1903.] ANATOMY OF THE JArAXi:SE SALAMANDER. 313 



parietes side by side in a symmetrically paired fasliion. It will 

 b6 seen that this account does not exactly agree with that of 

 Osawa. I saw only four intercostal veins, which are disposed 

 singly. Two of them lie at intervals from each other in front of 

 the first of the intercostal arteries ; then follow two aiteries ; 

 after this the third vein ; then a pair of arteries, and finally the 

 last of the intercostal veins. 



Arte7'ice comites aorta} *. — The unpaii'ed dorsal aorta for some 

 little distance before and after the origin of the subclavians is 

 accompanied, as shown in the drawing (text-fig. 34, p. 314), by 

 a slender aorta on each side, not distant more than 2 or 3 milli- 

 metres from the aortic trunk. At the origin of the subclavians 

 these vessels each give off a branch which accompanies in the 

 same way its own subclavian. These two vessels arise from the 

 aortic trunk between the union of the two aoi'tfe to form the 

 dorsal aorta and the origin of the subclavians, but miich nearer 

 to the latter. Each of these vessels, immediately after- its origin, 

 gives off a forwardly-running trunk which accompanies the aorta 

 on each side for a little distance, and is lost among the muscles of 

 the back. The main branches run posteriorly .and end upon the 

 transverse membrane wdiich ties the transversely- running portion 

 of the oviduct to the parietes. Osawa does not appear to have 

 studied these arteries. 



Each arteria comes aortm is reinforced during its course by 

 other branches aiising from the aoi-ta. That of the left side has 

 two such bupplementaiy roots. One springs from the aorta just 

 in front of tlie origin of the subclavian ; it ci'osses over (really, of 

 course, below) the subclavian, and runs for a considei^able distance 

 alongside of the first branch before joining it. A second branch 

 springs from the aorta on the posterior side of the subclavian. 

 This also i-uns alongside of the two above-mentioned vessels for a 

 considerable distance before joining them. The left-hand vessel 

 has only the first of the two accessory bi'anches, which arises 

 actually from the posterior margin of the subclavian. About 

 halfway between this point and the oviducal mesentery the artery 

 divides into two branches, of which the inner spreads out over 

 that mesentery, while the outer ends in a projecting reddish body 

 which is dealt with later (p. 315). 



The anatomy of this region of the body in Megalohatrachus is 

 in several respects by no means unsuggestive of the mammalian 

 diaphragm, or, to be more accurate, a portion of the diaphragm, 

 which is itself a complex structure. The transverse position of 

 the membrane supporting the end of the oviduct (found, it must 

 be remembered, in both sexes) is the first notable fact to be pointed 

 out. Secondly, the heart and a gi-eat part of the lungs lie in 

 front of this transverse septum, which from its position delimits 

 ^' a '"chest" not incomparable with the mammalian thorax. The 

 arteries, which have just been described as ramifying upon the 



*■ These arteries must not h^ rniirn'~crl \\ifh I he deep-runuing' vertehral arteries 

 figured bj' Hyrtl and just referred to by myself (p. 310). 



