TRANSACTIONS OF SECTION H. 697 



number of the ' Derbyshire Archreological Journal ' points out the close approxi- 

 mation of the heaviest specimen to the standard which Mr. Reginald Smith, of 

 the British Museum, had shown to be represented by a bronze weight found at 

 Neath (-4,770 grains), and another (of basalt) at Mainz (4,767 grains), and by the 

 normal weight deduced from that of a large number of iron bars found in the 

 purely British lake-village at Glastonbury and in other British sites. Mr. Smith's 

 conclusions entirely establish the soundness of the te.\t in Caesar, ' B. G.' v. 12, 4 : 

 ' Taleis ferreis ad certum pondus examinatis pro nummo.' Details are given by Mr. 

 Smith in his paper on the Ancient British Iron Currency ('Proceedings of the Society 

 of Antiquaries,' xx., 179, January 26, 1905), and in outline in the ' Guide to the 

 Antiquities of the Early Iron Age in the British Museum,' 1905, pp. 149f. Both 

 the Neath and the Mainz specimens exhibit the same cheese or barrel shape which 

 appears in four Melandra specimens (1, 2, 5, 12) ; each of the two is marked I on 

 the face, but the ilainz specimen has a further legend which no one yet has inter- 

 preted, I Q followed by a third sign, apparently a Q tilted to the left. 



The peculiar importance of the collection at Melandra is that we have 

 here represented certainly seven (including the unit), and quite possibly nine, 

 denominations of this standard, whose subdivisions have been hitherto entirely 

 unknown. 



The nature of the subdivisions is also interesting. Besides the duodecimal 

 principle (in Nos. 2, 3, 8, 25, and ?21), following that of the Roman libra and 

 uncia, to which Mr. May's article calls attention, we must recognise not less 

 clearly the quadratic (Nos. 2, 5, 8, ? 12, 20, 28, and ?21), giving us a division of 

 the unit into 4, 8, 16, 32, and ?96 parts. Nos. 2, 3, 5, and 21 could belong to 

 either, and 12 may just conceivably be Roman and represent 10 J drachmae or 

 seven times the weight of an Antoninianus. 



It would, of course, be possible to interpret all these weights as representing 

 so many ' British drachmae ' (if one may coin such a term for the sake of argument), 

 since 96 is a common denominator for both 12 and 16 ; but one seeks a reason for 

 the creation of weights to represent 6 and 12 ' British drachmae,' i.e. ^V and ^ of the 

 ' British pound ' respectively, if there was no other named standard than ^ of the 

 unit (' British uncia ') and ^ (' British drachma '). And that there was some 

 other such named unit weighing ^^ of this ' British pound ' (298'1 grains) seems 

 at least suggested by the markings on Nos. 12 and 20, which would then be the 

 weights of two and one such units respectively ; unhappily, No. 12 is nearly 

 8 per cent, under its proper weight on this hypothesis. It is also clear that the 

 markings on No. 8 vouch for the duodecimal system, as Mr. May points out. 

 But Nos. 20 and 28 are imimpeachable witnesses for the quadratic system. 



Can we conjecture from this that we have here the result of the imposition 

 of the Roman system of 12 ounces and 96 drachms upon a Keltic system of 

 dividing the pound into sixteen parts ? And that, therefore, the essential 

 characteristic of our modern ' avoirdupois ' measure goes back to the Early Iron 

 Age ? A similar case of the imposition of Roman divisions upon a local unit 

 occurs at Pompeii ; see ' The Mensa Ponderaria of the Naples Museum,' App. I. 

 to my edition of the remains of ' The Italic Dialects.' And examples more 

 important for Northern lands will be found in Appendix C of Professor Ridgeway's 

 ' Origin of Metallic Currency and Weights Standards.' 



No. 3, which has been considerably cut about, and does not correspond in 

 shape to No. 2, looks like a Roman weight cut down to the Keltic standard. 



It seems as if there must be some connection between this division and the 

 fact that in the utiffa of the Brehon Laws ( = 1 Roman uncia, or 432 grains) there 

 are 32 crosogs of 13-5 grains (Ridgeway, p. 396). But it is not quite clear 

 whether there is any relation between the actual standards of this Melandra 

 system and the older Irish system, which Professor Ridgeway (p. 406) suggests, 

 of 30 crosogs to an ounce of 405 grains. 



Some weights not yet publicly described, but said to correspond to the Neath 

 standard, have recently been found in Somersetshire, and are now in the Castle 

 Museum, Taunton. 



