RELATIONSHIPS OF LUNG-FISHES 127 
siren” of dealers, often of museums. It is the best known 
of Dipnoans, on account, partly, of the ease with which it 
may be transported alive. In the hardened mud cocoons 
with which it encases itself during the dry season, it is 
readily dug out of the stream bed and packed for exporta- 
tion. When placed in tepid water, the cocoon dissolves 
and the fish shortly revives. 
Relationships 
A review of our knowledge of Dipnoans gives but little 
satisfactory suggestion as to their relations as a group. 
They must certainly be looked upon as an advancing 
phylum from which the amphibia may early have diverged. 
Their many amphibian characters have been lately em- 
phasized by W. N. Parker. On the other hand, the 
evidences of the kinship of Dipnoans to the other types 
of fishes can only be interpreted as the common con- 
vergence of the ancient phyla toward the structures of the 
ancestral form of fish. Thus we find that the types of 
Devonian lung-fishes can only be distinguished from 
those of the contemporary Teleostomes by the pattern 
of arrangement of the plates of the head roof,* a condition 
which has led Smith Woodward to believe that these 
groups had already diverged before the appearance of 
dermal bones. 
Lung-fishes have unquestionably many structures which 
may have been derived from the more generalized condi- 
tions of the sharks; and as a group they may not unrea- 
sonably be looked upon as descended from the primitive 
elasmobranchian stem. Their ties of kinship to the sharks 
* The present writer regards this distinction as somewhat provisional; 
median head plates are nominally characteristic of Dipnoans (Fig. 124), but, 
as in the sturgeons and siluroids, they are also well known among Teleostomes. 
Protopterus has, moreover, a symmetrical arrangement of the head plates. 
