METEORIC AND ARTIFICIAL NICKEL-IRON ALLOYS. 
77 
undertaken on account of the comparative ease with which it could be performed. 
The average of the percentages of nickel obtained in this mechanically isolated taenite 
is approximately 27—almost identical with that obtained in the Cosby Creek taenite, 
which was probably also separated mechanically. It is remarkable that this 
percentage should agree so nearly with that suggested by thermomagnetic data to be 
contained by the eutectic ; but the interpretation from the present point of view 
is obvious. 
The chemical data can thus be explained simply if it is assumed that octahedral 
iron consists of two nickel-iron alloys, one (kamacite) nickel-poor but homogeneous 
and easily soluble in very dilute hydrochloric acid, the other (taenite) nickel-rich but 
heterogeneous and consisting of alloys nickel-richer and nickel-poorer, of which the 
poorer is much the more soluble in the dilute acid. 
§ 18. In the classification of the constituents of octahedral iron, proposed originally 
by Pi-eichenbach, it is usual to include a third constitutent, “ plessite.” It is 
distinguished by its mode of occurrence ; but, on account of the difficulty of isolation, 
no complete analysis has been made. Unless .the meteorite is comparatively rich 
in nickel, plessite is not abundant; in many octahedral meteorites it is scarce, and in 
some it is apparently absent (cf. Cohen, loc. cit. II., p, 219). 
This section of the paper in its original form included a discussion of the nature of 
plessite and of the point of view advocated by Osmond that plessite is the eutectic ; 
but this discussion, to some extent a digression and involving also the question of 
nomenclature, is omitted from considerations of space. 
§ 19. In discussing the question of the composition of meteoric iron from the point 
of view of the theory of solution, Osmond suggests that the curve showing the 
temperature of disappearance of magnetism in the irreversible alloys should be 
regarded as the true equilibrium curve marking the beginning of crystallisation 
(accompanied by the appearance of magnetism) in these alloys. On the assumption 
that the irreversibility would disappear if the cooling were excessively slow, and that 
the rate of cooling of the meteoric iron may have been such as to satisfy this 
condition for reversibility, Osmond infers that the eutectic temperature could be as 
high as 360° C., and also that the compositions of the constituents of the eutectic 
could be explained from the curves (cf fig. 23, p. 67) if it appeared subsequently 
(from chemical or other evidence) that the eutectic found in meteorites contained as 
much as 40 per cent, of nickel and was composed of alloys containing possibly 5 per 
cent, and 50 per cent, of nickel respectively. 
The work on carbon-iron alloys above referred to (Arnold and McWilliam, loc. cit.) 
shows incidentally that although slow cooling may be essential to the production of 
large-scale Widmanstatten figures, this is not the case—as far as mere form is 
concerned—when the figures are small. 
Further, it is doubtful (see below, Section VII., § 15, p. 103) whether infinitely slow 
cooling would cause crystallisation to proceed in the way which Osmond suggests. 
