( 148 ■) 



Barbus rothschildi. 

 GilDtb., Xnv. Z VIII p. 368, pi. 18, fig. u. 



Several s])eiimens from the Gum Erliiali. This .specie.s probably does not 

 attain to the same sine as B. reinii. The jiroloiigatioii of the anal rays is a sign of 

 maturity, or, at any rate, is not developed in very young specimens. The smallest 

 individual, in which I find these rays prolonged, is a male with undeveloped testicles, 

 measuring 5;J- in. Another male individual of the same size, but with the testicles 

 still less develojied, as well us all other individuals of smaller size, have short aniil 

 rays which do not rea<b the caudal tin. 



Barbus i-othschilt/i is a more slender species than Barbus frifischii, to which it 

 is most closely allied. The number of branched dorsal rays, which I stated to afford 

 an additional distinctive character, is, in fact, the same in both species — viz. nine. In 

 my original descri]ition fd' B. friixrJdi 1 was misled by the largest ftypicul) specimen, 

 in which the last ray is simjde, and not split to the base. All the other specimens 

 have nine rays, the last being double as usual. B. rothschildi has the dorsal sjjine 

 stronger than B.frifsrhii, but the difference is slight, hardly amounting to specific 

 distinctness. 



As already mentioned by Mr. Bouleuger, the two Capoetas recently described 

 by him from the Atlas {Ann. k Mag. N. H. 1902, IX p. 124) are so similar to the 

 Barbels described by me, that it must appear most desirable to obtain a more 

 complete series of these various forms than we possess at present. The jiresent 

 division of these fishes into species and genera can be taken as a provisional 

 measure ouly. 



NOTE REGARDING TESTJJDO ELEPIIANTOFUS. 



BY THE HON. WALTER ROTHSCHILD, Pu.D. 



DR. BAUR's statement that Dr. Giinther's Testtuh dephantopus was not the 

 T. clephantopas of Harlan (see antea,]}. 184) can be easily explained and 

 as easily refuted. The Academy of Natural Sciences of Philadelphia kindly lent me 

 the specimen examined by Dr. Banr, and believed to be Harlan's type. This specimen 

 was carefully examined by Dr. Giinther and myself, and dnly measured. On com- 

 paring the measurements with that <if Harlan, we found that the specimen was 

 certainly not Harlan"s, as the difference in length amounted to almost one-seventh ; 

 moreover, the specimen so kindly lent me for examination by the above institution 

 is undoubtedly a young specimen of Testudo vicina Giinth., which must have been 

 substituted at some time for Harlan's tyjie. Therefore it is evident why Dr. Baur 

 found the supposed T. elepknntopux Harlan to be dilferent from T. clephnxtojins 

 Giiuth. : it was, because he examined a substituted T. cicina, and not Harlan's 

 2\ elejihantopus at all. Although Harlan's type, and with it our sole jiroof of the 

 identity of the original 7'. elephaittopiis, seems to have disaj)i)eared, I think Dr. 

 Giinther's careful comparison of the original description with tiie specimens 

 identified by him as Testndn (dcphdiilopiis of Harlan's must convince all future 

 students that it is advisable to retain Harlan's name for these specimens. 



