38 



ANATOMY OF VERTEBRATES. 



Terminal caudal vertebra?, Sword-flsti. xxill. 



The interspaces of the neural arches -are occupied by a filjrous 

 aponeurosis— the remains of the primitive covering of the neural 



axis : but in 

 most fishes the 

 arches are ad- 

 ditionally con- 

 nected toge- 

 ther by articu- 

 lar or oblique 

 process e s 

 {zyrj apophy- 

 ses ) : in the 

 Pike the ante- 

 rior one, fig. 



32, z, is present, which barely touches the neural arch in advance ; 

 in Pohjpterus it overlaps that part. In the Perch a posterior 

 zygapophysis projects to receive the overlapping anterior one, 

 the relative positions being the reverse of those in most air- 

 breathing vertebrates. But, in some fishes, a second pair of 

 zygapophyses are developed, which resemljle the normal pair 

 in higher vertebrates in relative co-adaptation, but seem to 

 grow as exogenous processes, from the centrum itself, fig. 

 SI, z. It is also peculiar to fishes to have articular processes 

 developed from the parapophyses, as, e. g. in the abdominal region 

 of the Rays, and from the caudal vertebraj of the Sword-fish, fig. 



33, z. In the Tunny these processes are branched, and form a 

 network about the h;cmal canal. ' In Loricaria peculiar accessory 

 processes are sent out from the neural arch of the seven anterior ver- 

 tebra wliich abut against the lateral shields of the dermo-skelcton. 

 The parapophyses are short in some fis\\cs,( Salmo, Clupcca, Atnia), 

 of moderate size in many, and longest in the Cod-trilDC, fig. 34, 

 p, where they expand in the abdominal region and sustain 

 the air-bladder which adheres to their under surface. In one species 

 of Gddus, the bladder sends processes into deeper cavities of the 

 parapophyses, foreshowing, as it were, the imeumatic bones of 

 birds. The parapophyses gradually bend lower down as they 

 approach the tail, where, in many fishes, they unite to form the 

 lifcmal canal. In Lcpidosteus the canal is formed by the pleura- 

 pophyses: whilst these, in Amia, Thi/ini its, ami some others, are 

 appended to the parapophysial inverted arclies, like luvnial spines. 

 In Lepidosireu the elements p, fig. 41, which in the abdomen 

 represent either iilcurapophyses or long parapophyses, bend down 

 in the tail to form the luumal arch. Not until we reach tlie 

 Ijatrachia in the ascciisi\e comparison do we find true ' luvma- 



