The Galatian Shield without (symbol) Bronzes
Die Estimates

All photographs and descriptions are (C) by their respective owners.
All rights reserved worldwide. No copying, use, or distribution without written permission.
All photos used by permission.
Page layout, design, and related web content (C) 2011 by Daniel Wolf. Click on photo thumbnails for enlarged views.

Offprint of the Numismatic Chronicle Paper (online PDF)
All Web Supplemental Material For This Paper
Table 4 - Die Axis Distributions
Table 5 - Die Study Statistics
Table A2-1 - Die Axis Distribution Totals
Table A2-2 - Weight Distribution Statistics
Table A2-3 - Die Axis Statistics
Weight Histograms
Die Study with Images Supplement Page
Die Production Estimates Supplement Page
Die Axis Distributions Supplement Page
Weight Comparison of Sv610 Border Types Supplement Page
Online Calculators for Die Estimates and Die Axis Vector Summation

Die Production Estimates

The summary of data of the 225 coins is presented in Table 5. In most cases we were able to work with high quality photographs and some coins were available for physical inspection. In the interest of making conservative numerical estimates, about 15% of the coins (32) were rejected for die comparisons due to poor condition or photographs that were not clear enough. The statistical technique for estimating die production was published by Warren Esty (2006). The column labeled: D = Est. # of Dies shows the results of applying Esty’s method to the ratio of N (number of obverses suitable for comparisons) to d (the number of observed dies among the sample, N). Esty’s method also permits estimation of lower and upper 95% confidence intervals for the value of D, which are shown in the adjacent columns of Table 5. We applied the Esty method to each individual catalogue type as well as to sub-types (e.g. Sv 610-sub2 with linear reverse border) and to the categories discussed in our paper, as shown in Table 5.

Several of the die estimates and numerical count observations merit comment. We found that not only do the Sv 610 grouped together yield a smaller N/d than most other types, but that both subtypes (with the two different reverse borders) yield similarly small N/d ratios. The one exception to this pattern is Sv 624 with an N/d = 1.0000, but for which the sample of only 4 specimens is so small as to impel caution in comparing the low N/d to that observed for the large quantity of Sv 610.

We can also point out that the Sv 610 types listed in our die study (Appendix 1) yield another interesting quantitative observation, namely that the number of obverse dies is nearly the same as the number of reverse dies. For a total of 86 specimens of Sv 610 the ratio of reverse to obverse dies is only 1.038, while for the coins of West Greek style that ratio is 1.709 (107 specimens in Appendix 1, with 94 reverse dies for only 55 obverses). The reverse : obverse ratio for the subgroups of Sv 610, with the two reverse borders, are both close to 1 : 1.

These N/d and reverse : obverse die ratios support the distinction between Sv 610 and other types that we proposed on the bases of die axis distributions (see Die Axis Distributions ) and style. The near equality of obverse and reverse dies for Sv 610 may reflect a difference in production method that is consistent with the nearly fixed die axis we observed for those coins. One implication is that hinged (or otherwise keyed) dies were used and that loss (due to damage or wear) of either an obverse or reverse die led to the end of use of the pair. If that is the case then it is possible that the frequently cited estimates of production ‘per obverse die’ are high for this type and the total coinage might be smaller than otherwise estimated.

The varieties of West Greek style (hereafter termed simply West Greek types, for convenience) comprise slightly over half of the coins in our die study (107 specimens in Appendix 1) so we have a total recovery that exceeds the number of Sv 610, but it is distributed among seven variants distinguished by control and control configuration. The N/d values and reverse : obverse die ratios for each of those seven types are almost uniformly greater than either sub-variety of Sv 610. The implication is that the West Greek types were produced from fewer dies and by a different technology that led to greater reuse of obverse dies with various reverse dies.

For the coin type we’ve named Sv 610A, the small number of dies deserves special mention. All eight specimens we observed are from a single obverse die. Because these eight specimens are only about 7.4% of all the West Greek types included in the die study, the peculiarly large N/d for this one type does not substantially bias the overall N/d for the West Greek types grouped together. N/d for the West Greek types other than Sv 610A is 1.8333 (99 specimens, 54 dies); when the eight Sv 610A are included it is 1.9455 (107 specimens, 55 dies). In either case the West Greek types’ N/d is nearly twice that of Sv 610. Indeed each individual West Greek contributing type, except for Sv 624 (only four specimens examined), has an N/d somewhat larger than that for Sv 610. We have mentioned the caveat that production of Sv 610 per die might be a lower than is otherwise implied by their low N/d, a consequence of our observation of near equality of obverse and reverse dies for Sv 610. Nevertheless the die estimates and survival rate (nearly half the total) imply that this production was likely greater than for all the West Greek types combined. We confirm that die estimates and reverse : obverse ratios which distinguish Sv 610 from the other (West Greek) types of ‘Galatian shield’ bronzes are in agreement with other observations. s

We noted some similarities of ‘Galatian shield’ bronzes to the Hieron horseman bronze coins, however die study data for those coins that would permit comparison to our die study results are not available to us. The relationship of West Greek ‘Galatian shield’ and Hieron horseman die axes and weights are explored below.