Diatoms from Brazil: the taxa recorded by Christian Gottfried Ehrenberg

Abstract The flora of diatoms from Brazil has been studied by several authors from the beginning of the 19th up to now. Some of the old lists and descriptions are unknown or have been ignored by Brazilian researchers and the situation of the names cited was not assessed. Here we compiled a list of 101 taxa of diatoms from Brazil registered by Christian Gottfried Ehrenberg during the 19th century. We checked the current nomenclatural status of those taxa and lectotypified species from Brazil described by this author. For this, we accessed the Ehrenberg collection in the Museum für Naturkunde, Berlin, Germany, where 11 samples from Brazil studied by Ehrenberg are housed and published in different papers. Using these samples, we found 101 taxa (specific and infraspecific) published by Ehrenberg from Brazil. Five species (Eunotia bidens Ehrenb., Eunotia depressa Ehrenb., Eunotia elephas Ehrenb., Pinnularia microstauron Ehrenb., and Terpsinoe brasiliensis Ehrenb.) were new descriptions and were lectotypified here. The other species cited for Brazil were described initially from other places. However, 23 names were invalid and one illegitimate.


Introduction
Th e 19 th century was very important regarding the description of the biodiversity of algae and protists. At the fi rst haft of that century, material from diff erent parts around the world was analyzed by researchers from Europe and is today deposited in institutions on this continent. Christian Gottfried Ehrenberg was one of these early researchers (Lazarus and Jahn 1998). His collection comprises material from Africa, America, Europe, Asia and Oceania. However, the collection was mostly unavailable until 1990 (Lazarus and Jahn 1998). Th is inaccessibility and the absence of designated nomenclatural types of many taxa described by him, resulted in unclear taxonomic concepts of a number of these species." Brazil is among the regions from South America whose diatom fl ora was studied by Ehrenberg. Samples from the States of Amazonas, Rio de Janeiro, São Paulo, Minas Gerais and Santa Catharina were analyzed by Ehrenberg and the names of some species were published in fi ve publications (Ehrenberg 1839(Ehrenberg , 1841(Ehrenberg , 1843(Ehrenberg , 1851(Ehrenberg , 1854. Several of the taxa names published by him are in apparent disuse, and the current nomenclature of many of them is not known. Th is is true for the entire list of Ehrenberg's names. Th e validation of these names is guided by the ICZN (International Code of Zoological Nomenclature), as stated by Article 45.4 of the International Code of Botanical Nomenclature (ICBN) (McNeill et al. 2006), since diatoms were considered by Ehrenberg as polygastric animals.
Recently, Brazil is among the countries that have employed eff orts to repatriate data of plants collected by foreign researchers during 18 th , 19 th and 20 th centuries. Such eff orts have resulted positively in the creation of virtual herbaria (Zappi et al. 2012) and in the cataloguing of the fl ora thus increasing lists of species recorded in all its territory (Forzza et al. 2012). Several European collections have diatom material from Brazil (e.g. Ehrenberg, Grunow, Hustedt, Tempère & Peragallo, Krasske) which is still mostly unexplored. Such collections are very important for taxonomic studies which are the base to the knowledge of the biodiversity and, consequently, to the knowledge of tropical aquatic systems. Th us, this is the fi rst work that deals of repatriation of data about algae, specifi cally diatoms.
Th e aim of this study was to compile a list of taxa of diatoms from Brazil registered by Ehrenberg, to check the current nomenclatural status of those taxa, and to lectotypify specifi c and infraspecifi c taxa.

Material and methods
Th e Ehrenberg Collection (i.e., preparations, drawings and publications) was accessed at BHUPM (Museum für Naturkunde, Leibniz-Institut für Biodiversität-und Evolutionsforschung an der Humboldt Universität zu Berlin), where it is deposited and recorded under the following numbers of cases (Kästen) and card folders (Bücher): K. 18 B. 3-5, K. 19 B. 9, K. 52 B. 12, K. 53 B. 10;Jahn and Kusber (2004) have transcribed this for the use in databases; the material thus includes the following micas-trips: 180301-180316, 180401-180416, 180501-180516, 190901-190916, 521201-521216, 531001-531016. Th e data of the samples (i.e., original number of sample, locality, collector (leg.), date of sampling, study on which the analysis of such samples were published) are compiled in Table 1. Th e original number of sample is maintained by BHUPM as current control number.
Th e occurrence of the taxa cited for Brazil but originally described from other places than Brazil is provided in this study by the number of original samples (Table  1), except the sample from Tefé Lake (Ega-See) for which the Sample Index did not provide the number of the original sample and which will be referred to as Tefé Lake.

A brief history about the samples from Brazil
Ehrenberg never went to Brazil. However, he received samples from diff erent important persons of the history of such country. Th e main was Carl Friedrich Philipp von Martius (1794-1868), a researcher of the Brazilian fl ora, who sent samples from Coari (Amazonas State) to Ehrenberg (1839Ehrenberg ( , 1841Ehrenberg ( , 1843 (Ehrenberg 1843(Ehrenberg , 1854. On the other hand, the samples from Santa Catarina State were collected by Carl Pabst (1825-1863) who worked as a land surveyor in the old Dona Francisca Colonie (today Joinville city) and was an assistant of the botanist Franz Gustav Straube (1802-1853) (Straube 1992). However, these samples were sent to Ehrenberg by Carl Alfred Müller (1855-1907 (Ehrenberg 1854), aGerman bryologist.
Twelve samples from fi ve States from Brazil were sent to Ehrenberg (Fig. 1), but only eleven were cited by him in his studies, and one sample was not catalogued in the Sample Index (i.e., Tefé Lake, vide Table 1). Th e samples were stored in powder boxes (Figs 2-4) or old medicine boxes (Fig. 5) and are kept in the Museum für Naturkunde in a very good state. As well as the other Ehrenberg samples, the preparations of material from Brazil were made using mica discs, embedded with Canada Balsam (Lazarus and Jahn 1998).
From these samples, we cataloged 101 taxa (specifi c and infraspecifi c) published by Ehrenberg from Brazil, of which fi ve were new descriptions, 72 whose fi rst descriptions were made from other places than Brazil, and 23 were invalid and one illegitimate name.
Th e list of all taxon names is given below, and the species described initially from Brazil are here lectotypifi ed. Lectotype (designated here). Specimen in preparation 180404b, marked with yellow (g) ring, from sample 1099 "Rio Conigo in Rio de Janeiro", Ehrenberg Collection, in BHUPM (Museum für Naturkunde) (Fig. 6). Locality of the lectotype. Rio Conigo [Cônego River], Nova Friburgo, Rio de Janeiro, Brazil.
Locality of the syntypes. New-York, Andower "Conn." (Connecticut) Original description. "striata, ventre plano, dorso convexo medio exciso (bidentato), apicibus dilatatis truncatis. = E. diodon apicibus truncatis." Ehrenberg (1843, p. 373) cited the original locality as Brazil and USA (New York and Andower, Connecticut). In the material from New York (preparations 250401-250408, 260101-260112 and 260301-260308) we did not fi nd any reference to E. bidens, even under the name Himantidium that Ehrenberg (1843) related to the taxon on page 373. On the other hand, in the material from Andower (preparations 260201-260208) we found a reference to E. bidens in the preparation 260205d, marked with a blue ring, but it was not in a good condition to be photographed. Th erefore, we chose the material from Brazil as lectotype. Lectotype (designated here). Specimen in preparation 180403a, marked with yellow (g) ring , from sample 1099, "Rio Conigo in Rio de Janeiro", Ehrenberg Collection, in BHUPM (Museum für Naturkunde) (Fig 8).
Locality of the lectotype. "Rio de Concescao, Insula St. Catharina, Brasilien". Original diagnosis. "mit sehr kleinen Notenzeichen". T. brasiliensis was published the fi rst time by Ehrenberg in his book Mikrogeologie (1854). Several taxa of diatom published by Ehrenberg in this work are considered unavailable (invalid) according the Article 12 of ICZN (Ride et al. 1999) due to absence of a description, defi nition and indication of any illustration. Th is is not the case for T. brasiliensis and, maybe, could be the only case, in which Ehrenberg provided the following description "…und Terpsinoë brasiliensis, mit sehr kleinen Notenzeichen…" [and Terpsinoe brasiliensis, with very short musical notes]. Th is short description is considered to be enough by us, as well as several other descriptions published long ago by other authors (e.g., Agardh 1827).

First descriptions from other localities than Brazil
We provide a list of taxa cited by Ehrenberg as occurring in Brazil. Th is list is names based and has not been checked with respect to current taxonomy. Authors of combinations have been checked. [cited by Ehrenberg (1843Ehrenberg ( , 1854, sample 1087?] Gallionella procera Ehrenb., Abh. Königl. Akad. Wiss. Berlin, 1847: 270, 273, 279, 285, 287, 293, 299, 303, 317, 319, 396, 399, 442, 443, 445, 447, 448, 449, 450, 451, 452, 453, 455, 456, 458 , every new name published before 1931 must satisfy the provisions of Article 11 and must be accompanied by a description or a defi nition of the taxon that it denotes, or by an indication" (Ride et al. 1999). Subsequently, one of the meanings of the term "indication" and highlighted by us is the Article 12.2.7: "the proposal of a new genus-group name or of a new speciesgroup name in association with an illustration of the taxon being named, or with a bibliographic reference to such an illustration". Such conditions are not fulfi lled by some names published by Christian Ehrenberg.

Amphitetras antediluviana
Th e names of diatoms listed below are those cited specifi cally by Ehrenberg (1841Ehrenberg ( , 1843Ehrenberg ( , 1851Ehrenberg ( , 1854 from Brazilian samples and were considered by us as unavailable (invalid) (for cross code terms see David et al. 2012) because they do not fulfi ll Article 12 of the ICZN (Ride et al. 1999). On the other hand, in the case of Navicula obtusa Ehrenb. the name is available (valid), but illegitimate because it is a later homonym.

Discussion
Biological collections represent a set of organism that lived in some locality during some time. Th us, historical collections are a portrait of the biodiversity of some place that can never be faithfully restudied. Th is is due to natural or anthropogenic factors responsible for changes of biological diversity of an environment in short, medium or long terms, (Jackson and Sax 2009). Th erefore, historical collections of species are indispensable sources to study the biodiversity on earth (Shaff er et al. 1998).
Besides their scientifi c importance, such collections also perform a fundamental cultural role regarding the region in which the material was collected and for the development of science along the centuries. Th rough Ehrenberg's studies, for instance, it can be inferred which pathway some botanists such as Carl Friedrich Philipp von Martius, Karl Sigismund Kunth, Eduard Friedrich Poeppig, Édouard Louis Chavannes, and other personalities such as the Carl Pabst or Hermann Encke, took when traveling Brazil in the 19 th century.
Th e cultural and scientifi c importance of the Brazilian samples studied by Ehrenberg can be found in its historical context. At the beginning of the 19 th century, many regions in Brazil were occupied by natives whose social organization was diff erent of other native people from South America. In 1818, Prince Regent D. Joao VI of Portugal started an immigration policy that resulted in the arrival of 1,458 Swiss immigrants between 1819 and 1820. Th ese immigrants established themselves on the Cantagalo Farm and this settlement was known as Nova Friburgo. In 1824, Nova Friburgo received about 450 new people, this time German immigrants. Th us, it can be estimated that approximately 1,900 people lived in this region. In 1907, the Instituto Brasileiro de Geografi a e Estatística already estimated about 19,185 inhabitants and in 2011 there were estimates of 182,748 inhabitants (IBGE 1912(IBGE , 2011. In between, in 1850, Hermann Encke sent two samples to Ehrenberg. Although we have no accurate estimates of the population from Nova Friburgo in 1850, it is clear that the impact on biodiversity generated by less than 19,185 inhabitants (in 1907) is lower than the impact of the current 182,748 inhabitants. Th is example can be translated to the other samples collected in Brazil until the fi rst half of the 19 th century. Even if it refers to only few places, these materials may be the most accurate examples of the biodiversity of Brazilian diatoms. Moreover, historical samples like those can help to understand anthropogenic eff ects on the biodiversity of diatoms in tropical and subtropical regions.
Th e list of Brazilian diatoms published in Ehrenberg's studies allowed also to discuss the taxonomy and nomenclature of some diatoms described in the 19 th century. Although the taxonomy of the diatoms published by Ehrenberg are ruled by the ICZN, the conditions required for a taxon to be considered available (valid) are practically the same as those of the ICBN, that is, a description, a defi nition (diff erential diagnosis) or an indication (which can be a published illustration) are needed. Th is last issue justifi es, for instance, to ensure the availability of species such as Bacillaria australis, Gallionella crenata, G. procera, Pinnularia amphirrhina and P. vespa which have no diagnosis, but indications to an illustration was provided. Some unavailable names cited by Ehrenberg continued to be ruled by the ICZN which were available by him in later studies. Th is happened to Eunotia sphaerula and Pinnularia vespa which were cited by Ehrenberg (1854) and Ehrenberg (1851), respectively, but were made available by Ehrenberg (1870) and Ehrenberg (1854), respectively. Other species cited by Ehrenberg from Brazilian samples became ruled by the ICBN when validated by authors who treated diatoms as algae. Th is was the case withE. nonaria, E. octonaria and E. senaria, validated by Rabenhorst (1864), and G. longiceps validated by Ralfs (Pritchard 1861). Internal evidences enabled us to identify the author of the taxa by "Ehrenb. ex", linking them to the types from the Ehrenberg material.
Similarly, Navicula obtusa Ehrenb. (non Navicula obtusa Bory), the only illegitimate name among the 101 names published by Ehrenberg from Brazilian samples, was legitimated by Rabenhorst (1864, p. 197) with the new name Navicula appendiculata f. obtusa. Th is taxon was associated to "Ehrenb. (Verb. p. 131)". Th e term "Verb." mentioned by Rabenhorst (1864) referred to the paper Verbreitung und Einfl uss des mikroskopischen Lebens in Süd-und Nord-Amerika, published by Ehrenberg (1843). Th e paper that we had access has 154 pages, betweem the page 291 to 445 and, therefore, Rabenhorst (1864) referred to a page that does not exist on the paper that we analyzed, but maybe referes to a reprint version that we had not access. A similar fact is seen with Eunotia bidens Ehrenb. and Gomphonema cygnus Ehrenb. which were cited by Rabenhorst (1864, p. 74 and 286, respectively) and related to "Ehrenb. (Verb. p. 125 …)" and "Ehrenb. (Verb. p. 128 …)". Th e diff erence between the pages cited by Rabenhorst (1864) and the correct page on Ehrenberg (1843) in both cases is 288 pages. Th erefore, we can tell that Rabenhorst (1864) associated Navicula appendiculata f. obtusa with Navicula obtusa in Ehrenberg (1843, p. 419), the only taxa with the epithet obtusa on the page 419, even though he had not mentioned clearly N. obtusa.
According to the ICBN, an epithet of a latter homonym can be used in a diff erent combination and on a diff erent rank, if the epithet is available at this rank (Article 58.1, McNeill et al. 2006). In this case, the name is treated as new. Th us, the authorship published by Rabenhorst (1864) should be N. appendiculata f. obtusa Rabenh. and not N. appendiculata f. obtusa (Ehrenb.) Rabenh., which is in agreement with the Example 1 of the Article 58.1 of the ICBN (McNeill et al. 2006).
Another nomenclatural issue is about valid descriptions or defi nitions, which we have considered to defi ne valid publication of some diatoms published by Ehrenberg. Some taxa published in the 19 th century showed very short descriptions or diagnoses. Th is led us to consider the expression "…und Terpsinoë brasiliensis, mit sehr kleinen Notenzeichen…" provided by Ehrenberg (1854) as defi nition (diff erential diagnosis) of T. brasiliensis in relation to T. musica.
On the other hand, species such as Navicula gracilis, cited by Ehrenberg (1854) for Tefé Lake, and other species such as Gomphonema discolor, Navicula fusiformis, N. gibba, N. turgida, and N. uncinata, which were not cited for Brazil, showed one particularity. Th ey were defi ned initially by Ehrenberg (1832a) by only a range of measures of specimens from diff erent localities and can thus not be considered a valid description. According to the ICZN, description is "a statement in words" (Ride et al., 1999) and, therefore, only measurements do not constitute a description of some taxon. Th ereby, N. gracilis as well as Gomphonema discolor, Navicula fusiformis, N. gibba, N. turgida and N. uncinata have to be considered available (valid) only according to Ehrenberg (1832b), when they were provided with a valid description. Th is implies that the attribution of the author of N. gracilis to Ehrenberg (1832a), as found in Algaebase (Guiry & Guiry, 2012) is not correct and should be changed to Ehrenberg (1832b).
Despite similarities among the ICZN and ICBN of a valid name, one diff erence must be highlighted: the independence of the validity of the name of the genus and the name of the species. In the ICZN, a species name is available even if the genus name is not available (Article 11.9.3.1, Ride et al. 1999) while according to the ICBN this condition is not allowed (Article 43.1, McNeill et al. 2006). Th is refers to Discoplea comta (Silva 2003), a name cited for Brazil. However, we notice problems about the original sense of this species and its concept (Silva 2003, Kusber andJahn 2009). Nagumo (2003) also noticed a similar condition concerning Amphora libyca when carrying out lectotypifi cation of this species. Th us, the two cases showed that the names listed by Ehrenberg for Brazil do not necessarily correspond to the current concept of the taxa and highlight the need for future studies that reinvestigate the preparations and/or samples from Brazil and compare the types with the names cited according to Ehrenberg' concept. Even considering these limitations, it is noteworthy that only about 43% of the valid names recorded by Ehrenberg (1839Ehrenberg ( , 1841Ehrenberg ( , 1843Ehrenberg ( , 1851Ehrenberg ( , 1854 and compiled here were reported for Brazil by other researches. Th e other 57% include taxa of Eunotia, Pinnularia and other genera that even under their synonyms were not recorded for Brazil, which corresponds to 44 names. Th e 77 diatom names cited by Ehrenberg (1839Ehrenberg ( , 1841Ehrenberg ( , 1843Ehrenberg ( , 1851Ehrenberg ( , 1854 make him the most important diatomologist of Brazilian diatoms in the fi rst half of the 19 th century.

Conclusion
We catalogued 101 taxa recorded by Christian Gottfried Ehrenberg in a set of fi ve studies, among which 77 are available (valid) names, 24 names are unavailable (invalid) and one is illegitimate. Th e reason why these 24 names are considered as invalid is the absence of a description or a defi nition (diff erential diagnosis) or the indication of an illustration. Among the 77 valid names, fi ve were originally described for Brazil and are here lectotypifi ed. Only 34 taxa of this list had been recorded until now by other studies than Ehrenberg's. Th e other 57% of the valid taxa can be treated as fi rst citation of the name for the Brazilian diatom fl ora.
In the future, studies should be conducted to characterize all specimens of Brazilian diatoms described by Ehrenberg. Th is could be carried out through the sampling and analysis of recent material from the original locality of the lectotype.
With this compilation, we can go back to the Ehrenberg collection to compare Ehrenberg's identifi cation as well as the current and the original concept of the species of diatoms from Brazil recorded by him. For this, studies with new preparations from his original samples will be necessary. Finally, it will be possible to carry out new studies comparing the old and recent diatom fl ora and associate the likely changes with the historical variables. Th is could provide tools to understand changing mechanisms of the biodiversity of diatoms in tropical and subtropical habitats.