AND DEVELOPMENT OF PYROSOMA. 231 



If a line traversing the centre of the cyathozooid and the centre of tlie ovisac be 

 regarded as the axis of the whole foetus, then, in the present condition, the longest 

 diameters of the first and of the last ascidiozooids are parallel with that axis, and that 

 extremity of each, at which the elseoblast is situated, is directed away from the cyatho- 

 zooid. The long diameters of the intermediate ascidiozooids, on the other hand, cut the 

 axis of the foetus at a high angle, their elteohlastic ends being those which are nearer the 

 cyathozooid (fig. 13). 



As development advances, the fu'st and the fourth ascidiozooids retain their parallelism 

 to the axis of the foetus, while the whole series elongates, so that the fourth comes to l^e 

 situated close to the fii-st (Plate XXXI. fig. 14-), the four encircling the base of the cyatho- 

 zooid completely. This elongation of the whole series is effected, mainly, at the expense 

 of the isthmuses, which elongate so much as to be converted into slender cords, of which 

 the first connects the cyathozooid with the neural face of the first acidiozooid ; the second 

 connects the haemal region of the first ascidiozooid, at a point just opposite the endostylic 

 cone, ■with the neural face of the second ascidiozooid ; the third similarly unites the second 

 and the thu'd ; and the fourth, the third and fourth. 



But the elongation of the isthmuses is not merely sufiicient to allow the fourth ascidio- 

 zooid to come close to the third ; it is also enough to permit of a movement of rotation on 

 the part of the second and third ascidiozooids. The fkst and fourth, as has been seen, 

 early take up such a position that their long axes are parallel with the axis of the foetus ; 

 and, by degrees, the second and third revolve, their adjacent ends being allowed to sepa- 

 rate by the elongation of their connecting isthmuses, until their long diameters, from being 

 very obliquely inclined to that axis, also become parallel with it, and with the long dia- 

 meter of the first and fourth. Thus, at last, the long diameters of all four ascidiozooids 

 are parallel with one another and with the axis of the foetus, their similar ends being 

 turned the same way (fig. 14), while the isthmuses slope obliquely from the neural region 

 of one to the haemal region of the next. The long diameter of each ascidiozooid is at right 

 angles with its proper axis (which would be a line drawn from the oral to the cloacal aper- 

 ture), and, hence, the neural and hssmal sides of the body are at opposite ends of its 

 long diameter. The neural side is that, which is turned in the same direction as the 

 aperture of the cyathozooid, whUe the haemal side is the opposite. The mouth is at 

 that end of the true axis or short diameter of the body which is turned outwards ; while 

 the atrial aperture eventually makes its appearance at the other end of this diameter and, 

 consequently, on that face of the ascidiozooid which is adjacent to the ovisac and 

 cyathozooid. 



At the commencement of the series of changes here indicated the ascidiozooids arc, indi- 

 vidually, much smaller than the combined cyathozooid and ovisac ; but as development 

 advances, the latter diminish Avhile the former increase ; and as, by the increase of size of 

 the ascidiozooids, the interval between them becomes both relatively and absolutely less, 

 they, at last, completely hide the combined cyathozooid and ovisac from view, so that it is 

 not always an easy matter to find the latter (Plate XXXI. fig. 15). The test increases, 

 concomitantly with the ascidiozooids, enveloping them and filling up their intervals so as, 

 finally, to form a spheroidal investment for the entire tetrazooidal foetus (figs. 11. & 15). 



