May 30, 1902.] 



SCIENCE. 



871 



specimens, six (Nos. 16, 19, 20, 22, 23 and 25) 

 display paramorphie constitution, the amount 

 of marcasite varying from 3 to 74 per cent.; 

 this, too, notwithstanding that the intermix- 

 ture is invisible, and that it presents the 

 color and sometimes the crystalline form of 

 one or the other mineral. The possibility of 

 purity in some samples of pyrite had been 

 shown in 12 specimens of my paper (page 204) 

 and is confirmed by 5 out of 6 in another 

 series of Dr. Stokes's (page 39) ; but even there 

 the existence of paramorphism is proved by 

 the last, No. 33, which, with at least the white 

 color of marcasite, is found by Dr. Stokes to 

 contain 83 per cent, of pyrite. Abundant evi- 

 dence appears in both papers to prove the 

 variation in intermixture in specimens of both 

 minerals; the visibility of the impurity, when 

 in the form of mechanical intermixture, as 

 granules, films, etc.; and its invisibility when 

 present, even in large amount, in molecular 

 diffusion. 



2. It is Dr. Stokes's view 'that the density 

 affords no criterion of the composition,' in 

 opposition to my statement that 'the latent 

 constitution of these composite minerals is in- 

 dicated by variation in density, exactly pro- 

 portionate in most cases to the amount of 

 each constituent.' 



aside from the influence of their unquestion- 

 able intermixture. With the new evidence 

 from Dr. Stokes's experiments, I am the more 

 inclined to believe, at least, that these varia- 

 tions in density are caused by so many and 

 such irregular conditions that they may not 

 afford us reliable evidence as to the propor- 

 tion of one, the marcasite impurity, in those 

 specimens whose unusually low density, below 

 4.88, may indicate the influence of enclosures, 

 cavities or abnormal composition. But the 

 table of Dr. Stokes (page 39) has no bearing 

 on this conclusion, since in specimen No. 28 

 the composition was not deduced from density, 

 but, as in two others, from the abundant tar- 

 nish (page 204 of my paper), and Nos. 29 to 

 32 from abnormally low densities, 4.856 to 

 4.791 ; these were below that of marcasite, 4.88, 

 as shown by Dr. Stokes, and therefore plainly 

 due to enclosures of quartz, etc. The same 

 objection disposes of the relevancy of the last 

 four numbers in his table on page 13. All 

 that is of present interest to science is the 

 question whether there is any close relation 

 between the true densities of the two minerals 

 and those of their native specimens. Taking, 

 then, from his tables all the specimens of both 

 minerals of known densities which lie between 

 these datum-points, and for which Dr. Stokes 



Pyrite .... 



Marcasite . 



Pyrite . . . . 

 Marcasite . 



Locality. 



Col., Conn, and Utah . . . 



Dover Cliffs, Eng 



Galena, 111 



Linden Mine, Wis 



Galena, 111 



Hazel Green, Wis 



Weardale, Eng 



Webb City, Mo 



Crow Branch Mine, Wis 



Galena, 111 



Cumberland, Eng 



Density. 



018 

 023 

 041 

 881 

 891 

 901 



887 

 891 

 015 

 987 



Marcasite 

 Percentage. 



Stokes. Julien. 



Further study of etched surfaces, described 

 in my paper, had already led me, during the 

 fourteen years since its publication, to recog- 

 nize other conditions of structure and of com- 

 position which affect the density of pyrites. 



has ascertained the constitution by his chem- 

 ical method, the following table presents the 

 entire information on the subject now avail- 

 able. In the last column I have added the 

 proportion of marcasite founded on density. 



