216 D. M. S. Watson — A New Genus of Brachiopods. 



the interior of the ventral valve as determined from the casts seems 

 to me to be unworkable, and I prefer to leave any further comparison 

 to those who are able to re-examine the specimens. There is no 

 question of generic identity, as in Keyserlingina the ventral valve is 

 strongly convex and probably not completely attached. 



Afar more interesting comparison is with Leptodus {= Lyttonia). 

 Except that it is only attached by the urabonal region the young 

 specimen of L. americanus figured by Girty, 1908, pi. iv, fig. 8, 

 agrees extremely closely in its outline with our specimens, having an 

 axial slit in the dorsal valve, and on the right side at any rate, when 

 viewed from the dorsal surface, two lateral incisions. This specimen 

 also shows a short straight hinge-line in the form of a groove, and 

 distinct dental areas quite similar to those of Poikilosakos, so far as 

 they are not concealed by the dorsal valve. 



As Girty has pointed out, in his species the lateral " septa" of the 

 ventral valve very often take the form of long, narrow loops, which, 

 as shown in his pi. xxv, figs. 2 and la, are very like those of 

 Poikilosalcos. In L. americana the medium septum in the ventral 

 valve bifurcates towards the anterior end, and the dorsal valve is 

 always split by a medial incision from the anterior margin. 



The figures of Lyttonia nobilis given by Waagen, pi. xxx, figs. 6 and 9, 

 show that the muscular impressions on the ventral valve of shells of 

 this genus agree with those of our new genus in being long and 

 narrow in shape, very small in size, restricted to the extreme 

 posterior end of the shell, and widely separated from one another 

 anteriorly. These figures also suggest that there is some difference 

 in size of the muscles of the two sides, though such disparity does 

 not approach the remarkable condition seen in Poikilosalcos. 



These resemblances leave no-doubt that Poikilosalcos is a member of 

 the Lyttoniiclse and that it is an early and primitive member of this 

 group. Taken in connexion with L. americanum the comparison 

 makes it certain that, as Girty has already suggested, the " septa" of 

 Lyttonia both lateral and median are no true septa, but have arisen 

 by the growing together of the edges of a deep slit in what is the 

 functional and morphological, though not the real edge of the valve, 

 a process which in exceptional regions of certain specimens of 

 Poikilosalcos has already taken place. This origin affords a simple 

 explanation of the fact which JSToetling has recorded, that the " septa" 

 of the ventral valve of Oldhamina are double in structure. The fact 

 that the remarkable pinnate dorsal valve of Oldhamina has developed 

 by the excessive emargination of the shell, when considered in 

 connexion with the probability that in Poikilosalcos the lophophore 

 was a continuous band following the margin of the dorsal valve 

 .which supports it, allows us to deduce a hypothetical structure and 

 a theory of the functional arrangements of this remarkable type. 



It is pi'obable that in Oldhamina each pinna of the dorsal valve 

 supported a double band of lophophore running its entire length, and 

 that the median septum was provided with a strip of lophophore on 

 either side. JSToetling showed that each pinna of the dorsal valve 

 rests its anterior edge on a " septum" of the ventral valve, so that 

 the cavity of the shell forms a series of canals leading to a pair of 



