GRAll'OLITES OF NEW YOKK, PART 1 



637 



the Skiddaw slates Dichograpsus araiiea and D. sedgewickii. 

 We agree Avith Elles and Wood in considering the first of these names as a 

 synonym of D . octobrachiatiis, the differences being explicable by dif- 

 ferent modes of preservation. 



Hall noticed already that the number of branches is not absolutely 

 unvariable, but that suppression of a dichotomy may result in a smaller 

 number of branches. Elles mentions a septad type, where one branch of 

 the second order fails to divide dichotomously, and a hexad type, where 

 tAvo fail to divide. The only specimen of the hexad type noticed in our 

 material was foimd in the uppermost Deep kill horizon (with 

 Diplograptus dentatus), where the Dichograpti are rapidly 

 disappearing. 



Malaise observed a hexabrachiate foi-m in Belgium and proposed for 

 it the specific tei-m D. hexabrachiatus [loc. cit. p.44]. This hexad 

 variation occurs there also associated with octobrachiate forms, and in a high 

 zone Avith Didymograptus murchisoni and Diplograptus 

 p r i s t i n i f o r m i s (= d e n t a t u s ) , as in NeAV York. 



These occurrences shoAv that there existed a tendency to reduction, 

 Avhich has finally led to the production of species of Tetragraptus 

 [Introduction, p.567]. 



On the other hand we have observed, in graptolite bed 2, a form Avith nine 

 branches in Avhich one of the branches of the third order has divided again 

 dichotomously, thus producing two branches of the fourth order [pl.8, fig.7]. 

 While this form, by the strict application of the definitions of Logano- 

 graptus and Dichograptus, would have to be referred to the former stage, it 

 is, in the character of the thecae and branches, and its general habit, 

 a D. octobrachiatus. It is evident that hand in hand AAath the reduc- 

 tion of the number of branches other changes take place Avhich make a strict 

 application of this one criterion to the separation of the species impossible. 



