616 MR. 1{, II. WIIITEHOU.SE OX THE [Apr. 5, 



conl can be seen to extend iilong tlie wliole leiigtli of the last 

 liypuial. 



Cartilage enters very considerably into tliis caudal tin. Both 

 the anterior dorsal and anterior ventral fin-rays are supported 

 by cartilage which is somewhat iri-egularly disposed, and which 

 extends inwards even to the vertebral axis. It is the irregular 

 nature of these cai-tilages that has led nie to attach no importance 

 to them as homologues of ordinary supporting elements such as 

 i-adials, but merely to regard them, both here and frequently 

 elsewhere, as convenient secondaiy develoj)ments to support the 

 less important dermotrichia. 



I have been fortunate in obtaining a faii'ly good series of 

 developmental stages of Gohias, and the larvaj show exceedingly 

 Avell how a condition in which the urostyle is veiy much I'educed 

 in the adult, is preceded by stages in which this structui'e is very 

 prominent ; and how, by a secondary encioachment of the last 

 h\']5ural bone, the urostyle is incorporated with this hypural. 



The caudal fin of Gobius paganellus is a type of considerably 

 advanced homocercy. 



ECHENEIS NAUCRATES. (DisCOCejiliali.) 



A long well-developed urostyle is present, attached to which, 

 together with the last centrxim, are six hypural bones closely 

 fitting together. A well-developed neural arch is closely applied 

 to the doi-sal edge of the hist A'ertebral segment ; the penultimate 

 vertebra, possesses a hypural and an epural, and immediately pos- 

 terior to the latter are three dorsal caudal radials, slendeily built. 

 Tliis fin must therefore be considered a lowly specialized homo- 

 cereal caudal. 



ScoRr.ENA SCROFA. (Scleroparei, Scorpaenidie.) (Plate L, 

 fig. 27.) 



The veit('))ral axis ends in a well-marked and free urostyle. 

 Tluee undoubted hypural bones aie attached to the last vertebral 

 segment, while in a. direct line with the axis of the urostyle is a 

 slender lay-bearing bone which may be considered a hypui-al or 

 a radial. Between this bone and the epural (p/*.) attached to 

 the antepenidtimate vertebra are three dorsal caudal radials 

 ((Z.c.r.). The iieinul arch of the jienultimate vertebra is reduced ; 

 there is also a slender bone {n.a.^) immediately above, and 

 closely ap]iroxima.ted to the urostAle, which protects the delicate 

 ending of the spinal cord ; it is probably, therefore, a neural arch 

 of some vertebra now absorbed into the urostyle. Both the 

 secouil and third la.st vertebrie bear a strong liypural bone. This 

 caudal fin is an excellent example of a form inteiniediate between 

 a low)}' and a liigldy s|icciali/.c(l type. 



TiiicLA i.iNKATA. (Scleroparei, Triglida\) (Pl.ite L. fig. 28.) 



Tlic: urostyle here is practically non-existent, and traces of it 

 can be seen only after the tail has been treated with a clearing 



