in Cephalojyoda [Siphonopoda) . 303 



logous with the approximation of the later septa in tiie 

 senescent Nautilus. 



(/3) The compression of the anterior membranes in Sepia 

 is different ; its meaning we will now consider. I have pre- 

 viously (10) pointed out that the septa approach one another 

 postero-ventraliy, so that, in a vertical section of a small 

 portion of the shell, the last-formed appear closer than those 

 underlying. Further, I do not deny that in senescent shells 

 the last- formed septa may be closer than those preceding: 

 the law^s of geratology do not lapse when their action has 

 been extended. But these two appearances have been con- 

 fused by Riefstahl with another of very different nature. 

 Riefstahl states that membranes of chitin * continue from the 

 middle plate over the last-formed septa ; he implies that 

 sinuous partitions do not exist in these membranes. These 

 membranes are visible; but, as I pointed out (10), the partitions 

 pass through them and are seen on the surface (tig. 1, p. 305). 

 This observation was corroborated by Dr. Appellof, whose 

 paper, read on Sept. 14, 1887, was received in England just 

 two months later (11). He describes this structure as a mass 

 of chitinous membranes, traversed by partitions from the tirst. 

 In this the following changes take place : — cracks appear 

 between the membranes, at right angles to the partitions ; 

 these cracks widen, so that the membranes above and below 

 are pressed together, and in them calcification begins ', they 

 finally form the septa ; the intermediate membranes are fur- 

 ther se])arated by cracks, and form the " free-stretched mem- 

 branes." Since the partitions extend completely across the 

 space these membranes appear fastened to them. On liief- 

 stahl's hypothesis the free-stretched membranes should pass 

 through the partition- walls ; that they do so I could never 

 see [i![g. 2). My sections, unreconcilable with Riefstahl's 

 view, tully confirm that of Appellof: they bear out the sug- 

 gestion of the latter that the membranes are pressed together, 

 not separated, by calcification, so as to form septa ; the nacre 

 lies on either side of these compressed membranes, not be- 

 tween them. The structure just described bears no resem- 

 blance to approximated septa, and the changes are to be 

 explained in a different manner. 



The approximation of septa in Nautilus is therefore proved 

 to be not homologous with the coherence of the last-formed 

 membranes in Htpia. Riefstahl's argument for iutussuscep- 

 tional growth in Nautilus based thereon consequently falls to 

 the ground. 



* " Concbyoliu " is the term he employs ; it may be more correct, but 

 " chitin " is t<impler aud move usual. 



21* 



