Accessory Element in Rocks. 309 



Ities* have not been observed in any of the washings from 

 material obtained in situ. In one case the octahedral edges 

 are modified by faces of the diametric pyramid 1 -i not cited in 

 any of the descriptions of the mineral that are at hand for con- 

 sultation. At one locality the characteristic intergrowth with 

 zircon occurs and first led to a suspicion that certain minute 

 octahedral grains apparently too small for any attempt at miner- 

 al ogical identification might prove to be xenotime, a suspicion 

 that has been fully confirmed by microchemical and optical 

 tests, the last of which were kindly made by Dr. Eugene Hussak. 

 The crystals rarely attain a length of l mm , the usual size 

 being about J mm . In a few localities they are perfectly trans- 

 parent, yellow or colorless, with highly polished faces. Gener- 

 ally, however, the faces are roughened through alteration, which, 

 renders the whole crystal, with the exception of the extreme 

 edges, milk-white and more or less opaque. Owing to their 

 flattened pyramidal form they generally rest in the microscopic 

 preparation in such a position as to readily give the interfer- 

 ence cross in convergent polarized light and permit the deter- 

 mination of the positive character of the double refraction and 

 thus prevent confusion with anatase, a frequent companion 

 difficultly distinguishable in some of its forms from xenotime. 

 Samples from 21 different localities have been tested, of which 

 2 are in Ceara, 7 in Rio de Janeiro, 6 in Sao Paulo, 5 in Minas 

 Geraes, and 1 in Rio Grande do Sul. In this enumeration a 

 number of small dykes, all affording xenotime, in Sao Paulo 

 are excluded since they are presumed to be apophyses of a 

 large boss already included. Of these 21 localities 14, or 66% 

 per cent, gave xenotime always accompanied by zircon and in 

 all but three cases by monazite. Of the 7 which gave no 

 xenotine, 3 afforded such insignificant residues (even the omni- 

 present zircon failing in one) that a larger amount of material 

 should be tested before a positive conclusion is reached, while 

 two others are probably not true muscovite granites. One of 

 the latter is a small dyke near the Tingua mass of nepheline 

 syenite, from which it is possibly an apophesis and differs from 

 any authentic granite residue yet examined in the character of 

 its zircons, among which geniculated twins occur. The other 

 is a granitic mass in which the muscovite may be secondary 

 after biotite. Eliminating these and one xenotime-bearing 

 rock which appears to be a biotite gneiss, the proportion stands 

 as 15 to 13. That is to say, 86| per cent of the undoubted 

 muscovite granites that have been sufficiently tested have 

 afforded xenotime, which is thus as constant as any other acces- 

 sory except zircon. In the biotite gneiss above referred to 



* See Gorceix, Sur la xenotime de Minas Geraes (Bresilj, C. R., cii, 1886, pp." 

 1024-1026. 



