392 
THE VOYAGE OF H.M.S. CHALLENGEE. 
besides agrees with Atopogaster and the genera above it in most of the other points in 
its structure. 
At the point indicated by D. in the diagram (fig. 12) the axis of the Polyclinidse 
divided into two lines of descent, the one leading to an ancestral form in which the 
stomach -wall was areolated (see p. 177), while the second leads to a form F. in which the 
stomach- wall is thrown into well-marked longitudinal folds. The short caeca in the 
areolated condition are probably formed by the breaking up of longitudinal folds, since in 
Morchellioides affinis, a form derived from B., the caeca in the stomach-wall are distinctly 
arranged in longitudinal rows (see p. 179). 
At the point E. two lines of descent were formed, in one of which the branchial 
aperture of the Ascidiozooid became eight-lobed, while in the other it remained six-lobed, 
the condition in which it is found in most Compound Ascidians. The ancestors with 
eight-lobed branchial apertures gave rise (see fig. 12) to the genera Parascidia and 
Morchellioides (see p. 177) ; while the other line of descent has split up into Synoicum 
and the closely allied forms Sidnyum and Morchellium. 
The modification of the branchial aperture seen in Parascidia and Morchellioides is 
also found in some of the descendants of F. These forms, characterised by an eight- 
lobed branchial aperture, have given rise to the genera Fragarium and Circinalium, and 
the modification has apparently been formed independently in this branch, since both of 
these genera have well-marked longitudinal folds in the stomach- wall. The other line of 
descent derived from F. (see fig. 12) has given rise to the closely allied genera Apliddum 
and Amaroucium, which may be regarded as typical Polyclinidse, and some closely related 
forms. The genus Polyclinoides was probably derived from a side branch between F. 
and the ancestral form of Amaroucium. Sigillina seems to be an offshoot from the 
ancestors of Aplidium. Psammaplidium is an interesting form also derived from the 
old and well-known genus Aplidium, by a modification of the test, which has acquired 
the property of taking up and growing over sand-grains and other foreign bodies, so that 
they become a part of the investing mass (see p. 237). 
The second line leading from the point C. (in fig. 11, p. 388) gave rise first to the 
main axis of the Distomidse (see fig. 13), and here, just as in the case of the early 
Polyclinidse, a series of changes must have taken place, resulting in the formation of 
large colonies in which the Ascidiozooids were more or less completely imbedded in a 
common investing mass formed by the fusion of their tests. The union of the atria] 
apertures to form common cloacal cavities apparently did not take place so soon as in 
the case of the Polyclinidae, as in many of the Distomidae the atrial apertures of the 
Ascidiozooids are still found opening independently on the exterior of the colony. 
The line leading from the point C. (fig. 13) to the early Distomidas gives off a 
short side branch to the genus Diazona. This remarkable form, although not situated 
upon the main axis, is probably the nearest form now known to the common ancestor C. 
