36 EARLY STAGES OF TEREBRATULINA. 



composing the shell. From an extremely young specimen I ol>serve(l these same peculiar 

 acicular scales (fig. 45), and they probably indicate the presence of the crura. 



Nothing of a satisfactory nature was made out regarding the structure of the shell in 

 the earliest stage (fig. 2, plate i), though had the shell possessed the peculiar scale-like 

 structure and ca-cal perfoi-ations, they would have been visible. 



Growth of the Shell. Figs. 18, 19 and 20, plate i, severally show portions of the 

 haemal valve, neural valve, and the valves connected, from an early lingula-form stage. 

 In fig. 18, the hinge plate (hp) and the dental socket (ds) are well marked ; in the earlier 

 stages of their development the crura appear as slight projections. In fig. 19 the promi- 

 nence of the so-called teeth of the shell (ie) is given. 



Fig. 20 represents the valves joined ; the ca?ca are seen perforating the entire margin 

 of the shell about the foramen, which is very wide and rudimentary. 



On plate ii a series of stages are given to show the gradual development of the crura, 

 and parts surrounding the foramen. The upper line of figures in this series, marked N, 

 I'epresents the neural or ventral valve, and the lower line of figures, marked ii, represents 

 the hsemal or dorsal valve. In the youngest of these stages, figs. 48 and 49, a broad gap 

 in the neural valve indicates the peduncular foramen. It will be seen that the shell is 

 folded upon each side of the peduncular passage. These folds never meet in this species. 



Figs. 50 and 51 show a slightly more advanced stage, the difference being mainly in 

 the greater prominence of the crura. 



In the next stage, figs. 52 and 53, the calcareous loop is well advanced, though not yet 

 connected by a continuous shelly layer, though at all stages these portions are connected 

 b}^ a membrane charged with spicule. Figs. 54 and 55 represent similar portions of an 

 adult shell for comparison. It will be seen that the neural process of the loop becomes 

 connected, while the hsemal process does not become so connected, though in some speci- 

 mens it closely approximates. The jieduncular opening becomes more circular, and the 

 cardinal process [c])), which does not appear iji the earlier stages, is now present. 



Figs. 56, 57 and 58 are added, to exhibit the relative outlines of the more prominent 

 stages in the growth of the shell. Compare, also, fig. 2, plate i, where the shell is as 

 long as it is broad, and the hinge margin represents the greatest transverse diameter of 

 the animal. In fig. 56 the shell elongates, Avith the sides nearly parallel, and the 

 peduncular opening wide and gaping. (The line accompanying the figures represents the 

 natural size of the shell.) Fig. 57 represents a stage intermediate between the lingula 

 form and the adult ; the peduncular opening is moi-e contracted, the sides of the shell 

 are no longer parallel, though the hinge margin has nearly the same outline as in the first 

 stage. Fig. 58 represents the adult shell. Here the outline assumes entirely different 

 proportions, being nearly oval, while the hinge margin forms a sharp angle, and the fora- 

 men is quite circular. 



Relations with the Pohjzoa. The general affinities of the Brachiopoda with the Poly- 

 zoa, as indicated b}^ Milne Edwards, Huxley, Hancock, Agassiz, Hyatt and others, are 

 admitted by most zoologists. 



In the preceding examinations of the early stages of Terehratidinn. we have additional 

 evidence of their relationshij). In the evolution of the lophophore, from its circular char- 

 acter, surmounted \)\ a few cirri, to that of the hippocrepian stage, we have suggested, 

 first, the circular lophophore in the inferior, or GymnolaMuatous Polvzoa, and liually. the 



