Pisces.] SUBANTARCTIC ISLANDS OF NEW ZEALAND. 587 



that, by giving greater publicity to these discoveries, the family Galaxiidae will no 

 longer be included among those strictly confined to fresh waters and that students 

 of the geographical distribution of animals will be furnished with a clue to a problem 

 that has so often been discussed on insufficient data." 



There is little fear that anything of importance respecting such an interesting 

 subject as the distribution of the Galaxiidae will be overlooked, and, in addition to 

 the article above referred to and quoted, several other allusions to the marine habit 

 of G. hollansi have already been published. Dr. David S. Jordan* has given the 

 greatest possible publicity to the matter by reprinting Dr. Boulenger's article in his 

 " Study of Fishes," and in his paper above mentioned Mr. Regan has also referred 

 to Hutton's statement. 



I am very far from satisfied that G. brevivinnis is to be classed as a marine 

 species, and consider that the evidence is of the weakest nature, and I am surprised 

 that it should have been so generally accepted. The pool above described was quite 

 accessible to gulls and shags, and had I found Galaxias in their digestive tracts I 

 should certainly not have been surprised. Hutton seems to have taken it for granted 

 that the merganser is essentially a marine bird ; if this is so, it differs from its European 

 allies. I have seen the goosander {Mergus merganser) in England seventy miles from 

 the sea, and Yarrellf states that these birds " frequent fresh-water lakes as well as 

 the sea-shore and estuaries." Little appears to be known of the habits of the southern 

 merganser. HuttonJ states that " it does not frequent the coast and open waters, 

 but only the sheltered harbours." I have, of course, no evidence to prove that the 

 Galaxias does not occur in the sea ; but, on the other hand, I maintain that there is 

 no evidence to show that it does. 



I am penning my conclusions, though based on negative evidence, to draw 

 attention to what is probably a misreading of a simple incident connected with the 

 natural history of the species. I cannot claim that such conclusions will throw 

 much light on the question of antarctic land connections, for so many species of 

 Galaxias are now known to enter the sea. In addition to G. attenuatus in New 

 Zealand, we have records of G. maculatus in Chili, and this species and G. gracillimus 

 in the Falkland Islands. 



Some time after the foregoing was written I met Captain Bollons, and told him 

 that I had grave doubts as to the Auckland Island Galaxias living in the sea. I 

 asked him if he did not consider it likely that the merganser had obtained the fish 

 from some fresh-water stream close to the sea. " But why close to the sea ? " said 

 Captain Bollons. " I have never yet seen the bird feeding close to the sea ; it is 

 always picking about in the creeks." I was naturally gratified to find my surmise so 

 unexpectedly supported, and we may now safely dismiss the alleged marine habit 

 of Galaxias brevipinnis (hollansi) as incorrect. 



The following is a description of one of the examples secured : — 



B. VIII ; D. IV, 8 ; A. IV, 9 ; V., 7 ; P., 12 ; C, 15 + 4. Vert., 36 + 25 = 61. 



Length of head, 5-2 ; height of body, 7-5 ; and length of caudal, 6*0 in the length ; 

 diameter of eye, 5-8 ; length of snout, 3-4, and interorbital space 2-5 in the head. 



* Jordan, " Guide to the Study of Fishes," ii, 1905, p. 205. 



t Yarrell, " British Birds," 3rd ed., iii, 1856, p. 399. 



X Hutton and Drummond, " Animals of New Zealand," 1904, p. 321. 



