100 Proceedings of the Royal Irish Academy. 



of tin, did not show, on the contrary, any strong indications of tin ; and it 

 was estimated hy Mr. Pollok to contaui less than 0-5 per cent. To place the 

 matter beyond dispute, it was therefore decided to make a chemical deter- 

 mination of the actual tin in the specimen. 



Mr. Pollok finally reported : — " As I had been informed that the sample 

 W. 262 was supposed to contain about 2 per cent, of tin, I made two chemical 

 analyses of this sample, and found that, in point of fact, it contained 0'25 

 per cent, of metallic tin, which entirely confirms the spectrographic result." 

 It must therefore be finally accepted, Mr. Pollok adds, that W. 262 " con- 

 tains 0-25 per cent, of tin and not more." Some mistake must therefore have 

 occurred in the original analysis or in the printed paper. At first sight it 

 would seem as if the error was caused by a slip in the place of the decimal 

 point. But this is not so ; the results are uniformly given to two places ; 

 the total is correct, and from the text it is e^ddent that it was regarded as a 

 bronze. Moreover, Wilde quotes the analysis of this blade without com- 

 ment (p. 486). But the halberd is covered by a crust of brown- black patina 

 of oxide of iron, which does not dissolve in nitric acid. A portion of the 

 work may have been entrusted to a student ; and though the colour of the 

 precipitate should have indicated its nature, it is conceivable that the oxide 

 of iron was weighed in with the tin. The portion cut ofif for the original 

 analysis was e\idently quite large, judging from the present appearance of the 

 blade (PL III., ISTo. 30), and must have contained a considerable quantity of 

 the patina. Mr. Pollok found no less than 049 per cent, of iron oxide 

 crust in the portion, 2 grammes, analysed by him. Howe\ er it happened, 

 we can well understand that some mistake took place in the analysis at a 

 time, 1853, when the importance of the question involved was not appreciated. 

 There can fortunately be no doubt as to the identity of the specimen. It 

 still retains Wilde's original number, also a special label marked "Mallet," 

 and was the only halberd from which a piece had been cut off for analysis 

 prior to the present paper. There seems, however, to have been an error 

 in stating it was from Eoscommon. Wilde does not give any locality for 

 the specimen analysed by MaUet. I have gone into the subject of this 

 analysis in some detad., as it has been quoted in works of authority. 



Of the other seven halberds examined by the spectrographic method 

 Mr. Pollok says : " jSTone of them contained over 0-5 per cent, of tin; most 

 of them much less ; a number of them showed several lines of lead ; some 

 showed two lines of arsenic; and a number of them showed one line of 

 silver ; and one gave a faint single line of tin (W. 236). They are all nearly 

 pure copper, with small quantities of impurities named." The examples 

 examined were W. 271, W. 231, W. 233, W. 236, W. 247, R. 1978, and 



