62 GENERAL EVOLUTION". 



eniiinerated, the external or appendicular branchial vessels being 

 omitted, as belonging to the special serial deyelopment of the line 

 of air-breathing Anallantoidans. The division of the bulbus arte- 

 riosus into three instead of two may indicate a case of inexact 

 parallelism ; but, on the other hand, it may be that the pulmonary 

 partition is completed a little before the aorta-root partition, thus 

 passing through the batrachian permanent type. For explana- 

 tions of inexactitude see under Part II. No doubt the batrachian 

 type of hulhus arteriosus is passed by many serpents less extreme 

 and specialized than the Tropidonotus. 



The aortic and pulmonary divisions of the bulbus in the Cae- 

 cilia are not laterally placed, but one is dorsal and the other ven- 

 tral, the one passing a little spirally to the right of the other. So 

 the pulmonary division of the bulbus turns over to the right in 

 the Anura. When the septum of the true reptiles appear, it rises 

 on the anterior wall of the ventricle till it is seen, in Eunectes, to 

 meet the partition between the arteria pulmo7ialis and aorta-roots^ 

 and we have at once the right and left ventricles of the bird and 

 mammal structurally and functionally. Thus are the two ven- 

 tricles of man the same as the one ventricle of the fish, merely 

 divided by a septum.* 



In the fissure of the aortic bulbus, in the reptiles, a spiral turn 

 is again given, and in Testudo the one aorta-root issues behind 

 the other. In the crocodile the turn is still greater, and the 

 right aorta-root issues to the left of the left root, and vice versa. 

 In the birds we have lost the left root, and parallelism ceases with 

 this change. In the Mammalia the right root turns to the left, 

 so that in the comparison of these classes the rule of von Baer 

 above quoted is true ; no mammal at present known is identical 

 in a foetal stage with any fully grown bird, but with a foetus of 

 the same, up to a certain point. But for both classes the joaral- 

 lelism of those below them holds true. 



But it is with the exact parallelism or identity of genera that 

 we have to do in the present essay. That being established, the 

 inexact parallelism between the modern representatives of higher 

 groups follows by a process of reduction. 



* Agassiz, I. c , denies the homology of the ventricles of the turtle and mammal; 

 but it appears to me erroneously. He says : " The fact that the great blood-vessels 

 {aorta and art. puhnonalis) start together from the cavum venosum seems to prove 

 that the two cavities in the heart of turtles, which are by no means very marked, 

 do not correspond to the two ventricles in Mammalia and birds." 



