90 THE VOYAGE OF H.M.S. CHALLENGER 



"Toward its anterior part in the middle line the upper surface is perforated by a 

 transverse slit, the inhalent aperture (PL X. fig. 2, hr). The exhalent aperture is situated 

 in front of this, and at a lower level, proceeding from the wall of the body just above 

 the base as a short cylindrical tubular process (PI. X. figs. 1 and 2, at). The inhalent 

 aperture is enclosed by a pair of simple rounded lips, and is without tentacles. 



" The test forms a wide ca\dty, which extends freely into the capacious hollow conical 

 processes. The two apertures, inhalent and exhalent, form the only communication 

 between this cavity and the exterior. 



" A flat horizontal membrane is stretched across the test cavity in such a manner as 

 to separate off an upper chamber communicating with the inhalent aperture from a lower 

 communicating with the exhalent. This membrane, in the central region of the body, is 

 thick and of an opaque white colour (Fig. 10, g, on p. 93). This thickened central portion 

 runs out peripherally into eight processes directed to the intervals between the long 

 conical processes of the test. Opposite these intervals the processes become attached, or 

 ■give origin to bands of muscular fibres, which bands after a short radial course, in which 

 their fibres remain parallel, split into two halves (PI. X. fig. 1, m.h'.). The halves of the 

 bands diverge at an angle from one another, and proceed to the tijjs of the long conical 

 processes, where each is joined by the tip of a corresponding half-band from the next 

 adjoining process. The half-bands, as they run towards the tips of the conical processes, 

 give off a series of transverse muscular threads, which, passing from band to band, form 

 a series of loops one beyond the other, continued almost to the tip of the conical 

 processes. These radial muscles are apparently the homologues of the longitudinal 

 muscles of ordinary Ascidians. They would become longitudinal were the upper j^art 

 of the discoid body of the animal drawn upwards, so as to make the respiratory cavity 

 tubular instead of saucer-shaped. A second series of muscular threads lies beneath the 

 radiating bands just described, about their points of bifurcation, and extending thence 

 almost to their points of origin. This second set of muscles takes a circular direction 

 (PI. X. fig. 1, m.h.), and is continued round the entire circuit of the animal, the several 

 strands, of which there are about twelve, appearing to be continuous throughout their 

 length. These circular muscles lie beneath the radial ones ; and, were the respiratory 

 cavity elongated into a cylinder, the radial or longitudinal muscles w^ould thus be internal 

 in position, the circular external. In order to prevent confusion, and because of the 

 difficulty in drawing them clearly, not nearly the entire number of transverse and 

 circular strands is inserted in the figures in the accompanying Plate (figs. 1 and 2). 

 Their arrangement is shown in detail iu fio-. 5. 



" Over the muscular meshwork thus formed, and extending from it to be continuous 

 in all directions with its thickened central portion, the horizontal membrane is continued 

 as a thin and transparent lamina. Opposite the indentations in the margin of the 

 thickened central portion of the membrane, i.e. between the processes or thickened 



