40 
THE VOYAGE OF H.M.S. CHALLENGEE. 
more complicated system, where the branches at the right hand end have increased 
considerably and are beginning to fork in their turn. In E, finally, an anastomosis 
has been effected between two of the secondary branches at the right hand side of 
the system so as to produce an island of common test {x, y) surrounded by a ring 
of Ascidiozooids. Such an island (which is very commonly seen in large colonies of 
Botrylloides, see PI. I. fig. 4) is entirely diilerent from a simple system such as 
A or B, not only in mode of formation but also in structure. In the first place it 
contains no common cloaca, and in the second place the surrounding Ascidiozooids are 
placed so that their ventral edges face inwards, while in the simple system (A or B) the 
dorsal edges of the Ascidiozooids are nearest to the centre of the space. This 
arrangement is obviously a result of the mode of formation of the ‘‘ island,” which is 
morphologically outside the system although enclosed by a part of it. It is, therefore, 
not at all surprising to find that such islands usually contain numbers of “terminal 
knobs ” like those found in the colonial test outside the systems and forming the edges 
of the colony (see PL I. figs. 2, 5). Even more complicated conditions than that 
shown in E are found in old colonies of Botrylloides, but they are all produced in the 
manner which has been indicated by irregular branching and anastomosing. 
The other characters by which the typical Botrylloides differs from the typical 
Botryllus, viz., the shape of the Ascidiozooid, its position in the test, and the situation 
■of its apertures, may, I think be regarded as the natural result of the modification of the 
system which has just been traced. In Botryllus each Ascidiozooid in the system has 
its atrial aperture opening directly into the circular centrally placed common cloacal 
aperture, and therefore the body must be placed with its long axis directed radially, 
and the atrial aperture must be situated at the central (morphologically posterior and 
dorsal) end, far from the branchial aperture, in order that it may reach to the common 
cloaca. Hence in Botryllus the Ascidiozooids lie with their ovate bodies horizontal 
{i.e., the antero-posterior axis is parallel with the upper surface of the colony), and the 
apertures are distant. Now when the system became much elongated it would clearly 
be impossible for the Ascidiozooids at a distance from the common cloaca to have 
their atrial apertures opening directly into that cavity, so they came to communicate 
with canals continued outwards through the test, from the common cloaca, one in each 
branch of the system. Such an arrangement would not only do away with any 
necessity for radial extension in the body of the Ascidiozooid, but would, by bringing 
the common cloaca (in the shape of its prolongation the canal) closer to the body, tend to 
cause the branchial and atrial apertures to come nearer. As a result of this we find in 
Botrylloides the cylindrical Ascidiozooids placed more or less vertically {i.e., with the 
antero-posterior axis perpendicular to the surface of the colony), and with the apertures 
close together at the anterior end of the body. 
In the collections there are four species and a well-marked variety of Botrylloides; 
