67 333 



two latter lo the Solenostomidœ ' . do nut throw any light upon the question of 

 derivation from other forms, but they may prove to be of use in filling up some 

 of the gaps between the now existing genera. Leaving aside the laller point for 

 the present and only considering the existing forms reviewed above I lliink it will 

 be agreed, that the Syngnathids are the most aberrant and most strongly 

 specialised Solenichthyes. Compared with the other forms the Syngnathids have 

 acquired a more complete dermal armour and an elongated tail; at the same time 

 thev have lost Ihe first dorsal and Ihe ventrals, and reduced or lost the anal; the 

 second dorsal, taking up the function of locomotion, has generally been lengthened 

 and often encroaches upon a greater or lesser part of the trunk; the melapterj'- 

 goid is lost, the shoulder girdle has lost the postclavicle (and supracla viele), the 

 clavicle, expanding above, has been connected firmly to the anterior vertebrae, 

 while the pectoral skeleton is weakened like the branchial skeleton. Less trans- 

 formed are the Solenostomidœ which have kept the first (spinous) dorsal fin and 

 the ventrals (with a spine); the metapterygoid is lost, the shoulder girdle has lost 

 its postclavicle but is otherwise complete, while the pectoral skeleton is weakened, 

 tending greatly towards the structure found in Syngnathids, as also does the bran- 

 chial skeleton and the vertebræ in losing the articular processes. Upon the whole 

 the majority of characteristics combine the Solenostomids with the Syngnathids, 

 with which they have been placed since Cuviek established his Lophobranchii; the 

 correctness of his view in this respect may now I think be regarded as settled. 

 But on the other hand the Solenoslomidce show likenesses with different members 

 of the group Aulostomata; some of these likenesses may perhaps seem to be of 

 less importance — a matter, however, by no means easy to judge with any degree 

 of certainty. Thus, the extreme shortness of the tail in proportion to the trunk 

 and head, the lateral compression of body and head, the mandibular barbel remind 

 one of Aulosloma (whije the position of the mouth slit is more like that of Syngna- 

 thids and Centriscids). Further, there are some external likenesses with Centriscas 

 in the (dorsal) fins — as already alluded to by Regan (45 a and c) — **, in the 

 stiffness of the trunk, and, I might add, in the dermal ossifications: from a four- 

 sided shape with cruciform crests, which the large lateral plates show in the adult 

 Centriscus and all the scales in the young ones, the stellate dermal scutes in Sole- 

 nostomus — and besides also the plate-forms found in Sijngnathidœ (cfr. Hippo- 

 campus and those of young Syngnathus) — could easily be derived. 



In the group Aulostomata the family Aulostomidce approaches the group Lopho- 

 branchii {Solenostomidœ and Siingnathidæ) in the commencing reduction of the 

 branchial skeleton and in the fact that the upper parts of the rows of branchial 



■ In a forthcoming paper I liope to show that Calamostoma {breviculum Ag ) must belong to 

 the Solenostomidœ. That the genus Solenostomus itself is represented in the Mt. Boica formation, as 

 also Amphisile, I know from photographs, kindly sent to me from the Museo Civico in Verona. 



" BouLENGEK (4 c, p. 633) has also expressed the opinion that "the unique genus Solen, may be 

 regarded as in many respects intermediate between the Centriscidæ and the Syngnathidæ." 



43' 



