Nomenclature in Mesembryanthema ( Aizoaceae ) : the generic names by Rappa and Camarrone

A re-examination of the seven generic names used by Rappa and Camarrone reveals that three of them were validly published (Pentacoilanthus, Tetracoilanthus, Pteropentacoilanthus)\ the first one is lectotypified here. The other four names are invalidly published. Their synonymy is given nevertheless where possible.


INTRODUCTION
In a survey of all genera of the Aizoaceae (including the Mesembryanthema) (H .E .K .H .) it was found that the genera described by Rappa and Cam arrone (1953, 1955, 1960) have never been examined critically, and their validity and synonymy have never been established.Based on recent investigations in the group involved (V.B.), it is now possible to settle the nomenclatural uncertainties.Early investigations of flowers of Mesembryanthema led Rappa (1912) to the description and distinction of two different types of nectaries: lophomorphic ones, which are crest-shaped (Figure 1A,B), and coilomorphic ones, which are conchiform or shell-shaped (Figure 2A, B).Nectaries may also be absent, and consequently Rappa (1912) names three groups: Lofomorfi, Anettari, and Coilomorfi, which were later classified as subfamilies (Rappa & Camarrone 1953, 1960).They lack Latin descriptions, however, and therefore do not conform with article 36 of the Interna tional Code of Botanical Nomenclature (Greuter et al. 1988, abbreviated ICBN subsequently).The names are therefore not validly published.
The fundamental distinction between lophomorphic and coilomorphic types of nectaries has been confirmed repeatedly (e.g.Ihlenfeldt 1960).Bittrich (1987) stresses the fact that the subfamily Mesembryanthemoideae (= Aptenioideae Schwant.ex Bittrich & H. H artm .) is characterized by coilomorphic nectaries, and Bittrich & Hartmann (1988) mention them as a synapomorphic character for the subfamily.It is also worth noting that in Aptenioideae the number of carpels agrees with the number of perianth lobes as well as with the num ber of nectaries.N .E. Brown (1925) already used this character set in his keys to various genera.It must be rem em bered, though, that neither the number of carpels nor that of the nectaries is always constant within populations (e.g.Phyllobolus subg.Prenia, Bittrich 1987).Rappa & Camarrone (1953, 1955, 1960) use the number and shapes   (1955,1960) by the absence or presence of valve wings.In the first step, two genera based on five, respectively four, nectaries are distinguished: Pentacoilanthus and Tetracoilanthus (Rappa & Cam arrone 1953).Four more genera are added later after the importance of valve wings is recognized: Perapentacoilanthus (Rappa & Camarrone 1955), Pteropentacoilanthus, Pterotetracoilanthus, and Peratetracoilanthus (Rappa & Cam arrone 1960).The names indicate that a certain pattern is followed predicting particular character combinations for the genera.This approach can be traced from the recognition of informal groupings (Rappa & Cam arrone 1955) to the formal descriptions of taxa (Rappa & Cam arrone 1960).The mechanical and even predictive procedure followed is well illustrated by the description of Pterotetracoilanthus, with a character combination which is theoretically possible but which the authors failed to find in any real plant (see 6. below).This unconventional approach is certainly one reason why the descriptions of Rappa & Cam arrone have been neglected.Another reason is that, in most cases, the authors cite several species for their genera taken from various other genera previously described.Also, the new genera were not typified, making identification extremely difficult.Nevertheless, none of these shortcomings alone invalidate the names as such, and a detailed re-investigation was carried out to establish a reliable taxonomy for future use.Since each genus poses its own problems, the genera are discussed separately in chronological order.DISCUSSION 1. Pentacoilanthus Rappa & Camarrone (1953) The description conforms with the relevant requirements (art.3 2 -3 6 ICBN), and the name is therefore validly published.Four species are included, Mesembryanthemum aitonis, M. granulicaule, M. splendens and M. crystallinum.Unfortunately no type species is mentioned and a lectotype must therefore be chosen.
(1) Mesembryanthemum aitonis Jacq.can be excluded because, in contrast to the genus description of Rappa & Camarrone (1953) the nectaries of this species are not shell-shaped but tubular; these tubes may extend as dee ply into the receptacle as the locules (Figure 3).It seems unlikely that Rappa & Camarrone studied longitudinal sec tions and it is therefore assumed that they were unaware of the different nectary morphology of this species.Sur face views do not permit the assessment of the depth of the nectary (Figure 2A, B).M. aitonis is therefore not available for selection as a lectotype.At present, the spe cies is included in Mesembryanthemum L. subg.Mesem bryanthemum (Bittrich 1987).The remaining three species do possess shell-shaped nectaries, and the choice between them is not easy.
(2) Mesembryanthemum granulicaule Haw.(belonging to the genus Psilocaulon N.H.B r.) is a doubtful species.It is often synonymized with Psilocaulon articulation (Thunb.)N.E.Br. (the type species of the genus Psilocaulon) because of the rough papillose surface of leaves and stem; the latter species has white to pink flowers.The drawing of M. granulicaule, however, which is kept at Kew and is assumed to have been seen by Haworth (it would have to be chosen as the lectotype of the species in a revision), shows yellow flowers.The very few herbarium specimens with yellow flowers were aíl collected in a small area in the Orange Free State.They may well represent the true M. granulicaule as studied by Haworth.It seems extremely unlikely, though, that Rappa & Camarrone have investigated living or pickled material of this very rare species and herbarium material is useless for the investigation of nectaries.Furtherm ore, in species of Psilocaulon, four-and five-locular ovaries can occur within one species or even on one plant.Species of Psilocaulon can therefore not be associated unambiguously with either Pentacoilanthus or Tetracoilan thus.Considering the care with which Rappa and Camarrone led their investigations, it seems highly unlikely that they saw species, or even specimens, with both 4-and 5-lobed ovaries.
(3) Mesembryanthemum splendens [now included in Pliyllobolus subg.A in la n d (N .E. B r.) Bittrich. Bittrich (1987)| is a doubtful species because its identity is very difficult to determine.It belongs to a group of about twenty described species in the subgenus Aridaria most of which arc synonymous.Due to the insufficient knowledge of the group, however, the exact num ber and delim itation of the species is still uncertain.It is scarcely possible to identify any species of this group with the help of literature only.The most useful characters are form and sculpture of seeds, rarely mentioned in descriptions.It is unlikely therefore, that Rappa & Cam arrone investigated the true M. splendens.particularly since it is very rare in cultivation.Material kept under this name often belongs to the closely allied M. umbelliflorus Jacq.It would seem highly unwise to choose this species a s the lectotype.
(4) Mesembryanthemum crxstallinum L. is the only species figured by Rappa &. Cam arrone (1953).This species has not only been well known in Europe for a long time, it is also naturalized around the M editerranean and is sometimes used as a vegetable.The figure mentioned above shows w ith great certainty a flower of M. crystallinum.because the extremely large bladder cells of the epidermis of the receptacle are highly characteristic of this species.Therefore, and in contrast to the two other species with shell-shaped nectaries discussed above, it is most probable that M. crxstallinum is the plant studied by Rappa & Camarrone.Consequently, M. crxstallinum is chosen as the lectotype of the genus Pentacoilanthus.At the same time, M. crxstallinum is the type species of the genus Crxophytum (N.E. Brown 1926).By this lectotypihcation.the name Pentacoilanthus becomes superfluous and is illegitimate (art.63.1 ICBN).It should be stressed, though, that the choice was not influenced by this consequence hut is based on agreement of characters w ith the descrip tion and the probable identity of the original material.
It should be mentioned further that the name Pentacoilan thus would in any case be a synonym of another generic name, since all relevant species had been described earlier.I he name Pentacoilanthus crxstallinum, is not validly published because the basionym was not cited clearly and fully as required (art.33.2 ICBN).At present, the species is considered to belong to Mesembryanthemum L. subg.Crxophytum (N.E.Br.) Bittrich (Bittrich 1987).Rappa & Camarrone (1953) The description conforms with the relevant requirements (art.3 2 -3 6 ICBN), and the name is therefore validly published.Since only one species is included, this is accepted as the indication of the type (art.37.2 ICBN).
The combination T. cordifolius (L.f.) Rappa & Camar rone is not validly published, because the basionym is not indicated clearly and fully (art.33.2 ICBN).At present, the species is included in the genus Aptenia N.E.Br. (Bittrich 1987).Rappa & Camarrone (1955) The description of this genus is almost identical with that of Pentacoilanthus Rappa & Camarrone 1953 (see 1. above) except for the addition of 'valvis... quae, septis deficientibus.circa nudum axem consistunt'.This statement probably refers to the columella, the formation of which is very variable and changes from low to high with all intermediates.In Aptenioideae, the septa are always transformed into expanding keels in their upper parts, and consequently a remaining collumella will appear naked in the open capsule.These inconsistencies detract from the taxonomic value of the character.The two descriptions are consequently considered to be identical.

Perapentacoilanthus
Although Rappa and Camarrone did not state it explicitly, there is no doubt that they coined the name Perapentacoilanthus to replace Pentacoilanthus Rappa & Camarrone (1953).They had realised in the mean time that the absence or presence of valve wings is an important character.Consequently they were now planning to group species and name genera not only according to the number of nectaries, but also according to the absence or presence of valve wings, as described in the introduction above.This is borne out by the following: a, under the heading Pentacoilanthus.Genere privo di umcnoprosteci' (genus deprived of valve wings), Rappa & Camarrone (1955) indicate that they intend to alter the diagnostic characters of Pentacoilanthus, which indeed they did later (Rappa & Camarrone 1960: 13); b, in the 1960 publication they include all the species originally mentioned under Pentacoilanthus Rappa & Camarrone (1953) under Perapentacoilanthus.This action was obviously initiated with the aim to name the genera with five, respectively four, coilomorphic nectaries in parallel fashion (Pentacoilanthus matching Tetracoilanthus without valve wings; Pteropentacoilanthus-Pterotetracoilanthus with free valve wings; Perapentacoilanthus -Peratetracoilanthus with valve wings forming pockets; compare list in Rappa & Cam arrone 1955: 10).
To declare Perapentacoilanthus a superfluous and therefore illegitimate name (art.63.1 ICBN) requires the definite inclusion of the holotype of the earlier synonym Pentacoilanthus.This is not possible literally, since Pen tacoilanthus 1953 was published without the indication of a type, but the inclusion of all four species of the original Pentacoilanthus 1953 under Perapentacoilanthus (Rappa & Cam arrone 1960) makes it clear that the type would be included, whichever species had been chosen.As a consequence, Pentacoilanthus 1953 and Perapentacoilan thus 1955 must be considered synonymous, being names for the same genus.

Pentacoilanthus Rappa & Cam arrone (1960)
As stated above, the characterization of the genus differs from the Pentacoilanthus of 1953 in an important diagnostic character, namely the absence of valve wings, and an entirely different set of species is accordingly assigned to the genus.This name therefore represents a later homonym of Pentacoilanthus Rappa & Camarrone 1953 (although no type is mentioned in either genus) and is consequently illegitimate (art.64.1 ICBN).Further more, the lack of citation of a type at this date makes the name an invalidly published one (art.37.1 ICBN ).
All species included in Pentacoilanthus 1960 were taken from the genus Sceletium N.E.Br., and they are at present assigned to Phyllobolus N.E.Br. subg.Sceletium (N.E.Br.) Bittrich (Bittrich 1987).Rappa & Cam arrone (1960) Only one species is cited as an example with the description of the genus and this validates the name (art.37.2 ICBN).The genus is based on the same species as Halenbergia Dinter, namely H. hypertrophica (Dinter) Dinter.The name Pteropentacoilanthus is therefore a superfluous name and illegitimate (art.63.1 ICBN).

Pterotetracoilanthus
Perapentacoilanthus Rappa & Camarrone 1960, nom. illeg.,Note on the holotype: there is no doubt that the sheet cited and seen is the holotype since the description on the back refers to the original publication.The specimen agrees well with the description.The holotype was already recognized by both Norlindh and Glen in 1976, but there seems to be no published reference to this.In any case, the existence of the holotype supersedes the selection of a neotype by Preston & Sell (1988).Halenbergia Dinter: 200 (1937).

7.
Peratetracoilanthus Rappa & Cam arrone (1960)    Eight species are cited with the description, but no type is chosen, and the name is therefore invalidly published (art.37.1 ICBN).The new combinations are invalid as well because no basionyms are given (art.33.2 ICBN), and the name of the genus is invalidly published (art.43.1 ICBN).