Respiration in Invertebrate Animals. 247 



nective fibrous structures (PL VIII. fig. 19). In almost all 

 species of bivalve mollusks, the branchial bars more or less 

 closely approach the cylindrical in figure. To this rule of 

 structure those of the common Mussel form a remarkable excep- 

 tion ; they are here blade-shaped (figs. 17 & 20). The section 

 of the bar is frequently oval. In the genera Cardium, Unio, 

 Ostrea, &c. this form is exemplified. The subcylindrical canal, 

 circumscribed by the hyaline cartilages just described, is the true 

 blood-channel*. All naturalists have conjectured this fact; the 

 existence is now only for the first time proved. The cartilages 

 bounding these channels are now first announced. They do not 

 enclose the whole circumference of the vessel: they form a 

 third of the opposite halves (PI. VI. fig. 6^). The rows of cilia 

 correspond with their edges : the intervals between these edges 

 are membranous. The real osmotic movement of the gases con- 

 cerned in respiration is limited to these intervals. Along these 

 intervals, extending with beautiful regularity from one end of 

 the bar to the other, there travels a cilia-driven current. In 

 Mytilus the bars appear to swell (PI. VIII. fig. 17 o, o, o) at the 

 points at which they are joined together by the transverse 

 structures. The real blood-channel does not bulge. The car- 

 tilages of the bars at the base of the lamella are lost in and 

 identified with that embracing the trunk common to the whole 

 series {c, c). Traced carefully to the proximal border, they will 

 be observed to have this disposition : the cartilages of conti- 

 guous sides of adjoining bars form one piece, being so bent as to 

 become continuous at the proximal border of the lamella. The 

 bars are thus held firmly in situ and in relative connexion. 



At this point it becomes extremely interesting to inquire, 

 whether the lamella is composed of a single series (PI. VI. 

 fig. 4), laid side by side, of parallel bars, or of a double series 

 arranged in two separate planes? (fig. 5). The answer to this 

 question will implicate an important point of function. It is 

 difficult to convey clearly the idea of a double series of bars 

 constituting a single lamella. This undoubtedly is the disposi- 



* A very recent study of the minute structure of the gills in the Tunicates 

 and Ascidians has enabled me to resolve completely the homology of the 

 branchial bars in the bivalve mollusks, to explain demonstratively why it 

 is that in the gills of some Acephalans the blood-conduits are placed like 

 membranous channels between alternate bars (as is shown in PI. VII. 

 fig. 15, b, b, b), and that in others the blood-canal (as in PI. VIII. fig. 22) 

 occupies the axis of each bar. Though there exist in the gills of the 

 Tunicata a system of large transverse trunks, with which the parallel 

 ultimate blood-channels (the homologues of the " bars " in the Acephala) 

 openly communicate, in a supplementary note on this subject in the next 

 paper, it will be shown that the ultimate elements of the branchiae in 

 Tunicata and Acephala are really arranged on the same type. 



