THE COMPOSITION OF KULAITE 617 



rock orthoclase is associated with the lower Si0 3 , as well as with 

 a bytownite, while in the other leucite* occurs with higher Si0 2 

 and lower K 3 0, and with the plagioclase an andesine. In the 

 latter also the albite molecule is higher and nepheline corre- 

 spondingly lower, while the other constituents (except apatite) 

 remain the same. 



This is peculiar, but there seems to be no doubt of the facts, 

 and the difference is possibly to be connected with differences 

 in the conditions of solidification. The specimen of I came from 

 a depth of thirty-five meters in a lava flow, possibly of Kula 

 Devit, but more probably of an earlier volcano, while the leucite- 

 kulaite was from the surface of a late flow of Kula Devit. 



There would thus have been certain, even if slight, differ- 

 ences in pressure and rate of cooling, which might account for 

 the difference in mineralogical composition. It has already 

 been pointed out elsewhere 1 that these intermediate latitic 

 magmas seem to be in a nicely balanced state of chemical equi- 

 librium which only need very slight differences in conditions of 

 solidification to result in quite diverse mineralogical aggregates. 

 The facts here are also in accordance with the general observa- 

 tion that leucite is essentially a mineral of the effusive rocks, 

 while orthoclase occurs in either intrusive or effusive masses, 

 this tendency toward the formation of leucite in flows in contra- 

 distinction to the formation of orthoclase in domal eruptions 

 being especially well seen along the main line of Italian volca- 

 noes. 2 



We have already seen that the analyses correspond most 

 nearly with those of the tephrites. Accepting, however, the 

 general definition of a tephrite as an effusive rock composed 

 essentially of plagioclase and nepheline or leucite, with pyrox- 

 ene or other ferromagnesian components, it is very clear that 

 the rock represented by analysis I, which contains 23 per cent, 

 of orthoclase, cannot properly be put in this group. Indeed, 



1 H. S. Washington : Jour. Geol, Vol. V, p. 376, 1897. 

 F. L. Ransome: Amer. Jour. Sci., Vol. V, p. 370, 1898. 

 z Cf. Jour. Geol., Vol. V, p. 376, 1897. 



