Graptolite Taxonomy and Classification MU EN-ZHI Mu En-zhi: Graptolite Taxonomy and Classification. Bull. geol. Soc. Denmark, vol. 35, pp. 203--207, Co­ penhagen, July 1st, 1987. https://doi.org/10.37570/bgsd-1986-35-21 Graptolithina comprises chiefly six orders. Among them Graptoloidea and a part of Dendroidea known as Graptodendroids are planktonic in mode of life. Graptoloidea consists of three suborders namely Axono­ lipa, Axonocrypta and Axonophora. The families Dendrograptidae-Anisograptidae-Tetragraptidae and Didymograptidae-Isograptidae-Cardiograptidae-Diplograptidae-Monograptidae represent anagenetic grades. Some important evolutionary trends took place once again, representing cladogenetic divergen­ ces. All other families or subfamilies are offshoots of various grades. The suborder Axonocrypta is dis­ cussed in detail. Mu En-zhi, Nanjing Institute of Geology and Palaeontology, Academia Sinica, Chimingssu, Nanjing, China, August20th, 1986. General Consideration tinae, tri-radiate Anisograptinae and biradiate Adelograptinae) is derived from the floating Dic­ Graptolithina, a class of Hemichordata, com­ tyonema due to the loss of dissepiments (Mu, prises chiefly six orders known as Dendroidea, 1974), while the reclined Psigraptidae Lin (1981) Graptoloidea, Tuboidea, Camaroidea, Stolonoi­ with isolated autothecae is an offshoot. Recently dea and Crustoidea (Kozlowski, 1949, 1966; Bul­ Zhao & Zhang (1985) proposed a new family man, 1970). The thecae in Dendroidea and Grap­ Muenzhigraptidae with biform autothecae rep­ toloidea are regularly arranged in stipes, al­ resenting the direct ancestor of Psigraptidae. though Dendroidea has three kinds of thecae Graptoloidea first appeared in the late Xin­ (autotheca, bitheca and stolotheca), while Grap­ changian (X3) due to the loss of bithecae from toloidea has only one. In the graptoloid thecae, Adelograptinae and flourished in early Ning­ the proximal portion (protheca) and the distal kuoan (N1), marking another new stage in grap­ portion (metatheca) are homologous with the tolite history. Since then, Graptoloidea became stolotheca and autotheca respectively. the master of the quiet sea area in Ordovician Among these orders, Dendroidea is the ear­ and Silurian and even in Early Devonian. liest in appearance (M. Cam.?, Late Cam.) and the latest in disappearance (Late Carb.); Grapto­ loidea ranged from Early Ordovician to Early Subdivision of the Order Graptoloidea Devonian; the other orders are rare in materials and known to occur from Ordovician to Silurian. The subdivision of Graptoloidea has been treated All the Tuboidea, Camaroidea, Stolonoidea, by many graptolite workers. Two suborders, Crustoidea and most of Dendroidea are ben­ Axonolipa and Axonophora, proposed by Frech thonic in the mode of life, whereas Graptoloidea (1897) and emended by Ruedemann (1904, 1908) and a part of Dendroidea are planktonic. The have been used for a long time (Mu, 1950; Obut, floating dendroids with free sicula appeared in 1957; Mu, 1974; Yu & Fang, 1979). In fact, Axo­ the latest Cambrian (Fengshanian) and flour­ nolipa consists of Didymograpta and Dicello­ ished in Early Xinchangian (Tremadocian), in­ grapta of Lapworth (1880) and Axonophora con­ dicating a new stage in graptolite history. The sists of Diplograpta and Monograpta of Lap­ graptodendroid family Anisograptidae compris­ worth (1880). In the sixties, J,aanusson (1960) ing three subfamilies ( quadriradiate Staurograp- divided Graptoloidea into foursuborders namely Note: While the present volume was in press the editor received information on the death of Prof. Mu En-zhi. 204 Mu En-zhi: Graptolite Taxonomy and Classification Didymograptina, Glossograptina, Diplograptina from Isograptidae, a descendant of Didymograp­ and Corynoidina. Bulmann (1963) grouped Co- tidae (and Tetragraptidae?). rynoidina into the suborder Didymograptina and The suborder Axonophora comprises those divided the Diplograptina into two suborders, with virgula embedded in the median septum in Diplograptina and Monograptina. Mu & Zhan biserial form (Diplograpta) or in the dorsal wall (1966) established the suborder Axonocrypta in of the periderm in uniserial form (Monograpta). addition to Axonolipa and Axonophora based on In Diplograpta, the initial bud originates from the structure of the rhabdosome. Mu (1974) ten­ the porus on the distal part of metasicula and tatively used Didymograpta and Dicellograpta as even near the sicular aperture. The proximal the­ subdivisions of Axonolipa and Diplograpta and cae in early forms are curved upwards with three Monograpta as subdivisions of Axonophora re­ or more crossing canals; in some later ones, the spectively. crossing canals increase throughout the rhabdo­ Axonolipa comprises the forms with nema free some without dicalycal theca (aseptate). The vir­ and the stipes pendent to reclined just like float­ gella is well developed, beginning at the proximal ing dendroids. Those in Didymograpta are devel­ part of the metasicula even near the prosicula. oped in a primitive type. The initial bud orig­ The most primitive genus Glyptograptus of Di- inates from the poms on the ventral or dorsal plograptidae is most probably derived from Car­ side of the prosicula or the proximal part of meta- diograptidae of Axonocrypta due to the fusion of sicula. The proximal thecae grow downwards to the dorsal walls of two stipes, forming a median nearly horizontally with one or two crossing ca­ septum. Exigraptus seems to be a transitional nals, that is to say, the first or the second theca is form between Cardiograptidae and Diplograpti- dicalycal. The virgella is usually absent, or pres­ dae. In China, the earliest species of Glyptograp­ ent in advanced forms, it begins at the proximal tus, G. sinodentatus, direct ancestor of G. aus- or middle part of the metasicula. The most primi­ trodentatus, is more closely related to Exigraptus tive forms such as Didymograptus of Didymo- (Mu et al., 1979; Chen, 1982). It is questionable graptidae, Tetragraptus of Tetragraptidae and that G. austrodentatus is derived from Maeandro- bitheca-lacking Clonograptus of Dichograptidae graptus (Jenkins, 1980) and G. dentatus is related are the direct descendants of Adelograptinae of to Phyllograptus elongatus (Cooper & Fortey, Anisograptidae. In Dicellograpta, the initial bud 1983). arises from the porus on the middle or distal part In Monograpta the initial bud originates from of metasicula. The proximal thecae grow hori­ the sinus on the distal part of the metasicula near zontally to upwards with two or three crossing ca­ the sicular aperture. The proximal thecae grow nals, that is to say, the second or third theca is di­ directly upwards without crossing canal, i.e., no calycal. The virgella is well developed, beginning dicalycal theca. The earliest representatives of usually at the proximal part of the metasicula. Monograpta are Atavograptus (Rickards, 1974), Dicellograptus of Dicranograptidae is the earliest Monoclimacisl (Bjerreskov, 1975) and Pristio- derived from Didymograptidae. graptus (Li, in press) of Monograptidae, appea­ In Axonocrypta the nema (virgula) is enclosed ring abruptly in the beginning of Silurian (per- between the four to two scandent stipes. The sic- sculptus Zone) and representing another new ula is relatively long. The initial bud originates stage of graptolite history. It is believed that from the porus on the prosicula or the proximal Monograptidae is derived from Diplograptidae part of metasicula. It arises from the ventral, lat­ with discontinuity. Peiragraptidae and Dimor- eral or dorsal side of the sicula. The proximal phograptidae are offshoots. thecae grow downwards or curved distally, with one or two crossing canals. The virgella is usually present in advanced form. Those with four or On the Suborder Axonocrypta three stipes back to back (Phyllograptidae) are derived from Tetragraptidae, and those with two Axonocrypta linking with Axonolipa and Axo­ scandent stipes back to back (Cardiograptidae) nophora bears the fundamental characters be­ or side to side (Glossograptidae) are derived tween them in the structure of rhabdosome and Bulletin of the Geological Society of Denmark, vol. 35 1987 205 in the mode of development. There are three are different from Pseudotrigonograptus ensifor- families in Axonocrypta, known as Phyllograpti- mis as described by Rickards (1973, p. 600, figs. dae, Cardiograptidae and Glossograptidae (= 1-3). They are not conspecific. Cooper has Cryptograptidae). rightly distinguished these two species based on Phyllograptidae is directly derived from Tetra- the materials of New Zealand (Cooper, 1979, p. graptidae of Axonolipa with Tetragraptus phy- 91, figs. 83b-c). llograptoides and Phyllograptus cor as the tran­ The pore of the common canal in Pseudotrigo­ sitional forms (Cooper & Fortey, 1982; Cooper & nograptus uniformis corresponding to the "fore- Lindholm 1985). The Family Phyllograptidae was nic foramina" in Phyllograptus typus illustrated revised by Hsu (1934) and Mu & Zhan (1966) to by Cooper & Fortey (1982, figs. 71k, 74) is an ad­ be composed of Phyllograptus and Trigonograp- vanced feature. The Spitzberg materials de­ tus (= Pseudotrigonograptus) based on the struc­ scribed by Fortey (1971) and Cooper & Fortey ture of the rhabdosome. Recently, Cooper & (1982) as Pseudotrigonograptus ensiformis (4-sti- Fortey (1982) revised this family to comprise ped) and P. minor Mu et Lee (3-stiped) are pos­ Phyllograptus (s.s.) and a new genus Xiphograp- sible new forms representing a genus intermedi­