416 STUDIES IN EVOLUTION 



tory, therefore, shows a gradual increase of from one to six 

 turns of the lamella in each spiral. 



The transverse band connecting the primary lamellse also 

 undergoes a series of changes. It has been shown that the 

 centronelloid loop (figure 7) passes into one having the 

 lamellse joined by a posteriorly directed, transverse band 

 (figure 14). This form of loop is retained as a mature 

 feature in the brachia of Z. Ntcoletti, Z. Saffordi, and in the 

 lower Trenton varieties of Z. recurvirostris. Passing to the 

 specimens of the latter species, which are geologically later, 

 the band no longer joins the lamellse as far anteriorly as in 

 the older variety (figure 20}. The point of connection in 

 Z. modesta is variable (figures 25 and 26), but is usually 

 more posterior than in Z. recurvirostris, while in Z. Headi it 

 is manifestly more posterior than in any of the older species 

 of Zygospira. The transverse band is now no longer arched 

 backward, but is just the reverse (figure 24), while its posi- 

 tion is progressively more and more posterior, and the loop is 

 gradually shortened before the spirals make their appearance. 

 The gradual increase in the number of the whorls in each 

 spiral and the recession of the transverse band have gone on 

 together.* 



The family Atrypidse includes the genera Zygospira, Glassia, 

 Atrypa, and Dayia. It is easily distinguished from all other 

 families comprised in the sub-order Helicopegmata, since the 

 spirals are between the first descending branches of the 

 lamellse, while in the Spiriferidse, Nucleospiridse, and Athy- 

 ridse the primary lamellse are between the spirals. 



The gradual increase in the number of whorls in the spirals 

 and the pushing backward of the transverse band in the 

 Atrypidse is carried farthest in the species of Atrypa. In 

 Coelospira Barrandei and 0. mnrginalis the brachial supports, 

 as worked out by Davidson and Glass, consist of about five 

 volutions, and are similar to those of Zygospira, except that 



* The extreme anterior position of the transverse band in Z recurvirostris 

 is therefore of no more than specific value, and on this account Anazi/ga David- 

 eon cannot well be separated from Zjgospira. 



