THE SMITHSONIAN INSTITUTION. 159 



as follows : The rocks or minerals of these bodies are not of a sedi- 

 mentary character, nor such as are produced by the action of Avater. 

 This is obvious to any one who will examine them. A mineralogist 

 will also be struck with the thin dark-colored coating on the surface 

 of the stony meteorites ; but this is in most, if not in all instances, 

 the product of our atmosphere, and need not be further noticed .^ A 

 more interesting peculiarity is that metallic iron, alloyed with nickel 

 and cobalt, is of constant occurrence in meteorites, with but three or four 

 exceptions — in some instances constituting the entire mass, at other 

 times disseminated in fine particles through stony matter. The ex- 

 istence of this highly oxydizable mineral in its metallic conditionis 

 a positive indication of a scarcity, or total absence, of oxygen (in its 

 gaseous state, or in the form of water) in the locality from whence the 

 bmly came. 



Anotlier mineralogical character of significance is, that the stony 

 portions of the meteorites resemble the older igneous rocks, and par- 

 ticulai-ly the volcanic rocks belonging to various actije and extinct 

 volcanoes. It is useless to dwell on this fact ; the inference to be 

 drawn from it is very evident. It is highly significative of the igne- 

 ous origin of these bodies, and of an igneous action in other portions 

 of space similar to that now- existing in our volcanoes. 



Ever since the labors of Howard in 1802, the chemical constitution of 

 meteorites has attracted much attention, more especially the elements 

 associated in the metallic portion ; and although we find no new ele- 

 ments, still their association, so far as yet known, is peculiar to this 

 class of bodies. Thus nickel is a constant associate of iron in meteor- 

 ites, (if we except the Walker county, Alabama, and Oswego, New 

 York, meteorites, upon whose claims to meteoric origin there yet re- 

 mains some floubt ;) and although cobalt and copper are mentioped 

 only as occasional associates, in my examination of nearly thirty 

 known specimens (in more than one-half of which these constituents 

 were not mentioned) I have found both of the last-mentioned metals 

 as constantly as the nickel. With our more recent method of sepa- 

 rating cobalt from nickel, very accurate and precis-e results can be 

 obtained relative to the former. The copper exists always in such 

 minute proportion, that the most careful manipulation is required to 

 separate it. 



Another element frequently, but not always, occurring in associa- 

 tion with the iron is phosphorus. Here again an examination of thirty 

 specimens of this substance leads me to a similar generalization, namely, 

 that no meteoric iron is to be expected without it ; my examination 

 has extended as well to the metallic particles separated from the stony 

 meteorites as to the meteoric irons proper. It may be even further 

 stated, that, in most instances, the phosphorus was traceable directly 

 to the mineral Schreibersite. 



These four elements, then, (iron, nickel, cobalt, and phosphorus,) 

 I consider remarkably constant ingredients : first, in the meteoric 

 irons proper ; and secondly, in the metallic particles of the stony me- 

 teorites ; there ^eing only some three or four meteorites, among hun- 

 dreds that are kaown, in which they are not recognised. 



As regards the c6mbinatioa of these elements, it is worthy of re- 



