286 PAL/EONTOLOGY 



snapped by too great strain (as in Fig. 88, c) shows that 

 the polypary was a fairly stiff structure. 



All these forms, covering twenty-seven British species 

 alone, are habitually included in the one genus Didymo- 

 graptus. But the modern conception of a genus is that 

 of a number of species connected together and separated 

 from species of other genera by an immediate common 

 descent. That Didymograptus, in its accepted extent, is a 

 true genus has been rendered very doubtful by certain 

 observations made in 1895 by Marr and Nicholson. 



The genus Didymograptus is found in the Lower and 

 Middle Ordovician. In the Lower Ordovician only are 

 found graptolites which start growth like Didymograptus, 

 but branch again, forming four or eight stipes, and in the 

 Tremadoc beds (transitional between Cambrian and 

 Ordovocian) are still more complexly branched forms. 

 Marr and Nicholson pointed out that similarities of stipe- 

 disposition and thecal form can be traced through 

 the morphological genera Bryograptiis (or Clonograptus or 

 Loganograptus), Tetvagraptus and Didymograptus in these 

 successive periods, and they suggested that these indicate 

 the true lines of descent, while the mimber of stipes 

 underwent reduction from many to four, and at last to 

 two, in at least five independent lines of descent (Fig. 85). 



The reader may have been struck with the fact, which 

 has puzzled many geologists, that the earliest graptolites 

 (Tremadoc) are more complex than their descendants in 

 the Ordovician. Further, that in individual growth a 

 Tetragraptus must pass through a Didymograptus stage, 

 while a Clonograptus must pass through first a Didymo- 



