336 EXPEDITION' ()E THE " ALBATROSS," l,S99-l!l()(). 



c'luiractfr peculiar to the youiifi-, and if a constant coloi- character is of anv speci- 

 fic value this should he considered a distinct species and not the youno; of T. 

 intmacidafus. The disappearing black stripe on Jordan & Seale's Negros s])eci- 

 niens may he due to the action of the ]M-eservative rather than to age. 



Canthigaster solandri (Richardson). 



Jordan it Seale, Bull. V. S. Bur. Fish., 191)0, 25, p. 371. 



Tetraodon solandri. Richard-son, Zool, Voy. Sulphur. Ichth., l!S4.'i, p. 12.), j)!. .57, fi}j. 4-(). 



Nos. 05933, a specimen 3 inches long from Bora Bora, Society Islands, and 

 05823, M. C. Z. 294S7, a specimen 2| inches long from Makemo, Paumotu 

 Islands. 



These specimens are badly faded, but they agree very well with a specimen 

 fi-oin Samoa identified by Jordan & Seale, and with the descriptions of ('. sulamlri. 



Canthigaster constellatus, sj). nov. 

 Platf 7, fifiiirc 2. 



Type No. 65767 U. S. N. AL, 2 inches long; cotype No. 29396, AI. C. Z. 



No. 05922, nine specimens if to 2 inches long from Fakai'ava, Paumotu 

 Islands. These specimens are immature and fadetl and therefore difficult of 

 identification. But, so far as the color can be determined, they agree with the 

 figure of Tetrodon ocellatus Bennett in Fishes of Ceylon, pi. 21, in having 

 apparently the same sort of subdorsal ocellus and the same sort of lines about 

 the eye and snout and back, especially between the dorsal and caudal; the longi- 

 tudinal row of blue spots in Bennett's figure is indicated by a similar row of 

 white spots in the present specimens; our specimens show in addition irregular 

 dusky blotches, sometimes coalescing, forming a sort of band, sometimes barely 

 visible, extending from just above the gill opening to caudal; the lower edge 

 of these dark blotches sharply define the white ventral portion of the fish; in 

 one or two of our specimens the largest blotch, which is situated ,iust anterior 

 to a vertical line from the subdorsal ocellus, extends a short distance on the side 

 of the belly as a faint dusky bar and upward toward the back ; in several there 

 are faint traces of three blotches between the snout and the pectoral, the first 

 situated some distance back of the angle of the mouth, the second below the 



