TRANSACTIONS OF SECTION D. 493 
discontinuous it has the appearance of angular vesicles. The wall of the tube 
is less distinctly achromophile than in the luminous cells of Pyrosoma. It, too, 
is traversed by an achromophile reticulum in which are scattered many nuclein 
granules. This tube is also embedded in a scanty amount of liquid which fills 
the remainder of the cell space. 
Analogy with the conditions found in Pyrosoma, in which the distinctive 
cells are manifestly the only luminescent elements, affords justification for con- 
cluding that it is extremely probable that the cells in Cyclosalpa pinnata which 
contain an intracellular tube, with the very distinctive histological characters just 
mentioned, are luminous cells. Nevertheless, while these elements compose the 
entire cellular structure of the luminous organs of Pyrosoma, in Cyclosalpa blood- 
cells in every stage of development are also to be found in the lateral organs. 
These organs in C. pinnata play the double réle of being both luminous and 
hemopoietic. In P. gigdnteum, these two functions are carried out by different 
organs : light-production by the luminous organs of Panceri, which are situated 
symmetrically at the anterior or oral end of the pharynx in the pericoronary 
sinus; the hcemopdietic function by the dorsal organ (dorsal gland of authors) 
which, originally double and bilaterally symmetrical, becomes subsequently single 
and median, and situated in the dorsal sinus of the branchial sac. 
It would be interesting to ascertain whether the luminous cells of the lateral 
organs of the solitary form of Cyclosalpa are derived from the cell of the test of 
the egg, as is the case of the cells of the luminous organs of the four ascidiozoids. 
Without awaiting further results, it seems to be of sufficient interest to draw 
attention to the distinctive structure common to the luminous cells of Pyrosoma 
on the one hand, and to those cells which I believe to be the luminescent elements 
of the lateral organs of Cyclosalpa on the other. 
2. On the Life History of Echinocardium cordatum. 
By Professor E. W. MacBrinvz, LL.D., F.B.S. 
The results which this Paper recorded were obtained during a sojourn at the 
Biological Station of the West of Scotland Marine Biological Association at 
Millport last summer. 
Of the development of Spatangoidea very little was previously known. 
Spatangoid larve were described by Johannes Miller, and in particular the fully 
grown larva of Dehinocardium cordatum was described by Mortensen. The 
earliest stages of the development of this larva have been described by Ziegler, 
who reared them from the egg. Owing to the fact that I had at my disposal a 
rich culture of the diatom Nitschia, I was enabled to rear the eggs from the 
monient of fertilisation throughout their entire larval history until they meta- 
morphosed into young Spatangoids. This stage in their development was reached 
at a period of from eighteen to twenty-three days after fertilisation. 
The egg segments rapidly and forms a blastula which is ellipsoidal, not 
spherical : this blastula ruptures the egg membrane and rises to the top as a 
free-swimming larva twelve to eighteen hours after fertilisation. About ten 
hours later it has become converted into a gastrula, an invagination having taken 
place at one pole. There is an anterior tuft of specially long cilia. The skeleton 
appears in the typical manner as two calcareous ‘asters,’ situated right and left 
of the archenteron. By the outgrowth of specially long branches of these the 
formation of the postoral larval arms is initiated. The rods supporting these 
arms are latticed. The celom arises as an unpaired vesicle which is nipped off 
from the apex of the archenteron, and which then becomes divided into right 
and left halves. At the same time the stomodeum makes its appearance as a 
wide shallow pit which grows towards and meets the apex of the archenteron, 
and thus the alimentary canal is completed. The ciliated cells become restricted to 
a single loop, and the anterior tuft of cilia is incorporated in this, finding its place 
in the przoral portion of the loop. The antero-lateral arms grow out, supported 
by antero-lateral rods. At the aboral pole of the larva there is a mass of mesen- 
chyme cells, in which are embedded the distal ends of the body rods and the ends 
of the recurrent rods—these last-named being branches running back from the 
antero-lateral rods. In the midst of this mesenchyme an independent calcareous 
