214 NEW YORK STATE MUSEUM 



naturally bend away from each other in performing their function 

 and the widening of the plates would soon not only separate them, 

 but would place each in a position on the suture instead of at the 

 corner. The corner would thus soon offer a free position for the 

 protrusion of a branch b.v. which would increase the respiratory 

 area. Should any b.v. thus protruded remain at the corner it would 

 tend to stop the development of any new branches. It is very 

 likely, however, that more than one would seek the same corner and 

 their development would again insure their bending away from each 

 other. In this case greater freedom for function would be found 

 on a vertical suture and the growing corners of the plates would 

 soon push by the b.vs. and leave them in a fixed position on the 

 vertical sutures. Still newer b.vs. would be led to take positions 

 giving most room, which this time would probably be again on the 

 horizontal sutures. If the plates produced more stereom on the 

 vertical than on the horizontal sutural faces, the horizontal sutures 

 would become the longer, offer the most freedom for function, and 

 come to contain the greatest number of b.vs. A b.v. next to a mem- 

 ber of a first ])air would keep it from increasing its distance from its 

 neighbor and a new b.v. at the corner would have the same eft'ect 

 on the last previously formed. It would thus happen that the dis- 

 tance apart of these structures would become fixed and not increase 

 with growth though the number would increase. The distance apart 

 of the primitive pairs of b.vs. might easily become a specific 

 character. 



The newer b.vs. in our species are all smaller than the members 

 of the primitive pairs and they are set nearer each other. The 

 average distance of the three around the corner b.v. at the left end 

 of the suture between a and b, figure 2, is 0.19 mm. The 

 size and distance between the newer b.vs. are also likely to be 

 specific characters. 



A closer examination of the group of b.vs. just mentioned will 

 reveal several additional points of interest. The central one evi- 

 dently emerged before the others had become completely inclosed 

 by the growing plate corners (see text figure 3, which presents a still 

 greater enlargement of this area). These all seem to be branches 

 of the left-hand member of the pair between a and b. It sent oft* 

 five branches. The widening of the plate allowed two branches to 

 remain on the same suture, two on the vertical and one at the corner 

 with no newcomer to dispute for its territory. The b.vs. have been 

 numbered in the order of their distance from the angle of the plate. 



I 



