GRAPTOLITES OF NEW YORK, TART 2 8 1 



and a little further study shows that even the most extreme forms are con- 

 nected by an infinite number of transitional stages. It thus becomes 

 apparent that we have here one compact phylogenetic group of no more 

 than specific rank which, however, with a sudden outburst of variability 

 develops along various lines, some of which diverge widely enough to have 

 been designated as different varieties. 



In the succeeding Utica formation the species, or a mutation of it, per- 

 sists, but the wealth of appendical forms has disappeared, and the onto- 

 genetically earliest and phylogenetically oldest plain, double-spined form 

 alone survives. 



I have here arranged a number of the most characteristic forms from the 

 Normanskill shale on plate A to bring out synoptically the probable lines of 

 development of the varieties. The elements of these modifications are the 

 two spines, arising from the outer walls of the first two thecae, the virgella 

 or principal spine of the sicula and the alate extensions of the spines. 



The simplest stage is the bicornis stage, occupying the lowest tier of 

 the plate. In this only the two lateral or thecal spines are developed. The; 

 young of all the other varieties pass through this stage and the form shown 

 by most of the young and the Climacograpti of the preceding horizon is to 

 be considered as the most primitive. This is the one represented at the 

 base of the plate, with two slightly curved, widely diverging thin spines. 

 These appendages then begin to vary by (i) the degree of their diverg- 

 ence and (2) their curvature. They become either straight as in the lines 

 g and e, recurving as in f or curving inward and becoming semilunar as in 

 the other lines a-d. 



The degree of divergence of the spines may be influenced by the 

 direction of compression and some of the apparent modification is hence 

 due to this cause. Since with the exception of the line a, in which the 

 prongs have been subparallel, the angle of divergence has not been given 

 any importance and the appendages would naturally tend to settle on this 

 flat side, the direction of compression is evidently a factor that can be here 

 neglected. 



