Intricate Networks in Nomenclature: Cases of Naming in Arthrocaulon, Arthrocnemum, and Salicornia (Amaranthaceae)

The nomenclatural status and typification of the names Arthrocaulon macrostachyum, Salicornia fruticosa, S. fruticosa var. deflexa, S. fruticosa var. glaucescens, S. fruticosa var. intermedia, S. fruticosa var. humilis, S. fruticosa var. pachystachya, S. fruticulosa, S. glauca, S. lignosa, S. macrostachya var. virescens, S. macrostachya var. glaucescens, S. perennis, S. radicans, S. radicans var. caespitosa, S. sarmentosa, S. sempervirens, and S. virginica, as well as an unnamed β-variety of S. fruticosa proposed by A. Bertoloni, are investigated. Concerning A. macrostachyum, we document that the type indicated in literature (G00177362) is not a holotype, and that lectotypification is necessary. A specimen from G (G00687638) is here designated as a lectotype. On the level of variety, Arthrocnemum fruticosum var. macrostachyum is an earlier legitimate name for Salicornia fruticosa var. pachystachya. Furthermore, Piirainen et al. are incorrect when citing Forsskål’s “Salicornia” from Alexandria as “S. virginica Forssk.”; it is not a new name and should be cited as S. virginica auct. non L., as published in Forsskål’s Flora Aegyptiaco-Arabica. Like with numerous other parallel cases in Flora Aegyptiaco-Arabica, Forsskål’s designation of “Salicornia virginica” for an Arabian plant is to be considered a misapplication of the earlier Linnaean name for an American plant. Arthrocnemum glaucum (a nomen illegitimum of Ungern-Sternberg), was listed as type species of Arthrocnemum by the Names in Current Use project; the basionym, Salicornia glauca Delile, is here lectotypified and identified as Arthrocaulon meridionale, published by Ramirez et al. Updated synonymies of Arthrocaulon macrostachyum, A. meridionale, Salicornia fruticosa, and S. perennis are proposed. Salicornia sempervirens is an invalid name according to Art. 36.1a of ICN. No original material was found for S. radicans var. caespitosa. This paper also refer to lecto- or neotypifications on specimens deposited at BM, G, LINN-HS, LY, MPU, NAP, and PAL, and their current taxonomic positions are suggested in a taxonomic part of the paper.

The species of Arthrocaulon are sometimes still referred to Arthrocnemum, for example by de la Castroviejo [5] (p.526-527), de la Fuente et al. [6] (p. 1249), and Ramírez et al. [7] (on p. 1422 and nearly all following pages), where the typification of the generic name Arthrocnemum as Salicornia fruticosa (L.) L. by Pfeiffer [8] (p.279) is rejected.However, Pfeiffer's typification has the priority of designation (ICN Art. 9. 19) and is cited as acceptable according to Art. 7.11, Ex. 17 of the Shenzhen Code [4], thus making Arthrocnemum a synonym of Salicornia unless that genus is very narrowly defined.For this reason, Piirainen and G. Kadereit published the new generic name Arthrocaulon.In this study, we accept Arthrocaulon and Pfeiffer's typification of Arthrocnemum (for details of the various typifications of Arthrocnemum, see 'Section 3.1.7Salicornia glauca' of the present paper).
As a whole, the genera Arthrocaulon and Salicornia are difficult from the taxonomical point of view due to their low number of morphologic characters, their high phenotypic variability, and recurring hybridization [1][2][3].This has caused a proliferation of names over time, sometimes leading to nomenclatural problems (see [4] for an example).
As part of the ongoing research on Salicornioideae [9][10][11][12][13], here we present a range of nomenclatural and taxonomic notes regarding names that are used to refer mostly to perennial species belonging to the genera Arthrocaulon and Salicornia, names which mostly have been previously untypified.

Material and Methods
This study is based on the analysis of the relevant literature (i.e., protologues of the names investigated and works in which these names have been listed or discussed).It is also based on a search for and the examination of specimens preserved at the herbaria BM, C, CGE, G, E, FI, K, LINN, LINN-HS, LY, MA, MANCH, MNW, MPU, NAP, OXF, P, PAL, RO, SLBI, SWA, and UPS (acronyms following those of Thiers [14]).
Nomenclatural articles and references to the Glossary, as cited throughout the text, are given following the Shenzhen Code [4].
The abbreviations of author names follow the guidelines of IPNI (https://www.ipni.org/,accessed on 12 June 2024).
The information shown in the label transcriptions is given between double quotation marks.The studied names are listed alphabetically according to epithets.

Results and Discussion
3.1.Publication and Typification of the Names 3.1.1.Salicornia anceps Castroviejo [15] (pp.212-213) discussed Lagasca's name Salicornia anceps [16] (p.52) and stated that "Typus: Se cría en Roquetas y Cabo de Gata, en donde le encontró don Simón de Rojas Clemente (MA 29474)" (note that Lagasca's herbarium and types were destroyed, but duplicates are preserved at MA according to the HUH-Index of Botanist [17]).Although the phrase "designated here" (or an equivalent) was not reported by Castroviejo [15], the typification (lectotype available at https://imagenes.rjb.csic.es/herbarioV/visorVCat.php?img=MA-01-00029474, accessed on 20 June 2024) is to be accepted according to Art. 7.11 of ICN (Castroviejo's statement was published before 1 January 2001).The MA specimen is a sterile and terminal part of one plant and, therefore, it cannot be identified with certainty as any of the Salicornia perennial species.This was also the conclusion of Castroviejo [15], who considered this name to be ambiguous.We agree with Castroviejo [15] on this point.Based on our preliminary check, the issue is still quite complicated, and although we accept Castroviejo's lectotypification, we here decide to postpone the further identification of type.Conclusive identification of the type material might involve designating an epitype, collected from the locus classicus.
Based on the synonyms and the literature cited by Bertoloni [18] (p.18), it is clear that the Italian author, with his Salicornia fruticosa var.β, had in mind to propose a new taxon for prostrate plants occurring in coastal areas of eastern Italy, spanning from the north ("litora Ravennatia, et Venetiis"-Ravenna and Venezia are two cities of, respectively, the Emilia-Romagna and Veneto regions, in NE-Italy) to the south and to the cliffs of Gargano's promontory [". . .rupibus di Viesti prope Garganum", where Vieste is a small town of the Apulia region (SE-Italy) located on the Gargano promontory] and occurring in Sardinia and Corsica ("... Bonifacio, et Portovecchio .... Bastia).To fully understand Bertoloni's concept of his unnamed variety β, we first checked the Herbarium BOLO, where Bertoloni's herbarium and types are mainly preserved [26].We found the following four sheets:
The identification of the BOLO specimens is not a simple issue since the ranges of the diagnostic characters between the related Arthrocaulon macrostachyum and A. meridionale, given by Ramírez et al. [7], appear to be partially overlapping and as the reliable diagnostic character is essentially determined by whether the plants are diploid or tetraploid.Therefore, as a valid and workable distinction between these two species, we rely on their chorology.We therefore identify BOLO100045, BOLO100046, and BOLO100046 (both collected from the Adriatic coast of Italy; Emilia-Romagna, Veneto, and Apulia regions, respectively) as A. macrostachyum, whereas BOLO100048 (the plant from the Sardinia region) is identified as A. meridionale.
Regarding the synonyms cited by Bertoloni [18] (p.18), we can make the following remarks:
The name appears to be untypified based on our literature search (see discussion under 'Section 3.1.6Salicornia fruticulosa').
[=Moricand] is proposed.Therefore, when Koch stated that he followed Bertoloni's work, he must mean that he followed Bertoloni's treatment of Moricand's S. macrostachya in full synonymy under S. fruticosa, and not that he only referred to Bertoloni's unnamed variety β, although both Bertoloni and Koch have a "var β".Therefore, Bertoloni's var.β and Koch's var.β are different and Koch's "S.fruticosa var.β. pachystachya" includes both Moricand's Salicornia macrostachya as a synonym of S. fruticosa in the sense of Bertoloni and Bertoloni's var.β.Because of the existence of the earlier name at the rank of variety [Arthrocnemum fruticosum var.macrostachyum (Moric.)Moq., proposed by Moquin-Tandon [30] (p.112), who cites Salicornia macrostachya Moric. in synonymy], S. fruticosa var.pachystachya was a nomenclaturally superfluous and illegitimate renaming and is to be typified by original material of the name S. macrostachya (see 'Section 3.1.5Salicornia fruticosa var.pachystachya' and the discussion of synonymy in 'Section 3.2 Taxonomic treatment' below).

Salicornia fruticosa Varieties by Michele Tenore
Tenore [31] (p.582) classified Salicornia fruticosa into three varieties, namely, var.glaucescens Ten.(diagnosis: "elata, caulinibus lignosis, rami patentibus, articuli valde remotis"), var.intermedia Ten.(diagnosis: "radicans glauca"), and var.humilis Ten.(diagnosis: "virescens, caulibus procumbentibus radicantibus ramisque divaricatis").No information about these three varieties was found in volume IV of Tenore's Flora Napolitana [31] (p.5), where S. fruticosa was listed with only the var.macrostachya.According to Tenore [31] (p.5), the localities of S. fruticosa, and presumably also the localities of the varieties published the following year, were "In inundatis salsis.Fusaro, Maremorto, Lago salso".We traced two sheets at NAP (where Tenore's herbarium and type are mainly deposited; see [32]), i.e., NAP0000051 and NAP0000052.NAP0000051 bears three parts of plants and the following two labels: "Salicornia fruticosa virescens varietas" and "Salicornia fruticosa humilis virescens".Since the two labels are not clearly associated with any of the three parts of plants, we suppose, according to the diagnosis of var.humilis, that the part in the center of the sheet (which includes roots and is clearly procumbent) can be referred to as this variety.The sheet NAP0000052 bears two parts of plants and the label "Salicornia fruticosa glaucescens ...|Fusaro", where Fusaro is a coastal lake occurring in the Bacoli Municipality (Province of Naples, Campania region, southern Italy), as reported by Tenore [31] in Flora Napolitana (see above).Unfortunately, no date of collection was reported in these two NAP sheets and, therefore, we cannot be sure that they are ante-1831 collections.Therefore, we prefer to avoid their use as lectotypes (Art.9.3 and 9.4 of ICN [4]).Since no further sheet of original material was found, neotypifications are required according to Art. 9.8 of ICN [4]).We designated NAP0000051 (plant part on the center of the sheet) as the neotype of S. fruticosa var.humilis and NAP0000052 as the neotype of S. fruticosa var.glaucescens.Finally, as regards var.intermedia, we traced two specimens at LY (barcodes LY0517535 and LY0517536).These were useful for neotypification since they were collected at "lago fusaro", as reported in an original label, and Pellanda s.n.(LY0517535) is here designated as a neotype of S. fruticosa var.intermedia.NAP0000052 is identifiable as S. fruticosa (an erect and large branch with many terminal spikes, up to 4 cm long).The other types (NAP0000051, LY0517535, and LY0517536) cannot be identified according to De La Fuente et al. [33].In fact, based on these authors, NAP and LY specimens can be referred to as Saronornia perennis (Mill.) A. J. Scott.(currently Salicornia perennis) or Sarcocornia alpinii (Lag.)Rivas Mart.(Salicornia alpinii s.s.according to [2]), but seeds are lacking in these types.Further investigations (field surveys) are necessary to reach a taxonomic conclusion about these two names that were given by Tenore.As a consequence, we prefer to avoid synonymizing these two names, which are presented separately (see 'Section 3.2 Taxonomic Treatment').
We traced just one specimen that was part of the original material for Salicornia fruticosa var.deflexa, i.e., a Corbière's collection in Saint-Vaast was preserved at LY (barcode LY0745272).This specimen is here designated as its lectotype (Art.9.12 of ICN [4]).According to the current concept [1,11], LY0745272 is identifiable as S. perennis.
[nobis]" and the dates of collection ("7bre [settembre = September] 1827" for PAL58796 and "7bre 1826" for PAL58797).Since the label of PAL58780 does not include the collection date, we cannot be sure that the plant was collected before 1827 (year of the original publication) and, therefore, we prefer to exclude it from the lectotypification.Among the other two specimens (which are clearly part of different gatherings), it is important to note that PAL58796 includes two parts of plants which are referrable to different species, i.e., Salicornia perennis (plant on the left, creeping and rooting at nodes and with its few terminal spikes being very short, up to 0.5 cm) and S. fruticosa (plant on the right, with many long spikes up to 4 cm), whereas PAL58797 cannot be identified since fertile branches are lacking [6,27].Therefore, we here designated PAL58796 (plant on the right) as the lectotype of Salicornia fruticulosa.

Salicornia glauca
Delile [36] (p.49) published Salicornia glauca Del., citing "Salicornia virginica FORSKAL." after his own new name.Delile described the S. glauca in order to correct Forsskal's misidentification of the Linnaean S. virginica (see 'Section 3.1.16Salicornia virginica').The abbreviation "As." was also cited; this was taken from Forsskål's explicatory notes [page "L" and LIX (Roman page numbers) on the first part of Forsskal's Flora Aegyptiaco-Arabica] and it is not part of the name; it refers to the provenance and status of the species: "As.= Alexandriae spontaneae".Based on the discussion in 'Section 3.1.16 Salicornia virginica' and on the ruling of ICN Art.41.7 Note 3, Delile's name is to be considered new to science, and S. glauca should replace Forsskål's misidentified Linnaean name.However, S. glauca is an illegitimate and later homonym (Art.53.1 of ICN [4]) of the identical name published the year before by Stokes [37] (p.8), who discussed if his species S. glauca could be identical with S. arabica L. Note that Salicornia arabica is a name that was recently proposed for rejection [9].Note also that the name of a new but previously misidentified species should be typified with original material relating to the new name, not automatically with material related to Forsskål's misapplied name (ICN Art.41.7 Note 3).
According to the HUH-Index of Botanists [38], Delile's Egyptian collections (period 1798-1801) are preserved in several herbaria.We traced three specimens of Salicornia glauca, one at LINN-HS (no.20-13, https://linnean-online.org/29388/#?s=0&cv=0&z=0.0657,0.3685,0.6281,0.7607,accessed 22 June 2024), one at P [barcodes P04918422 (specimen to the left on the sheet; image available at http://mediaphoto.mnhn.fr/media/1441381909236kOTJBPm8nnr55Zwf, accessed 22 June 2024), and one at P05234345 (image at http: //mediaphoto.mnhn.fr/media/1441396286253vyQsUv9BFK48R0xt,accessed 22 June 2024).These samples were all collected by A.F. Delile in Egypt (as reported on the original labels).These specimens were clearly part of the material used by Delile to describe Salicornia glauca.According to the current concept in Salicornioideae [2,11], these specimens belong to the genus Arthrocaulon; this is based on the cymes being free and protruding.However, only the LINN-HS specimen is identifiable with certainty due to the presence of visible seeds that are glabrous, black, and shiny, which are diagnostic features of Arthrocaulon.Using species rank and based on the diagnostic features indicated in Ramírez et al. [3], LINN-HS20-13 can be identified as A. meridionale (spikes 5-6 cm long vs. up to 4 in A. macrostachyum).The occurrence of A. meridionale in Egypt is also congruent with the chorology of the species (south Mediterranean basin).We here designate the specimen LINN-HS20-13 as the lectotype of the name Salicornia glauca Delile since it appears to be much better preserved and richer in flowers than the two specimens at P. The features that can be seen on LINN-HS20-13 are important in the identification of the Salicornia species [6,27].
Arthrocnemum glaucum Ung.-Sternb., nom.illeg.(basionym Salcornia glauca Delile), was designated as being the type of Arthrocnemum in the enumeration of generic names in current use by Greuter et al. [39] (p.86, see also the online version of the enumeration [40]).The designation is in agreement with the requirements in ICN Art.7.11.According to ICN Art.10.1, the type of the name of a genus is the type of the name of a species, and the lectotypification of A. glaucum and the identity of the type is therefore relevant.Greuter et al. [39] cites the name of Arthrocnemum glaucum Ung.-Sternb.without a basionym, as does IPNI [41], but the name is frequently cited in international databases as Arthrocnemum glaucum (Moq.)Ung.-Sternb., for example in POWO [42] and GBIF [43].Arthrocnemum glaucum (Moq.)Ung.-Sternb.must be based on Arthocnemum fruticosum var.glaucum Moq.[30] (p.112), published in 1840, where the informal name "Salicornia glauca plerumque Auct." is cited as coming from Egypt and Syria without specified localities, collectors, or collections.However, according to the information published when the new combination Arthrocnemum glaucum was published by Ungern-Sternberg [44] (p.283), its correct name is Arthrocnemum glaucum (Delile) Ung.-Sternb., basionym: Salicornia glauca Delile.Therefore, it must be typified with Delile's original material of this species.It must, as shown above in our identification of Delile's original material, be a heterotypic synonym of the much younger but still legitimate Arthrocaulon meridionale.
However, as mentioned in the general note in the introduction to this paper about our acceptance of the name Arthrocaulon, Pfeiffer's lectotypification of Arthocnemum [8] (p.279) with Salicornia fruticosa (L.) L. must stand.Pfeiffer's typification is acceptable according to Art. 7.11, Ex. 17, and it has the priority of designation as the first lectotypification of a previously untypified name, which is to be followed according to ICN Art.9.19.Standley [45] (p.81) lectotypified Arthrocnemum, with "Arthrocnemum fruticosum Moq." also cited as the accepted type of Arthrocnemum Moq. by the Missouri Botanical Garden's database TROPICOS [46], but Standley's typifications are rejected as using a mechanical method of type selection (ICN Art.10.7), as is the report on mechanical methods for lectotypification by McNeill et al. [47] (p.1447).

Salicornia lignosa
Woods [48] (p.31), in discussing Smith's Salicornia radicans [25], proposed a new species (S. lignosa J. Woods) from Hailing Island (Hampshire, UK) that, according to the author, "somewhat resembles S. radicans in its diffuse mode of growth".The diagnostic characteristics distinguishing between S. lignosa and S. radicans would be the "thickness and very firm structure of the lower part of the stem" (vs."least solid stem").Moreover, Woods [48] (p.31) compared S. lignosa with S. fruticosa.These differ in terms of the length of the spike ["one inch or a little more long, about one-sixth this width" (s.lignosa) vs. "relatively longer" (S. fruticosa)].
Piirainen et al. [2] (p.123) indicated that the specimen G00177362 was the holotype of S. macrostachya.However, Moricand [50] (p. 2) did not specify any holotype (see Art. 9.1 of the ICN [4] and the considerations by McNeill [56]) and cited a syntype; lectotypification was therefore necessary.We examined G00177362 and verified that M. Ph.Moricand (the grandson of M.-E.Moricand) gave it to the Herbarium G in 1908 (as reported in the printed label on the top of the sheet).The handwritten label (top-right of the sheet) states "Salicornia macrostachya Moric.|exipso!|Des environs de Venice malamocco 7 bre [September]".This specimen was collected by M.-E.Moricand, as clearly indicated in the printed label ("Les étiquettes non signees, se rapportant à des plantes d'Europe, proviennent des herborisations personnelles de Moricand" = "The unsigned labels, relating to European plants, come from Moricand's personal collections").Although the collecting locality cited in the protologue is mentioned on the label, the information about G00177362 does not include the year of collection (only the month is reported).Therefore, we cannot be sure that it is an ante-1820 addition to Moricand's collection and may not part of the original material of the name Salicornia macrostachya.Fortunately, we traced another specimen (G00687638) with a plant collected in the "environs de Venice" by Moricand in 1818 (with that year and also "Salicornia macrostachya N. [Nobis]" written on the specimen).Thus, the specimen in G-DC, G00687638, is certainty part of the original material.
All things considered, the typification by Piirainen et al. [2] (p.123) is not correct, their search for original material was not adequate, and Art.9.10 of ICN [10] (correction of the term holotype to lectotype) cannot be applied.G00687629 is not identifiable with A. macrostachyum (Moric.),published by Piirainen and G. Kadereit [possibly the Australian endemic S. arbuscula R.Br.(currently known as Tecticornia arbuscula)].Therefore, G00687638 is the only extant material clearly part of the original material and useful for lectotypification, and it is here designated as the lectotype of Salicornia macrostachya.This lectotype morphologically matches Moricand's description, being a fruticose and fleshy plant with ascending branches and sessile and cylindrical inflorescences.Concerning the current identity of G00687638, we note that Moricand's species is currently accepted under the genus Arthrocaulon Piirainen and G. Kadereit as A. macrostachyum [2] (p.123).

Salicornia macrostachya Varieties by Michele Tenore
As mentioned in 'Section 3.1.3Salicornia fruticosa varieties by Michele Tenore', Tenore [31] (p.5), in volume IV of his Flora Napolitana, recognized Salicornia macrostachya as a variety of S. fruticosa.As regards habitat and provenance, he reported "Ibidem", referring to the localities cited just above under S. fruticosa (sensu stricto), i.e., "In inundatis salsis.Fusaro, Maremorto, Lago salso".One year later, Tenore [24] (p.582) accepted S. macrostachya at the species rank and classified it into two varieties (described as new), i.e., var.virescens Ten.(diagnosis: "elata, caulis valde lignosis (digitalis crassitici bipedalis altitudinis) ramisque strictis.Planta laete virens fere arborescens nunquam gregarie crescens") and var.glaucescens Ten.(diagnosis: "patula depressa, caulibus fruticulosis (1-3-linearis crassitici pedalis altitudinis), ramisque divaricatis prostratis.Planta viridi-glauca suffruticosa passim radicans"); no further data were given by Tenore [24] (p.582) for these varieties, and he made no direct or indirect references to the varieties descibed in 1827 under S. fruticosa.We traced two specimens at NAP [32] that bore original labels annotated with "Salicornia macrostachya virescens ...|Miseno" (barcode NAP0000047) and "Salicornia macrostachya glaucescens ..." (barcode NAP0000048); NAP0000048 also bore a second label, reporting "Salicornia fruticosa|Mare morto".Note that Miseno refers to a coastal lake included in the territory of Bacoli Municipality (Province of Naples, Campania region, southern Italy) and that Mare morto is a local name for this lake (E.Del Guacchio pers.comm.).These localities match those reported for S. fruticosa in volume IV of Tenore's Flora Napolitana, as indicated above.Unfortunately, no date of collection was reported in these two NAP specimens and, therefore, we cannot be sure that they are ante-1831 collections.As a consequence, we prefer to avoid their use as lectotypes (Art.9.3 and 9.4 of ICN [4]).Since no further specimen of the original material was found, a neotypification was required according to Art. 9.8 of ICN [4] and we designated NAP0000047 as the neotype of S. macrostachya var.virescens and NAP0000048 as the neotype of S. macrostachya var.glaucescens.Both the neotypes were identifiable as Arthrocaulon macrostachyum (we observed undivided hollows in the segments where the flowers had fallen) according to the current concept [2,11].Note that Tenore [31] indicated "passim radicans" (= sometimes rooting) for var.glaucescens (this is a character which could link his concept to Salicornia perennis); however, the neotype does not appear to creep and, therefore, it possible that Tenore observed (but did not collect) the first internodes near the main root.

Salicornia perennis
Miller [57] ("Salicornia no. 2 ('perenne')" and in the corrigenda "Salicornia 2, lege (Perennis)") described S. perennis Mill.from plants growing in Sheepy Island.Guilló et al. [58] (p.333) proposed a neotype for Miller's name on a sheet deposited at K (K000450665).The material on K000450665 is represented by two sterile plants and, according to the current concept [6,27], sexual character is almost always essential for the identification of taxa of Salicornia.Therefore, an epitypification might be desirable (Art.9.9 of ICN [4]).However, the neotype is a suffruticose plant, with green, diffuse stems; rooting at the nodes, with numerous short and creeping sterile branches (even underground); short ascending stems, which are slightly lignified; and fine, slightly thickened segments.These morphological features allow us to identify the plant on K000450665 as Salicornia perennis.

Salicornia radicans
Smith [25] (unpaginated text opposite Tab.1691) proposed his Salicornia radicans Sm. with a short diagnosis ("Stem woody; procumbent and taking root at the base.Joints compressed, notched; interstices nearly cylindrical.Spikes oblong.Style deeply divided.Stamens two").An illustration (Table 1691) was also published, and it is part of the original material.However, Smith [25] also reported two collections (syntypes according to Art. 9.6 of ICN [4]).One was collected by the reverend C. Sutton (in September 1798) from "sea coast at Holm, Norfolk"; the second one was taken by Mr. W. Borrer (no date indicated) from the "harbour at Shoreham, Sussex".The locality of Weymouth (Dorset, UK) is also reported.We traced the two syntypes at LINN-HS (where Smith's herbarium and types are preserved [59]).One sheet (LINN-HS20-6) bears one plant and the following annotation (on the bottom-left corner of the sheet): "Coast of Sussex.M r Borrer.1805".The other sheet bears two plants (barcoded as LINN-HS20-5-1, LINN-HS20-5-2) and is annotated with the numbers "1" and "2".This links the specimens to the notes on the bottom-left corner of the sheet: "Holm Norfolk.Rev. C. Sutton (D.D.) 1798" (no. 1, barcode LINN-HS20-5-1) and "Norfolk coast.M r Borrer.1806" (no.2, barcode LINN-HS20-5-2).According to the protologue, of those displayed on this sheet, only LINN-HS20-5-1 is a syntype, whereas LINN-HS20-5-2 cannot be since, as reported in the protologue [25], the specimen cited in the protologue was collected by M r Borrer in Norfolk County, not in Sussex County.Both syntypes (LINN-HS20-6) and (LINN-HS20-5-1) are identifiable as S. perennis according to the current concept [2,11].We here designate LINN-HS20-6 as the lectotype for S. radicans since the other syntype (Suttons' collection) is represented by a small fragment of the terminal part of the inflorescence, whereas Borrer's specimen is complete.Although Smith's illustration (Table 1691) is part of the original material, ICN [4] states that when syntypes occur, the lectotype must be chosen among them (Art.9.12 and Ex. 12 of ICN [4]).In the glossary of the ICN [4] (p. 202) a "syntype" is defined as including "Any specimen cited in the protologue when there is no holotype, . ..".
Unfortunately, no specimen that would be useful for the lectotypification purpose was traced at LY, where Rouy's herbarium and types were preserved (M.Thiebaut pers.comm.).Therefore, a neotypification would be desirable (Art.9.8 of ICN [4]).However, Rouy [34] (p.60) did not specify any locality, providing a wide distribution area (Great Britain and the Iberian Peninsula).It is therefore difficult to select a collection made at a reasonably narrowly defined locus classicus as the neotype.The original description is quite vague but is based on these characteristics: "Plantae relativ courte, en touffe ou buissonnante" (= plant relatively short in tufts or shrub).Thus, Rouy's variety falls within the circumscription of Salicornia perennis according to the current concepts [2,6].Concluding, we here refrain from the typification of Salicornia radicans var.caespitosa but propose it to be placed in the synonymy of S. perennis.

Salicornia sarmentosa
Duval-Jouve [60] (p.174) described S. sarmentosa Duval-Jouve, providing a detailed description.Information on habitat and provenance ("Bords de étangs et des marais salants, dans les lienx très-humides et inondés hiver ... Aigues Mortes ... à Carnon ... entre Palavas et Maguelonne ... à Vic; à Frontigna; aux Onglous ... Saint Vaast (Manche), d'où M. le docteur Lebel") was also provided, and it was suggested ("an") that it was close to but not a synonym of Tenore's Salicornia fruticosa var.humilis (the latter would have made S. sarmentosa an illegitimate renaming).We found sixteen relevant specimens at MPU (where Duval-Jouve's collection is mainly preserved) and P (where there are duplicates of the MPU specimens) [61].The nine specimens found at MPU were MPU227866, MPU259991, MPU1316700, MPU1316694, MPU1323504, MPU1323505, MPU1323510, MPU1323511, and MPU1323513, and the collections were made at Vic (4), Maguelonne (2), and Palavas (1), which are localities cited in the protologue.However, six out of these nine specimens were collected after the date of publication (1869, 1877, and 1887), whereas the other ones were syntypes (Art.9.6 of ICN [4]) and original material for Salicornia sarmentosa (years of collection: 1862 and 1868).Seven specimens were found at P (barcodes P06739611, P05158076, P05235503, P00724221, P00724223, P00724224, and P00724226) but they were all collected after 1868.Therefore, the P specimens cannot be considered to be part of the original material for Salicornia sarmentosa.Among the three MPU specimens that are original material, one was collected at Maguelonne in 1842 (MPU1316694), with the other two (MPU1316700 and MPU1323511) collected at Vic in 20 September 1868.We here designate the MPU1316700 specimen, which is part of "HERBIER DE LA FLORE DEL MONTPELLIER PAR Lt. [Loret] et B [Borrandon]" (as indicated in the original label on the bottom-right corner of the sheet), as the lectotype of the name Salicornia sarmentosa.According to De La Fuente et al. [33], its type cannot be identified.According to these authors, MPU1316700 can be referred to either Saronornia perennis (Mill.)A.J. Scott.(currently Salicornia perennis) or Sarcocornia alpinii (Lag.)Rivas Mart.(Salicornia alpinii s.s.according to [2]), but seeds are lacking.Further investigations (field surveys) are necessary to reach a taxonomic conclusion about the identity of the name.As a consequence, we prefer to avoid synonymizing these names, which are presented as separate (see 'Section 3.2 Taxonomic Treatment').

Salicornia virginica
A plant called Salicornia virginica was first validly described and named by Linnaeus [63] (p.4) in the 1st Edition of Species Plantarum.Salicornia virginica L. is currently considered the correct name for a species native to N. America (S.Alaska and coastal areas on the eastern and western sides of Canada and USA), N. Mexico, and the Caribbean [64].However, note that Ball [27] (p.384) accepted S. depressa Standl.as the name for this American plant, with S. virginica as a possible synonym, and that Piirainen et al. [2] (p.124) indicated S. depressa as a doubtful heterotypic synonym of S. virginica L., whereas POWO [64] and Southeastern Flora [65] reported S. depressa as a sure synonym of the Linnaean plant.Information about the previous lectotypification of S. virginica is shown below.
In Flora Aegyptiaco-Arabica [66], brought to the press by an anonymous editor 12 years after Forsskål died on the expedition to Egypt and Arabia, the name S. virginica is listed in two places in the text with reference to a plant from Egypt.This includes Forsskål [66] (p.LIX, No. 3): " [Salicornia] virginica, farinosa.-As."Here, the typography indicates that "virginica" is an epithet, and that "farinosa" is a descriptive term, and "As" is an abbreviation for the locality and status, i.e., "Alexandria spontanea"].This also includes Forsskål [66] (p. 2, No. 2): "SALICORNIA virginica.Arab.Chraesi."This gives the Latin name of the species and the vernacular name in Arabic, followed immediately by a description of the species.Forsskål's S. virginica was considered a species new to science by Christensen [67] (p.10) and several subsequent authors.If this interpretation is correct, then "S.virginica Forssk." is a later homonym of the Linnaean S. virginica and, hence, Forsskål's name would be illegitimate under Art.53.1 of ICN.This is also how POWO [68] previously interpreted the situation, accepting Salicornia virginica Forssk.as an illegitimate heterotypic synonym of Arthrocaulon macrostachyum.Piirainen et al. [2] (p.123) cited Forsskål's name of "Salicornia virginica Forssk., Fl.Aegypt.-Arab.: 2. 1775, non L. 1753" as a heterotypic synonym of Arthrocaulon macrostachyum.However, S. virginica sensu Forssk. is a misidentification in the sense of the Shenzhen Code [4], (p.124), and it is, according to the Recommendation 50D of the ICN, to be indicated by the words "auct.non", followed by the name(s) of the original author(s) and the bibliographic reference of the misidentification.This was originally demonstrated by Hepper and Friis [69] (p.101), who cited Forsskål's name as "Salicornia virginica sensu Forssk.1775: 2 (LIX no.3; Cent I No 2), non L. (1753)", which is different from the format of ICN but has the same meaning.
In order to document this point, we have re-examined the conclusion by Hepper and Friis [69] (p.101), not only regarding Forsskål's use of S. virginica, but also other parallel cases of Linnaean names for American plants listed as accepted names in Flora Aegyptiaco-Arabica.We know that Forsskål, on the expedition to Egypt and Arabia, worked with a very limited number of books (Hepper and Friis [69] (p.25)).Moreover, Hepper and Friis ([69] (pp.[25][26]) observed that "he expected to find American and Asian plants in Egypt and Arabia . . .He seems often to have accepted an identification if the plant matched a description in the Linnaean works he had with him, no matter where the species was [originally] described from.He used the botanical books of Linnaeus as a world flora.".This conclusion is documented by statements made by Forsskål himself on the journey, for example in a letter to Linnaeus from Yemen: "Here I have found a lot of American, Indian, and new plants . ..".A total of 188 cases of Forsskål's misidentifications of Linnaean names can be found among the names listed by Hepper and Friis [69].As a further argument for considering Forsskål's identification of his Egyptian plant with S. virginica.L. as a misidentification, it should be noted that Forsskål's [66] (p. 2) description of S. virginica included part of Linnaeus's diagnostic phrase for S. virginica L. almost verbatim ["Articulis . . .apice compressiusculi, emarginato bifidi" (in Flora Aegyptiaco-Arabica) vs. "articulis apice compressis emarginatis bifidis" (in Species Plantarum)].
However, does Forsskål's plant in fact differ from the Linnaean S. virginica and, if so, how, and what is its identity?We studied the original material in Forsskål's herbarium at C [70], which contains three specimens of S. virginica sensu Forssk., which are also referred to here with their old collection numbers used in the work of Hepper and Friis [69] (p.101), links to images, and notes about the presence and length of inflorescences (Figure 2): It should finally be noted that the wide distribution of A. meridionale, indicated here and outlined in [3], is not currently accepted in POWO [73], which only accepts the distribution on Sicily.It should also be noted that Arthrocnemum/Arthrocaulon macrostachyum has a much wider distribution than that recorded by Ramírez et al. [3].It reaches south through most parts of Egypt [71] (p.108), along the Red Sea coast in Sudan [74] (p.277), in Eritrea [75] (pp.289-290), in Yemen [76] (p.84), along the Gulf of Aden down to Socotra, and in the inland localities of southern Somalia [77] (p.131).Moreover, POWO [78] indicates that in addition to the above distribution, A. macrostachyum is recorded in Saudi Arabia, Jordan, Oman, United Arab Emirates, Iran, and Pakistan.The identity of the plants from these areas was not discussed by Ramírez et al. [3], and more statements about the distribution of Arthrocaulon macrostachyum and A. meridionale must be postponed until further studies have been conducted on the taxonomy of these two species and their distribution.
In conclusion, Forsskål did not consider his S. virginica to be a name for a hitherto unknown species, and thought that it should be referred to as Salsola virginica auct., non  These three specimens were all collected by Forsskål near Alexandria in Egypt and are part of the material used by him to write the description of his S. virginica.
Although Forsskål's plants clearly belong to the genus Arthrocaulon, it is, as for Bertoloni's specimens discussed above in 'Section 3.1.2Salicornia fruticosa var.β by Antonio Bertoloni', not a simple issue of identifying them as either Arthrocaulon macrostachyum or A. meridionale.The quantitative diagnostic features given by Ramírez et al. [7] partially over-lap or are very near to each other.The best characteristic for this seems to be inflorescence length, but Forsskål's plants have very few intact and apparently unripe inflorescences, as outlined above.In the lists of Ramírez et al. [3] with diagnostic features of A. macrostachyum, the inflorescence length is given in cm as "(2.5) 2.9 (4)", and for A. meridionale it is given as "(2.5) 3.8 (5.5)".As such, the preserved spikes on Forsskål's plants are shorter than the intervals given for both species, but perhaps verge towards A. macrostachyum.The real diagnostic feature according to Ramírez et al. [3] is that A. macrostachyum is diploid and A. meridionale is tetraploid.Not being able to test the ploidy of Forsskål's plants, as with the identification of Bertolini's collections in 'Section 3.1.2Salicornia fruticosa var.β', one may have to rely on chorology.From this, Forsskål's plants could be identifiable as A. meridionale, a species which according to Ramírez et al. is distributed throughout North Africa [3].However, Ramírez et al. [3] do not cite identified specimens from Egypt, where A. macrostachyum is widespread according to Boulos [71] (p.108), nor from the coasts of the Red Sea and the Gulf of Aden, from where all material was identified as A. macrostachyum before a distinction emerged between this and A. meridionale (see below).In '.2.Taxonomic treatment', we have tentatively referred Forsskål's Egyptian plants in the synonymy of either A. macrostachyum or A. meridionale.
It should finally be noted that the wide distribution of A. meridionale, indicated here and outlined in [3], is not currently accepted in POWO [73], which only accepts the distribution on Sicily.It should also be noted that Arthrocnemum/Arthrocaulon macrostachyum has a much wider distribution than that recorded by Ramírez et al. [3].It reaches south through most parts of Egypt [71] (p.108), along the Red Sea coast in Sudan [74] (p.277), in Eritrea [75] (pp.289-290), in Yemen [76] (p.84), along the Gulf of Aden down to Socotra, and in the inland localities of southern Somalia [77] (p.131).Moreover, POWO [78] indicates that in addition to the above distribution, A. macrostachyum is recorded in Saudi Arabia, Jordan, Oman, United Arab Emirates, Iran, and Pakistan.The identity of the plants from these areas was not discussed by Ramírez et al. [3], and more statements about the distribution of Arthrocaulon macrostachyum and A. meridionale must be postponed until further studies have been conducted on the taxonomy of these two species and their distribution.
In conclusion, Forsskål did not consider his S. virginica to be a name for a hitherto unknown species; he thought it could be identified with the Linnaean S. virginica from North America.It should therefore be referred to as Salsola virginica auct., non L.: Forsskål, Fl.Aegypt.Arab.: LIX; 2 (1775).The many other cases where Forsskål identified his Middle Eastern and Arabian plants with names previously proposed by Linnaeus should be treated in the same way as such issues were resolved here for S. virginica.The following infraspecific names cannot currently be attributed to current taxa and are not synonymized here: Salicornia fruticosa L. var.intermedia Ten., Syll.Pl.Fl.Neapol.: 582.1831.

Figure 6 .
Figure 6.Lectotype of Salicornia fruticosa (Herb.Burser XVI(2): 22, UPS, plant to the right.Reproduced under the terms of the Creative Commons Attribution License [CC BY 4.0], Museum of Evolution, Uppsala University.

Plants 2024 , 29 Figure 10 .Funding:
Figure 10.Lectotype of Salicornia sarmentosa (MPU!).Author Contributions: Conceptualization, D.I.; Methodology, D.I.; Investigation, D.I., I.F. and M.I.; Resources, D.I. and M.I.; Data curation, D.I. and I.F.; Writing-original draft, D.I.; Writing-review & editing, D.I., I.F. and M.I.All authors have read and agreed to the published version of the manuscript.Funding: This research received no external funding.Data Availability Statement: Data are contained within the article.Acknowledgments: Thanks are due to the Directors and Curators of the Herbaria quoted in this paper for their support during visits or loan of specimens/photographs, with special thanks to L. Loze and F. Stauffer (Herbarium G), to R. Vallariello (Herbarium NAP), to Olof Ryding (Herbarium C) and Mats Hjertson (Herbarium UPS) for the permission to reproduce, respectively, the lectotype of Salicornia macrostachya, the neotypes of Tenore's varieties of S. fruticosa and S. macrostachya, one of