1903,] MONSTROSITIES IN FISHES. 17 



place close to the anus. Yentral convergence of the sagittal planes 

 of the twin bodies is always marked, and is often extreme enough 

 to allow the notochords to unite anteriorly to the place of union 

 of the spinal cords and of the dorsal edge membranes. The 

 notochords are always double opposite the vent and for a consider- 

 able distance behind it, but I have not met with any cases in 

 which they remained sepai-ate to their extreme tips. The dorsal 

 aoi-t88 unite at the same level as the notochords, and the caudal 

 veins a little in advance of the aortse. On the other hand, the 

 spinal cords, while they usually unite far back in the caudal 

 region, in extreme cases may remain separate altogether. 



The ventral edge membranes tend to u.nite earlier than those 

 of the dorsal edge, but in cases of piu'e ventral union there may 

 be two composite ventral edge membranes at opposite corners of 

 the quadrangular composite body, the other corners of which 

 caiTy normal dorsal edge membranes. Apparently, during the 

 concrescence of the twin bodies, the blastema for the ventral edge 

 of each was kept apart in two halves, each of which, continuing 

 to occupy a lateral position, met and fused with a corresponding- 

 part derived from the other twin. 



In all of my specimens except one (PI. III. fig. 26) the vent is 

 single, and in all of them the intestines are united close to the 

 vent. Franz Schmitt, however (t. c. p. 53) refers to a case in 

 which, apparently, no such union had taken place. 



Very great variation is found in the arrangement of the ureters, 

 bladders, and urinary pores in this gToup. All my specimens have 

 two bladders, which sometimes communicate with one another 

 and sometimes are quite separate. In all cases, the right ureter 

 of one twin and the left lu-eter of the other open into one of the 

 bladders, while the two remaining ureters go to the second 

 bladder. Thus, each bladder receives a right and a left ureter 

 derived from different embryos, and, except in cases of symmetrical 

 ventral union, the ureters which go to the one bladder may be 

 recognised as the inner or adjacent pair, while those which go to 

 the other may be recognised as the outer pair. In such cases, 

 the first bladder lies anterior and ventral to the second, with 

 which also it frequently communicates, especially when destitute 

 of an extei^nal opening. 



Attention may be drawn to figs. 23, 24, 25, 26, on Plate III., 

 which are reconstruction diagrams illustrating the principal 

 variations referred to above. Figure 23 was drawn from a 

 specimen in which the bladder (-SX') in connection with the inner 

 or adjacent pair of Wolffian ducts has no luminary pore, bvit opens 

 into the bladder {£L) in connection with the outer pair of 

 "Wolffian ducts {WDa, WBb). Figure 24 illustrates a case in 

 which the first bladder {BL') had no opening and was enor- 

 mously expanded, as were also the lower ends of its ureters. 

 Figure 25 is from a specimen in which the bladders {BL' and 

 JBL) were separate and had virinary pores which opened in the 

 mid-ventral line, one behind the other. It will be seen that 



Proc. Zool. Soc— 1903, Yol, II, No. II. 2 



