Vol. 54.] THE GRAPTOLITE-FAUNA OF THE SKIDDAW SLATES. 531 



capable of arrangement into natural phylogenetic groups, of which 

 those of Nicholson & Marr must be looked upon as the first sug- 

 gestions. 



I agree with these authors (1) that these resemblances are of 

 genetic origin, and therefore (2) of systematic value; and 

 further (3) that in any natural group the forms with relatively 

 fewer branches were developed from the more complex forms, so 

 that it follows from this (4) that the so-called ' genera ' of the 

 usually accepted classification of the Dichograptidse are far more 

 of a chronological than of a zoological significance. But 

 my own work among these forms has led me to the conclusion 

 that the forms in question are most probably the result of develop- 

 ment along certain special lines, and I would suggest that 

 such is their real origin. The group-relationships which seem to 

 follow from this suggestion are indicated below. 



In saying that the apparently simpler types have been derived 

 from the more complex ones, I do not mean to imply thab there 

 have been no other modifications : for example, that every Didymo- 

 f/raptus is necessarily derived from a different species of Tetragrajotus. 

 Some are, no doubt, merely modifications of previously existing 

 Didymograptus-ioTVCLQ. There does, however, appear to be a certain 

 group-relationship. I consider, for instance, that all the ' tuning- 

 fork ' Didymograpti have been derived from what might be termed 

 the \fruticosus-ty^Q^ of Tetragraptus, though not all from T. fruti- 

 cosus (Hall), as Nicholson & Marr seem to consider (Geol. Mag. 

 1895, p. 535). And on the other hand, the Didymograpti in 

 which the stipes form with each other an angle of 180° probably 

 originate from Tetragrapti of the quadribrachiatus-type, though not 

 all from T. quadribrachiatus (Hall) itself. Examples will, perhaps, 

 make this clearer. 



Thus, starting with Bryograptus ramosus var. cumbreTisis, we 

 have a Bryograptus-ioxxo. which resembles a Didymograptus with 

 compound branches coming off from the celluliferous side of its 

 main stipes. Observation shows that the branching in this case is 

 certainly the result of division of the common canal. Suppose now 

 that, instead of having various compound branches, the division of 

 the common canal takes place only once on each side of the sicula, 

 and near the proximal end of the stipes, the resulting form will be 

 identical with Tetragraptus pendens, while the absence of branching 

 altogether gives a form like Didymograptus indentus (Hall). 



Nicholson & Marr have shown similar cases, though I confess 

 that Didymograptus furcillatus, Lapw., seems to me more likely to 

 be the direct descendant of Tetragraptus fruticosus (Hall) than 

 Didymograptus Murchisoni (Beck) ; this is, however, a minor point. 



On the other hand, starting with Clonogyxiptus sp., the reduction 

 of indefinite dichotomy results in Loganograptus sp., and a further 

 limitation in branching yields Dicliograptus sp. Every stage between 

 a typical Dicliograptus octobrachiatus (Hall) and Tetragraptus 

 quadribrachiatus (Hall) can be traced, for there occur in the 

 Skiddaw Slates, in addition to the typical * octobrachiate ' form with 



