V ■ 

 Dr. Gerhard Holm — On the Structure of GraptoUies. 483 



as part of a third branch (but which also might' cause it to be 

 supposed that still another branch existed, concealed in the rock)." 

 This appendage of the sicula is, however, quite certainly not 

 a branch, but is formed by the unusually long, thread-like extension 

 of the initial part of the sicula, as shown by an isolated specimen of 

 my own from the grey glauconitic Vaginatenkalk near Halludden, 

 in Oland. The length of this in the specimens in question is over 

 7 mm., while the length of the branches is only 3 mm. It tapers 

 gradually for the whole of its length towards the apex. 



In connection with the genus Didymograptus may be mentioned a 

 singular form of graptolite (PI. XIII, Figs. 4-6), the figure of which 

 is drawn from a damaged specimen, but subsequently two complete 

 ones were obtained. The structure of the polypary is very simple, 

 as besides the sicula it is formed only of the left theca and the 

 connecting canal. A comparison between the figures of D. minntus 

 (Mut.) and the present form, placed side by side and viewed in 

 the same position, shows that in the latter the development of the 

 parts mentioned is completely similar to that in Didymograptus, 

 with the exception that from the bud of the sicula is developed 

 besides the connecting canal only a left theca, but no left stipe. 

 The connecting canal ends also near the right side of the sicula, 

 without even giving origin to any right theca in case the connecting 

 canal itself does not function as such. In the figured specimen the 

 free apertural margin of the connecting canal projects slightly at its 

 upper corner from the right side of the sicula. This, however, is 

 not the case with the other two specimens. The sicula is provided 

 with a true apertural spine, which originates in the periderm some 

 distance behind the aperture, in the same way as the spines in 

 JDiplograptus, sp. (p. 438), and it appears to be of the same nature, 

 but curiously enough, contrary to what is usually the case, it lies on 

 the right side of the theca. The aperture of the left theca is turned 

 straight out, and the lines of growth on the anti-sicula side show the 

 same course as in the common canal — or the proximal pai't of the 

 second theca on the left branch — in D. minntus, Mut. (Figs. 3 

 and 6). 



That, however, the left theca is certainly terminal, and that con- 

 sequently the left branch cannot develop, is clearly shown by the 

 fact that the course of the growth-lines is everywhere quite regular, 

 and parallel with the apertural margin ; therefore no indication 

 whatever of bud formation can be perceived on the proximal part 

 of the left theca, either in this way or in surface relief. To this 

 may be added what we have seen concerning the growth of the left 

 branch in D. minutas, Mut. : the formation of the common canal and 

 left theca takes place simultaneously, so that an interthecal wall 

 cannot originate in a completed part of the stipe. 



It has been observed in many cases that the same thing occurs at 

 the distal end of the stipe : PI. XIV, Figs. 6, 13, 14. It follows, 

 therefore, that in the present form no left branch can be supposed 

 to develop from the aperture of the left theca. The development of 

 a right theca, or a right branch, from the connecting canal does 



