ADDITIONAL COMMF.Nl S ON THE GF.OI.OGY OF IHF JIAN FERNANDEZ ISLANDS 6r 



active x'olatile phases at tlie time of eruption" (Linnean Soc. of X. S. \\'.. I'roc. 51, 

 1926, p. 617). 



Finally 1-J)\\.\KI)S in his paper 'The formation of iddin^site' says: "It is 

 concUuletl, tiierefore, that iddingsite forms durin<^ or after extrusion, accordinj^ 

 to the temperature of the magma at the time of extrusion, and that, if the magma 

 has cooled sufficiently before extrusion, not enough iron oxides are left in the 

 residual volatiles for iddingsite to be formed" (Am. Min. 23, 1938, p. 281). 



The formula for iddingsite is gi\cn b\- Ross and Shannon as Mg( J • I-'eg^^a " 

 3 Si(^2'4 HgO with magnesia replaced in part by CaO, approximately in the ratio 

 I : 4. The calculated composition according to this formula includes 7.90 "o HzO' 

 ami 7.90% HgO" (ibid. p. 17). 



All the authors, now quoted, come to the same conclusion that the forma- 

 tion of iddingsite cannot be regarded as the result of normal weathering processes 

 but that the mineral must be of deuteric origin, formed during the later stages 

 of a cooling magma. The de\elopment of abundant iddingsite in the lava flow 

 from Masafuera leads to the same conclusion. I may also remark that most of 

 the photomicrographs, representing different stages of alteration from olivine to 

 iddingsite, published in PI. I — II in Ross and Shannon's paper, correspond in 

 nearly ever\- detail with equivalent alterations in the lava from Masafuera (Fig. 

 18). As iddingsite according to the formula of Ross and Shannon contains 

 15.80% H2O, we may conclude that an appreciable amount of the 3.46 °o H2(), 

 according to the anah\sis of the la\"a in question (Table I, \o. 7), enters the 

 iddingsite molecule. 



Another singular alteration in the iddingsite basalt in question is also worthy 

 of notice. The slender labradorite laths in the groundmass as well as some 

 phenocrysts of the same mineral are found to be more or less completely 

 altered to an isotropic substance, though retaining their crystal habit perfectly 

 intact. I had observed this circumstance in my former publication on the Geology 

 of the Juan Fernandez Islands (12, p. 278] and then came to the conclusion that 

 the invading material was glass, formed through fusion of the feldspars in the 

 volcanic vent (Fig. 19). 



Dr O'lTo Mellls has kindl}- undertaken a renewed investigation of this 

 isotropic alteration product. At first he also came to the conclusion that it con- 

 sisted of glass, formed b\- fusion of the feldspars during a period of high-temperature 

 autometamorphism. A more detailed study, however, led to dift'erent conclusions. 

 Dr Mellis gives the result as follows: "Repeated determinations showed that the 

 refractive index of the isotropic material lies between 1.49 and 1.50. These values 

 must exclude the presence of glass, formed by fusion of a labradoritic feldspar. 

 The index of refraction should according to Franco and Schairer in such a case 

 lie between 1.53 and 1.54 (Journ. of Geology, Vol. 59, 195 1, p. 266). A further 

 evidence in this respect is that, when treated with HCl, the isotropic matter 

 readil}- gelatinized, which is not in accordance with what one can assume to be 

 the case with glass, formed by fusion of a basic plagioclase. Opal must for the 

 same reason be excluded as a possible component. Continued investigations 

 resulted in the conclusion that the alteration of the feldspars in the iddingsite 



