M.CYONARIA. 11 



in da in--.-. ih.it tin-re was a great deal of excuse for my original mistake. I wish 



-- my thanks t. Dr. Vcrsluys for his very valuable advice ami assistance in 

 l.-timir_ f the position of this sjHM-ies. 



Tii.- .mis I'ritHHiflld is usually defined as simple or unbranchcd, but as 

 i\- (\:\) has shown that Karelin divtiriatta of Studer must be transferred to 

 tin- <_-.-!ui> I'riinnivllii, and this species is " sparlich verzweigt " (Studer 10, p. 643), 

 the brain -hing habit of our new species is not quite exceptional. 



The branching is irregular and approximately in one plane, so that a large 

 colony would probably be flabelliform. Some of the branches appear to divide 

 <li< hotomously, but in general the smaller secondary branches arise alternately from 

 the large main branches. The colony as a whole has some general resemblance in 

 its mode of branching to a species of Cnliyorgia, figured by Gray (1, p. 37) under the 

 name Callicella eltgaiis, and to Ctdii/tn-yin tulterculntti, as figured by Versluys(13, pi. vi., 

 fig. 15). About one-third of the colony is devoid of cojnenchym and calices. 



The niiml.-r of calices in a whorl varies. On the largest branches there are 

 eight or nine ; on the terminal branches from six to four. On the thicker branches 

 the whorls are in places scattered, and there are some calices isolated. On the terminal 

 branches the whorls are about 1 mm. apart. 



The calices are 1 mm. in height, and many of them show a swelling on the axial 

 si.li', due to the presence of a gonad. They are protected by three or four rows of five 

 or six orifli, which often overlap when the calyx is dried, but in spirit specimens, and 

 fi.'.jufiitly when dried, appear to be separated by considerable intervals of skin, as 

 shown in fig. 8. The opercular scales are very small, and the next circle of scales, 

 called by Versluys the " Randschuppen," or marginal scales of the calyx, are not much 

 larger, but do not overlap the opercular scales. The opercular scales are usually 

 provided with a short spine, fig. 9, but the marginal scales are rarely, and the other 

 scales of the calyx never pointed in this way. The opercular scales are about 

 0*3 x O'l mm., and the other scales about 0*5 x 0*3 mm. in size ; but they vary 

 a great deal in size, according to their position. 



As in other species of Primnoella the ccenenchym is protected by an armature of 

 overlapping scales, but these are exceptional in \ic\ng irregular in arrangement, and 

 they are not provided with a dentate margin (fig. 10). These scales are usually nearly 

 square in shape, with rounded edges, and may reach a size of 0*5 mm. across 

 diagonally. The surface of each scale is ornamented by a series of small tubercles as in 

 Primnitella, Underneath the scales there are rows of small spicules or " sclerites " 

 about 1 mm. in length, very similar to those drawn by Versluys for Primnuetta 

 austrataxise (13, fig. 60, p. 53). ^ 



The axis is horny, but contains a considerable amount of calcium carbonate. 



The species is related to Primnoella in the following characters. The scales are 

 thin and covered with small tubercles, but not sculptured nor marked by prominent 

 lines extending from the nucleus to the margin. The scales of the ccenenchym overlap. 



2 B 2 



