214 APPENDIX. 



during the two voyages of Kotzebue : a reduced one of the 

 whole group may be seen in Krusenstern's Atlas, and in 

 Kotzebue's Second Voyage. The group consists (with the ex- 

 ception of two little islands which probably have had their 

 lagoon filled up) of a double row of 23 large and well-charac- 

 terized atolls, from the examination of which Chamisso drew 

 up his well-known account of coral formations. I include in 

 this group Gaspar-Bico, or Cornwallis Island, which is de- 

 scribed by Chamisso (Kotzebue's First Voyage, vol. iii. p. 179) 

 ' as a low sickle-formed group, with mould only on the wind- 

 ward side.' Gaspard Island is considered by some geographers 

 as a distinct island lying N.E. of the group, but it is not en- 

 tered in the chart by Krusenstern ; left uncoloured. In the 

 S.W. part of this group lies Baring Island, of which little is 

 known (see Krusenstern's Appendix, 1835, p. 149). I have 

 left it uncoloured ; but Boston Island I have coloured blue, as 

 it is described (ibid.) as consisting of 14 small islands, which, 

 no doubt, inclose a lagoon, as represented in a chart in the 

 CoquilWs Atlas. — Two islands, Aur Kawen and Gaspar Rico, 

 are written in the French chart with capital letters ; but this 

 is an error, for from the account given by Chamisso in Kotze- 

 bue's First Voyage, they are certainly low. The nature, posi- 

 tion, and even existence, of the shoals and small islands north 

 of the Marshall group are doubtful. 



New Hebrides. — Any chart, on even a small scale, of these 

 islands will show that their shores are almost without reefs, 

 presenting a remarkable contrast with those of New Caledonia 

 on the one hand, and the Fidji group on the other. Never- 

 theless, I have been assured by Mr. G. Bennett, that coral 

 grows vigorously on their shores ; as, indeed, will be further 

 shown in some of the following notices. As, therefore, these 

 islands are not encircled, and as coral grows vigorously on their 

 shores, we might almost conclude, without further evidence, 

 that they are fringed, and hence I have applied the red colour 

 with rather greater freedom than in other instances. — Matthew 's 



