— 422 — 



riiriu to which it is at least very close. We have 

 available for comparison several specimens of P, 

 pusarum Starks, cotypes. This species is described 

 as having large abdominal j^ranular scales, but they 

 are the same size as those of the material Ave call 

 guacari, and a small pisarura, of the same size as 

 our small guacari Avith abdominal scales, lacks such 

 scales. The pusarum material rather consistent- 

 ly differs from our guacari, however, in a some- 

 what ampler dorsal which extends further back in 

 relation to the adipose when depressed. As here 

 understood, then, guacari is a widely distributed 

 variable forai, which may consist of a number of 

 races, or possibly species, but as far as our mate- 

 rial goes it is not possible to recognize or describe 

 any such. 



PlecosloiiHis variipielus von Ihering. 



I place here a specimen'? — inches to base 

 of caudal from the State of Bahia, another of the 

 same size, and two of five inches from the State of 

 São Paulo. These lish have the upper row of 

 scutes, and the lower one between pectoral and 

 ventral angled, scarcely keeled. The Bahia speci- 

 men has 28, the others 27 scutes. The former 

 has lost ail color, two of the others show more or 

 less veriniculate white niarks on the dorsal. There 

 is a single plate behind the supraoccipital, and 

 in the two larger specimens the posterior border of 

 the supraoccipital is aluiost straight, merely in- 

 denting the plate behind it a little in the center, a 

 condition different from that shown by allied species 



of Plecostornus. There is also a specimen of A — 

 inches from the State of Minas Geraes of which we 

 give a description, as there is some doubt if this 

 material can be properly placed in vnriiji'Ctus. 



Length to base of caudal 111 mm.; head con- 

 tained 2.9 limes in tliis measure; depth 5.5; snout 

 1.7 in head; interorbital 2.7; tirst dorsal ray 1.0; 

 pectoral spine 1.2; ventral 1.4; adipose 3.7; eye 



