14 



MEMOIRS NATIONAL ACADEMY OF SCIENCES. 



rvoL-xiv. 



Phosphorus not determined in either case. 



Smith fm-ther determined the stony portion to consist of 69 per cent soluble in aqua regia 

 and 41 per cent insoluble, giving analyses of each, from which the bulk analysis given below 

 was calculated. 



1 Analyses recalculEted by Farrington. Smith reported also traces of lithium and copper. 



> Not determined. 



'Trace. 



Five grams of the finely pulverized stone were boiled in distilled water for an hour, 

 resulting in a solution yielding 0.062 per cent SO3 and 0.012 per cent CaO. A portion of 

 the SO3 probably came from the decomposed troilite, rendering any calculations imcertain, 

 while the amoimt of lime (CaO) is too small to make the results moi'e than suggestive of 

 the presence of a minute quantity of oldhamite. A second 5 grams were boiled for half 

 an hour in acetic acid of 15 per cent normal strength. The solution yielded 0.08 per cent 

 P2O5 and 0.122 per cent CaO. Inasmuch as the bulk analysis shows 0.23 per cent P2O5, it is 

 evident that a complete solution of the phosphate was not accomplished by the acetic acid. 

 Be this as it may, the relative proportion of acid to base is such as to render it unsafe to draw 

 definite conclusions.' 



It is difficult to account for all the discrepancies between the two analyses. The difference 

 of some 3 per cent between the amoimt of stony matter and troilite may perhaps be accoimted 

 for on the supposition that Smith worked, as is so often the custom, on verj^ small amounts 

 that did not correctly represent the stone as a whole. (Whitfield had 19 grams of selected 

 material.) The analyses of the metallic portion, it will be noticed, agree fairly well excepting 

 that Whitfield reports no copper. In the bulk analyses, however, we find a difference of 

 3 per cent (in round numbers) in the total silica, nearly 7 per cent in the ferrous iron, 3.92 per 

 cent in the almnina, and 2.25 per cent in the lime, with minor differences, mainly due to 

 omissions elsewhere. The totals for iron and magnesia do not differ more than might be 

 anticipated from analyses on separate portions, made by even the same analyst. That Smith 

 did not determine the nickel and cobalt in the silicate portion is not strange, it being customary 

 in his day to regard these elements as constituents of the nickel-iron only. The phosphoric 

 acid, amounting to 0.23 per cent, should in this day certainly not be overlooked. 



It does not seem in the least probable that the phosphate to which I have referred above 

 as evident to the unaided eye can be the white mineral mentioned by Smith. Nevertheless, 

 a most careful examination of all the material in the Museum and Shepard collections reveals 

 nothing that is even suggestive of his doubtful unisilicate. 



I Since the above was written, Dr. E. T. Wherry ( Amer. Mineralogist, vol. 2, No. 9, 1917) has complimented mo by suggesting that the problem- 

 atic phosphate be given the name merrillUe. Had I been consulted in the matter I should have suggested a postponement until a more definite 

 statement of its composition could be given. Incidentally, it may be stated, I had considered the use of Shepard 's name, apetoid (Amer. Journ. 

 Boi., vol. 2, 1846), but abandoned it because of his definite statement that his mineral contained no phosphorus. 



