diiicAi' i'ii'i:fisii. 



HCil 



that according to his opinion liad alreiuly been represented by lloN- 

 DELET (De Pise, lib. VIII, cap. Illl, p. 229) in distinct figures, but 

 combined in one chapter (as one species). The first of tliese figures 

 evidently represents the species called Ti/plile in Hei.on (iN'a^., Divers. 

 Poiss., p. 4-t6); the second (lower) figure should thus correspond to 

 the species called by LinN/KUS Si/nynatliiix urns, assuming that this 

 species occurs in tlie Mi'ilitcrriiooan, from which locality RoNDEI.KT 

 hail priicurtd his lish". .Xftcr .ViMKiu's observation: "Nnnicrus iu- 

 cisuraruni sen crcniiiniu transversaruMi, in hoc genere, oninino alten- 

 Hendus est", when we employ tlie Linniran name, we must first be 

 guided by tlu' character which LiNN.Klis gives, and which is drawn 

 from the number of thi' rings on the body. Ii(.th Fries (I. c.) and 

 SUNPEVAI.T, (1. c.) have adciiuately shown that Lin'n.hu.S otherwise con- 

 founded these two species, and cannot be rcgiirdcd as an anihorily on 

 this question. As Sundkv.m.i, also points out, L)onova.\, in his descrip- 

 tion of Si/m/ii. Ii/iililf, was the first after .Vktkdi clearly to define the 

 distinction between the two species, aiitl the first to give to the name 

 of aciis the employiiiei.t \\hich lias subsiv|uenlly and rightly bi^en 

 observed by the Kiiglish faunists. 



Tlic (ire.'it Pijiofish atttiins a, leiio-ili of tiliout half 

 a metre (Storm). From DAiiekil (Boliuslan) the Royal 

 Museum lias received through Mr. C. \. H.w.ssox .spe- 

 cimens U]) to 14i' 111111. long. 'V\u' eloiigateil, wliiji-like 

 liodv sliows a distinct lireak, cspecialh' in adult speci- 

 mens, at the l)oundar\' liet^'een the trunk and tlie tail: 

 and tlie truid< is thickest, as usual, in the females. The 

 greatest dej)tli of the Ijody, at the middle of the trunk, 

 ma\' sometimes rise to nearly 5 % of its length, liut is 

 usually only ■) or 4 % thereof, or in the females about 

 13'4 — 1.5 %, in the males about 11 — Ti % of the length 

 of the trunk''. The greatest breadth is as a rule about 

 ^ ., of the greatest depth. .\t the beginning of the tail 

 the de|jtli of tlie body in old specimens is at most 

 about ;> %, sometimes onh' 2^/^ %' of its length, and 

 from this point it decreases regularh- to the shallow 

 base of the caudal fin. The length of the trunk Itcliiiid 

 the gill-cover is ab( ,it 2S — 2G % of that of the liody; 

 and the length of the tail, including the caudal fin, 

 varies between about .59 and (jl % of that of the body. 



The plate-armour is distributed in such a manner 

 that :21 or 22, sometimes, in Aoung specimens, 20 rings 

 (including tlic ring of the shoulder-girdle) belong to 

 the trunk, and 43 — 4.5 to the tail. Here, as in the 

 other Si/iif/iiatlii, these plates are striated with trans- 

 verse ridges and grooves between them, starting from 

 the middle carina, which is finely dentated. The ujiper- 



most plates, which meet at the middle of I he litick, are 

 almost stpiarc and bent at a .slighth obtuse angle, ren- 

 dering the dorsid side Hat or slightly concave. In the 

 middle lateral row on the trunk the plates ai-' more 

 distinctly hexagonal ov even octagonal, and bent at a 

 ver\- obtuse angle. In the lowest lateral row on the 

 trunk the angU; of curvature varies according to llie 

 gri'ater or less distension of the bellw but the furiu of 

 the plates is more like that of those in the middle 

 lateral row. The middle \('iitral row also consists of 

 hexagoutd or octagouiil plates. The two hindiiiost rings 

 oil the trunk are without ventral plate, its jdace being 

 occujiied ]'\ the anal region. On account of tlie lorni 

 of the plates two rows of diamond-shaped patches, eo- 

 \cred onh with skin, are left on each side of the trunk 

 and also on its \entral side; but on the back there is 

 only one row of these rhomlis, which generally extend 

 across the whole dorsal plane, but sometimes, especially 

 in front, are indistinet, the contiguous plates being jjer- 

 fecth' rectangular. The middle lateral row of the trunk 

 ends in tlie hindmost ring of this region, as a rule 

 ol)li(jue]y lielow the beginning of the row that forms 

 behind the dor.sal fin the upper lateral margin of the 

 tail, and below the jjosterior part of the tlorsal fin (be- 

 hind the vent) ti middle lateral row belonging to the 

 tail. But sometimes the last-mentioned row wedges 

 itself into the hindmost ring on the trunk, wliicli ring- 

 may thus contain four plates on each side. The dif- 

 ference in this respect is, however, so irregular that 

 we sometimes find the former to be the case on one 

 side of the body, the latter on the other side. The 

 four sides of the tail behind the dorsal Hn are con- 

 structed in the same manner tis the dorsal jdane ot 

 the trunk; but the ventral side is broader than tlu; 

 dorsal, which is parallel to it, the difference being 

 greatest in old specimens. The marsupiiim of the males 

 extends under 24 — 26 rings. 



The length of the head measures 12 — 13'/, % of 

 that of the body. At the occiput it is only slightly 

 shallower than the beginning of the trinik and not 

 very sharply marked off therefrom. Ih^rc as in the 

 Si/iK/iiafhi in general, the trunk is furnished at the 



" Whether the Atlantic .Si/iir/iiatliKs :icus also occurs in the Mediterranean, is not yet decided with certainty. Mohk.u; had no know- 

 ledge of it from this locality, and Ca.nESTIUNI (Frtanii D'ftnlia. Pesci) does not include it among the fishes ot Italy. On the other hand, 

 MoREAU gives Si/tignalhiis ethojt, ,and Canestrini >Si/gii. tn-iiionutKS. which seem to stand in the same relation to tlie Mediterranean •Si/iii/n. 

 rubesceus, as iSyngn. rostellatns to Sijngn. arus. 



* In young specimens about l' , dm. long the greatest depth of the body may be only about 2' ., ". of its length and 10 ". of tlie 

 length of the trunk. 



'■ In young specimens still less, sinking at least to 2' . "» of the length of the body. 



