207 
through the conus.” Preceding this Bruner states: “The conus of the 
Salamandrina shows the same general structure as that we found in the 
conus of the Salamandra. <A spiral valve is distinctly recognizable in the 
lungless form.” (Salamandra has lungs; Salamandrina has none.) 
This point in the comparative anatomy of the amphibian circulation 
I hold to be an excellent objection to the described course of the blood 
through the frog heart. 
4. Experimental evidence on the amphibian circulation leaves much 
still to be done. Mayer found that if the tip of the ventricle was cut off 
the blood issued in two distinct streams. This, in addition to the colora- 
tion in the beating frog heart, seems to hold for a segregation of the 
venous and arterial blood in the spongy ventricle. But Gompertz’s experi- 
ments also seem to indicate that even if this be true a mixing must occur 
in the vessels. 
The step from the amphibian to the class of Dipnoi is not a very great 
one, and still we find something which may throw light on the character 
of blood circulating in the fish. According to Wiedersheim ‘in Ceratodus 
the conus arteriosus is provided with eight rows of valves and begins to 
be divided into two chambers. In Protopterus this division is complete, 
so that two currents of blood, mainly arterial and mainly venous re- 
spectively, pass out from the heart side by side. The former comes fiom 
the pulmonary vein, from which it passes into the left atrium, thence into 
the left portion of the ventricle, and thence to the two anterior branchial 
arteries. The venous current, on the other hand, passes from the right 
portion of the ventricle into the third and fourth afferent branchial ar- 
teries and thence to the corresponding gills, where it becomes purified; it 
reaches the aortic roots by means of the efferent branchial arteries. The 
paired pulmonary artery, like the corresponding vessel in the crossopteryg- 
ians, arises from the fourth efferent branchial in Ceratodus, and from the 
aortic root in Protopterus and Lepidosiren.” 
There appears to be a physiological flaw in this description unless the 
fish blood behaves quite differently from that in other animals. Under 
the assumption that the blood in the fish becomes fully oxygenated in its 
passage through the gills, the blood carried to the lungs from the efferent 
branchial artery would already be charged with oxygen, and in this case 
the lungs would only be functional when the fish is hibernating in the 
dried mud. Under the assumption that the fish blood is not fully oxy- 
genated in its passage through the gills, the lungs would be accessory to 
