ANTARCTIC FI«HE8 OF THE SCOTTISH NATIONAL ANTARCTIC EXPEDITION. 291 



slender form, the smaller head, etc. My specimens, 53-55 mm. in total length, were of 

 the same size as the smallest examples of G. attenuatus, but, bearing in mind the extra- 

 ordinary larval history of Anguilla, Albula, etc., I wrote: "Possibly this species may 

 be based on a larval form of G. attenuatus, but if so it is remarkable that it has been 

 recorded only from South America, and that larval forms of other species have not been 

 described." A series of Galaxias attenuatus from the Falkland Islands, since received 

 from Mr Vallentin, includes specimens of 20 to 30 mm. which agree with those of 55 to 

 60 mm. in form, size of head, etc., and show pretty conclusively that G. gracilli'mus does 

 not represent a stage in the life-history of this species. Mr Vallentin's collection also 

 includes some young examples of G. smithii, hitherto known only from the type from 

 Sir Andrew Smith's collection ; these are yellowish, with numerous brownish irregular 

 vertical stripes. 



The South American species of Galaxias are seven in number, viz. : — 



1. Galaxias attenuatus, , J euyns, 1842. 

 S.E. Australia ; Tasmania ; New Zealand and neighbouring islands ; Chile ; Pata- 

 gonia ; Tierra del Fuego ; Falkland Islands. 



2. Galaxias gracillimus, Canestrini, 1864. 

 Chile ; Falkland Islands. 



3. Galaxias maculatus, Jenyns, 1842. 

 Chile ; Patagonia ; Tierra del Fuego ; Falkland Islands. 



4. Galaxias al'pinus, Jenyns, 1842. 

 Alpine lakes of Tierra del Fuego. 



5. Galaxias hullocki, Regan, 1908. 

 Ann. Mag. Nat. Hi»t. (8), i. p. 372. 



Temuco, Chile. 



6. (jralaxias platei,^t&ui^., 1897. 



Galaxias titcumhi, Everm. and Kendall, Proc. U.S. Nat. Miis., xxxi., 1907, p. 92, fii;. 

 Patagonia ; Argentina. 



7. Galaxias smithii, Regan, 1906. 

 Falkland Islands. 



It should be noted that only the marine species occur both at the Falkland Islands 

 and on the continent of South America, and there can be little doubt that Haplochiton 

 zebra, with this distribution, will prove to be marine. 



The conclusion that the Galaxiida3 are originally marine and are establishing them- 

 selves in fresh water is strengthened by their relationship to the Osmeridaj; their distribu- 

 tion has little bearing on the question of a former extension of the Antarctic Continent. 



The expense of the publication of this Memoir is defrayed from the Government 

 Publication Grant administered by the Royal Society of London. 



