498 PROCEEDINGS OP THE AMERICAN ACADEMY 



tions are cleavage crystals, broken off probably by the impact of the 

 mass against the atmosphere. 



Secondly. That these masses show cleavages parallel to the planes 

 of all the three fundamental forms of the isometric or regular system, 

 namely, the octahedron, the cube, and the dodecahedron. 



Thirdly. That the Widmanstattian figures and Neumann lines 

 are sections of planes of crystalline growth parallel to tiie same three 

 fundamental forms of the isometric system. 



Fourthly. That on different sections of meteorites Widmanstat- 

 tian figures and Neumann lines can be exhibited in every gradation, 

 from the broadest bands to the finest markings, with no break where 

 a natural line of division can be drawn. 



Fifthly. That tlie features of the Widmanstattian figures are due 

 to the eliminations of incompatible material during the process of 

 crystallization. 



This investigation throws no new light upon the origin of mete- 

 orites, except so far as it strengthens the opinion that the process of 

 crystallization must have been extremely slow. The occurrence of 

 large masses of native iron occluding hydrogen gas, and containing 

 nickel, cobalt, phosphorus, sulphur, etc., implies a combination of 

 conditions which the spectroscope indicates as actually realized m 

 our own sun and in other suns among the fixed stars, and the most 

 probable theory seems to be that these masses were thrown off 

 from such a sun, and that they very slowly cooled, while revolving 

 in a zone of intense heat. 



In this paper we have not taken into consideration a number of 

 iron masses, whose meteoric origin has been generally accepted, which 

 show no Widmanstattian figures and not even any Neumann lines. 

 A considerable proportion of these are certainly not meteoric. In 

 the Harvard cabinet there are two specimens, labelled respectively 

 Campbell County, Tennessee, and Hominy Creek, North Carolina, 

 which are evidently nothing but cast-iron, and a third, labelled Ta- 

 rapaca Hemalga, Chili, which is probably of similar material. We 

 could find on the specimens of this class in the Harvard collection 

 no distinct evidences of crystallization; but also we could find no 

 features incompatible with that unity of structure which it has been 

 the chief object of this paper to illustrate. 



