294 28 



Ihe foremost "rings" of the trunk have I found it necessary to enter into details 

 below in treating of the endoskeietal parts closely connected with them. 



The arrangement of the dermal plates has been described by Kröyer (29 p. 

 683), Peters (41 p. 104), Duméril (12a p. 143, 12b p. 478), E. Moreau (36 p. 28), 

 Schaff (5ü), Lilljehokg (37 p. 437), Smitt (54 p. 675), Huot (19 p. 2Ü2), * Duncker (13 

 p. 18, p. 62), a. o.; the shape and structure of the single plates, their relation to the 

 layers of the skin etc. are especially mentioned by Schaff, an account of their 

 first development is given by Kasanzeff (24 p. 854). 



The presence of "lateral-line organs" was first shown by Duncker (13 p. 22, 

 PI. 3j; but as these organs — the existence of which I can confirm — are not en- 

 closed in any canal, and some of them may appear on places, where the typical 

 lateial line does not occur (e. g. on the operculum), it seems open to doubt if they 

 really represent the true lateral line or some of the other dermal sensory organs 

 found in teleosts. 



Endoskeleton. 



The vertebral column consists of 56 vertebrae; the anterior 19 I count as 

 abdominal, as the front outUne of the caudal armour, behind the anal opening, lies 

 just beneath the middle of the 20th vertebra ; this vertebra besides supports the anal 

 fin and sometimes its transverse processes are joined to form the foremost inferior 

 arch; but sometimes this is the case with the next vertebra. The three anterior 

 abdominal vertebrae are immovably joined together (PI. IV, fig. 4), their neural 

 arches being firmly connected basally through sutures, with long dentations, while 

 their corpora simply meet in the same way as those of all the remaining vertebrae. 

 When macerated (even in hydrate of potassium) the three anterior vertebrae, there- 

 fore, always cohere, whereas the others drop off". The immobility is strengthened 

 by the expanded clavicle, which is firmly united with the transverse processes of 

 the two foremost vertebrae. 



The first vertebra (PI. IV, fig. 3, 4) is shorter than any of the following; in 

 front it carries a pair of strong processes (a), articulated with the skull and behind 

 continuing as wing-shaped transverse processes; the outer margin of the latter is 

 embraced by folds of the anterior part of the clavicle. The spinous process is a 

 low, elongated crest. The base of the neural arch has behind a deep triangular 

 incision on each side, into which fits a long process from the next arch, corre- 

 sponding to the process of the first vertebra; in this way the stiff connection men- 

 tioned above is brought about. The second vertebra carries on its middle a stout 

 transverse process, somewhat bifid at the outer end, which is firmly bound by 

 ligament to the clavicle. The neural arch and its spine are similar to those of the 

 first vertebra. 



The transverse process of the third vertebra is longer and more slender, the 



* The small "intermedial" scutes which occur in most genera of Syngnathids, e. g. Siphonostoma, 

 Hyngiiathus, Nerophis, have been overlooked as such by Huot and mistaken for lateral-line organs. 



