DOUBLE MONSTROSITY— PARALLEL UNION 29 



which they exhibit, is that in fishes posterior duplicity is rare, and when it does occur, is always of 

 the imperfect hemididymous kind. By contrast, the other two groups both provide excellent 

 examples of posterior doubling. In this connection, it should be added that no example of posterior 

 triplicity is known among fishes, while undoubted instances of this condition have been recorded 

 among mammals (see p. 39). 



We have seen that the imperfect doubling characteristic of katadidymus in fishes arises in all 

 probability from the circumstance that inasmuch as the posterior end of the embryo remains 

 continuous with the edge of the blastoderm, the inner sides of a posteriorly dividing embryonic 

 rudiment abut on the yolk, and, accordingly, are unable to obtain the peripheral cell material 

 normally employed in building up the lateral and ventral body walls. For want of this, the two 

 diverging limbs of the divided axis are unable to complete themselves by becoming bilaterally 

 symmetrical in the full degree. 



In birds, on the other hand, the groove of the sickle early becomes obliterated, the blastoderm 

 spreading rapidly behind it. The posterior end of the primitive streak (and afterwards of the 

 growing embryonic axis) is thus removed from the margin of the blastoderm, and lies entirely within 

 the area of the latter. Should the posterior end of the primitive streak (or of the embryonic axis) 

 for any reason undergo fission or dichotomy, the new growth-systems thus produced will find their 

 inner or adjacent, as well as their outer, sides abutting equally on the various blastodermic layers. 

 Nothing will then hinder the systems in question from completing themselves so as to become 

 bilaterally symmetrical, in virtue of that power and tendency towards the realisation of the whole, 

 which, as we saw reason above to believe, may be manifested at very early stages by all embryonic 

 growth-systems should opportunity and necessity arise. 



On the whole, the reptiles will tend to resemble the fishes as regards the types of monstrosity 

 which they exhibit, since in them the primitive streak and the posterior end of the embryo maintain 

 a modified continuity, for a considerable time at least, with the margin of the blastoderm. It is true 

 that, as far as I am aware, no examples of the hemididymous condition have been recorded among 

 reptiles, but neither (with one very doubtful exception x ) are there any instances of true posterior 

 duplicity. On the other hand, anterior duplicity is far from being rare, considering the relatively 

 small numbers of reptilian ova and embryos that come under observation. 



CLASS IX. 

 Longitudinal or Parallel Union. 



My own material has not provided me with any examples of this class, and indeed the sole 

 instance on record in the Salmonidae appears to be Barbieri's (6), which has reference to an 

 egg of Salmo irideus fixed seven days after fertilisation and showing a double embryo 3 mm. in 

 length. 



There is a marked tendency for the organs of the ventral part of this embryo to show 

 duplicity in a greater degree than the dorsal ones. 



For example, the brain cavity in cross section looks like a three-rayed star. One of the rays 

 is directed dorsally, while the other two point downwards and to the sides, and represent the 

 ventrally-doubled central cavity. The spinal cord is triangular in section, and has a notochord 

 below each basal corner, while ventrally and mesially a series of united inner mesoblastic somites is 

 to be found. There are two pairs of optic vesicles, the inner ones being more or less fused together. 

 One of the twin sides is rather smaller and less well developed than the other, the mesoblastic 

 somites in particular being deficient in size. Two vesicles of Kupffer can be made out, the one 

 belonging to the smaller twin being situated a little in front of the other. In the figures there is no 

 sign of duplicity on the part of the gut. 



It will be remembered (see p. 10) that one of the types into which Lereboullet classified 



1 A young blastoderm of Lacerta ocellala showing on a single embryonal shield two primitive streaks converging gently 

 in front (Tur, quoted from Schwalbe 222, II. p. 30). 



