Respiration in Inverteh^ate Animals. 253 



cally, they do not enclose two systems of interlarnellar water- 

 tubes^ but only one. The mechanical problem presented by the 

 gills of Cai'dium proved extremely difficult of solution. When 

 understood^ it challenged any living mechanism for beauty and 

 perfection. In Mytilus the loops of the two lamellse are soldered 

 into union at the free borders : they stand vertically (PI. VIII. 

 figs. 17 & 24) : they circumscribe a deep intermediate gutter. In 

 Solen (PI. VIII. fig. 23) the loops expand. In Venus they also 

 somewhat exceed in diameter that of the bars, of which they are 

 the bend. In Mytilus, the inner and outer gills exhibit the same 

 formation. In the Ostreadse (PI. VIII. fig. 21), the loops at the 

 margins of the gills are so closely packed together horizontally, 

 as to appear like a continuous membrane bounding an angular 

 groove. Numerous other varieties in the mere shape and size of 

 the loops occur in different families of Bivalves — the type of 

 structure never changes. 



The proximal or attached border (PL VI. figs. ] & 2 «, 6) of the 

 gills occurs under many varieties of anatomical plan. Pholas 

 exemplifies one type. The two lamellse* are attached to the pallial 

 tunic. All the vascular bars terminate in a common trunk 

 (PI. VI. fig. 2 a, b) which runs at right angles to their axes, and 

 parallel with the length of the entire gill. There are two of 

 these trunks, one afferent, the other efferent. They occupy re- 

 spectively the proximal margins (PI. VI. fig. 1 a, b) of the two 

 laraellse of which each gill is composed. In Pholas these trunks 

 are supported by the framework of solid structure [c c and d d) 

 which occupies the interlamellar spaces. In Solen (PI. VIII. 

 fig. 23 a) and Mytilus (fig. 17 A) another plan of formation is 

 observed. Here the proximal border of the superior lamella of the 

 upper gill, and inferior lamella of the under gill, are unattached, 

 floating in the mantual cavity. In such instances the inter- 

 lamellar framework is wanting. The vascular bars at this border, 

 for some distance up the breadth of the gill, are tied together 

 by means of a continuous membrane {e, e) . Here the interlamellar 

 scaffolding, and the water-tubes which the former assist to form, 

 exist only where the two lamellae are adherent; viz. over the two- 

 thirds of the breadth of the gill nearest the free border. 



* It should be clearly explained that the word lamella, as applied to 

 the gill of the Acephalan Mollusk, should signify, one of the two plates of 

 which the gill, whether double or single, is composed. The gill is the whole 

 organ. In those instances in which (as shown in fig. 5, PI. VI.) the bars 

 stand vertically on the same lamella, then of course each lamella would 

 be composed of two plates, or finer lamella?. I am not quite certain that 

 such an arrangement exists in nature. In several gctiera — in Cardlum espe- 

 cially — when care is taken to avoid pressure upon the margin, such is 

 the true position of the loops, if not of the bars ])rocccding backwards 

 from them. 



