CERATODUS, PROTOPTERIJS, AND CHIMERA. 497 



segment of the longitudinal (spiral) valve is seen projecting from the left-hand side of 

 the divided wall of the lower part of the conus (compare fig. 7, A and c). The wall 

 of the lower part of the conus is continued downwards to form the arch-like auriculo- 

 ventricular valve. The free depending portion has a perfectly smooth surface ; but this 

 is not the case higher up. In fact the wall of this lower segment of the arterial cone 

 is richly provided with peculiar pocket valves, which have not hitherto been described. 

 These valves are so delicate that they might easily escape detection, or, in a specimen 

 which was less excellently preserved than the one now examined, might be destroyed. 

 An enlarged view of this portion of the wall of the arterial cone, pinned out flat and 

 advantageously illuminated for the purpose of showing the valves in question, is given 

 in fig. 5. It will be observed that there are three rows of these valves, of which the 

 lower is the most complete, exhibiting as many as eight separate flaps, whilst these, 

 and more especially the flaps of the middle series, exhibit a tendency to reduplication 

 by transverse division. The three rows of valves correspond to three folds on the 

 concave surface of the spiral longitudinal valve. The valves thus disposed consist of 

 delicate membranous flaps (some as large again as, others smaller than, the area of a 

 common pin's head), which are fixed to the wall of the cone by their lower borders ; 

 and most of them hang freely by the other three sides, except for the presence of more 

 or less numerous chordae tendineee, which attach them loosely at various points (see 

 fig. 6). Some of the larger valves are attached by the two sides right and left, as well 

 as by the base, so as to form regular watch-pocket valves like those in the upper limb 

 of the cone. 



The delicate and diminutive nature of some of these valves suggests very forcibly 

 that we have here the remnant of an ancestral condition in which more powerful 

 pocket valves were present thi'oughout the extent of the conus, and that these have 

 dwindled in proportion as the longitudinal spiral valve, characteristic of Dipnoi, has 

 developed. That this is the case is further pointed to by the fact that in Ceratodus 

 pocket valves are still strongly developed in the upper part of the conus, into 

 which the longitudinal valve does not reach, whilst, on the other hand, in Lepidosiren 

 and Protopterus, where the longitudinal valve does extend into the upper vertical limb 

 of the conus (fig. 7 b), no pocket valves are present in that region. 



Seeing that the delicate valves of the lower part of the conus of Ceratodus were 

 such as to escape attention, it occurred to me as possible that similar delicate pocket 

 valves might exist in the corresponding part of the conus of Protopterus ; and accord- 

 ingly I searched for them in that position and found them (fig. 12). 



Before leaving the heart of Ceratodus and passing on to that of Protopterus, there 

 are three additional points which a comparison of my own notes with Dr. Giinther's 

 description leads me to consider as worthy of mention. First, as to the number and 

 form of the pocket valves in the upper vertical segment of the cone in my specimen. 

 Dr. Giinther examined the hearts of two specimens of Ceratodus ; and, knowing the 



