Coffey — Irish Copper Halberds. 



99 



analysis. Detailed analyses of the following five specimens were made by 

 Mr. James H. Pollok, D.sc, F.C.S., Assistant Chemist in the Eoyal College of 

 Science, Ireland ; and I have to express to him my thanks for the care he has 

 taken in a somewhat troublesome matter — one of no very exciting nature to 

 the chemist. 



Mr. Pollok's analyses are set forth in the following table; the samples 

 taken were mostly too small for the accurate determination of traces, and in 

 some cases, as W. 248, were a good deal oxidised. The specimens analysed 

 are all figured,^ and are indicated by the word " analysed." 







Copper. 



Tin. 



Anti- 

 mony. 



Arsenic. 



Lead. 



Silver. 



Iron. 



Bismuth, 

 Nickel, Zinc. 



1 



King's Co. 



Day Coll., No. 25, 



99-02 



0-22 



Nil 



Nil 



0-19 



0-26 



0-04 



Nil 



2 



Antrim, 



1903,235, No. 9, 



97-31 



0-31 



0-14 



0-18 



Nil 



Nil 



Nil 



Nil 



3 



Gal way, 



W. 241, No. 19, 



98-06 



0-22 



Nil 



Nil 



0-58 



Nil 



0-17 



Nil 



4 



Cork, 



R. 459, No. 7, . 



98-30 



0-30 



0-27 



0-37 



Nil 



Nil 



Nil 



Nil 



5 



W. 248, No. 28, . 



97-24 



0-18 



Nil 



1-54 



Nil 



0-25 



Nil 



Nil 



These analyses show that the metal of the copper halberd blades is in no 

 way different from that of the copper celts analysed in my former paper. 

 Mallet's analysis, however, still stood in the way, causing me to suppose that 

 a higher percentage of tin might be found in some of the specimens which 

 had not been analysed. Mr. Pollok, therefore, made a spectroscopic analysis 

 of eight additional specimens, including that previously analysed by Mallet, 

 with a view of determining which, if any, showed strong tin lines, so that a 

 quantitative analysis could be made of them if necessary. 



It may be well to explain that the method involves no injury to the 

 specimen whatever. It consists of using the specimen as one of the electrodes 

 of a Euhmkorff coil, and photographing the spectrum of the spark. The 

 spectrograph is then compared with the spectrographs of a known series of 

 alloys of copper and tin — in this case from 0*5 per cent, to 8*0 per cent, of tin ; 

 and from the comparison of the number and strength of the lines seen in the 

 spectrum a close approximation of the composition of the metal can be made. 



The spectrum of the specimen W. 262, believed to contain 2*78 per cent. 



1 The portion taken for analysis was in cases somewhat larger than M'onld be inferred, as 

 unfortunately an accident happened to some of the results, necessitating a second analysis, 



