50 Whitney, on the Metamorphosis of the Tadpole. 
point at the opposite end of the crest ; while the venous or 
efferent vessel, beginning as a mere radical, gradually en- 
larges, and thus becomes the trunk that conveys the blood 
out of the gill to its ultimate destination. Calling this vessel 
the upper branchial vein as long as it remains in contact with 
the gill, we subsequently change its name when it leaves the 
gill and winds upwards for distribution to the head, and then 
designate it the cephalic artery. The middle branchial artery 
and vein proceed in like manner in connection with the 
middle crest, and the lower artery and vein in connection, 
with the lower crest. The middle and lower venous trunks , 
having reached the extremity of the crests, curve downwards 
and inwards, and then leave the gill. The former trunk, 
converging towards the sjnne, meets its fellow, and with it 
forms the ventral aorta. The latter gives origin to the pul- 
monary artery , and supplies also the integuments of the 
neck. 
We must now return to the external gills , for the sake of 
offering an explanation of the vascular change which accom- 
panies (and apparently causes) the shrinking and ultimate 
disappearance of those organs, while, concurrently, we observe 
the progressive development of the internal gills. 
It is evident that the three arterial trunks arising from the 
heart convey the blood to the outer gills for aeration during 
the period that these organs are in full development and play. 
It is evident, also, that during the same period the incipient 
inner gills exist in the form of small tufts (fig. 4), which at a 
little later period may, as I have said, be seen with the blood 
traversing the single loop of blood-vessels contained in each of 
the digit-like processes composing these tufts (fig. 12). As in 
the tadpole we can distinctly see the pulmonary vessels in the 
incipient lung of the future frog, so we cannot doubt that 
this loop of blood-vessel (however small) exists in the gill tuft 
in the first stage (fig. 4). I have already shown how the 
loop becomes the crest , and how the latter is connected with 
the large branchial vessels (page 49). For the same reason 
that I infer the existence of the loop in the first stage of the 
tuft, I infer also the existence of its vascular connection with 
the main trunks or branchial vessels which are then con- 
veying the blood to the outer gills ; and as, at a later period, 
the blood-vessel loop is visible, so, at a little later period, the 
vascular connection with the main trunks is also demon- 
strable, Hence, if it be granted that the visible vascular 
connection is but the enlarged development of the incipient 
state, I think we may understand “ the change or conver- 
sion from external to internal gills ” with the aid of the 
