202 MEMOIRS NATIONAL ACADEMY OF SCIENCES, VOL. XIII. 



Especially in the middle portion of the block rounded nodules of graphite are found, in addition to the troilite, 

 which attain a length of 8 mm. and a breadth of 6 mm., but which as a rule are of only small size. Occasionally, too, 

 graphite occurs as a scaly envelope of larger troilite nodules, while an intimate admixture of both minerals, such as is 

 not infrequently seen in other meteoric irons, is not present in this. 



The nickel-iron phosphides can be discerned with certainty only by the aid of a strong glass. Rhabdite appears 

 to be entirely wanting, and only here and there are small tin-white granules of schreibersite visible. In a few places 

 are to be seen extremely thin laminae, as much as 5 mm. in length, which appear under a strong glass merely as hair 

 lines. They would be easily overlooked were they not for the most part surrounded with a small aureole of rust. This 

 arises apparently from the fact that iron chloride appears on the borders of the leaflets, while the surrounding compact 

 nickel iron, which is free from accessory constituents, remains entirely fresh. According to the analogy of Locust 

 Grove these lamellae may be regarded as nickel-iron phosphides; although they are much finer in Forsyth, shorter and 

 scarcer, and therefore not, as there, characteristic of the meteoric iron. 



Likewise, only in isolation and upon a few places in the meteorite are to be found peculiar spindlelike forma- 

 tions, which attain a length of 2.5 mm. and a thickness in the middle of 0.25 mm. Something similar may be observed 

 in the Cape iron, where the formations are larger and sometimes sharp at tie one end only, while the other is cut 

 straight off. They resemble hemimorphic prismatic forms terminated at one end by a base, at the other by a pyramid. 

 In the case of the Cape iron, the color of the forms in reflected light suggests iron sulphide; here they are so small that 

 only a conjecture concerning their nature can be expressed. Many of the lamellae and spindles are surrounded with 

 a fine light etching zone. 



All these inclusions show a preference for the amorphous portion of the meteorite. In the granular part, graphite 

 is entirely wanting; schreibersite, which must be present from the content of phosphorus shown in the analysis, does 

 not occur macroscopically, and troilite occurs only in isolation. Most of the nodules of graphite and troilite, as in the 

 case of Locust Grove, lie in proximity to the original crust of the meteorite. 



Forsyth belongs to those meteorites which, in consequence of their content of chlorine, rust easily and abundantly. 

 The first is evidenced by the fact that a freshly etched section (especially of the granular variety) quickly acquires 

 specks of rust from lying in the air, the last is inferred from the abundant coating of rust on specimens submitted by 

 Stiirtz and from a comparison of the given weights. According to de Schweinitz, the block weighed originally about 

 22.5 kg., while Stiirtz estimated the weight after the exfoliation of rust-crustat20.300kg. Asde Schweinitz mentions 

 only a thin coating of rust, it must have formed more than 2 kg. of rust within a comparatively short period of time. 



For the purpose of ascertaining the chemical composition of the iron, pieces of the granular and the dense por- 

 tions, as widely separated as possible, were taken and submitted to exactly the same method of analysis; both pieces 

 were free from visible accessory constituents. The test of the granular portion for manganese and chromium gave a 

 negative result, and hence, in the case of the amorphous portion, was neglected. 



Analyses (Sjostrom): 



Fe Ni Co Cu C S P CI 



Granular 94.18 5.56 0.60 0.02 0.04 0.05 0.19 0.17 =100.81 



Amorphous.... 94.03 5.55 0.53 0.02 0.02 0.03 0.23 trace =100.41 



The chemical composition of both varieties, with exception of the content of chlorine, is as good as identical; 

 that the chlorine is less in the dense portion than in the granular, leads to the supposition that the latter rusts very 

 much easier than the former. 



From the above data are obtained the following mineralogical composition: Granular. Amorphous. 



Nickel-iron (kamacite) 98. 33 98. 42 



Phosphides 1. 23 1. 49 



Troilite 0. 14 0. 08 



Lawrencite 0. 30 0. 01 



100. 00 100. 00 

 The specific gravity of a section of the granular variety was determined by Leick as 7.3357, and that of the amor- 

 phous variety as 7.4954. Since these values are unusually low, the first determination was repeated, but with the 

 same result. The abnormal specific gravity may be referred to the porosity of the mass, and in fact the granular sec- 

 tion after two and a half hours developed bubbles, when it was submerged in water under the air pump to remove the 

 adhering atmospheric air. In this connection it must be remembered that some time previous a lower specific gravity 

 than had been hitherto found in meteoric iron, was obtained for Lick Creek on which a few porous places were 

 observed with the unaided eye, although this was heavier than that of the present case. 



If the accessory constituents be disregarded, the specific gravity of the nickel-iron of the granular variety is 7.3872, 

 that of the dense variety is 7.5066; in reality, the values should be somewhat higher, since the rust coating adhering 

 to the small quantity of the section could not be brought into the computation. 



Since the chemical composition of the two varieties is essentially the same, the difference in specific gravity, as 

 well as the varying resistance to the etching acid, can be occasioned only by the divergence in structure. 

 Three determinations of the chlorine in the rust coating were made as follows: 



After boiling with water 3. 55 per cent. 



After treatment with cold, dilute nitric acid 3. 48 per cent. 



After heating with dilute sulphuric acid 4. 99 per cent. 





