LARVAL OR SWIMMING STAGES 27 



The anterior end is first manifested by the earliest swimming movements of the 

 larva; it is the end which precedes in locomotion, the end which bears the prototroch. 

 The opposite or posterior end is often somewhat narrow, but not necessarily pointed or 

 bent. I do not believe the shape is sufficiently set and constant to mark out what is 

 dorsal or ventral, left or right. It requires a great deal of preliminary observation on 

 nvunerous specimens to decide what form is normal and what is not, for there are many 

 deformities. It seems to me that at this period a normal type is almost symmetrical 

 about the longitudinal axis, which extends from the centre of the prototroch to the 

 posterior end. Organs on the longitudinal axis can not serve to determine the dorsal 

 or ventral surface, but those like the blastopore or shell can, since they occur on but 

 one surface. 



It is usual and natural to regard the surface upon which the blastopore opens as 

 being ventral, and with that the other surfaces become fixed. Brooks believed that the 

 shell originates on the same side as the blastopore, and, as the shell is imdoubtedly 

 dorsal, then the blastopore must be dorsal. In assigning the shell to the 

 side of the blastopore there is no doubt that he made a mistake, and it very likely 

 came about through his not being aware that a shell-gland is formed by invagination, 

 similarly, but on a smaller scale, to the archenteron. In the larva at the stage of his 

 figure 32 he knew the blastopore, but did not recognize the depression opposite to it as 

 possibly the beginning of the shell-gland. His fig. 36, which exhibits the first appear- 

 ance of the shell, he of course oriented by comparison with fig. 32, taking the trans- 

 verse depression to be the same as g, whereas it is most probably the same as the de- 

 pression on the other side, i.e., the shell-gland. The blastopore in fig. 36 being closed, 

 as he believed, can easily have lost its transverse furrow, while the shell-gland was at 

 its best (compare Horst's fig. 11). The ends of the two transverse furrows (blastopore 

 and shell-gland) approach towards the same point on both sides (Brooks' fig. 32; Horst's 

 figs. 11, 12), so that the minute separate valves might easily have been judged connected 

 with the wrong groove. That this is Brooks' own more modern view is shown by his 

 work of 1905, Plate IV, fig. 7, where he reproduces fig. 32 (or one like it turned end 

 for end) and marks g of the original as St in the new, and the figured but un-named de- 

 pression opposite g is now designated Sg. 



Horst says: "If one, however, compares fig. 32 of his work with my figs. 9, 10, 

 and 12, I believe that it must be admitted as highly probable, that what Brooks has 

 taken for the blastopore is nothing but the opening of the preconchylian gland." 

 Brooks did not know this gland, Horst did, and yet Horst thinks that g in fig. 32 is the 

 shell-gland. Now fig. 33 (reproduced in my fig. 23) is an optical section of fig. 32, 

 and if g is the preconchylian gland instead of the gastrula-mouth then where is the 

 archenteron, which ontogenetically as well as phylogenetically should precede the pre- 

 conchj^lian gland and would in this stage be present with it? (The condition in the 

 peculiarly specialized fresh-water Unionidse, where the large shell-gland develops before 

 the small coeleuteron, is doubtless secondary and does not hold for the more primitive 

 marine Lamellibranchia) . 



We have to consider illustrations as well as statements. When the former are 

 pictures of actual natural objects they are more likely to be trustworthy than the latter, 

 but when they are merely diagrams to illustrate the thoughts they are just as likely to 

 be modified by theory. Fig. 32 is the picture of a natural object; whether it was a 

 fortunate selection from among millions is open to question. Fig. 33 may be partly dia- 

 grammatic, illustrating what the author believed he saw, but both possible and prob- 

 able. Brooks believed g to be the blastopore and that the shell-valves originated in 

 relation to it. He did not discuss the depression opposit; to it, but, if he thought about 

 it at all, may have considered it the beginning of the future mouth. Horst, in inter- 

 preting the figure, believed that g was the shell-gland, and did not discuss the position 

 of the blastopore or of > he future mouth, but may have considered them to be represented 

 in the opposite depression. If so, fig. 33 proves that it cannot be the blastopore since 

 it is not connected with an archenteron, and similarly it is not the future mouth as there 

 is no e.c.osed archenteron into which it can open. 



My own view is that Brooks' figures are correct representations of the object, but 

 that both he and Horst misinterpreted them in that both thought the shell originated 

 at g in figs. 32-35. The whole difficulty seems to turn upon Brooks' fig. 36, which 

 doubtless should be inverted, when it would fall naturally between 32 and 37, and m of 

 37 would be the definitive mouth, in the position of the original blastopore, correspond- 

 ing to g of figs. 32-35, but closed even to the loss of its transverse furrow in 36, where 

 its position would be on the side opposite to the shell and transverse furrow of the 

 shell-gland. This perhaps accounts for the posterior end (anal papilla) being bent in a 

 direction opposite to what it is in the preceding figures. According to this 



