GRAPTOLITES OF NEW YORK, PART 2 I 13 



from the facts that the branches cross and recross so that they are alternately 

 superjacent at the intersections and that the thecae appear alternately on 

 the outside and inside of the branches [see I), furcatus, text fig. 49]. 

 It is here shown that also the branches of Dicellogr. gurleyi [see 

 species description] grew in a wide and loose spiral and the same may be 

 claimed for D. cad vice us Lapworth. It is further demonstrated in this 

 paper that in D. smith i a tendency to an apparent light torsion or a 

 long spiral growth of the branches already apparent in I), sextans var. 

 tortus is further developed and quite distinctly intermediate between the 

 extremely long spiral of the first named variety and the more contracted 

 spiral of D. furcatus. But in looking over my material and the draw- 

 ings of foreign species, I can venture the general statement that wherever 

 longer branches of the rhabdosomes are preserved thev exhibit the following 

 two features: (1) a final convergence [for example Dicr. ramosus, 

 pi. 21, fig. 6, or D. nicholsoni, in Elles & Wood pi. 25, fig. ia]. 

 however straight and divergent the)- may have been at first, and (2) a 

 gradual wandering of the thecae from one side of the branch to the other 

 [see Elles & Wood ; D i c e 1 1. morris i, pi. 2 1, fig. 6b, c, d ; I ) . m o f f a t- 

 e n s i s , ibid. pi. 23, fig. ia ; D i c r . b r evicaul i s, ibid. pi. 24, fig. 3a ; 

 D. r a m o sus, ibid . pi. 24, fig. 8a ; D . n i c h o 1 s o n i , ibid. pi. 25, fig. 

 1 a and pi. 20, fig. 3, 5 of this memoir]. These two features, however, point 

 both to the same inference, viz, that of an arrangement of the branches in 

 two long spirals, passing equidistally on the surface of an imaginary double 

 cone. 



The advantages of such an arrangement are quite apparent ; they con- 

 sist first in the possibility of increasing the length of the branches without 

 pushing the distal part of the rhabdosome too far away from the center of 

 the colony and second in imparting the elasticity of a spiral to the rhabdo- 

 some, thereby protecting it against being broken or torn off. 



It can further be easily seen how the two spirals of the branches would 

 when disturbances occurred have exerted a special strain at the distant point 

 where they were united, i. e. at the sicular end ; hence the tendency to 



