522 Proceedings of the Eoiial Irish Academy. 



A large number of early bronze celts would require to be analysed before 

 it could be definitely determined wliether there was a gradual increase of tin 

 in their composition during an experimental stage; but as the flat celts 

 analysed by Mallet (No. 1) and by Gowlaud (No. 2) both proved to be 

 rich in tin, such a stage appears improbable. It is possible that the 

 comparatively small amounts of tin in the socketed celts analysed by 

 Mallet (No. 5) and Phillips (No. 7) may have been due to these celts having 

 been recast from worn-out or broken implements, with a consequent loss of 

 tin, due to oxidation on remelting.' 



At the same time, it must be borne in mind, as Professor Gowland^ has 

 pointed out, that no single alloy of bronze would be equally suitable for all 

 the weapons and implements of Bronze-Age man. A sword or dagger would 

 require to be harder than a celt ;. and while 11 to 14 per cent, of tin might 

 be required for them, a smaller amount of, say, 10 per cent, or less would be 

 quite satisfactory for a celt. Variations in the alloy might, therefore, be due 

 to attempts at discovering the most suitable composition for various 

 implements rather than as experiments in hardening copper by the addition 

 of tin. 



In the absence of any dii-ect proof to the contrary, the writer does not 

 incline to the belief that the secret of alloying copper was discovered by a 

 process of experiment ui Ireland. The improbability of the art of making 

 bronze alloys having been discovered independently in different countries has 

 been pointed out by Professor Boyd Dawkias, who, relying on tables of 

 analyses of bronze implements published by hun, which show that in British 

 implements the percentage of tin varies between 5-09 and IS'31, and in 

 French between 1-50 and 21-5, concludes that " The uniformity of the com- 

 position of the cutting implements of the Bronze age implies that the art of 

 compounding tin with copper was di.scovered in one place, . . . Had it 

 spread from difi'erent centres, this uniformity would have been impossible."' 



The frequent references to Spain that occur in early Irish literature 



' But see on this point, Gowlaud, Journdl Hoyal Anthropological Institute, xxxvi, 

 p. 27. (He does not think there would be much loss of tin on remelting.) 



-Journal Royal Anthropological Institute, xlii, p. 243. (On the other hand, the 

 addition of a large proportion of tin, say up to some twenty-seven per cent. , while it 

 increases the hardness of the alloy, also increases its brittleness. The alloy of bronze in 

 use at present, as being the most suitable for parts of machinery, is the same as that of 

 prehistoric times, i.e. an alloy constituted of about ninety per cent, of copper and ten 

 per cent, of tin. See M. v. Schwarz, Prathuitorische Zeitschrift, ii, pp. 421, 422.) 



^ Early Man in Britain, p. 410. (Rice Holmes, Ancient Britain and the Invasions of 

 Julius Caesar, p. 125 n., quotes the above, and adds: "The uniformity which subsists 

 between 5-09 and 1S'31, and between 1-50 and 21-5 is remarkable.") 



