OF THE MARSIPOBRANCH FISHES. 
427 
and also its essentially double nature. It is, now, separated from the ethmoidal flange 
by a narrow tract of soft cartilage, which, however, expands at the median line, thus 
giving two roots , as it were, to this large leafy tract. The fore edge is ragged, and there 
are points of cartilage, where broad lobes will be in a week or two (see also Plate 19, 
figs. 1, 2, c.tr.). At this stage the “flying buttress” is much more like that of the 
late Tadpole of the Bull Frog (op. cit., Plate 4, figs. 8, 9), than in the next stage 
(Plate 19, figs. 1, 2), for the whole framework is more outspread, and the prepalatine 
(pr.pa.) points forwards, leafy and flat. 
The subocular fenestra (s.o.f.) is made reniform by the basi-cranial flange ; its more 
arched outer boundary is made by the postpalatine ( pt.pa .) passing into the 
pterygoid ( pg .), which runs upwards into the pedicle ( pd.), and downwards into the 
epi-hyal ( e.hy .), without a trace of the quadrate lobe and condyle, and without a sign 
of segmentation. If we compare these stages of the skull of P. marinus, one-third 
grown (Plate 18); 5 inches long (Plate 19, figs. 1-3); 4 inches long (Plate 10, figs. 
4, 5), we shall see that the auditory capsules are relatively larger the younger the 
specimen. Also that they are short-oval in the youngest, then reniform, and then 
short-oval again. At first the semicircular canals make no difference in the outer 
form ; then the two wide tubes bulge the capsule out; and this is lost, afterwards, 
by the thickening of the walls of the capsule. The meatus internus (Plate 18, 
fig. 5, VIII.), is pyriform and smallish in the largest of the three, then relatively large 
in the next (Plate 19, fig. 3), and in the smallest (Plate 10, fig. 4), it is a large, gaping 
space. In making these comparisons we have to keep the particular species in view, 
for in P. jluviatilis, the capsules of the Ammocoste (Plate 19, figs. 4, 5) are already 
reniform. 
My work has been done according to what was possible with the materials at hand, 
first one species and then another; the reader must make the proper specific equation. 
Solid sections of tlte head and branchial region of the adidt Petromyzon fluviatilis. 
A.— Vertical sections. 
The various sectional views have purposely been made bald and diagrammatic, for 
the skeletal parts are very complex, and their relations to the surrounding parts very 
manifold and confusing; an exhaustive memoir on the anatomy of this type would 
have to be tenfold the bulk of this selected piece of work. 
In the large vertical section (Plate 23, fig. 1) the segmental muscles and then- 
deciduous septa are seen to have mounted up and overlapped the cranium as far as 
the external nasal opening ( e.n .) ; below, the ventral muscular bands are merely 
indicated in the figure. The mvelon and cerebrum ( my.f C 1-3 .) are seen enclosed in 
their fibrous sheath (theca vertebralis and dura mater), and below them is the 
* The line from my. is too low in this figure. 
