272 MR. ST. GEORGE MIVART ON THE [Apr. 28, 
but nevertheless show indications of separation exactly corresponding 
to the distal ends of the transverse processes to which they are at- 
tached, having two superimposed articular surfaces when the last- 
mentioned processes have such, and presenting a longitudinal groove 
in front and hehind each rib, the two grooves being most marked at 
the proximal end of the rib, and vanishing distally. 
The number of trunk ribs which thus bifurcate proximally is 
sometimes as many as eighteen* on each side; but the number is 
subject to some variation even in the same species. Thus in some 
individuals of Salamandra maculata (fig. 12) I have found three ribs 
(on each side) thus bifureating, in others six ; and in Triton cristatus 
twelve pairs of ribs will sometimes be thus conditioned. It is always, 
however, towards the anterior and posterior ends of the trunk that 
this bifurcation tends to disappear, and mainly towards its posterior 
end, the bifurcation generally commencing at the second vertebra 
when the first one has its proximal end undivided. The sacral rib 
rarely bifureates}, and it generally differs from the preceding ones 
by its greater stoutness, though it is usually short. The rmbs in 
general maintain a tolerable equality of length throughout the 
trunk ; very rarely (only in Taricha) the ribs increase markedly in 
length towards the middle of the body, and then as obviously be- 
come again shorter. In Avolotl the second, third, or fourth is the 
longest pair, and thence backwards they markedly decrease in length 
to the sacral, which is suddenly much longer again. Occasionally 
they are very short indeed, being much less in length than the ver- 
tebree to which they are attached ; this is the case especially in Pro- 
teus, Amphiuma, and Siren, where they are rudimentary (the poste- 
rior ones being quite so) and, as has been said, few in number. In 
Spelerpes, Amblystoma, and especially Plethodon, they are very short 
indeed, never, except very slightly, exceeding a single vertebra in 
length. Generally they do not equal in length two midtrunk-ver- 
tebree of the same individual; but P/eurodeles differs from all the 
other Urodela in the length and strength of its ribs, the longer ones 
considerably exceeding the length of two of the longest vertebree of 
the body. 
As has been said already, ribs are not unfrequently developed be- 
hind the sacrum. I have seen two such pairs in Glossoliga poireti, 
and traces of such in Plethodon and Amblystoma; one at least is 
sometimes present in Menobranchus, and often two and sometimes 
three (if not more) pairs are present at the root of the tail in Meno- 
poma and Cryptobranchus. These postsacral ribs, however, never 
bifurcate at their proximal ends, and they are always short and more 
or less rudimentary. 
The first rib of all is generally stouter as well as shorter than the 
ribs of the trunk which follow after (fig. 11). Not unfrequently it 
developes a process from its outer or upper side towards its distal - 
end ; this process projects outwards and somewhat upwards, aud very 
rarely is so considerable as almost to equal in size the remainder of 
* Spelerpes rubra, British Museum. 
t It does so in Spelerpes rubra and Amblystoma punctatum in Brit. Mus. 
