designated a collection in herbarium TO, where the original material was deposited, as the lectotype of the name.We got access to an image of the TO collection and agree with the synonymy.Furthermore, as the name E. martii had priority over E. falcata due to its date of publication, as specified in Article 11 of the ICN (Turland et al. 2018), Martins and Tozzi (2015) proposed to conserve E. falcata since it was a very well-established name for the species.The proposal was approved by the Nomenclature Committee for Vascular Plants of the ICN at the XIX International Botanical Congress (Applequist 2016;Turland et al. 2017;Wilson 2017).( 2) Spegazzini and Girola (1910) published the variety E. crista-galli var.inermis, which was doubtfully synonymized under E. falcata by Krukoff (1938), who did not see the type specimen.The synonymy was later confirmed by Lozano and Zapater (2010).Furthermore, Gutiérrez et al. (2002) investigated the type collections by Spegazzini in herbarium LP and found that only wood samples were collected for some species.As those samples could not be found anywhere in the LP collections, the holotype for this name is presumably lost.The protologue also could not be found online, but we had access to an image of the isotype's exsiccata at LIL and designated it here as the lectotype.
Etymology.The specific epithet "falcata" is derived from Latin, meaning "curved" or "sickle-shaped", and was presumably chosen due to its corolla falcate shape, especially the standard and keel petals.

Erythrina fusca
(1) Willdenow ( 1801) published E. glauca from Caracas (Venezuela), but did not mention any type specimen.A collection from Caracas labeled as E. glauca was found in Willdenow's type specimens' section in herbarium B with the same description given by him, and it was thus designated here as the lectotype.The name has been considered a synonym of E. fusca since Krukoff and Barneby (1974).Walpers (1851Walpers ( , not 1850) ) (1974), but as the authors mistakenly cited it as "E.fusca Lour.var inermis", the correct name is designated here as a new synonym.( 6) Rock (1920) published the variety E. fusca var.inermis from a specimen being cultivated in Hawaii after seeds brought from Manila (the Philippines), based only on the absence of spines, which is a character with well-documented morphological plasticity in Erythrina species (Guedes-Oliveira et al. manuscript in preparation).His photograph was designated here as the lectotype of the name, which has been considered a synonym since Krukoff and Barneby (1974).
Etymology.The specific epithet "fusca" is derived from Latin, meaning "dark" or "dusky", and it was presumably chosen due to the dark-orange color of the petals in some individuals, described as "fuſco-ruber" in the protologue of the species.It is important to point out that the color of the petals varies a lot in this species, from shades of light-yellow to dark-orange and even vinaceous-red (Guedes-Oliveira et al. manuscript in preparation).
Vernacular names.According to herbaria labels, E. fusca is generally known in Brazil as "mulungu", and also as "alecrim" in the state of Acre; "açacurana" (and spelling variations) or "assacu branco" in Amazonas; "assacurana" (and spelling variations) in Amapá; "eritrina-da-baixa" or "sumaúma" in Bahia; "abobinha" or "flor-de-aboboreira" in Mato Grosso; "abobreiro" in Mato Grosso do Sul; "assacuhy", "parica" or "pau angico" in Pará; and "assacurana" in Rio de Janeiro.Notes.Blanco (1837) published E. picta from the Philippines with a short description but without mentioning any type specimen.However, as Linnaeus (1763) had already validly published it before, it is an illegitimate name.The name was synonymized by Krukoff and Barneby (1974) as E. fusca, but could not be confirmed and thus remains here as a doubtful synonym.Notes.Ridley (1911) published E. atrosanguinea from Malaysia with a full description and type specimens, but without mentioning any herbaria.Krukoff (1939) considered this a synonym of E. fusca, but he could not see the type specimens.As Ridley described its flowers as "deep red black", and the type collections were not found on digital databases, it remains as a doubtful synonym.(Martius 1859).He included the name E. mulungu, given by Martius, citing a collection from the state of Minas Gerais but without mentioning any herbaria.Krukoff and Barneby (1974) located a collection by Martius in herbarium M and designated it as the type specimen.However, as there were two different exsiccatae in the mentioned herbarium, Martins and Tozzi (2018) correctly selected one of them as a second-step lectotypification (Fig. 5).Some databases give the authorship of the name only to Martius, but Bentham is the author of this species in Flora Brasiliensis.Kuntze (1891) published Corallodendron mentioning E. mulungu as a synonym of C. mulungu, but the genus was later synonymized under Erythrina in Engler and Prantl (1894).Additional material: F (neg. 6302, negative of M0240565), IAN001759 (photo of F neg. 6302).

Erythrina mulungu
(1) Spegazzini published E. chacoënsis in Spegazzini and Girola (1910), but the original protologue could not be found online.However, Lillo (1924Lillo ( [reprint from 1917]]) published Notas sobre el Herbario Venturi, mentioning the name and the original collection of Venturi in Argentina.An exsiccata from the same location was found at herbarium LP, which has the inscription "Erythrina chacoënsis Speg.(n.sp.)" and follows the description given by Lillo; hence it is believed to be the original material cited by Spegazzini.The name has been considered as a synonym of E. dominguezii since Krukoff (1938), who believed E. mulungu was a synonym of E. verna Vell., which was later found to be a different and accepted species by Martins and Tozzi (2018).( 2) Hassler (1922) published E. dominguezii mentioning three collections, one from Argentina (Jörgensen 3215) and two from Paraguay (Rojas 2061 and 2122), but without citing any herbaria.Lozano and Zapater (2010) then correctly designated the collection from Jörgensen as the lectotype of the name.The name was considered accepted until Martins and Tozzi (2018), who correctly synonymized it under E. mulungu.As for the remaining syntypes, some collections are cited here with what was found on digital herbaria and believed to be the original ones.Additional material: MO (MO-1624248, photo n.v.), P (P02934647, photo of GH00066286).( 3) Ducke (1922) published E. xinguensis from Brazil, with a full description and correctly citing the type specimen.However, as there were two exsiccatae in herbarium RB with the same number but without any information regarding a possible division of the same collection in different sheets (e.g.part 1 of 2), one of them was designated here as the lectotype.This name has been synonymized under E. ulei Harms since Krukoff (1938), and it caused much confusion in herbaria as the subsequent taxonomists could not properly identify the species.Ducke's collections, in fact, correctly represent E. mulungu (Guedes-Oliveira et al. manuscript in preparation) and the name is thus synonymized here.Additional material: F (neg. 2379, negative of B), IAN001764 (photo of F neg. 2379), MO-1680465 (photo of F neg. 2379).
Etymology.The specific epithet "mulungu" is derived from the most common vernacular name applied to all Erythrina species in Brazil.The origin is unknown, and it has many different meanings in African languages, most referring to a deity or a god creator of everything, father of all gods (e.g.Frankl 1990).
Vernacular names.According to herbaria labels, E. mulungu is generally known as "mulungu" in Brazil, and also "pau-de-tiriça" in the state of Minas Gerais; "abobreira" (and spelling variations) or "maleitoso" in Mato Grosso do Sul; "açacurana" (and spelling variations) in Pará; and "bico-de-papagaio" in São Paulo.Notes.Walpers (1851Walpers ( , not 1850) ) published Micropteryx with a new species from Peru (M.poeppigiana), which was based on Erythrina micropteryx, a name given by Poeppig that was never published.This mention also resulted in the publication of E. micropteryx itself, although it is considered a nomen nudum.Urban (1899) validly published the name E. micropteryx from Poeppig again.As the genus Micropteryx proposed by Walpers was later synonymized under Erythrina in Engler and Prantl (1894), Cook (1901) published E. poeppigiana based on M. poeppigiana by Walpers, which is the name that has been used for the species ever since.Krukoff (1938) designated the collection of Poeppig in herbarium B as the lectotype, but it was presumably destroyed in the bombing raid in 1943 (Botanischer Garten und Botanisches Museum Berlin 2022).Thus, Martins and Tozzi (2018) correctly designated the remaining available material in herbarium NY as the new lectotype (Fig. 6).Skeels (Galloway 1912) again published E. poeppigiana based on M. poeppigiana by Walpers, but as the name was already validly published by Cook (1901), it is considered illegitimate.Additional material: F (neg. 2373, negative of B), IAN001761 (photo of F neg. 2373), MO-1684973 (photo n.v.).

Erythrina poeppigiana (
(1) Spruce ( 1859) published E. amasisa from Peru, but did not mention any type specimen.Krukoff (1938) cited a collection from Spruce in herbarium K as the type, but as there were three different exsiccatae in the mentioned herbarium, one of them was designated here as a second-step lectotypification.
(2) Posada-Arango (1909) published E. pisamo from Colombia but did not assign any type specimen.It was subsequently correctly designated by Martins and Tozzi (2018) as his illustration in the same publication.The name was already synonymized under E. poeppigiana by Krukoff (1938).( 3) Standley (1916) validly published E. darienensis from Panama with a full description mentioning the type specimen, and the name was already synonymized under E. poeppigiana by Krukoff (1938).Additional material: P (P02951340, photo of US00004481), W (W19390013217, photo of US00004481).( 4) Steyermarkii and Lasser (1981) published the form E. poeppigiana f. redmondii from Venezuela based only on some specimens with yellow flowers, and correctly cited the type specimen.As no other known characteristics support the distinction between forms (Guedes-Oliveira et al. manuscript in preparation), nor is this variation reported from elsewhere, it is designated here as a new synonym.
Etymology.The specific epithet "poeppigiana" was a homage to Eduard Friedrich Poeppig (1798-1868), a German botanist, zoologist, and explorer who collected the type specimen attributed to the species.
Vernacular names.There are no other known vernacular names for E. poeppigiana in Brazil besides the commonly used "mulungu".Notes.There are no nomenclature issues with E. similis, as the name was validly published and the type was specimen correctly cited (Fig. 7).However, all databases and studies of the genus state that the protologue was published in 1939.Still, according to the journal's website, the publication date is October 1938.Additional material: P (P02960062, photo of MO-2050072).
Etymology.The specific epithet "similis" is derived from Latin, meaning "similar to", and was chosen due to E. similis similarity with E. amazonica, especially in dried specimens.
Vernacular names.There are no other known vernacular names for E. similis in Brazil besides the commonly used "mulungu".Notes.Andrews (1807) published E. speciosa, a name communicated by Lambert from a cultivated specimen in British greenhouses, with a short description and a well-made illustration of a leaf, inflorescence, and flowers.As Andrews did not mention any type specimen, only citing that the species is "supposed to be a native of South America", Martins and Tozzi (2018) correctly designated his illustration as the lectotype (Fig. 8).

Erythrina speciosa
(1) Brotero ( 1824) published E. poianthes from a cultivated specimen in the Jardim Botânico de Lisboa (Portugal), with a complete description and two very detailed illustrations of leaves, inflorescences, and dissected flowers, but without mentioning any type specimen.However, his description was published in the previous year by Tilloch and Taylor (1823) as a resumed version for the Proceedings of Learned Societies.This publication takes place due to the principles of priority as specified in Article 11 of the ICN (Turland et al. 2018), meaning the name published by Brotero is illegitimate.The name has been treated as a synonym since Krukoff (1938), and Martins and Tozzi (2018) correctly designated one of Brotero's illustrations as the lectotype.Hasskarl (1855) published Stenotropis berteroi based on E. poianthes by Brotero, but the genus was later synonymized under Erythrina in Engler and Prantl ( 1894).
(2) Lindley ( 1833) published the variety E. poianthes var.subnermis, with a short description and a well-made illustration of a leaflet and inflorescence from a cultivated specimen in a greenhouse in London (England).The variety was based solely on the absence of spines and the leafy habit in anthesis, which are characters with well-documented morphological plasticity in the species (Guedes-Oliveira et al. manuscript in preparation).The name was already synonymized under E. speciosa by Krukoff (1938).Lindley did not assign any type specimen, so Martins and Tozzi (2018) correctly designated his illustration as the lectotype for the name.
(3) Presl ( 1834) published E. reticulata from a specimen in Rio de Janeiro (Brazil), with a full description and a detailed illustration of leaves, inflorescence and dissected flowers, but without mentioning any type specimen.The name was already synonymized under E. speciosa by Krukoff (1938), but as he also did not designate any type specimen, Martins and Tozzi (2018) correctly designated Presl illustration as the lectotype.Walpers (1851, not 1850) published Micropteryx mentioning E. reticulata as a synonym of M. reticulata, but the genus was later synonymized into Erythrina in Engler and Prantl (1894).Kuntze (1891) published Corallodendron mentioning E. reticulata as a synonym of C. reticulatum, but the genus was synonymized under Erythrina in Engler and Prantl (1894) as well.(4) Mattos (1967) published the variety E. speciosa var.rosea from a cultivated specimen in the Jardim Botânico de São Paulo (Brazil), based on the pinkish color of its flowers.The original protologue could not be found online, but the type specimens were seen.As there are no other morphological characters to support this variety besides the corolla's color, which is a common mutation observed in cultivated specimens of E. speciosa (Guedes-Oliveira et al. manuscript in preparation), the name was already synonymized under E. speciosa by Martins and Tozzi (2018).
Etymology.The specific epithet "speciosa" is derived from Latin, meaning "handsome" or "splendid", and was presumably given by horticulturists who were amazed by the showy appearance of its flowers when the species was introduced to the United Kingdom.

Erythrina ulei
Etymology.The specific epithet "ulei" was a homage to Ernst Heinrich Georg Ule (1854Ule ( -1915)), a German botanist who collected one of the syntype specimens attributed to the species.

Erythrina velutina
Brazil.Pernambuco: Main island, scattered bushes near the village and in the Sapate.One full-grown tree in the cocoa-nut plantation at Sueste, [1887?],Notes.Willdenow (1801) published E. velutina with a complete description of a specimen from Venezuela, but without mentioning any type specimen.Martins and Tozzi (2018) then designated, from the Willdenow's type specimens in herbarium B, a collection from Humboldt composed of three sheets as the type of this name.However, as the authors mistakenly stated it as a holotype, the typification is corrected here to lectotype (Fig. 10).Walpers (1853Walpers ( , not 1854) ) published Chirocalyx mentioning E. velutina as a synonym of C. velutinus, but the genus was later synonymized under Erythrina in Engler and Prantl (1894).Kuntze (1891) published Corallodendron mentioning E. velutina as a synonym of C. velutinum, but the genus was also synonymized into Erythrina in Engler and Prantl (1894).Additional material: F (neg. 2378, photo of BW13100010), IAN (IAN001755, photo of F neg. 2378), MO (MO-1624337, photo n.v.;MO-1624338, photo n.v.).
(1) Desfontaines ( 1804) only mentioned the name E. aculeatissima in the Tableau de l'École de Botanique du Muséum d'Histoire Naturelle without giving any description, hence it is considered a nomen nudum.The name was considered doubtful by Krukoff (1938), but then was treated as a synonym of E. velutina by Krukoff and Barneby (1974), where the authors designated a collection in herbarium P as the type specimen.(2) Diels ( 1937) published E. splendida from Ecuador mentioning a collection from the province of Guayas in herbarium B (Diels 1230) as the type specimen, which was presumably destroyed in the bombing raid in 1943 (Botanischer Garten und Botanisches Museum Berlin 2022).No duplicate or any other collection by Diels could be found on digital databases, so a collection from the same locality was designated as the neotype.
(3) Ridley ( 1890) published E. aurantiaca from a specimen in the archipelago of Fernando de Noronha (Pernambuco, Brazil), with a full description and an illustration of a leaf, inflorescence, dissected flowers, fruit and seeds.The name was later synonymized under E. velutina f. aurantiaca by Krukoff (1938) based solely on a different coloring of the seeds of some specimens in the archipelago.As there were no other morphological characters to support this form, as the seeds vary in color both in the archipelago and on the mainland in Pernambuco and other Brazilian states (Guedes-Oliveira et al. manuscript in preparation), the name was synonymized into E. velutina by Martins and Tozzi (2018).
Etymology.The specific epithet "velutina" is derived from Latin, meaning "velvety", and was presumably chosen due to the abundance of trichomes in the species, especially on the petiole, abaxial leaflet surface, peduncle, pedicel, and calyx.

Erythrina verna
inflorescences and dissected flowers that could be easily mistaken for some other Erythrina species, Martins and Tozzi (2018) also correctly designated an epitype to represent the species better (Fig. 11).
(1) Herzog (1909) validly published E. flammea from Bolivia with a complete description and mentioning the type specimen.The name was already synonymized under E. verna by Krukoff and Barneby (1974).
Etymology.The specific epithet "verna" is derived from Latin, meaning "related to spring", and it was presumably chosen due to the association of the flowering period of the species to the beginning of the spring season in the state of Rio de Janeiro, Brazil (September).
Etymology.The specific epithet "fluminensis" was based on the Portuguese word "fluminense", which is derived from Latin meaning "inhabitant of river".It was presumably chosen as a homage to the denomination of people born in the state of Rio de Janeiro, Brazil, where the hybrid species was first cultivated in the country and still exists to the day of this publication.
Vernacular names.There are no other known vernacular names for E. × fluminensis in Brazil besides the commonly used "mulungu".
However, Lima (1995) mentioned this name as being authored by Vellozo and placed it as a synonym of E. speciosa.The description and illustration given by Vellozo undoubtedly depict the unique morphological features of E. speciosa, but according to Recommendation 50D of Chapter VI of the ICN (Turland et al. 2018), "misidentifications should not be included in synonymies but added after them", so we exclude the name here.Type.Brazil.São Paulo: "silvis mediterraneis transalpinis prope praedium Boavista", s.d., s.leg. s.n. (lectotype, designated here: illustration in Biblioteca Nacional Digital Brasil [mss1198656_107]; also in Vellozo 1831, tab. 103).

Erythrina mediterranea
Notes.The species was poorly described in Vellozo ( 1829) and the accompanying illustration in Vellozo (1831) depicts a sterile branch from a specimen growing in the municipality of Cunha, state of São Paulo (Pastore et al. 2021).According to Krukoff (1938), it could represent either E. crista-galli L. or E. falcata Benth.As it is impossible to even confirm it as Erythrina, the name remains unplaced.Notes.Colla (1834) mentioned E. adansonii Hortul as a doubtful synonym of E. crista-galli L. However, as no other valid publication or description of this name was found elsewhere, it was excluded here as a nomen nudum.Notes.Miquel (1855) mentioned E. argentea by Blume as synonym of E. ovalifolia Roxb.without any further information.As there was no other valid publication of this name elsewhere, it was excluded here as a nomen nudum.Notes.Bull (1871) briefly described the aesthetic characteristics of a cultivated plant named E. compacta by horticulturists.Later in the same year, Koch (1871) mentioned the specimen cited by Bull, but only repeated his description.It was later also described by Carrière and André (1882), who observed that the specimen was just a more compact cultivated variety of E. crista-galli.According to Article 38 of the ICN (Turland et al. 2018), a name cannot be effectively published using a description of purely aesthetic features, which is the case for all three publications related to this name, thus all of them are excluded here as nomen nudum.Notes.Krukoff and Barneby (1974) mentioned "E.fusca var.inermis" as a synonym of E. fusca, resulting in the publication of this name.However, as Pulle (1912) did not base his variety on E. fusca Lour.but on E. ovalifolia Roxb., the name is excluded here as misapplied.Notes.Guillaumin and Beauvisage (1914) cited E. moelebei by Vieillard in a list of species from New Caledonia, but without giving any additional information.

Erythrina moelebei
As there was no description of this name elsewhere and the type specimen cited (Vieillard 60) could not be found, the name remains unplaced.Krukoff (1939) treated it as a "hyponym" of E. fusca Lour.without seeing the type specimen as well.

Conclusions
The present work highlights the importance of comprehensive and detailed work regarding scientific nomenclature, showing that there are still many issues to address even in a relatively well-known genus like Erythrina.Despite the treatments proposed by Krukoff (1938) and Krukoff and Barneby (1974), and the more recent works by Martins (2014) and Martins and Tozzi (2018), in this revision it was possible to designate several new synonyms and type specimens for names related to Brazilian Erythrina.We also highlighted the knowledge gaps that remain and should be addressed in future works.Moreover, the ever-growing importance of virtual herbaria, online databases and digital libraries with up-to-date and reliable scientific information is noteworthy mentioning.Virtually all type specimens and original protologues analyzed in this work were accessed via digital resources, which must remain available to any user through any electronic device with a reasonable internet connection.

Erythrina crista-galli var. laurifolia Tod., Index. Sem. Panorm. 1861: 32. 1862
Jacquin (1768)l protologue for E. laurifolia(Todaro 1860)could not be found, it is not possible to determine if this name was validly published and thus illegitimate, as it was already validly published byJacquin (1768), or it can be considered as a nomen nudum.Moreover, the protologue is cited on digital databases as being published in 1860, but according to TL-2 it was published either in 1859 or 1862.Todaro (1862) later published the variety E. crista-galli var.laurifolia based on his previous E. laurifolia, but made no additional informative description that could help confirm the identity of the mentioned specimen.
Walpers (1851arneby (1974) E. glauca as a synonym of C. glaucum, but the genus was also synonymized into Erythrina inEngler and Prantl (1894).Additional material: F (neg. 2372, photo of B-W13101-010), IAN (IAN001757, photo of F neg. 2372).(2)Roxburgh(1814)mentionedE.ovalifolia from India, but did not describe the species, so this name was first considered a nomen nudum.However, he fully described the species in Flora Indica(Roxburgh 1832), although no type specimen was assigned.Wight (1839, not 1840) published a redrawing of Roxburgh's unpublished plates of species described in 1832, and his illustration for E. ovalifolia was designated here as the lectotype.The name has been considered a synonym of E. fusca sinceKrukoff and Barneby (1974).Walpers (1851Walpers ( , not 1850) )published Duchassaingia mentioning E. ovalifolia as a synonym of D. ovalifolia, but the genus was later synonymized under Erythrina in Engler and Prantl (1894).(3) De Candolle (1825) published E. patens based on a plate made by Sessé and Mociño for the Flora Mexicana, later published by Alph.De Candolle (De Candolle 1874).As stated by Krukoff and Barneby (1974), the species does not occur in Mexico and must have been collected somewhere in the Caribbean.The authors mentioned a collection by Sessé, Mociño, Castillo and Maldonado as the type, but did not mention any herbaria.Three exsiccatae of this collection were found in herbarium MA and one of them was designated here as the lectotype, in a second-step lectotypification.Kuntze (1891) published Corallodendron mentioning E. patens as a synonym of C. patens, but the genus was later synonymized under Erythrina in Engler and Prantl (1894).(4) Blanco's description (Blanco 1845) of E. caffra from the Philippines matches E. fusca, but as the name was already validly published by Thunberg (1800), Blanco's publication was considered illegitimate.Martins and Tozzi (2018) correctly designated his illustration in Flora de Filipinas 3 rd edn.(Blanco et al. 1883?) as the lectotype, but according to TL-2 (Stafleu 1976), both its publication date and plate number remain doubtful.(5) Pulle (1912) published the variety E. ovalifolia var.inermis from Indonesia based only on the absence of spines, a character with well-documented morphological plasticity in Erythrina species (Guedes-Oliveira et al. manuscript in preparation).The exsiccatae found in herbarium L undoubtedly place the name as a synonym of E. fusca.The variety was already synonymized in Krukoff and Barneby Engler and Prantl (1894)mentioning E. glauca as a synonym of D. glauca, but the genus was later synonymized under Erythrina inEngler and Prantl (1894).Kuntze Figure 4. Lectotype of Erythrina fusca Lour.(1790: 427), designated by Martins and Tozzi (2018: 399).Source: Missouri Botanical Garden -Peter H. Raven Library via Biodiversity Heritage Library, available at https://www.biodiversitylibrary. org/page/187502.(1891) published

in Du Puy et al., Legum. Madagascar: 516. 2002, non Lam., Encycl. 2: 391. 1786. Notes.
Puy et al. (2002)tioned by DuPuy et al. (2002)as a synonym of E. fusca Lour. in their treatment of Erythrina in The Leguminosae of Madagascar.The protologue is cited by the authors as being published in 1944, but according to BHP these publications by Viguier ranged only from 1927 to 1936.Moreover, some databases state that this name was used to refer to E. fusca Lour.As the original protologue could not be found, it remains unplaced.