THE BRACHIOPODA OF THE SCOTTISH NATIONAL ANTARCTIC EXPEDITION. 159 



same stages depicted l)y Friele in the ontogeny of the type species, M. cranium 

 (Friele, 1877, pi. i. figs. 2-2i), the only point of difference being in the possession of 

 a less number of spinules on the septal pillar. 



The number of pores per square millimetre in this specimen ranges from 120 to 124. 

 The scale-like structure of the shell-mosaic is very clear. 



The second specimen from the same deposit, though slightly damaged, is interest- 

 ing as showing a very advanced terebratelliform (or pre-adult) stage (PI. II. fig. 18). 

 The dorsal valve of this example measures : length, 875 mm. ; breadth, 8-5 mm., and 

 shows the final development of the loop almost achieved. The descending branches are 

 broad, and possess two internal triangular apophyses indicating the position occupied by 

 the transverse (jugal) band attached to the septum in the terebratelliform stage ; also 

 very prominent spinules at the recurvation. The stage of this specimen is almost 

 equivalent to that of M. cranimn figured by Friele (1877, pi. iii. fig. 10) and 

 Beecher (1895, pi. ii. fig. 1 : 1901, pi. xxiv. fig. 1). 



The descending lamellfe are supported at their origin by vertical, slightly converging, 

 crural plates ; no cardinal process is yet present. 



The median septum is only very slightly visible and takes its rise immediately below 

 the apex of the valve, whence it runs a distance of half the length of the valve and 

 then ceases midway between the tiansverse band and the point of recurvation of 



the loop. 



In none of the fully adult specimens of M. diamantina which I have examined is 

 there any trace of the connecting bands on the descending branches, though the spinules 

 at the recurvation are still apparent but much reduced. The median septum, which is 

 only feebly developed, is also much reduced in length and confined to the umbonal 

 reoion, where it supports the rather prominent cardinal process. On either side are two 

 short parallel median septa supporting the convergent, but not united, crural plates. 



The ventral valve of the above specimen (length, 10 ram.) shows a large foramen 

 with rudimentary deltidial plates, beneath the anterior angles of which are the two 

 rather prominent teeth. 



On the exterior of both valves several conspicuous growth-halts are visible ; the 

 shell-punctse= 112 per square millimetre (middle of the ventral valve). 



Thouo-h the material at my disposal is so scanty, it does not seem improbable 

 that, judging from the stages just described, the intermediate phases in the meta- 

 morphosis of the loop will show considerable similarity to those described by Friele 

 in M. cranium. 



The correct relationship of M. diamantina with the sub-family Dallininas, a group 

 so characteristic of the northern hemisphere, is thus clearly established by the trans- 

 formations undergone by the brachial support. 



This fact, which is, I believe, the first recorded instance of the " Dallinoid " type of 

 development in austral waters, is of great importance, as it has hitlierto been considered 

 that the two phyla, of common origin, of the section Terebratella, i.e. the sub-families 



(ROY. SOC. EDIN. TRANS., VOL. XLVIIl., 381.) 



