Fueled by economic interest in phycocolloids, the study of Gracilaria, an important source of agar, has spread rapidly throughout the world, resulting in numerous proposals for taxonomic and nomenclatural change. Several lines of investigation have been followed, including the delineation of taxa in a particular locality or region, the search for new specific characters, the assessment of morphological variability in the genus and the concomitant reduction of the number of recognized taxa, the search for characters that may be used to break up this large and highly variable genus into subgeneric taxa or segregate genera, and the determination of the taxonomic relationship between Gracilaria and other red algae.
Dawson (1949a: 40) segregated Gracilariopsis from Gracilaria on the basis of two cystocarpic characters, namely, a small-celled, broad-based gonimoblast and the absence of nutritive filaments (= nutrient tubular cells of Sjöstedt, 1926; traversing filaments of Kraft, 1977c; absorbing filaments of B. Xia & Abbott, 1985, 1987; gonimonemata of V. Krishnamurthy & Rajendran, 1986b, 1987a; tubular nutritive cells of Fredericq & Hommersand, 1989a, 1989b, 1990b) connecting the gonimoblast with the pericarp. Papenfuss (1967: 98--101) examined British material determined by him as Gracilaria verrucosa (Hudson) Papenfuss and found that nutritive filaments were present in some plants, absent in others. Moreover, he found no fundamental difference in the size of the gonimoblast cells between Gracilaria verrucosa and Gracilariopsis sjoestedtii (Kylin) Dawson (treated as a taxonomic synonym of Gracilariopsis lemaneiformis (Bory de Saint-Vincent) Dawson, Acleto, & Foldvik (q.v.)) and therefore merged the two genera. On the basis of a painstaking morphological study of several populations of each of these species, Fredericq & Hommersand (1989b) resurrected Gracilariopsis, which they distinguished from Gracilaria by the lack of tubular nutritive cells, the transformation of inner pericarp cells into a nutritive tissue, the lack of secondary expansion of the primary fusion cell by incorporation of additional gametophytic cells in the floor of the cystocarp, and the superficial position of spermatangial mother cells, which are initiated in pairs or groups of three by outermost cortical cells rather than being produced in filaments from intercalary cortical cells lining a depression or pit (conceptacle). In having found that representatives of these two genera can grow in mixed stands in Atlantic Europe, Fredericq & Hommersand (op.cit.: 240) provided an explanation for Papenfuss's inconsistent observations on tubular nutritive cells. Recent analyses of nucleotide sequences of the nuclear small-subunit rRNA gene (Bird et al., 1994) and of the RUBISCO spacer region (Goff, Moon, & Coleman, 1994) support the recognition of both genera.
Building on the work of Dawson (1949a), who demonstrated that the development and form of spermatangial sori was of the utmost significance in delimiting certain species of Gracilaria, Yamamoto (1975) studied male plants of 15 Japanese species and was able to distribute them among three subgenera: Gracilariella, with continuously superficial spermatangia; Textoriella, with shallowly depressed sori; and Gracilaria, with deeply sunken sori or conceptacles. Meanwhile, in a publication unknown to Yamamoto, Chang & B. Xia (1963) reported their finding of compound (multicavitied) spermatangial conceptacles in several species of Gracilaria, which they segregated as the genus Polycavernosa. In these species the clusters of spermatangia are derived from intercalary inner cortical or outer medullary cells. A second character of Polycavernosa given by Chang & Xia is the production by the gonimoblast of darkly staining, irregularly shaped, basal filaments that penetrate the underlying gametophytic cells. Filaments extending upward to the pericarp are absent, as in Gracilariopsis. Fredericq & J. Norris (1985) showed that while in Gracilaria the gonimoblast is initiated directly by the fusion cell, in Polycavernosa it is initiated by a tissue derived from the fusion cell. B. Xia & Abbott (1987) added the observation that in Polycavernosa the spermatangial conceptacles are clustered rather than randomly distributed over the surface of the thallus. They also showed that Hydropuntia urvillei Montagne (1842a: 7), the type of its genus, is referable to Polycavernosa, but neglected to adopt the earlier generic name. This nomenclatural oversight was pointed out by Wynne (1989a). After finding Gracilaria-type spermatangial conceptacles in the youngest branch portions and Hydropuntia-type spermatangial conceptacles in older parts of the same frond of an undescribed gracilarioid and in other western Pacific species previously assigned to Hydropuntia, Abbott, Zhang, & B. Xia (1991) reduced Hydropuntia to the synonymy of Gracilaria. The dust has not settled, however, and proponents of Hydropuntia may be expected to respond.
Determination of the correct names for the two stringy gracilarioids from the northeastern Atlantic has been an arduous task. For more than a century the two species were confused under the name Gracilaria confervoides (Linnaeus) Greville (1830: liv, 123), based on Fucus confervoides Linnaeus (1763: 1629; type locality: England). The intended basionym, however, is not priorable for two reasons: first, it is a later homonym of F. confervoides Hudson (1762: 474; type locality: York, England) (Rhodomela confervoides (Hudson) P. Silva, 1952a: 269); and second, it is superfluous since Linnaeus cited Fucus teres rubens ramosissimus of Ray (1724: 51, no. 52), which is the type of Fucus purpureus Hudson (1762: 471) (Cystoclonium purpureum (Hudson) Batters, 1902: 68). Combinations that attempted to use Fucus confervoides Linnaeus as a basionym, including Ceramium confervoides Wiggers (1780: 91) and Sphaerococcus confervoides Stackhouse (1797 [1795--1801]: xxiv), are similarly superfluous. Gracilaria confervoides Greville (1830: liv, 123) is superfluous even though Greville excluded the Cystoclonium element from the protologue since he cited as a synonym Fucus verrucosus Hudson (1762: 470--471; type locality: England). The latter name had been placed in the synonymy of F. confervoides Linnaeus since the time of Withering (1796: 114--115). Without regard to its being a later homonym or a superfluous name, Fucus confervoides Linnaeus was abandoned by Papenfuss (1950b: 195), who adopted F. verrucosus as the epithet-bringing name for this species. Accordingly, he made the combination Gracilaria verrucosa (Hudson) Papenfuss, a name that was widely applied by subsequent authors to stringy gracilarioids in various parts of the world.
Following the revelation by Fredericq & Hommersand that two species, in fact two genera, occurred in Britain, it was essential to determine, if possible, the generic placement of the alga going under the name Gracilaria confervoides, which had been designated by Schmitz (1889: 443) as type of the generic name. Disregarding the fact that Fucus confervoides Linnaeus, as a superfluous name for F. purpureus Hudson, must be typified with the type of the latter name in accordance with Art. 7.5, Steentoft, L. Irvine, & Bird (1991) sought the type of the Linnaean name in his herbarium (LINN). They found two sheets labeled F. confervoides that bear gracilarioid algae, and one of the specimens was designated lectotype. This specimen, which cannot serve as lectotype, proved to be representative of Gracilariopsis. Accordingly, those authors proposed that the generic name Gracilaria be conserved with a conserved type, G. compressa (C. Agardh) Greville (treated as a taxonomic synonym of G. bursa-pastoris (S. Gmelin) P. Silva (q.v.)). The proposal was approved by the Committee for Algae and accepted by the Fifteenth International Botanical Congress (see P. Silva, 1994: 263).
The next step in the search for the correct names for the two North Atlantic stringy gracilarioids was to locate authentic material of Fucus verrucosus. In the absence of Hudson specimens, L. Irvine & Steentoft (1995) designated as lectotype a specimen in the Dillenius Herbarium (in OXF) that is the voucher for Fucus marinus purpurascens parvus, caule & ramulis seu foliolis teretibus of Ray (1724: 50, no. 50), which was cited by Hudson. An examination of this lectotype revealed that it is representative of Gracilariopsis, thus confirming the traditional synonymy and supporting the decision to designate Gracilaria compressa the conserved type of the generic name.
To find the correct name for the North Atlantic stringy Gracilaria, Steentoft, L. Irvine, & Farnham (1995) investigated species of Fucus that had been proposed in the late 18th and early 19th centuries and subsequently referred to Gracilaria. They found that Fucus gracilis Stackhouse (1801 [1795--1801]: xxx, 100--101, pl. XVI; syntype localities: England [``Padstow, Cornwall, and elsewhere'']) was representative of Gracilaria, as determined by a lectotype specimen (in BM ex K), and therefore made the combination G. gracilis (Stackhouse) Steentoft, L. Irvine, & Farnham.
Believing that the adoption of Fucus verrucosus as the epithet-bringing name for the North Atlantic Gracilariopsis would cause confusion in view of its long-time association with Gracilaria, L. Irvine & Steentoft (1995) proposed it for rejection in accordance with Art. 56.1, which permits rejection of ``any name that would cause a disadvantageous nomenclatural change''. In anticipation of the approval of this proposal by the Committee for Algae and its acceptance by the next International Botanical Congress, Steentoft, L. Irvine, & Farnham looked for an epithet-bringing name applicable to Gracilariopsis and found that Fucus longissimus S. Gmelin (1768: 134, pl. XIII; syntype localities: Belgium; Sheerness, Kent, England) was available. They designated as lectotype a specimen in the Dillenius Herbarium (in OXF) that is the voucher for Fucus teres rubens minus ramosus in longum protensus of Ray (1724: 51, no. 53, from Sheerness), which was cited by Gmelin. Having confirmed its generic placement, they made the combination Gracilariopsis longissima (S. Gmelin) Steentoft, L. Irvine, & Farnham.
Ceratodictyon and Gelidiopsis, which Kylin (1956: 256) included in the Gracilariaceae, were moved to the Rhodymeniales by Price & Kraft (1991). (See Note to Ceratodictyon.)