202 Observations on the DeveIoj)ment [^jomiL,^.''^m:'^ 
the blood in its passage : lastly, the operculum closes, expelling most 
of the remainder. To support the delicate branchiostegous mem- 
brane, four bony processes are given off from the Os Hyoides. The 
blood-currents through the gills are next subdivided and broken up 
into an innumerable series of small vascular loops, by which means 
the blood is more thoroughly exposed to the purifying action of 
the water. It is only the commencement of this development which 
can be seen by the microscope, as pigment cells grow rapidly and 
obscure the parts beneath. The first step is the budding of the 
branchial arch into a set of tubercles, as at a, Fig. 35. These 
tubercles support a set of small vessels or viUi, little vascular 
loops, which spring from the branchial artery, and transmit the 
blood-corpuscles in single file. The tubercles increase in size, and 
form a broadish leaf-like structure, of which there are a great 
number, springing from each arch. These leaflets support the sub- 
divisions of the branchial artery : along the edge of each leaflet 
the artery is split up into many capillary vessels ; these coalesce to 
form a vein containing aerated blood. The branchial artery divides 
lengthwise (Fig. 37) into artery and vein, the former tapering from 
the heart extremity to the aortic one, and giving off continually a 
branch to each leaflet as it passes along the branchial arch, the vein 
being parallel with it, but tapering in the other direction, being 
smallest where the artery is largest, and becoming larger as it 
receives the coalesced capillaries and advances past each leaflet to 
the aortic circle. Fig. 35 : a, the leaflets as they are seen at first 
(Fig. 36, a, being another view) ; h, the villi, as first seen ; c, the 
two together ; d, pectoral fin ; e, the intestine ; /, liver ; ^, the yelk- 
ball, much shrunk by absorption of the vitellus. Fig. 36, h, c, d, 
the capillary loops ; Fig. 37, diagram of a, the branchial artery ; 
h, the branchial vein, with the intervening capillaries. The first 
villi were seen about the 28th April, 15th day. They then showed, 
as simple loops, the blood discs ascending, curling round, and de- 
scending. On the 30th the loop appeared spread out, almost circu- 
larly, and was bent, and sometimes twisted on itself (c, d, 36). 
They were increasing in length, and were beginning to subdivide, 
as seen at h, 33. Pulsation was evident in the loop. After this 
period the rapid increase of pigment prevented further observations 
on this point. On the 27th, in the auditory capsule, I observed 
three semicircular canals (Fig. 38), and an ampulla (a), into the 
dilated extremity of which two ends of one and one end of a second 
may be seen to enter. At the base of the ampulla is an otolith. 
The relations of the other otolith and the other terminations of the 
canals cannot be made out. 
This brief abstract being now concluded, I have only to recapi- 
tulate the chief points observed by me. 
The blood-corpuscles flow to the heart before the heart has 
