Test cuts occur on staters from all mints in Cilicia Trachea with the exception of Aphrodisias, see Table 1. However, for Holmoi staters they occur only on coins from Group 2 and for Pseudo-Kelenderis staters only on Type 2 (represented by a single specimen so far). Usually the test cuts occur singly, but a significant number of coins have multiple test cuts (see Coin Corpus).

Mint	Coins in corpus	Coins with test cut(s)
		Count	Percentage
Anemourion	2	1	50.0%
Aphrodisias	5
Holmoi, Group 1	9
Holmoi, Group 2	7	4	57.1%
Kelenderis, Group 1	14	4	28.6%
Kelenderis, Group 2	229	8	3.5%
Kelenderis, Group 3	246	18	7.3%
Nagidos, Group 1	138	23	16.7%
Nagidos, Group 2	246	5	2.0%
Nagidos, Group 3	198	51	25.8%
Pseudo-Kelenderis	21	1	4.8%
Total	1115	115	10.3%

Table 1: Overview of mints

From the point of view of time sequence, it could undoubtedly have happened that the risk of accepting a plated coin was excluded by a test cut and the coin verified in this way was subsequently countermarked, but it could also have happened that the countermarked coin travelled to another region and was tested there by means of a test cut. Evidence of the latter possibility is, for example, the countermarked fourrée Lederer 1931, 1c (a countermark would not have been placed on a coin in which a test cut revealed a bronze core). Nevertheless, the relationship between test cuts and countermarks needs to be investigated. As shown in Table 1 above and in Table 2 in the Section Countermarks, a sufficient number of observations of both test cuts and countermarks are available only for the Kelenderis Group 3 and the Nagidos Group 3.

The frequencies of the presence/absence combinations of test cuts and countermarks are given in Tables 2 and 3. These contingency tables are supplemented with row and column percentages. For both Kelenderis and Nagidos, we can see that the relative frequency of coins with a countermark is higher for coins with a test cut than for coins without a test cut (22.2% versus 14.5% for Kelenderis, 33.3% versus 14.3% for Nagidos). Similarly, we can see that the relative frequency of coins with a test cut is higher for coins with a countermark than for coins without a countermark (10.8% versus 6.7% for Kelenderis, 44.7% versus 21.2% for Nagidos). This suggests that the occurrences of test cuts and countermarks are not independent of each other, but are positively correlated, especially for Nagidos. For Kelenderis, the correlation coefficient and Jaccard index are equal to 0.056 and 0.078, respectively, for Nagidos the correlation coefficient and Jaccard index are equal to 0.212 and 0.236, respectively. The rates of interdependence are low for Kelenderis, which is confirmed by Fisher’s exact test, which does not reject the hypothesis of independence of the occurrences of test cuts and countermarks (p-value equal to 0.325). In contrast, for Nagidos, Fisher’s exact test rejects this hypothesis (p-value equal to 0.006).

Test cut(s)	Countermark(s)
	Yes	No	Total	Yes	No	Total	Yes	No	Total
Yes	4	14	18	22.2%	77.8%	100.0%	10.8%	6.7%	7.3%
No	33	195	228	14.5%	85.5%	100.0%	89.2%	93.3%	92.7%
Total	37	209	246	15.0%	85.0%	100.0%	100.0%	100.0%	100.0%

Table 2: Kelenderis, Group 3 – test cuts versus countermarks

Test cut(s)	Countermark(s)
	Yes	No	Total	Yes	No	Total	Yes	No	Total
Yes	17	34	51	33.3%	66.7%	100.0%	44.7%	21.2%	25.8%
No	21	126	147	14.3%	85.7%	100.0%	55.3%	78.8%	74.2%
Total	38	160	198	19.2%	80.8%	100.0%	100.0%	100.0%	100.0%

Table 3: Nagidos, Group 3 – test cuts versus countermarks

Contingency Tables 2 and 3 do not take into account other characteristics of the coins, in particular the individual coin types, and therefore conclusions based on these aggregated data may be biased by the influence of the Yule–Simpson effect (also referred to as Simpson’s paradox). The decompositions of Group 3 for Kelenderis and Nagidos by coin types are therefore given in Tables 4 and 5. For Kelenderis, a possible positive correlation would be considered for Types 3.7 and 3.8, however, the number of occurrences of countermarks is very low, and therefore it does not make sense to carry out a more detailed analysis of the Kelenderis coins for the time being. In contrast, for Nagidos Type 3.4, 10 and 11 observations of test cuts and countermarks, respectively, are available. The contingency table for this coin type is shown in Table 6. Fisher’s exact test rejects the hypothesis of independence (p-value less than 0.001) and also the correlation coefficient and Jaccard index take high values of 0.698 and 0.615, respectively.

Table 7 lists all 13 Nagidos coins of Type 3.4 with a test cut or a countermark. For better clarity, coins with both a test cut and a countermark are highlighted in light red, coins with only a test cut in light green, and coins with only a countermark in light yellow. Interestingly, with the exception of the coin from Fritz Rudolf Künker’s Auction 174 (27 September 2010, Lot 408), all coins on which both the test cut and the countermark occur are characterized by the same short deep cut, which does not go through the entire thickness of the coin. In contrast, the two coins with a test cut without a countermark (Lederer 1931, 46a and 51) have different types of test cuts. It therefore appears that all coins of Type 3.4 with both a test cut and a countermark (perhaps with the exception of the coin from Fritz Rudolf Künker’s Auction 174) were subjected to the cut test at the same place. And because there are two types of countermarks on these coins (five of the bull and crescent type, three of the eagle and trident type), the tested coins were probably already countermarked and thus were tested in a different region than where the countermarking occurred.

This observation, however, cannot of course be generalised to other groups of coins where test cuts and countermarks occur simultaneously. Each coin type, or a differently defined group of coins, could have had a different fate within the monetary circulation and would have required a separate investigation.

Type	Coins in corpus	Coins with test cut(s)		Countermarked coins
		Count	Percentage	Count	Percentage
3.1	33			3	9.1%
3.2	18			1	5.6%
3.3	1			1	100.0%
3.5	4			2	50.0%
3.7	54	7	13.0%	3	5.6%
3.8	40	5	12.5%	3	7.5%
3.9	25	3	12.0%
3.12	9	1	11.1%
3.13	10			1	10.0%
3.14	5	1	20.0%	2	40.0%
3.15	14	1	7.1%	11	78.6%
3.16	15			10	66.7%
Total	246	18	7.3%	37	15.0%

Table 4: Kelenderis, Group 3 – test cuts and countermarked coins according to coin types

Type	Coins in corpus	Coins with test cut(s)		Countermarked coins
		Count	Percentage	Count	Percentage
3.1	14	2	14.3%	5	35.7%
3.2	13	2	15.4%	8	61.5%
3.3	14	4	28.6%	11	78.6%
3.4	49	10	20.4%	11	22.4%
3.6	29	9	31.0%
3.7	45	17	37.8%	2	4.4%
3.8	11	4	36.4%	1	9.1%
3.9	4	3	75.0%
Total	198	51	25.8%	38	19.2%

Table 5: Nagidos, Group 3 – test cuts and countermarked coins according to coin types

Test cut(s)	Countermark(s)
	Yes	No	Total	Yes	No	Total	Yes	No	Total
Yes	8	2	10	80.0%	20.0%	100.0%	72.7%	5.3%	20.4%
No	3	36	39	7.7%	92.3%	100.0%	27.3%	94.7%	79.6%
Total	11	38	49	22.4%	77.6%	100.0%	100.0%	100.0%	100.0%

Table 6: Nagidos, Type 3.4 – test cuts versus countermarks

Type	Reference
3.4a	Fritz Rudolf Künker GmbH & Co. KG, Auction 143 (6 October 2008), Lot 240 (acsearch.info URL). Rev.: countermark (bull walking r., crescent above).
	CGB Numismatics Paris, MONNAIES 36 (23 October 2008), Lot 200 (acsearch.info URL). Ex CGB Numismatics Paris, MONNAIES 31 (21 June 2007), Lot 77 (acsearch.info URL). Ex CGB Numismatics Paris, MONNAIES 25 (26 January 2006), Lot 107 (acsearch.info URL). Rev.: countermark (bull walking r., crescent above). Test cut(s).
	Gorny & Mosch Giessener Münzhandlung, Auction 244 (6 March 2017), Lot 335 (acsearch.info URL). Ex A Munich art dealer, February 1981. Rev.: countermark (bull walking r., crescent above).
	Salon Numizmatyczny Mateusz Wójcicki, Auction 14 (28 September 2024), Lot 5006 (acsearch.info URL). Ex Classical Numismatic Group, Auction 124 (19 September 2023), Lot 245 (acsearch.info URL). Ex The New York Sale, Auction 40 (11 January 2017), Lot 1117 (acsearch.info URL). Ex Numismatic Fine Arts (Beverly Hills, USA), Mail Bid Sale (18 October 1990), Lot 370 (Gallica URL). Ex Numismatic Fine Arts (Beverly Hills, USA), Mail Bid Sale (15 January 1982), Lot 245 (Gallica URL). Rev.: countermark (bull walking r., crescent above). Test cut(s).
3.4b	Gerhard Hirsch Nachfolger, Auction 256 (5 May 2008), Lot 323 (acsearch.info URL). Obv.: countermark (bull walking r., crescent above). Test cut(s).
3.4c	Lederer 1931, 47 = BMC 21, p. 114, 24 = The British Museum (London), museum number TC,p187.1.Nag Rev.: countermark (eagle standing r., upward trident before it). Test cut(s).
	Savoca Coins, 20th Silver Auction (1 January 2018), Lot 204 (acsearch.info URL). Rev.: countermark (bull walking r., crescent above). Test cut(s).
	Fritz Rudolf Künker GmbH & Co. KG, Auction 174 (27 September 2010), Lot 408 (acsearch.info URL). Rev.: countermark (eagle standing r., upward trident before it). Test cut(s).
3.4g	The J. Paul Getty Museum, Villa Collection, Malibu, California, USA (provenance: donated to The J. Paul Getty Museum in 1980 by Lily Tomlin, Santa Monica, California, USA), Object Number 80.NH.152.86 (Getty’s Collection Online URL). Obv.: countermark (unidentified).
	SNG Levante, 13 = Fritz Rudolf Künker GmbH & Co. KG, eLive Auction 59 (25 March 2020), Lot 8108 (acsearch.info URL). Rev.: countermark (eagle standing r., upward trident before it). Test cut(s).
3.4h	Lederer 1931, 46a. Test cut(s).
3.4i	Dr. Busso Peus Nachfolger, Auction 384 (2 November 2005), Lot 335 (acsearch.info URL). Ex Dr. Busso Peus Nachfolger, Auction 380 (3 November 2004), Lot 576 (acsearch.info URL). Ex Gerhard Hirsch Nachfolger, Auction 173 (19–22 February 1992), Lot 464. Rev.: countermark (bull walking r., crescent above). Test cut(s).
3.4o	Lederer 1931, 51 = Classical Numismatic Group, Electronic Auction 560 (17 April 2024), Lot 162 (acsearch.info URL). Ex Naville Numismatics, Auction 73 (8 May 2022), Lot 155 (biddr.com URL). Test cut(s).

Table 7: Nagidos, Type 3.4 – coins with test cuts and countermarks