In all the types above Fishes considerable changes are effected in the primitive arrangement of the arteries in the visceral arches.
In Amphibia the piscine condition is most nearly retained[229]. The mandibular artery is never developed, and the hyoid artery is imperfect, being only connected with the cephalic vessels and never directly joining the dorsal aorta. It is moreover developed later than the arteries of the true branchial arches behind. The subclavian arteries spring from the common trunks which unite to form the dorsal aorta.
In the Urodela there are developed, in addition to the hyoid, four branchial arteries. The three foremost of these at first supply gills, and in the Perennibranchiate forms continue to do so through life. The fourth does not supply a gill, and very early gives off, as in the Dipnoi, a pulmonary branch.
The hyoid artery soon sends forward a lingual artery from its ventral end, and is at first continued to the carotid which grows forward from the dorsal part of the first branchial vessel.
In the Caducibranchiata, where the gills atrophy, the following changes take place. The remnant of the hyoid is continued entirely into the lingual artery. The first branchial is mainly continued into the carotid and other cephalic branches, but a narrow remnant of the trunk, which originally connected it with the dorsal aorta, remains, forming what is known as a ductus Botalli. A rete mirabile on its course is the remnant of the original gill.
The second and third branchial arches are continued as simple trunks into the dorsal aorta, and the blood from the fourth arch mainly passes to the lungs, but a narrow ductus Botalli still connects this arch with the dorsal aorta.
In the Anura the same number of arches is present in the embryo as in the Urodela, all four branchial arteries supplying branchiæ, but the arrangement of the two posterior trunks is different from that in the Urodela. The third arch becomes at a very early period continued into a pulmonary vessel, a relatively narrow branch connecting it with the second arch. The fourth arch joins the pulmonary branch of the third. At the metamorphosis the hyoid artery loses its connection with the carotid, and the only part of it which persists is the root of the lingual artery. The first branchial artery ceases to join the dorsal aorta, and forms the root of the carotid: the so-called carotid gland placed on its course is the remnant of the gill supplied by it before the metamorphosis.
The second artery forms a root of the dorsal aorta. The third, as in all the Amniota, now supplies the lungs, and also sends off a cutaneous branch. The fourth disappears. The connection of the pulmonary artery with both the third and fourth branchial arches in the embryo appears to me clearly to indicate that this artery was primitively derived from the fourth arch as in the Urodela, and that its permanent connection with the third arch in the Anura and in all the Amniota is secondary.
