|Constancea 83, 2002|
University and Jepson Herbaria
P.C. Silva Festschrift
The genus Inkyuleea Choi, Kraft, & Saunders (2000) previously tentatively included in the Ceramiales is here referred to the Ceramioideae on the basis of ceramialean post-fertilization stages and presence of whorl-branches on the vegetative thallus. The genus Irtugovia Perestenko (1994) originally created for three North Pacific species of Antithamnionella with distichous thallus ramification is cladistically supported, but should be expanded to include Antithamnionella floccosa, A. nigricans, and A. alternans. Irtugovia is, however, antedated by Haplocladium Nägeli (1862) and usage of the former name requires its conservation. The genus Liagorothamnion Huisman, Ballantine & Wynne (2000) originally considered to be closely related to the Dohrnielleae is excluded from the Ceramioideae and the Ceramiales, since the type species L. mucoides lacks essential ceramialean features such as procarps, auxiliary cells, and connecting cells, and possesses connecting (2-celled) filaments and nutritive cells issued from hypogynous cells. A cladistic analysis of six primitive ceramiaceous genera indicates that the ceramialean ancestor was probably similar to Warrenia, exhibiting a bilaterally ramified thallus with lateral filaments of unlimited (or latent) growth, transverse apical divisions, gland cells, carpogonial branches borne on intercalary axial cells, procarpic post-fertilization stages, and tetrasporophytes with cruciately-decussately divided tetrasporangia.
The critical study of the Ceramiaceae begins with Nägeli's (1862) monograph where twenty-three genera and eighteen subgenera are formally described. Yet, Nägeli introduced binomial combinations for both, irrespective of rank (Silva 1970, p. 942), and therefore at least two of his subgenera have now been accepted (or conserved) with generic rank. Nägeli's particular interest for the Ceramiaceae is revealed in two previous publications where his enormous descriptive capacity is both literally and illustratively displayed (Nägeli 1847, Nägeli and Cramer 1855). His drawings show for the first time the presence of primary pit plugs in red algae, and allow unequivocal identification of his material with distinct species. Nägeli was also first to describe the procarps of the Ceramiaceae (Nägeli 1862, pl. 1, figs. 28, 29), later studied in greater detail by Bornet and Thuret (1876, 1880), Schmitz (1883), Zerlang (1889), Phillips (1897, 1898), and Oltmanns (1898).
Nägeli's (1862) classification of the Ceramiaceae was less successful, as he included several other genera presently assigned to the orders Acrochaetiales and Gigartinales. A more modern classification of the family was given by Schmitz (1889) who recognized thirty-nine genera within fourteen tribes. Later, Schmitz and Hauptfleisch (1897, p. 481504) increased the number of tribes to twenty-five and the number of genera to forty.
Kylin (1930, p. 76; 1937, p. 105) did not pay much attention to all these tribes, but he divided the Ceramiaceae into two lineages, including the Crouanieae, Wrangelieae, and Ceramieae in the first group, and the Spermothamnieae, Griffithsieae, Monosporeae, Callithamnieae, and Ptiloteae in the second one. He pointed out the presence of whorl-branches (lateral filaments of limited growth) in the first group, where he regarded the position of the carpogonial branch on fully developed or reduced whorl-branches as an evolutionary transition.
Feldmann-Mazoyer (1941, p. 243) adopted Kylin's division of the Ceramiaceae, adding in the first group the Dohrnielleae, Spyrideae, Ptilocladiopseae(1), and Callithamnieae, and maintaining in the second group the Ptiloteae, Spermothamnieae, and Griffithsieae, together with the Compsothamnieae and Sphondylothamnieae. Like Kylin, her phylogenetic considerations were limited to Mediterranean taxa, but compared to him (or even Nägeli), Feldmann-Mazoyer was a lumper. She retained Antithamnion in the Crouanieae to include species previously referred to Antithamnion Nägeli, Antithamnionella Lyle, and Pterothamnion Nägeli (= Platythamnion J. Agardh). Itono and Tanaka assigned a fourth genus, Balliella Itono & Tanaka to the tribe for members of Antithamnion having gland cells connected distally to periaxial cells. Each of these four genera is presently assigned to a separate tribe.
In his monograph of the red algae, Kylin (1956) maintained Feldmann-Mazoyer's broad concept of Crouanieae, including twelve genera in this tribe, although he recognized Antithamnion, Antithamnionella, and Platythamnion as being generically distinct. Kylin accepted only ten other ceramiaceous tribes (of the twenty-five recognized by Schmitz and Hauptfleisch) but increased the number of genera to sixty-nine (from forty).
In his thesis, Hommersand (1963) also divided the Ceramiaceae into two groups, but most significantly erected the Antithamnieae to include nine former crouanioid genera, viz. Acrothamnion J. Agardh, Antithamnion, Antithamnionella, Ballia Harvey, Bracebridgea J. Agardh, Heterothamnion J. Agardh, Platythamnion, Ptilocladia Sonder, and Warrenia Harvey. He also considered the families Rhodomelaceae, Delesseriaceae, and Dasyaceae as independent evolutionary lines arising from the (paraphyletic) Antithamnieae (Hommersand 1963, p. 343, fig. 52).
Wollaston (1968, 1971, 1972a, 1972b, 1972c, 1974, 1976, 1977a, 1977b, 1978, 1979, 1980, 1984, 1990) considerably expanded our knowledge of Indo-Pacific members of the tribes Crouanieae, Antithamnieae, Dasyphileae, and the newly established Heterothamnieae, describing several new exotic genera, and soon more colleagues followed her in studying the Ceramiaceae from other parts of the world (see Moe and Silva 1979, 1980, 1983; Stegenga 1986, Itono 1977, Norris 1987).
All these taxonomic studies made it possible to attempt a phylogenetic analysis of the subfamily Ceramioideae, after reexamining representative material of over hundred taxa (Athanasiadis 1996). A second significant step to accomplish the phylogenetic analysis was the extensive biosystematic studies on species of Antithamnion, Antithamnionella, Scagelia Wollaston and Pterothamnion (Rueness and Rueness 1973, 1975; Rueness 1978; Sundene 1959, 1962, 1964, 1975; Lee and West 1980; Notoya and Yabu 1980; Boo and Lee 1983; Athanasiadis 1983, 1985, 1986; Jacobsen et al. 1991; Athanasiadis and Rueness 1992). These studies clarified species concepts and made possible the circumscription of taxa as natural units.
Cladistic analysis of the Ceramioideae indicated the presence of several monophyletic groups, that were accepted as distinct tribes. The results also supported Hommersand's evolutionary proposal of a paraphyletic Ceramioideae with respect to the remaining ceramialean families (Athanasiadis 1996).
In the past eight years, several new taxa have been added to the Ceramioideae (or Ceramiales), the most significant additions being the monogeneric new tribe Liagorothamnieae (Huisman et al. 2000) considered to be related to the Dohrnielleae, the new genus Inkyuleea Choi et al. (2000) to accommodate certain species of Ballia (the type species being segregated into the new order Balliales), the new genus Irtugovia Perestenko (1994) to accommodate the distichously ramified species of Antithamnionella, and the new genus Elisiella Womersley (in Womersley and Wollaston 1998a, p. 193) of the Heterothamnieae. Moreover, several new species have been added to the genera Antithamnion, Antithamnionella, and Perithamnion.
In this paper I will assess whether these new taxa should be retained in the Ceramioideae.
Indeterminate axes occur in the thallus of all red algae whether organization is uniaxial or multiaxial and whether growth is monopodial or sympodial. These (axial) filaments exhibit continuous apical and intercalary cell divisions, the former adding to the length of the filament and the latter producing all other peripheral filaments including new indeterminate axes. At least in the Ceramiales, the first indeterminate axis is readily recognized a few divisions after the germination of single spores and remains detectable in the thallus due to its unique position (always central) and specialised function (continuous apical and intercalary cell divisions). New indeterminate axes arise in three possible ways: (a) at regular or irregular intervals from indeterminate axial cells, either via apical (subdichotomous) or intercalary (lateral) divisions, (b) through further apical growth of lateral determinate axes (see below), or (c) adventitiously from cells of whorl-branches (see further below). With regard to their origin, Wollaston (1968, p. 219, 220) distinguished between main axes, axial branches and lateral indeterminate branches, but she also recognized that in all these indeterminate axes the growth process was identical.
Determinate axes or potentially indeterminate branches (sensu Gordon 1972, p. 6) develop laterally from indeterminate axes and are identical to them except in the temporary inhibition of their apical growth. In other words, determinate axes are lateral indeterminate axes with latent growth and this is confirmed by the identical structure and function that these two filament types display, as soon as determinate axes resume growth in older parts of the thallus (Fig. 1), or i]f damage occurs to the growing apex (Gordon 1972, p 6). Within the Ceramiales, determinate axes occur in the most primitive tribes Delesseriopseae and Warrenieae, but also in all other ceramiaceous tribes, such as the Ptiloteae, Callithamnieae, or Rhodocalleae(2), which do not belong to the Ceramioideae.
Like determinate axes, whorl-branches develop also laterally from indeterminate axes, but differ both functionally (being of limited growth and producing adventitious indeterminate axes from their cells) and structurally (displaying specialised ramification). Originally these filaments were named whorl-branchlets (Wollaston 1968, p. 220), but Moe and Silva (1980, p. 2) adopted the term whorl-branches because the ramification is primary with respect to the indeterminate axes that bear them and because they can be ramified producing branchlets up to the fourth order. It is apparent that in the Ceramiales whorl-branches represent an evolutionary modification of determinate axes, because these the two filament types occupy the same position on the thallus and are mutually exclusive. Therefore, it is possible to consider determinate axes and whorl-branches as two independent states of the same character, which however provides the same phylogenetic information as if these states are scored as independent characters (see The Ceramialean Ancestor). In most tribes of the Ceramioideae (such as the Pterothamnieae, Dohrnielleae, Scagelieae, Antithamnieae, Perithamnieae, Heterothamnieae), whorl-branches display an open organization becoming diminutive in Amoenothamnion Wollaston and Leptoklonion Athanasiadis and at least in one species of Ceramium (i.e., C. shepherdii Womersley), before they completely condense in the advanced families Rhodomelaceae, Dasyaceae, and Delesseriaceae. Yet, throughout these advanced groups, whorl-branches retain their unique function and structure which makes possible their identification as a synapomorphy.
Lateral branches of limited growth are also evident in certain species of ceramiaceous genera not included in the Ceramioideae, such as in Callithamnion Lyngbye of the Callithamnieae. In particular, the lateral branches of Callithamnion tetragonum (Withering) S.F. Gray may display a limited growth with adventitious production of new axes from intercalary cells (e.g., see Rosenvinge 19231924, fig. 232 B; and personal observations), and the same may be true for other members of this genus as also for Aglaothamnion Feldmann-Mazoyer and Pleonosporium Nägeli. Yet, the fact that these species belong to genera or tribes that also accommodate members having determinate axes as laterals, suggests that their branches of limited growth have developed independently and should not be considered homologous to whorl-branches. This hypothesis is also supported by the fact that lateral branching in these genera or tribes results through subdichotomous (or oblique) and not lateral divisions, and that opposite ramification is generally lacking on their vegetative thallus. In other cases where lateral branches of limited growth have been described (e.g., Haloplegma Montagne and Lasiothalia Harvey; see Womersley and Wollaston 1998b, p. 284, and below), the available information is not sufficient to allow unequivocal identification of filament type.
Whorl-branching is apparently also evident in several genera of the Batrachospermales and Gigartinales (e.g., Acrosymphyton Sjöstedt and Atractophora Crouan frat.), where the lateral filaments (often termed cortical filaments) display specialised ramification and a limited growth. Moreover in Acrosymphyton, adventitious development of new axes from whorl-branch cells has been demonstrated(3), which indicates that the whorl-branching in these remote taxa should be regarded as a case of parallelism.
The term determinate filaments was used by Hommersand et al. (1998) to define the cortical filaments of Rhodocallis. It is, however, not possible to apply comparatively, because the ceramialean thallus exhibits several other distinct kinds of determinate filaments, such as whorl-branches, trichoblasts, or spines, structures that are not homologous to cortical filaments, while the cortical filaments of Ceramium are condensed whorl-branches and not homologous to the determinate filaments of Rhodocallis indicating that cortication developed independently several times in the Ceramiales.
While the presence of periaxial c.bs on whorl-branches is unequivocally a synapomorphy for the Ceramioideae, it is logical to assume that periaxial c.bs originally appeared on determinate axes. Yet, this hypothesis implicates other primitive genera as the potential sister-taxon of the Ceramioideae, as is discussed below (see The Ceramialean Ancestor).
LASIOTHALIEAE Womersley in Wollaston and Womersley (1998a, p. 37; type: Lasiothalia Harvey)
COMMENTS: The monotypic tribe Lasiothalieae was established to accommodate Lasiothalia Harvey, but the nature of several characters of the generitype L. hirsuta Harvey remains unclear. In particular, the lateral filaments of the main thallus were said to be whorl-branchlets borne opposite to each other in unequal pairs (Wollaston and Womersley 1998a, p. 38); fig. 11 B of the same publication, however, shows laterals of unequal length and development, suggesting that these whorl-branchlets should rather be considered as laterals of unlimited growth (i.e., determinate axes). Similarly, the nature of the procarp-bearing filaments (said to be specialised, pilose cortical filaments sensu Wollaston 1990) is unclear. As suggested before (Athanasiadis 1996, p. 13), it is possible that these pilose cortical filaments are new adventitious axes and not filaments of limited growth. Therefore Lasiothalia appears to lack whorl-branches and transverse ramification on the main thallus (Athanasiadis 1996, p. 113), and since none of the other ceramioid synapomorphies occur in this genus a position within the subfamily Ceramioideae cannot be supported. Moreover, the presence of spermatangial heads and tetrahedrally divided (subspherical) tetrasporangia are advanced characters and therefore a primitive position in the Ceramiaceae can be ruled out. If the development of carpogonial branches on specialised ('pilose') filaments (of limited growth) remote from the main axes is correctly interpreted, it would provide a relationship with Haloplegma and Spongoclonium Sonder, where carpogonial branches are also said to develop on lateral (surface) filaments of limited growth (see Womersley and Wollaston 1998b, p. 284; Womersley and Wollaston 1998c, p. 287).
LIAGOROTHAMNIEAE Huisman, Ballantine & Wynne (2000, p. 515; type: Liagorothamnion Huisman, Ballantine & Wynne)
Liagorothamnion Huisman, Ballantine & Wynne (2000, p. 507; type: L. mucoides Huisman, Ballantine & Wynne)
Liagorothamnion mucoides Huisman, Ballantine & Wynne (2000, p. 507, figs. 119; type locality: Leeward Media Luna Reef, La Parguera, 1.0 m, Caribbean Sea; holotype: in MICH)
COMMENTS: The monotypic tribe Liagorothamnieae was established to accommodate the new genus and species Liagorothamnion mucoides from the Caribbean Sea. Although a position in the Ceramioideae was not explicitly stated, the authors considered the Liagorothamnieae to be most closely related to the Dohrnielleae. Yet, Liagorothamnion lacks essential reproductive features to qualify for a position in the Ceramiales. In particular, the production of spermatangial systems in whorls on a series of 316 axial cells (Huisman et al. 2000, abstract and fig. 9) is not previously reported in the Ceramiaceae, where, in the Ceramioideae, spermatangial systems are located either on (specialised) branchlets (of limited growth) or on outer whorl-branch cells (see Athanasiadis 1996, p. 28, fig. 3), while in the Callithamnioideae they occur either terminally on short determinate axes (e.g., in Gymnothamnion J. Agardh; see Kajimura 1989b) or on specialised branchlets forming cushions or heads (e.g., see Gordon 1972, Maggs and Hommersand 1993).
The periaxial position of single 3- or 4-celled carpogonial branches is analogous to that in several rhodophyte orders, including the Ceramiales, but the production of more ('sterile') cells from hypogynous cells (Huisman et al. 2000, figs. 12, 14) is not known in any ceramialean member. Yet, such cells are well-documented in gigartinalean families such as the Naccariaceae (e.g., see Kylin 1928, fig. 8 D), and a nutritive function has been suggested (see Abbott 1985, p. 559). Production of two groups of ('sterile') cells from the supporting cell itself is known to occur in a variety of families, including members of the Dasyaceae, Rhodomelaceae, and Delesseriaceae, but it is unlike what occurs in the Ceramiaceae. At least in the Ceramioideae, such cells are simply the remains of reduced whorl-branches (hence the absence of several sterile cells in those genera where the periaxial cell supports a single vegetative branch). Even more anomalous, by ceramialean standards, is the apparent lack of auxiliary cells and the putative transfer of the zygote from the carpogonium to contiguous periaxial cells (i.e., basal cells of opposite or transverse branches), via connecting (2-celled ?) filaments. Failure of auxiliary cell development is definitely a rare event in the Ceramiales and in the few species or genera recorded, it is generally interpreted as a secondarily derived condition (through reduction). Yet, in the monotypic Liagorothamnion it could equally represent a plesiomorphous condition. The fact that the single fertilization results in the diploidization of three or four [periaxial cells = non accessory auxiliary cells] (Huisman 2000 et al. 2000, p. 509), indicates that in Liagorothamnion the post-fertilization is non-procarpic and definitely not homologous to that in the Ceramiales. Finally, production of quadripartite carposporangia is presently recorded in a variety of families (i.e., Gelidiaceae, Liagoraceae, Phyllophoraceae, Palmariaceae, and Achrochaetiaceae) and, while such structures are unknown in the Ceramiales, their presence can not provide evidence for excluding or including Liagorothamnion in the Ceramiaceae.
Several vegetative features of Liagorothamnion are considered to be potentially homologous to those of tribes in the Ceramioideae, but this is not well-documented for two fundamental characters recognized as apomorphies for the entire subfamily. Lateral filaments in Liagorothamnion arise in successive whorls on axial cells, each whorl including one indeterminate axis and 14 whorl-branch[es]. Yet, no production of adventitious axes is described from whorl-branch cells, while the statements that many axial cells bear whorl-branch[es] of varying lengths..., generally simple, 115 cells long suggest that these filaments may be homologous to determinate axes. Conversely, if these filaments are indeed whorl-branches, they can also be homologous to those occurring in members of the Gigartinales such as Atractophora hypnoides Crouan frat. (see Bornet and Thuret 1876, pl. 17; Kylin 1928, 5 A-D), Dudresnaya Crouan frat. (see Robins and Kraft 1985, fig. 50), or Acrosymphyton (see Millar and Kraft 1984).
Moreover, the presence of rhizoidal outgrowths, small periaxial cells, and papilliform branch cells can hardly be considered as diagnostic characters to support a position near the Dohrnielleae, when significant reproductive features are alien to those known in the Ceramiales and the Ceramioideae in particular. Liagorothamnion is here excluded from the Ceramioideae, the Ceramiaceae, and the Ceramiales, and a possible affiliation with the Naccariaceae or another family of the Gigartinales is suggested.
Inkyuleea Choi, Kraft & Saunders [2000, p 285; type species: I. ballioides (Sonder) Choi, Kraft & Saunders]
Inkyuleea ballioides (Sonder) Choi, Kraft & Saunders (2000, p. 285)
Inkyuleea mariana (Harvey) Choi, Kraft & Saunders (2000, p. 285)
Inkyuleea beckeri (Mazza) Choi, Kraft & Saunders (2000, p. 285)
COMMENTS: The genus Ballia, including the type, B. callitricha Harvey, B. pennoides Wollaston, and B. nana Kraft & Saunders has recently been transferred to the new family Balliaceae within the new order Balliales (Choi et al. 2000). The molecular evidence presented by these authors is strongly supported by the occurrence of domed pit plugs (see Archer 1876) and the unusual linear chromosomes observed in the generitype (Athanasiadis 1996, fig. 5). It could be added that the lateral filaments of B. callitricha are not of determinate growth (as explicitly stated in the diagnoses of the new taxa or the emended genus; Choi et al. 2000, p. 278), because these filaments gradually develop into new axes (and hence are homologous to determinate axes). In Ballia callitricha, this feature was originally observed by Archer (1876, pl. XXVIII, fig. 15, the branches...become secondary rachises [i.e., axes]) and more recently confirmed in culture material of this species (Athanasiadis 1996, p. 12, 201).
The southern Australian species Ballia mariana Harvey and B. ballioides (Sonder) Wollaston, together with B. beckeri Mazza from South Africa were assigned to the new genus Inkyuleea, which was tentatively retained in the Ceramiales. In considering whether Inkyuleea qualifies for a position in the Ceramiales, and within the Ceramioideae in particular, it is significant to stress that although it is uncertain whether the auxiliary cell branch (sensu Wollaston 1974, p. 23) is a de novo structure, or if it represents the first cell of the gonimoblast with the attached remnants of the connecting cell, or if it is an aberration, the genus fulfils the ordinal reproductive criteria, possessing: (1) 4-celled periaxial carpogonial branches, (2) development of an auxiliary cell after fertilization, and (3) development of a connecting cell (or tube) linking the fertilized carpogonium with a single (auxiliary) cell. Moreover, considering the nature of the auxiliary cell branch in Inkyuleea, it should be mentioned that Athanasiadis and Kraft (1994, fig. 11) have also illustrated a two-celled auxiliary branch in Pterothamnion, where the auxiliary cell cuts off apically a smaller cell that receives the zygote (or the male nucleus) from the trichogyne (Fig. 12). The nature of such auxiliary cell branches (whether rare or aberrant forms) needs to be confirmed by further studies, since in the Ceramiales it is the carpogonium itself that generally produces the connecting cell as an extension tube (Fig. 13). Yet, Inkyuleea also differs from the remaining Ceramiales by developing adaxial carpogonial branches on periaxial cells. This unusual position is alien to all ceramialean tribes and families, where carpogonial branches are consistently borne either abaxially or laterally on supporting cells (e.g., see Athanasiadis and Rueness 1992, figs. 13). The ontogeny of adaxial carpogonial branches in Inkyuleea may be a unique (aut)apomorphy or it may be homologous to the origin of adaxial gland cells in Balliella (Figs. 8, 9) (4) in which the presence of a series of ancestors having both adaxial and abaxial carpogonial branches on single periaxial cells (Fig. 14) (5) may be assumed.
As regards the structure of the vegetative thallus, the lateral filaments of Inkyuleea are of limited growth and display specialised ramification. In particular, the lateral filaments of Inkyuleea mariana are dimorphic and analogous to Trithamnion Wollaston (Dohrnielleae), having one major filament (tristichously ramified as occurs in Pterothamnion subgenus Platythamnion) opposite a pair of minor ones. In Inkyuleea ballioides, the ramification of the major lateral filament is distichous (as in many species of Pterothamnion). All these specialized patterns of ramification indicate that the lateral filaments of Inkyuleea are homologous to whorl-branches, although development of adventitious axes from whorl-branch cells has not been reported.
In conclusion, the whorl-branch organization of Inkyuleea, along with the presence of transverse ramification, and procarps with periaxial carpogonial branches collectively support a position in the Ceramioideae, if the following assumptions can be made:
Irtugovia Perestenko [1994, p. 204; type: I. shimamurana (Nagai) Perestenko]
COMMENTS: The first cladistic analysis of the Dohrnielleae indicated that the genus Antithamnionella Lyle is paraphyletic with regard to the other four members of this tribe, i.e., Callithamniella Feldmann-Mazoyer, Dohrniella Funk, Trithamnion Wollaston, and Acrothamniopsis Athanasiadis & Kraft (Athanasiadis 1996, p. 102). Parallel to this work, Perestenko (1994) established the new genus Irtugovia to accommodate three boreal species of Antithamnionella, viz. I. shimamurana (Nagai) Perestenko, I. spirographidis (Schiffner) Perestenko, and I. pacifica (Harvey) Perestenko. Athanasiadis (1996, p. 123) included I. shimamurana and I. pacifica (both as Antithamnionella) in the list of potential synonyms of Antithamnionella floccosa (O.F. Müller) Whittick, that together with Antithamnionella alternans Ricker, Antithamnionella nigricans (Gardner) Athanasiadis and Antithamnionella (Irtugovia) spirographidis, formed a distinct group in the strict consensus cladogram of the parsimony analysis. This clade was supported by a unique apomorphy (the presence of distichous thallus ramification) and one homoplasy (the presence of successively divided tetrasporangia) (Athanasiadis 1996, p. 88, cladogram 5). Recognition of Irtugovia as a distinct lineage within the Dohrnielleae is therefore cladistically supported, although the fate of the remaining paraphyletic species of Antithamnionella remains unclarified and needs further studies.
Nomenclaturally, Irtugovia is antedated by Haplocladium Nägeli (1862, p. 377) which is based upon its type (and only) species Haplocladium floccosum (O.F. Müller) Nägeli (1862, p. 378). Although Nägeli originally intended Haplocladium to be a subgenus of Pterothamnion, he introduced the combination H. floccosum which could be interpreted as an elevation to generic rank (Silva in Whittick 1980, p. 78). This policy has been apparently adopted for at least one more (sub)genus described by Nägeli, viz. Dasythamnion Nägeli (Silva 1970, p. 19), while Pleonosporium Nägeli is listed as a nom. cons. in the ICBN. The case of Haplocladium Nägeli is, however, complicated by the usage of this name for a genus of Bryophyta with nearly hundred species (i.e., Haplocladium J.C. Müller 1896; see Whittick 1980, p 78). A formal proposal to conserve Haplocladium J.C. Müller 1896, or Irtugovia Perestenko 1994, against Haplocladium Nägeli 1862 remains to be made.
Antithamnionella longicellulata Perestenko [1994, p. 204; type locality: Sidimi Gulf, Gulf of Peter the Great; holotype: Mare Japonicum, sinus Petri Magni; sinus Sidimi, 5 m alt. in arena, 18 VI 1957, K.L. Ivanova (Vinogradova) legit (in LE)]
Antithamnionella nagaii Perestenko [1994, p. 204, tab. XXII figs. 8, 9; type locality: Commander Islands; holotype: Insulae Kommandorenses (Praefectoriae), insula Beringii, E.A. Kardakova legit (in LE)]
COMMENTS: The descriptions of these two new species given by Perestenko are brief and do not include reproductive characters from gametangial plants, which is essential information to confirm the generic position. Antithamnionella nagaii may belong to the same entity described by Nagai (1941, pl. V fig. 4, pl. VI figs. 12, 13) as Antithamnion sp. Recognizing the cold temperate genus Irtugovia as distinct from Antithamnionella, the latter genus accommodates seventeen species mainly distributed in the tropical and warm temperate regions of the world (see Athanasiadis 1996, p. 101).
Antithamnion aglandum Kim & Lee in Kim and Chah (1996, p. 204; type locality: Cheju Island, 127 E 33 N, Korea; holotype: SNU9003241)
Antithamnion verticale (Harvey) J. Agardh
COMMENTS: Antithamnion aglandum from Korea was morphologically considered to be closely related to Antithamnion pectinatum (Montagne) Athanasiadis & Tittley (Kim and Chah 1996, as A. nipponicum Yamada), while a comparison of rbcS sequences by Lee et al. (2001) indicated a closer relationship with Antithamnion callocladum Itono. These three species belong to the subgenus Pteroton Athanasiadis (1996, p. 142) that is distinguished by an orthostichous thallus ramification. Antithamnion aglandum lacks gland cells and in this respect represents a clear case of parallelism along with remote glandless members of the genus from the Mediterranean Sea [i.e., A. heterocladum Funk and A. tenuissimum (Schiffner) Hauck]. Previously tentatively included in the list of synonyms of Antithamnion hanovioides (Sonder) De Toni (Athanasiadis 1996, p. 153), Antithamnion verticale was reinstated by Wollaston and Womersley (1998b, p. 105) who pointed out that its ramification is strictly distichous (in contrast to the decussate arrangement of whorl-branches of A. hanovioides ). Therefore Antithamnion verticale belongs to the subgenus Pteroton, while A. hanovioides belongs to the subgenus Antithamnion section Enallassomena Athanasiadis (1996, p. 149). The molecular study of Lee et al. (2001) supports the distinction between these two subgenera. The genus Antithamnion currently accommodates thirty-three species within two subgenera, viz. the subgenus Antithamnion widely distributed in the tropics, warm and cold temperate regions of the world, and the subgenus Pteroton with a clearly Indo-Pacific (and apparently post-Tethyan (6)) origin.
Macrothamnion acanthophorum (Kützing) Womersley in Wollaston and Womersley (1998b, p. 127, figs. 52F, 55 B-D)
COMMENTS: This is the fourth species of Macrothamnion, a genus so far reported only from southern Australia. The relationship of M. acanthophorum to Antithamnion dendroideum G.M. Smith & Hollenberg from California requires further study (see Athanasiadis 1996, p. 158).
Elisiella Womersley in Womersley and Wollaston [1998a, p. 193; type: E. arbuscula (J. Agardh) Womersley]
COMMENTS: Athanasiadis (1996, p. 184) retained only Tetrathamnion Wollaston and Heterothamnion J. Agardh in the Heterothamnieae, but Womersley and Wollaston (1998a, p. 156) restored to this tribe several genera that were cladistically placed in the Dohrnielleae, Perithamnieae and Ceramieae (see Athanasiadis 1996, p. 19). No major changes were proposed in Heterothamnion by Womersley and Wollaston (1998a, p. 158), who however divided Tetrathamnion Wollaston creating Elisiella for two (of the three Australian) species having tetrasporangia and spermatangia on specialised branchlets. Womersley and Wollaston selected Elisiella arbuscula as the generitype [including in its synonymy Tetrathamnion densum (Wollaston) Athanasiadis] and also recognized Elisiella dispar (Harvey) Womersley as a congeneric species. Apart from its generitype, Tetrathamnion lineatum Wollaston, Tetrathamnion includes T. myurum (Suhr) Athanasiadis from Peru that also possesses spermatangia on specialised branchlets, which suggests a position in Elisiella. Tetrathamnion myurum remains known only from its type material.
Perithamnion J. Agardh
Perithamnion muelleri (Harvey) Womersley in Womersley and Wollaston (1998b, p. 203, fig. 95)
COMMENTS: The transfer of Crouania muelleri Harvey to Perithamnion expands the circumscription of this genus and the tribe Perithamnieae, since P. muelleri lacks subdichotomously ramified main axes (that are present in the generitype P. ceramioides J. Agardh) and also possesses terminal carposporophytes which suppress growth of the fertile axis (not previously known in Perithamnion or in the second genus of the tribe Perithamnieae, Scageliopsis Wollaston). Perithamnion muelleri and P. ceramioides appear as sister-taxa within the Ceramioideae if the characters of P. muelleri (Womersley and Wollaston 1998) are entered into the data matrix of Athanasiadis (1996, p. 15).
COMMENTS: The transfer of Amoenothamnion Wollaston and Leptoklonion Athanasiadis to the tribe Ceramieae was not accepted by Womersley and Wollaston (1998a, pp. 184, 190) and is poorly supported cladistically (the decay index being only a single step; Athanasiadis 1996, Cladogram 2C). However, it remains a working hypothesis, based on the most parsimonious distribution of the characters studied so far. The presence of at least one species of Ceramium, C. shepherdii Womersley (1998, p. 384), with non-condensed whorl-branches is a clear indication of the evolutionary transition between the open whorl-branch organization and the transformation of the whorl-branches to a cortex within this tribe. Womersley (in Womersley and Wollaston 1998a, p. 190) reduced the generitype of Leptoklonion, L. elongatum (Wollaston) Athanasiadis, to a synonym of L. fastigiatum (Harvey) Womersley but a formal study of the latter remains to be made.
CROUANIEAE Schmitz & Hauptfleisch
Crouanophycus Athanasiadis [1998, p. 517; type: C. latiaxis (Abbott) Athanasiadis].
COMMENTS: The crouanioid genus Crouaniella Athanasiadis (1996), erected for two tropical species of Antithamnionella, has been renamed Crouanophycus Athanasiadis (1998) because the generic name is a later homonym of Crouaniella (P.A. Saccardo) Lambotte.
Phylogenetic analyses based on molecular sequences have so far concentrated on the inter-relationships between families and orders of the red algae, and the hitherto investigations at the level of tribe, genus or species are few and the results modest. Phylogenetic relationships have been based on comparisons of rbcL, SSU rRNA and LSU rDNA nucleotide sequences, but collectively the results are largely contradictory for members of the Ceramiales (see Saunders et al. 1996, Jong et al. 1998, Choi et al. 2000, Harper and Saunders 2001). Moreover, what uniformly characterizes these studies is the limited number of taxa compared so far, and in particular the lack of primitive ceramiaceous members or several representatives of the subfamily Callithamnioideae.
Five ceramiaceous taxa are included in each of the studies of Saunders et al. (1996, fig. 2) and Jong et al. (1998, fig. 1), while eight taxa are compared in the paper of Choi et al. (2000, fig. 2). The results of these studies are here depicted in Fig. 15. The phylogenetic relationships advertised by Saunders et al. (1996) are congruent with those based on morphological data (Athanasiadis 1996). Yet, these early results are partly incongruent with those of Jong et al. (1998), although the subdivision of the Ceramiales into two groups (Callithamnioideae and Ceramioideae) and the inclusion of the advanced families Rhodomelaceae, Dasyaceae and Delesseriaceae in the Ceramioideae are supported. In the molecular study of Choi et al. (2000), no members of the Callithamnioideae are examined, while the genus Inkyuleea appears as remotely related to the Ceramiales. More recently, Harper and Saunders (2001, figs. 1, 2) have also studied relationships based on LSU rDNA nucleotide sequences, but only three ceramiaceous species of the genera Pterothamnion, Centroceras, and Spyridia are compared (Fig. 15).
Considering the large amount of literature that describes the biology of Ceramiales, and despite the uncertainties surrounding the form variation of many characters and the inconsistent application of terms in describing homologous (or not) structures, some generalizations about the nature of the ceramialean ancestor can be made.
All phylogenetic studies suggest or point to the Ceramiaceae as the most primitive family in the order and in particular to members exhibiting: 1) a bilaterally ramified thallus with filaments of unlimited (or latent) growth dividing by transverse apical divisions, 2) carpogonial branches on axial (or periaxial) cells of filaments of unlimited growth, and 3) cruciately-decussately divided tetrasporangia. Genera that possess these features are found in the tribes Warrenieae and Delesseriopseae.
I have here selected two typical members (i.e., Warrenia and Balliella) of these two tribes and four members of the subfamilies Ceramioideae (i.e., Pterothamnion and Inkyuleea) and Callithamnioideae (i.e., Dasyptilon G. Feldmann and Plumariopsis De Toni) that also possess most of these primitive characters (see Moe and Silva 1983; Athanasiadis 1996; Choi et al. 2000), to investigate possible patterns of early phylogenetic relationships in the order. Apart from the characters mentioned above, I have also included those considered as synapomorphies for the Ceramioideae. These characters have been treated either in a binary way (which assumes that each one has evolved independently; Table I) or as mutually exclusive states (Table II). The analysis of these two data matrices with the PAUP version 3.1 (Swofford and Begle 1993) gave identical results. Assigning Warrenia as the outgroup, the following relationships were recognized: (((Pterothamnion-Inkyuleea)(Dasyptilon-Plumariopsis))Balliella)Warrenia (Fig. 16). In terms of classification, these relationships are reflected by recognizing the suborders Warreniineae and Ceramiineae, the families Ceramiaceae and Delesseriopseae, and the subfamilies Ceramioideae and Callithamnioideae.
The six characters that the two phylogenetic analyses assigned as synapomorphies for the Ceramiales have persisted in Warrenia indicating that the ceramialean ancestor exhibited a filamentous bilaterally ramified thallus with laterals of unlimited growth and transverse apical divisions. Carpogonial branches occurred on intercalary cells of filaments of unlimited growth and tetrasporophytes possessed cruciately-decussately divided tetrasporangia. Moreover, the ceramialean ancestor possessed gland cells. Some of these features are also evident in the sporophyte phases of the Bonnemaisoniales and in certain members of the Gigartinales (i.e., Nemastomataceae and Schizymeniaceae), orders that are closely related to the Ceramiales (see Choi et al. 2000, fig. 2). Therefore it is not unreasonable to assume a filamentous ancestor for the entire Inner Cap Layer-free lineage (i.e., membrane-only lineage sensu Choi et al. 2000, fig. 2), which assumption also fits the probable ancestor of the second major rhodophyte lineage (i.e., the Outer Cap Layer lineage that includes the Acrochaetiales, Nemaliales, Batrachospermales).
Thanks are due to James Norris for arranging a loan of isotype material of Liagorothamnion (US #198087). Gerry Kraft drew my attention to the paper of Millar and Kraft (1984), commenting on a previous manuscript concerning the diagnoses of Balliales, Balliaceae, and Ballia.
1. The original orthography Ptilocladiopsideae (Schmitz and Hauptfleisch 1897, p. 485, 503) is incorrect because the genitive case of the feminine word opsis is opse-os.
2. Kützing (1847, p. 36) did not explain the etymology but most likely Rhodo-callis is a compound word; callis probably derives from the Greek neuter kallos, that as a second compound forms epithets such as peri-kalles. The alternative etymology could be the Latin masculine word callis (footpath). In both instances the genitive case of Rhodocallis is Rhodocall-ous (Greek) or Rhodocallis (Latin), and hence the orthography of Rhodocalleae.
3. In Acrosymphyton taylorii Abbott new axes may develop from outer whorl-branch cells (Millar and Kraft 1984, fig. 5), which is analogous to the exogenous development of new axes that occurs in members of the Rhodomelaceae, Dasyaceae, and Delesseriaceae.
4. The adaxial position results if the gland cell (or young carpogonial branch) is attached on the upper part of an axial cell which is dividing laterally to produce a periaxial cell (the gland cell or carpogonial branch follows the part of cell wall that is displaced adaxially forming the new periaxial cell).
5. It could be added that records of two carpogonial branches on single periaxial cells have been reported in Griffithsia (Baldock 1976, fig. 8; Kajimura 1989a) and in Scagelia (Athanasiadis and Rueness 1992, fig. 3).
6. Members of Pteroton are restricted to Japan, South Australia, South Africa, and Pacific Central America (Antithamnion pectinatum (Montagne) Athanasiadis & Tittley and A. amphigeneum Millar are recent introductions to the Mediterranean and the North Atlantic; see Athanasiadis 1996, p. 141). This distribution suggests an origin subsequent to the closure of the Tethyan passages to the Mediterranean or to the Caribbean from the Pacific.