Ectocarpus siliculosus (Dillwyn) Lyngbye and E. confervoides (Roth) Le Jolis have been recognized either as separate but closely related species or as morphological phases of the same species. In either treatment, the nomenclature has been consistently misinterpreted.
The former name was firmly established by Lyngbye (1819: 131) at the time that he proposed Ectocarpus as a new genus. The lanceolate-acuminate plurilocular organs considered typical of the species are clearly illustrated in his pl. 43C, based on material from the Danish island Fyn. As basionym, Lyngbye cited Conferva siliculosa Dillwyn (1809 [1802--1809]: 69, suppl. pl. E). Dillwyn's description was based on two English collections supplied by W.J. Hooker, one from Cromer in Norfolk, the other from Hastings in East Sussex. Plurilocular organs are shown in pl. E, which Dillwyn accredited to Hooker. While the quality of this drawing is well below that of Lyngbye's, it is clear that the two authors had the same species in hand. Lyngbye followed Dillwyn in citing as a taxonomic synonym Ceramium confervoides Roth (1797: 151--153, pl. VIII: fig. 3). Roth's description was based on a collection from Eckwarden on Jadebusen, an inlet of the North Sea in the Oldenburg region of Niedersachsen, Germany. The ``capsules'' were said to be oval-acuminate, lateral, and pedunculate, but were not figured.
No one objected to the use of an apparently later synonym prior to Le Jolis (1863: 75), who adopted Roth's name, making the combination Ectocarpus confervoides (Roth) Le Jolis. In a monograph of Scandinavian Ectocarpaceae, Kjellman (1872: 67 ff.) recognized six forms of E. confervoides, including forma siliculosa (Dillwyn) Kjellman. With regard to plurilocular organs, in the nomenclaturally typical form (forma confervoides) they were said to be ovoid-oblong or ovoid-conical, erostrate, and pedunculate. Those of forma siliculosa were said to be subulate or linear-subulate, often rostrate, and sessile or subsessile. Kjellman's concept of forma confervoides must have been based solely on Roth's description, as reproductive organs were not figured by Roth and Kjellman did not indicate having seen an authentic specimen. While Kjellman's concept of forma siliculosa, which prevails at present, is supported by the excellent drawing in Lyngbye, no one to my knowledge has ever attempted to locate the syntype specimens.
In a later work, Kjellman (1890) assigned four of the six forms of E. confervoides to three other species, including E. siliculosus. Various authors, including Kuckuck (1891), De Toni (1895b), and Hamel (1931b), followed Kjellman in recognizing both E. confervoides and E. siliculosus. Dixon & Russell (1964) cast a pall over both names by pointing out that Roth had cited Conferva littoralis Linnaeus (Pilayella littoralis (Linnaeus) Kjellman (q.v.)) in the protologue of Ceramium confervoides, thus rendering the latter name superfluous and illegitimate. Moreover, a fact not mentioned by Dixon & Russell is that Ceramium confervoides Roth is a later homonym of Ceramium confervoides Wiggers (1780: 91), based on Fucus confervoides Linnaeus (see Gracilariaceae, Taxonomic and Nomenclatural Notes (q.v.)). Dixon & Russell reasoned that Conferva siliculosa Dillwyn is superfluous since Ceramium confervoides Roth was cited in synonymy. I pointed out, however, that Dillwyn recognized both Conferva littoralis Linnaeus and C. siliculosa, so that the citation of Ceramium confervoides should be interpreted to mean ``Roth excl. syn.'' (see Russell 1966: 275, footnote).
While the name Ectocarpus siliculosus is thus legitimate, the concept of Ceramium confervoides Roth appears to lack a legitimate name. Because Roth's name is superfluous and hence not priorable, the binomial Ectocarpus confervoides must be attributed directly to Le Jolis as a new name rather than a new combination. Le Jolis, however, cited Conferva siliculosa Dillwyn, so that Ectocarpus confervoides is also superfluous. If the concept of E. confervoides is to constitute an infraspecific taxon of E. siliculosus, as in the treatment by Russell (1966), it must be described as new.