Acknowledgements. I want to express my sincere gratitude to the Director of the Heraklion Museum, Dr Y. Sakellarakis, for the permission to study these roundels and to the epimeletria E. Grammatikakis for all her help to me during my studies in the museum. To the Managing Committee of the British School at Athens and the Ministry of Culture and Sciences Department of Foreign Schools in Greece I owe my warmest thanks for permission to publish these roundels. I also want to thank warmly Dr Judith Weingarten who has read the manuscript and suggested several improvements. Finally, I want to express my gratitude to the Danish Research Council for the Humanities, for the scholarship which made this study possible.

The following abbreviation is used in addition to the standard ones: GORILA 1–5 = Godart, L. and Olivier, J.-P., Recueil des inscriptions en Linéaire A (Études Crétoises 21, 1–5) (Paris 1976–1985)Google Scholar. Drawings and photographs are by the author except FIG. 1 KN Wc 26 and 29, FIG. 2, KN Wc 3 and 30 which are from GORILA 2, and PLATE 5, KN Wc 26 and PLATE 8, KN-Rb which are reproduced from CMS 10 no. 120. All roundels are reproduced (both in figures and plates) in scale 1:1, while the seal impressions are reproduced in scale 4:1.1 have not wanted to anticipate the forthcoming CMS drawings of the seal impressions, and have therefore chosen to give sketches in scale 1:1 and two detailed photographs. In PLATES 5–7 the roundels have been given in three columns: first the obverse, second the reverse, and third in perspective (or profile).

¹ See forthcoming paper by the author: ‘The Function of Seals on the Minoan Roundel’, in Fragen und Probleme der bronzezeithlichen ägäischen Glyptik (CMS, Beiheft 3) (Marburg 1987).

² They are all identified thanks to the seal impressions, Gill, M., ‘The Knossos Sealings: Provenance and Identification’, BSA 60 (1965) 70–1Google Scholar, who enumerates the HM numbers 413 (Wc 3), 342 (Wc 23), 343 (Wc 25), 344 (Wc 29), 345 (Wc 30), and 346 (Wc 41). Hereafter Gill.

³ Evans, A. J., ‘The Palace of Knossos. Provisional Report for the Year 1903’, BSA 9 (1902–1903) 51. Hereafter Evans 1903.Google Scholar

⁴ Evans, , The Palace of Minos at Knossos i–iv (London 1921–1935)Google Scholar (hereafter PMi–iv). PM i. 620 with fig. 456 (only roundel).

⁵ Gill, 71 (La).

⁶ Evans 1903, 55.

⁷ Evans 1903, 54.

⁸ Gill, 70 (L 18–19).

⁹ AE/NB 1900,56.

¹⁰ Gill, 73, O.3 (PLATE 12).

¹¹ Mühlstein, H., ‘Ein Linear-A-Dokument’, Minos 6 (1958) 7 n. 2.Google Scholar

¹² None of them can be identified, neither on the inscription nor on the impression.

¹³ The number seem to be read M 137 or M 127 and it was stored in a box where the compartment for this roundel had the number T 384.

¹⁴ The compartment in the box where this roundel was stored had the number PA 380.

¹⁵ The spiral on KN Wc 46 and the rope design on KN Wc 45 may fit Evans's description of 1–6 (Evans 1903, 54): ‘Decorative designs with spiral and curvilinear patterns’, while the remaining one on 45 may be listed among those too imperfect to be made out with certainty.

¹⁶ The seal impressions can never be more than uncertain indications since seals may have circulated for long periods of time; see Niemeier, W.-D., ‘Probleme der Datierung von Siegeln nach Kontexten’, in Studien zur minoischen und helladischen Glyptik (CMS, Beiheft i) (Marburg 1983) 91–103.Google Scholar

¹⁷ By Betts in CMS 10 no. 120.

¹⁸ Same design as found on Rh and Rn. The cross sign was also especially noted by Evans 1903, 90 fig. 61 and PM i. 515, fig. 374. Parallels to plain crosses on seals are extremely rare. One is known from tholos Tomb II at Lenda, CMS 2.1 no. 200, and a second one will be referred to below.

¹⁹ Gill, 70 (L 18–19). The motif reoccurs on KN Wc 23. A similar design was found among the Zakro sealings described as ‘three rosettes’ with conventionalized lotus blooms between’. Hogarth, D. G. ‘The Zakro Sealings’, JHS 22 (1902) 88 no. 132.CrossRefGoogle Scholar Two from the Ashmolean Museum are illustrated in Kenna, V. E. G., Cretan Seals (Oxford 1960) pls. 16.23.S, 17.27.S.Google Scholar

²⁰ Published in Evans 1903, 59 fig. 38; PM ii. 831 fig. 547; PM iii, 465 fig. 324; also Gill, 70 (L 47).

²¹ Raison, J. and Pope, M., Corpus transmuméré du linéaire A (Cabay 1980) 232.Google Scholar

²² CMS 10 no. 120.

²³ Gill, 71 (La) is perhaps Evans 1903, 55 no. 34. Well-known type except that the goat is looking backwards.

²⁴ Gill, 71 (L 27). A parallel motif, also from the Temple Repository, is found on HM 399; PM i. 696 fig. 518c. To judge from the drawing given in PM this one is not from the same seal.

²⁵ Gill, 70 (L 8–12). The ‘architectural design’ is very popular in the Neo-palatial period.

²⁶ On PLATE 12 the impressions are given to the same scale. Although none of the crosses is completely preserved and are partly disturbed by inclusions in the clay it can be seen: These two observations positively indicate that the two impressions Rg and Rn were done by the same prism.

²⁷ Both the dimensions of the seal and the dimensions of the cross.

²⁸ Probably Evans 1903, 54 nos. 1–6. The spiral or recumbent S is occurring both with and without decorative elements on both three-sided prisms (for example, CMS 13 no. 89 and CMS g no. 18) and stamps (for example, CMS 4 no. 92).

²⁹ I can point out no obvious parallels for this design.

³⁰ Gill, 70 (L 16–17); Evans 1903, 54.

³¹ The designation ia + b and 2a + b means that the finger was held in the same position during the sealing procedure–but that the different pressure on the roundel gave slightly different outlines of the finger.

³² For proportions on rings: Hallager, E., The Master Impression (SIMA, 69) (Göteborg 1985) 25 n. 1.Google Scholar

³³ Gill, 73 (O.3), PLATE 12. Popular design in Neo-palatial period.

³⁴ Intermingled rope-designs appear to have been popular in the MM period as, for example, seen in Phaistos, CMS 2·5 nos. 152–66.

³⁵ The elaborated running spiral is not usual as a main motif on seals, although known as, for example, seen in CMS 4 no. 118.

³⁶ Evans, , Scripta Minoa i (Oxford 1909) 185–6 (hereafter SM i)Google Scholar

³⁷ I am grateful to Dr Jean-Pierre Olivier for, in the Heraklion Museum, calling my attention to the two documents KN Wc 42 and 43 which are here included among the inscribed roundels from Knossos.

³⁸ Olivier, J.-P., Les Scribes de Cnossos (Incunabula Graeca 17) (Rome 1967) tables xix and xlvi.Google Scholar

³⁹ e.g. is the re-mirrored version on TL Za 1.a without the small bent stroke inside the sign, GORILA 4, 58–9.

⁴⁰ Hallager, E. and Vlasakis, M., ‘Two New Roundels with Linear A from Khania’, Kadmos 23 (1984) 8.CrossRefGoogle Scholar

⁴¹ GORILA 5, 155 6.

⁴² GORILA 2, 84–5, Brice, W. C., Inscriptions in the Minoan Script of Class A (Oxford 1961) iii. 1–6 and iii. 17Google Scholar; Raison and Pope, op. cit. (see n. 21 above) 221–36.

⁴³ About 44 per cent of the entire inscribed corpus of roundels have signs incised which may with reasonable certainty be considered ideograms. Hallager, op. cit. (see n. 1 above).

⁴⁴ PH (?) 31a.2.3, while the CAP sign is also found on HT Wc 3014 and possible KH 81.3.

⁴⁵ Side A has the same inscription as found on KH Wc 3007 (GORILA 2, 74) which makes it reasonably certain that the inscription on the B side is a separate one.

⁴⁶ Here the ‘word-divider’ on the B side and the fact that a similar inscription is found on HT 54a.2 makes it most reasonable to continue the reading from side A to side B.

⁴⁷ GORILA 5, 182, 194.

⁴⁸ GORILA 2, lvi.

⁴⁹ On KN Wc 23, 25, 26, 29, 42, 44, and 45.

⁵⁰ Hallager, op. cit. (see n. 1 above).

⁵¹ For shapes and motifs used on other roundels, Hallager op. cit. (see n. i above) table 8.

⁵² Evans, , SM i. 19–22, 144–8, 263–72.Google Scholar A further sealing from the Hieroglyphic Deposit with impressions from ordinary seals and prisms with hieroglyphs, Yule, P., ‘On the Date of the “Hieroglyphic Deposit” at Knossos’, Kadmos 17 (1978) 1–7Google Scholar, with corrections by I. Pini, ‘Eine Tonplumbe aus Knossos im Ashmolean Museum’, ibid. 21 (1982) 1–4.

⁵³ Evans, SM i, 263–8.

⁵⁴ Meriggi, P., ‘Das Wort “Kind” in den kretischen Hieroglyphen’, Kadmos 12 (1973) 144–33.CrossRefGoogle Scholar

⁵⁵ Evans, , SM i. 265.Google Scholar

⁵⁶ Also argued by Poursat, J.-C., ‘Fonction et usage des sceaux en Créte à l'époque des premiers palais’, forthcoming in Fragen und Probleme der bronzezeitlichen ägäischen Glyptik (CMS, Beiheft 3) (Marburg 1987).Google Scholar

⁵⁷ Evans, , SM i, 174, 177.Google Scholar

⁵⁸ Scholars have here followed Evans, SM i. 230.

⁵⁹ In addition to the above, it should also be considered whether or not it would be likely that ‘decorative elements’ would suddenly appear in the middle of an inscription (central position). To my knowledge such a phenomenon is unknown in any other script. And that the S-sign has often additional decorative ornaments attached need not worry us since this is also quite often seen on other hieroglyphs, for example the trowels on P 28d and P 30b, Evans, , SM i, 155, 156.Google Scholar

⁶⁰ HT Wa 1033–1121; GORILA 2, 10–16.

⁶¹ In Khania the cross is found as one of the two inscriptions on the sealings KH Wa 1013–1016, GORILA 3, 108–10.

⁶² One is found on Levi' Zakro no. 160; Weingarten, J., ‘Two Inscribed Sealings from Zakro’, Kadmos 22 (1983) 107–8.CrossRefGoogle Scholar On the second inscription is likewise seen a plain cross, where, however, two more lines may suggest a different reading.

⁶³ E. Grumach, ‘Zwei hieroglyphische Siegel’, ibid, i (1962) 153–4. If, however, the ‘croix pommetée’ may be considered a forerunner for the plain cross found on the Knossos roundels there would exist an even stronger case for its use by the ‘hieroglyphic’ administration as shown by the finds in Mallia. In Quartier Mu, for example, this motif is found on the same type of administrative seals as those with the hieroglyphs, Poursat, , ‘Sceaux et empreints de sceaux’, in Le Quartier Mu. II (Études Crétoises, 26) (Paris 1980) nos. 244, 247Google Scholar; also Poursat, op. cit. (see n. 56 above). I am grateful to Judith Weingarten for drawing my attention to this.

⁶⁴ Hallager op. cit. (see n. 1 above).

⁶⁵ GORILA 5, 316–17 as two different fractional signs A 715 and A 703—a combination which would not make sense on one roundel. The same reading was suggested by Caskey, J. F., ‘Inscriptions and Potters’ Marks from Ayia Irini in Keos', Kadmos 9 (1970)Google Scholar no and also by Raison and Pope, op. cit. (see n. 21 above) 150 who, however, consider the reading uncertain.

⁶⁶ GORILA 5, 144–5 as part of inscription (sign AB 06) while Raison and Pope, op. cit. (see n. 21 above) 278 suggest AB 02 as an isolated sign on Wc 39 (their PH W 35) and AB 06 as a possible part of a full inscription on Wc 40 (their PH W 36).

⁶⁷ e.g. CMS 2.2 no. 310 and Evans, , SM i. 153, P 24.Google Scholar

⁶⁸ Hogarth, op. cit. (see n. 19 above) 89, Types 136, 137, 140.

⁶⁹ Levi, D., ‘Le Cretule di Haghia Triada’, ASAtene 8–9 (1925–1926) 86 fig. 29, Type 2.Google Scholar

⁷⁰ The problems on inscriptions on roundels has been shortly treated in Hallager, op. cit. (see n. 1 above) where it has been made clear that the roundels often carry ‘additional inscriptions’ compared to the commodity treated on the roundel. This, for example, is the case on MA Wc 5 where the commodity is GRA and where the inscription must be an additional information. A more elaborated hypothesis in Hallager, ‘The Roundel in the Minoan Administrative System’, forthcoming in Proceedings of the Symposium ‘Problems in Greek Prehistory, Manchester 1986 (Bristol Classical Press 1987).

⁷¹ e.g. CMS 5 no. 180.

⁷² The results of a police investigation of this roundel will be found in Hallager, , ‘On the Track of the Minoan Bureaucrats and their “Clients”’, forthcoming in Festschrift in Honour of Prof. N. Platen (Heraklion 1987).Google Scholar

⁷³ Hallager op. cit. (see n. 1 above).

⁷⁴ Only in one other case where roundels are concerned have I observed a similar phenomenon. In the Haraklion Museum I had the opportunity to investigate two identical (unpublished) roundels probably from Mallia (MA Wc 13 and 14), each of which had two different motifs impressed. One was decorative (rosettes) the other pictorial (perhaps daemon); Hallager, op. cit. (see n. 1 above) n. 35. Other multi-seal-impressed documents without hieroglyphic prisms are also known. Apart from the Zakro sealings, Weingarten, The Zakro Master and His Place in History (SIMA pocketbook, 26) (Göteborg 1983), the most impressive is the triangular clay bar from Khania with twelve impressions, ten of which were different, Hallager, op. cit. (see n. 32 above) 11–12 fig. 2. Further a now lost sealing from Hagia Triada with three different impressions, HT Wa 1021 bis), Levi, op. cit. (see n. 69 above) 83 no. 3 fig. 25, and some two-sided noduli from Zakro, Weingarten, , ‘Some Unusual Minoan Clay Nodules’, Kadmos 25 (1986) nos. 14–16CrossRefGoogle Scholar, and Knossos, Kenna op. cit. (see n. 19 above) 144, S 7, pl. 16.

⁷⁵ I am grateful to my wife, Birgitta, for suggesting this idea which will be further explored.

⁷⁶ See catalogue under KN Wc 30, KN-R1 with n. 24.

⁷⁷ If this explanation is correct it reminds us of the situation at Zakro where many almost identical impressions always appear together with one or two more impressions on a sealing. Judith Weingarten, however, has kindly informed me that this explanation does not work at Zakro, while she agrees that the habit of multiple stamping extends from Knossos to Zakro.

⁷⁸ The general use of the roundel including the entire corpus has been treated in Hallager op. cit. (see n. l above). Here it will be seen in the different tables that the Knossos roundels, by and large, fall into the general picture: Number and type of roundel Table I Associated finds Table 2 Size-categories Table 3 Use of space on edge for impressions Table 4 Reading of roundels Table 6 Type of seals used by individuals Table 7 (here Table 1) Decoration on seals used by individuals Table 8

⁷⁹ Hallager, ‘The Roundel in the Minoan Administrative System’ (see n. 70 above).