Chaetognatha
Systematics

All chaetognath genera remained for a long time in a single group until Tokioka (1965) proposed a new classification of the phylum. He defined two classes: Archisagittoidea for the fossil species Amiskwia sagittiformis Walcott, 1911, and Sagittoidea Claus and Grobben, 1905 for the extant species. The former is now considered probably a Nemertina (but see chapter on nemerteans in this volume), as previously indicated by Owre and Bayer (1962). In the Sagittoidea, Tokioka distinguished two orders:

—Order Phragmophora
Provided with transverse muscles and various kinds of glandular structures on the body surface, comprising the littoral and benthic family Spadellidae (genus Spadella), and the pelagic family Eukrohniidae (e.g. genus Eukrohnia and genus Heterokrohnia).

—Order Aphragmophora
Without transverse muscles, with scarce glandular structures on the body surface. It comprises two suborders: Ctenodontina, grouping the families Sagittidae Claus and Grobben, 1905 (genus Sagitta sensu lato), and Pterosagittidae (genus Pterosagitta); and Flabellodontina (genus Krohnitta), the more highly evolved suborder. An interesting proposal was to split the genus Sagitta s. l. into nine smaller genera, since it concerns most of the commonest planktonic species: Sagitta s.s., Zonosagitta, Serratosagitta, Parasagitta, Aidanosagitta, Mesosagitta, Solidosagitta, Caecosagitta and Flaccisagitta. Such a division of Sagitta s.l. had already been envisaged by Colman (1959), but he thought that "this would only create confusion". This view was perhaps shared by Tokioka himself since he never used his new genera again. In fact, this splitting has two main disadvantages. On the one hand, whereas some of these small genera cover a group of closely related species hitherto recognized in Sagitta s.l. (e.g., the genus Serratosagitta, which comprises the species of the serratodentata group, see below), others overlap more than one group, like Flaccisagitta, which includes species of the lyra and hexaptera groups, whose only shared feature is the flaccidity of the body. An additional drawback in this system is that it ignores the fact that species belonging to the same genus must descend from a common ancestor; thus, while members of Serratosagitta are probably closely related phylogenetically, those of Flaccisagitta are not. Furthermore, it is sometimes difficult to identify the juveniles of some Tokioka's genera. However, some workers do still use this classification system, notably Bieri (1991), who rightly divided the genus Flaccisagitta (see above), but also removed some species from Tokioka’s genera reassigning them elsewhere. Clearly, Colman’s fears as far as splitting of the genus Sagitta is concerned are totally justified.

A slight modification of Tokioka's system was proposed later (Casanova, 1985), the chaetognaths being assigned to two subclasses: Syngonata (with ducts between genitals in trunk and tail), and Chorismogonata (without such ducts). The former contains the single order Biphragmophora (transverse muscles in trunk and tail), with the family Heterokrohniidae (genera Heterokrohnia, Archeterokrohnia and Xenokrohnia). The latter includes two orders: Monophragmophora (transverse muscles in trunk only), with the families Spadellidae (genera Spadella, Paraspadella, Hemispadella) and Eukrohniidae; and Aphragmophora as defined above. A full account of this subject was given by Bieri (1991).