246 PROFESSOR MARSHALL AND EDWARD J. BLES. 



which he does not specially mention, and has described a connection 

 between the lingual artery and the vessels of the mandibular arch 

 which does not really exist. That the former alternative is, however, 

 quite possible, is shown by the very different development of the 

 mandibular vessels in the two species. Maurer's description, which 

 on this point leaves no room for doubt, shows that in Bana esculenta 

 there is a complete vessel in the mandibular arch at a stage when 

 the dorsal aorta has not yet formed ; that the mandibular vessels 

 are in fact among the very earliest to appear. In Bana temporaries,, 

 on the other hand, we find no trace of a vessel in the mandibular 

 arch until long after the aorta is completed, and at no period do we 

 find this vessel in continuity with the heart. 



The general conclusions to be drawn concerning the vessels of the 

 mandibular and hyoidean arches appear to be that in both arches 

 vessels are found which in mode of development and in general 

 relations conform to the type of the vessels in the branchial arches. 

 In Bana tenvporaria the hyoid vessels appear earlier, and correspond 

 more closely to the branchial type than do the mandibular vessels ; 

 while in Bana esculenta, according to Maurer's description, the reverse 

 appears to be the case. 



The establishment of direct communication between the afferent 

 and efferent vessels of the branchial arches requires some further 

 notice. There is no point in our investigations which has given us 

 greater trouble, or about which we have obtained such conflicting 

 evidence from different specimens. It is not always possible to 

 determine the presence of a small hole of communication between two 

 closely adjacent vessels by examination of a series of sections ; nor 

 is it possible in every case in which such a hole or passage is 

 detected to make quite certain that it is natural and not artificial. 

 We, therefore, desire to speak on this point with considerable 

 reserve, fully acknowledging that confirmation of our results is 

 needed before their acceptance can be claimed. We have found it 

 a good test to compare the two sides of the body in any doubtful 

 case : an opening which is present on both sides in identical positions 

 may fairly be assumed to be natural. 



According to our own observations detailed above, the direct 

 communications between the afferent and efferent branchial vessels, 

 which become established in tadpoles of about 12 mm. length (cf. 



