EEPORT ON THE TUNICATA. 
271 
If a thin section is soaked for from six to twenty-four hom-s in water saturated with 
carbon dioxide, it is found that the spicules are either entirely dissolved out or only 
partially so, according to the length of the treatment. Their remains when visible are 
always found inside the large spherical or polygonal vacuoles, the boundaries of which 
stain readily with eosine or andine blue, and more faintly with picrocarmine. In a few cases 
after the treatment with carbon dioxide the vacuole was distinctly stellate in form, and 
not very far outside the rays of the partially dissolved spicule. 
The chief difference between the sections decalcified by means of carbon dioxide 
and those treated with hydrochloric acid was that in the former, in nearly all cases, 
the more or less stellate membrane which had apparently been attached to the spicule, 
and was set free by its dissolution, lay distinctly inside the margin of the vacuole (see 
PI. XXXV. fig. 6, where 1 shows the margin of the vacuole, and 2 the stellate membrane 
inside which the spicule lay). 
These results seem to indicate that the calcareous spicule is formed by a group of 
modified test cells, which constitute, when the spicule has attained a fair size, its mem- 
branous investment.^ Possibly the angular thickenings, which stain more deeply than 
the rest (see PI. XXXV. fig. 5), may be the nuclei or the protoplasmic remains of the 
constituent cells. This membrane is apparently in the natural condition in close contact 
with the test matrix externally, and with the spicule internally, with the exception of 
occasional small chinks at the bases of the rays in some specimens (PI. XXXV. figs. 8, 9). 
Slow decalcification with carbon dioxide sets it free from both test matrix and spicule 
(PI. , XXXV. fig. 6), while the more rapid action of hydrochloric acid causes the membrane 
to assume a polygonal or spherical form, and seems to press it outwards against the 
bounding test matrix (PI. XXXV. figs. 5, 7, 10), probably as a result of the evolution 
of carbon dioxide. It is not, however, evident to me why, after treatment with water 
saturated with carbon dioxide, the vacuole in the test matrix should assume a more or 
less spherical form (PI. XXXV. fig. 6), as in this case no gas is evolved. 
The branchial sphincter and the muscular system generally are well developed. The 
invaginated test lining the branchial siphon is distinctly visible in the vertical sections 
(see PI. XXXV. fig. 2, hr.) and contains a number of calcareous spicules. A few longi- 
tudinal muscle bands start from the posterior end of the branchial sphincter and radiate 
outwards over the anterior part of the body. 
The stigmata in the branchial sac are notable for their considerable length. The 
ciliated cells are distinct, and are pointed at the free ends. The tentacles are very 
distinctly of two sizes (PL XXXV. fig. 2, tn. aud tn'.), which as a general rule are placed 
alternately. 
^ The observations which were made on the test and the discoid spicules of the genus Cystodytes point to a similar 
process of formation (see p. 139). 
