334 Mr. C. T. Regan on the 



however, are reduced in number to 18, owing to the fusion of 

 the first 3 and the next 2. The ventral fin is supported by a 

 series of 12 baseosts, exactly similar to those of the anal, 

 which also articulate with a series of axonosts, which in this 

 case are 8 in number, owing- to the fusion of the anterior 5. 

 In both anal and ventral the cartilages of the "baseost" 

 series, or radials, show a tendency to segment into 3, thus 

 forming proximal, median, and distal series of segments, 

 whilst external to the last, and completely overlapped by the 

 dermal rays, are a series of short " marginal " cartilages. 

 In the specimen described the anal fin is 23 mm. in length 

 and is composed of 70 dermal rays supported by 21 baseosts, 

 whilst the ventral is 11 ram. in length and is composed of 

 38 rays supported by 12 baseosts, a proportionate correspond- 

 ence sufficiently close to be remarkable. 



I would submit, then, that the extremely similar structure 

 of the anal and ventral fins in Psephurus can only be 

 explained on the theory of a directly similar origin, and that 

 the theory that the structure of the anal is primitive, whilst 

 that of the ventral is derived in some way from a biserial 

 archipterygium, is fantastic and entirely unsupported by 

 evidence. Thus, in an actual living species we have clearer 

 and more complete evidence of the similar origin of the 

 median and paired fins than in the extinct Gladodus^ which 

 has been considered so important. 



The pectoral fin of Psepliurus is more specialized than the 

 ventral ; the baseosts are 7 in number, the anterior 3 being 

 attached to the large coraco-scapular cartilage, which repre- 

 sents the iused anterior axonosts and which underlies a 

 membrane-bone, the cleithrum. The posterior axonosts are 

 also fused to form a single cartilage, the metapterygium. In 

 other living Chondrostei the pectoral fin is very similar to 

 that of Faej^hurus, whilst in the ventral fin the extent of fusion 

 of the axonosts and the number of the baseosts show some 

 variation. In the Palaioniscida?, so far the earliest and most 

 generalized Chondrostei known, the ventral fins often had an 

 extended * base and were composed of numerous rays. In 

 one genus, the Liassic Coccolepis, a series of baseosts have been 

 discovered. The axonosts have not so far been distinguished, 

 but there is every justification for believing that in this 

 generalized family tins so similar to those of Psephurus 

 had their supports arranged in the same primitive manner. 

 As regards the pectorals, the coraco-scapular cartilage with 



* It is interesting to note that in the Devonian genus Cheirolepis the 



veutial tin is lonp;er than the aual. 



