64 



The digestive cavity lengthens, its walls remaining thick. 

 It curves forwards and upwards, and reaching the upper 

 surface immediately in front of the neural cavity opens out 

 into the mouth (see fig. 6) ; behind it lengthens out with one 

 or more windings, and the anus (a) (generally somewhat on 

 one side) is now distinctly visible. It is highly probable that 

 the anus is identical with the primary orifice of the diges- 

 tive cavity, but its identity cannot be traced throughout 

 satisfactorily. 



The epidermis becomes thin and flattened out over the 

 whole of the body except at the extreme front, where it is 

 developed into three prominences, varying in shape in 

 different species, and serving for the purposes of attachment. 

 The cells beneath the epidermis multiply rapidly, being in 

 some places very numerous. 



The axis band of the tail undergoes remarkable changes. 

 The roof of cells seen in fig. 4 increases very much in length, 

 the tail becoming in consequence very much curved. Each 

 cell possesses a well-marked nucleus. Very soon, along the 

 middle line of the axis band, particles of some highly refrac- 

 tive material make their appearance at the junction of the 

 cells. These increase very rapidly in size, pushing the cells 

 on one side and gradually taking their place, the new refrac- 

 tive body appearing between every two of the old cells. 

 AVhen two enlarging refractive bodies meet they coalesce, 

 and continue to do this rejjeatedly, the whole of the axis 

 band thus becoming made up of a mass of homogeneous re- 

 fractive material forming a case, the original cells being left 

 in a reduced condition at the outside only, where they con- 

 stitute a sheath. In this sheath elongated nuclei are visible, 

 but whether they are the transformed nuclei of the original 

 cells is not quite clear. 



This mode of origin of the axis band quite corresponds with 

 the development of the notochord of Amphioxus. There is 

 in both cases the same primary strand of nucleate cells, and 

 the same appearance of refractive particles which invade and 

 eventually replace the primary cells. The chief difference 

 is that, whereas in the ascidian the new material finally forms 

 an unbroken mass, in Amphioxus it exists in the form of 

 j)lates applied closely to each other. 



The rounded cells lying in the tail between the notochord 

 and the epidermis are gathered round the former in a mass 

 which ultimately becomes transformed into a tubular muscu- 

 lar coat. This strong likeness supports the view that the 

 axis band of the tail of Ascidians is really a notochord, and 

 thus we have in the Ascidian larva the three great vertebrate 



