998 REPORT OF COMMISSIONER OF FISH AND FISHERIES. [18] 



resultant forces being alternately reversed as often as the tail is 

 vibrated in opposite directions, will tend to throw up the end of the 

 axis of the body, and thus i^roduce a heterocercal condition, or one in 

 which the end of the notochord will at least tend to be bent upwards. 

 In case the dorsal and ventral rays of the tail are equally developed, 

 there is no disturbance or upward displacement of the notochord, as 

 may be seen in the cases of Chimccra, Frotojjterus, and Geratodus, be- 

 cause here the interacting forces, if their lines of action are traced, may 

 be shown to have a tendency to promote a preservation of the original 

 or lophocercal symmetry, in which the chordal axis is straight i^oste- 

 riorly. 



The extent of the development of the ventral lobe is highly variable, 

 but it usually presents the greatest amount of surface laterally if the 

 line of curvature of the urostyle, which is the true caudal axis, be used as 

 the line dividing the ventral and dorsal lobes, and the curvature of this 

 line is the index of the degree of heterocercality. It follows from this 

 principle also that there is a correspondence between the proportion of 

 the dorsal and ventral lobes to each other, and the degree to which the 

 extremity of the chorda or axial skeleton is bent upwards posteriorly, 

 if the prin<'ii)le of .mechanical evolution here traced has any meaning. 



An examination of a large series of types beginning with the Chon- 

 drostei, Selachii^ GrossnpterygU, Rhomhoganoidei, Cycloganoidei, and end- 

 ing with the most highly differentiated members of that series, the Tele- 

 osts, shows that the second anal or lower lobe of the caudal is not 

 serially in the same relative position in different types ; that in some 

 forms it arises near the end of the axis of the embryo, in others a long 

 distance in front of the termination of the axis of the body. Upon the 

 location of the lower lobe, therefore, depends the number of segments in 

 the dorsal or axial lobe of the tail from the point where it is flexed up- 

 wards to its termination. In AlopicH tbere are about two hundred seg- 

 ments in the long dorsal lobe of the tail, and in some Teleosts there may 

 be no vertebrte at all developed behind the point of upward flexure, as 

 in GasterosteuSj for example, where the axis of the dorsal lobe is repre- 

 sented only by the urostyle and modified posterior half of the last ver- 

 tebrae. The range of modifications which the tails of fishes undergo is, 

 therefore, largely conditioned by this variation in the position of the 

 ventral lobe, which may be placed near the end of the axis or far in* 

 front of it, and thus involve a smaller or larger number of the terminal 

 segments of the axis in the upward flexure of its terminus. 



Thepointof flexure of the chordal axis ought, if the mechanical hypoth- 

 esis here outlined is at all probable, to begin at the anterior margin of 

 the lower lobe, which is what we find to be the case. From the type m 

 which the caudal axial skeleton is not flexed, as in Gerafodus, to the 

 extremest type known, namely, Alopias^ this rule holds. The cases of 

 Anguilla and Gnidoglanis are somewhat difiicult to reconcile, yet in 

 truth their heterocercy is but slight, the caudal being reduced to eight 



