488 Dr. Gerhard Holm — On the Structure of GraptoUtes. 



circles, and the four rows of thecse are therefore of equal height. It 

 follows from this that the genus PhyUograptus must be considered 

 to have arisen from four branches, like those of a Tetragraptus, with 

 their backs to each other, and that, since the sicula in Tetragraptus 

 can branch dichotomously, the first thecse of the branches ought to 

 be of the same height. 



We have seen above that the sicula does not branch either in 

 Tetragraptus or in PhyUograptus, but that without this it gives 

 origin to thecae which, both in Didymograptus and BipJograptus, 

 are developed from each other, and therefore occupy a different 

 height in the polypary. Tornquist's supposition of a circular 

 arrangement of the thecse is therefore incorrect. The figures show 

 that the arrangement changes conspicuously (PI. XIV). This is 

 due to the fact that the thecse in different rows are separated by 

 longitudinal septa, and are developed completely independently of 

 each other, and therefore, since the height of the thecas can change, 

 any very regular arrangement is impossible. 



Tullberg has done greater service than any other in emphasizing 

 the near generic relationship between Tetragraptus and PhyUo- 

 graptus. In his Classification of Graptolites, mainly based on that 

 of Lapworth, which he justly takes to be the most natural, up-to- 

 date, analytical table of the more important Graptolite genera, he 

 has partially introduced some emendations. One of these is the 

 suppression of the section Tetraprionid^, originally proposed by 

 Hopkinson for the genus PhyUograptus and accepted by Nicholson 

 and Lapworth. Tullberg rightly places PhyUograptus with the 

 Dichograptida9 as a close ally of Tetragraptus. Tornquist (Siljan. 

 Grapt.) again returns to the old way, and has placed PhyUograptus 

 in a separate family, Phyllograptidse, allied to Dichograptidse. 

 Tetragraptus and PhyUograptus are therefore separated from each 

 other by Didymograptus, which Tornquist places in the Dicho- 

 graptidae. 



Phyllograptus angustifolius, Hall. PI. XIV, Figs. 1-12. 



B}' comparing the figures of T. Bigshyi, Hall, and this species, 

 the close agreement between them is quite evident, both in internal 

 and external structure ; with the necessary reservations due to the 

 fact that the coenosarcal th'^eads of the four branches in the Tetra- 

 graptus polypary are in PhyUograptus disposed near each other, and 

 instead of four independent periderm walls form a single, cruciform, 

 four-winged, longitudinal septum. The oblique position of the sicula 

 and of the proximal end of the left theca in relation to the median 

 plane of the polypary stands out very conspicuously, as it lies com- 

 pletely revealed on the sicula side. Only the apertural edge of the 

 sicula, which is turned, not only to the right side but also to the 

 anti-sicula side, is visible on the latter. The connection between 

 the sicula and the left theca takes place near the proximal end of 

 the sicula: PI. XIV, Fig. 9. The foramen between the left theca 

 and the connecting canal has the same position as in Tetragraptus. 

 From the right end of the connecting canal the right theca arises ; 



