GIBSON— CEPHALOCHORDA : " AMPHIOXIDES." 219 



The ratio of length to breadth and that of total to post-anal length both dimiuish 

 with age, within the period of growth represented, and show characteristic differences in 

 the two species. The youngest animal obtained, measuring 3*9 mm., showed extreme 

 juvenility in the position of its anus, wliich lay close to the end of the tail. 



Mesenchymatous Elements of the Fins. 



In this category may be included : — 



(1) The radiating cell-strands of the tail and the forwardly directed strands of the 

 dorsal and ventral fins. 



(2) The so-called " lymph-canals " of the rostrum. 



(3) Connective tissue, described by Goldschmidt as existing in the form of cellular 

 lamellae in the tail, rostrum, and cheeks. 



In reality, the first two of these elements almost certainly form parts of a single tissue — 

 that which Goldschmidt (1908) has since described in detail, for the adult Branchiostoma 

 lanceolatum, under the name of " Haut-skelett." Tliis consists of an anterior and a 

 posterior system of anastomosing cell-strands, presenting arrangements adapted to 

 mechanical needs, which clearly indicate a supporting function. Most previous authors 

 regarded it as a lymphatic system: notably Andrews (1893) described i^ ^^ such in 

 Asymmetron. Van'Wijhe(i902),however,adopted the view since followed by Goldschmidt, 

 and both authors agree in denying the existence of any connection between the irregular 

 spaces found in the strands and other cavities : Andrews, on the other hand, regarding 

 the strands as lymph-canals, stated that they open into tiie dorsal fin-i-ay boxes. 



(1) Goldschmidt described these strands as solid, and originally separated them on 

 that account from the so-called lymph-canals; but he has since recognized both as 

 belonging to the " Haut-skelett." My sections show apparent cavities in the cell-strands, 

 which, however, I do not doubt to be artefact, since epithelium can be seen on one side 

 only. 



(2) Goldschmidt describes these as fine backwardly-directed canals springing from the 

 "dorsal rostral canal" — itself a space overlying the rostral extension of the notochord 

 (separated from it by the " subdorsal rostral canal "), and connected posteriorly with the 

 anterior extension of the fin-ray boxes and, on the right side only, with the cavity of the 

 first myotome. 



If this account is a true one, the lymph-canals would seem to provide an exception to 

 the rule that the cavities of the skeletal tissue do not communicate with other cavities. 

 I cannot, however, corroborate the existence of the connections described, and I believe 

 .that the rule still holds good so far as Amphioxides is concerned. 



The arrangement of the lymph-canals is seen in figs. 3 and 4. Trom sections they 

 certainly appear — particularly in A. vuldivm — to possess cavities continuous with that 

 of the " dorsal rostral canal." The latter, however, I cannot find to be connected either 

 with the system of fin-ray boxes or with the cavity of the first myotome. In. A. valdivia; 

 it is seen clearly enough as a laterally flattened space lymg above the subdorsal canal. 

 Just in front of the brain, as we pass backwards, it becomes constricted into two cavities. 



