276 REPORT—1846. 
centres of the vascular system (heart and lungs) ; according to which types, 
for example, it may be either closed below by the meeting of the sternal ribs 
one $i or by the intervention of a single or divided sternal bone 
hemal spine). And, further, since in fishes, as the lowest class of vertebrata, 
the vegetative character of repetition of forms, proportions and composition 
in the successive segments of the skeleton prevails in a greater degree than 
in any of the higher classes, so we may conclude that this hemal arch pre- 
sents, by its articulation with the epencephalic neural arch, its normal position; 
and that the whole occipital vertebra here manifests its veritable and typical 
character. 
As the hemal arches in the trunk of fishes commonly support diverging. 
appendages, which project freely outwards and backwards, but are hidden and 
buried in the muscular masses to which they give attachment, so the occipital 
arch, also, commonly supports its diverging appendages. They are absent 
in Gymnothorax and some other Murenide. The appendage is present in 
the form of a single multiarticulate filament in the eel-like protopterus* and 
lepidosirent ; it is modified by that mode of vegetative repetition which 
results in adding to the number of similar filaments directly articulated to 
the supporting arch; and is further complicated by the expansion or conflu- 
ence of the proximal joints in different degrees as they recede from the sup- 
porting arch, so as to constitute definable segments of the appendage}. 
Such is the condition of the part in most osseous fishes, and such is shown 
in the diagram of the base of the appendage in figure 5 ; where the proximal 
segment consists of two broad and flat bones (54 and 55), the next segment of 
five narrower and shorter but thicker bones (56), and the last segment of 
more numerous bones of the primitive filamentary form and multiarticulate 
structure, which bifurcate and radiate as they recede from the centre of at- 
tachment. 
We may connect the tendency to extreme and variable development in the 
peripheral parts of a vertebral segment, with the freedom which is the neces- 
sary consequence of their position: they are attached by one end only, they 
have not, therefore, that physical restraint to growth which may arise out of 
the fettering by both extremities, which characterizes the more central ver- 
tebral elements entering into the composition of the neural and hemal arches. 
Even in these we find the disposition to luxuriant growth or vegetative sub- 
division greatest in the peripheral elements, viz. the neural and hemal spines : 
much more, therefore, might it be expected in the less constant, diverging, 
and commonly freely projecting appendages of the vertebral arches. Although 
here the polarizing forces which tend to shoot out particle upon particle after 
the pattern of dendritic corals, plants or crystals, are so controlled by the 
antagonizing principle of adaptation, that the radiating growth is always 
checked at that stage and guided to that form which is suited to the wants 
and required by the mode of life of the species. 
Since, however, we are able to retain firmly and with certitude our recog- 
nition of the special homology of the diverging appendage of the occipital 
hzmal arch, through all its modifications, from the single ray of the lepidosi- 
ren to the hundred-fold repetition of the same elements with superadded 
dichotomous bifurcations sustaining the enormous pectoral fins of the 
broad and flat plagiostomous fishes thence called ‘ rays’ par excellence, so 
we can retrace, with equal certitude, the serial homology of this appendage, 
when it is so plainly manifested by its simple form as well as connections in 
* Linnean Transactions, vol. xviii. pl. 23, fig. 4, w. ; 
~ tT Bischoff, Lepidosiren paradora, Ato, pl. 2, fig. 4, ¢. 
} Hunterian Lectures on Vertebrata, figs, 27, 40, 41, 42, 43, 75. 
