Geology and Mineralogy. 257 



hollow, so that frora the sea the island seems to be uninhabited. 

 The central hollow is drained by filtration through the cracked 

 and porous coral rock. Kwaiawata Island, which is from one 

 and a half to two miles in diameter, showed precisely the same 

 form and structure, and in Gawa Island there is a still more 

 perfect instance of a raised atoll. The coral wall in the last 

 instance rises so abruptly to the height of about 400 feet that 

 part of it has to be climbed by ladders, and the plateau repre- 

 senting the old lagoon is nearly 100 feet below the level of the 

 edge. Iwa, another adjacent island about a mile in diameter is 

 of the same kind, only the gently sloping border has been worn 

 away, and the coral cliff meets the sea nearly all round. — Proc. 

 Roy. Geograph. Soc, June, 1892. 



2. Tlie Origin, of Igneous Rocks. — Mr. J. P. Iddings, in a 

 paper read before the Philosophical Society of Washington in 

 June last (Bull., vol. xii), has made a very important contribution 

 to science on the Origin of Igneous Rocks. After mentioning at 

 length the opinions that have been presented on the subject, he 

 considers the intimate relations of the various igneous rocks in 

 mineralogical constitution, and especially in chemical composi- 

 tion, illustrating the subject with a large array of facts and 

 tables of analyses. Further, the associations or groupings of the 

 different kinds of rocks in various regions of eruption and the 

 order of succession in origin or outflow in each region are 

 reviewed. After a careful and judicious survey of these subjects 

 in their various relations, the author presents the following con- 

 clusions. 



The differences in kinds of igneous rocks are not due to the 

 existence of two or more subterranean zones of unlike magmas, 

 or of zones of unlike rock-material which under physical changes 

 might become such magmas ; but they arise from the local differ- 

 entiation of a common magma ; and the series, in any region of 

 eruption, usually commences with a kind having the mean com- 

 position of the serios and ends with rocks of one or both 

 extremes. Mr. Iddings remarks that this law, while it has its 

 exceptions, holds for all the localities that he had personally 

 studied, and for the order of eruption described by Prof. Judd 

 for the lavas of the Lipari Island, " which began with rocks of 

 intermediate composition and has reached the stage where rhyo- 

 lite and basalt are being thrown out." Other conclusions are : 

 that the variation in the composition of the rocks of a series of 

 eruptions at any volcanic center is the result of the chemical 

 differentiation of a magma of mean or intermediate composition ; 

 that molten magmas are essentially solutions, as put forth by 

 Bunsen — a point illustrated in the experiments by Barns and 

 Iddings described on a preceding page ; that in each case those 

 portions of the magma which were the later to crystallize may 

 be considered as having been a solvent for the other portions — a 

 solvent not for the silicates necessarily but for their constituents ; 

 that the differentiation in any case is due chiefly to differences in 



A\r. Jour. Sct.— Thtrd Series, Vol. XLIV, No. 261.— September, 1892 

 17 



