¹ See p. 225. Incidently it might not be amiss here to point out that part of what is said below has some relevance to fields far beyond that of terracotta figurines, e.g. to the study of much of the mould-made plastic decoration on clay vases, lamp disci etc., and to the classification of terracotta reliefs.

² On archaic Attic terracottas. I would like at this point to record my gratitude to the Committee of the British School of Archaeology at Athens, whose generosity has made my terracotta studies possible.

³ Hermes LXIII 391 ff.; Corolla Curtius 95 ff.

⁴ Neutsch, , Studien zur vortanagräisch-attischen Koroplastik 3.Google Scholar

⁵ On the insuperable difficulties met in fitting a mechanical classification to a typological framework see pp. 224–52.

⁶ See p. 221.

⁷ Kleiner, , Tanagrafiguren 2 ff.Google Scholar

⁸ The need for a universally accepted range of terms covering the different types is very pressing, particularly for the earlier periods.

⁹ The restrictiveness of a typological classification is particularly evident when applied to material of archaic date.

¹⁰ Kleiner, op. cit. See especially the review in AJA LIV 440 ff.

¹¹ JHS XLIX 38 ff., ARV 892 ff.

¹² The mere physical union of two heads on a Janiform vase is in itself, independently of stylistic criteria, naturally no certain evidence for a common attribution of the two originals, and this is particularly the case where one or both of the heads is not of the first generation (see pp. 219–20).

¹³ BSA XXXII 23 ff. This article is also of importance for the advances it makes in dealing with the problem of derivative foreign production (see p. 221).

¹⁴ Op. cit. pls. 14 and 15. The application of our rather different criteria suggests the need for some minor changes. Pl. 13, 3 and perhaps also pl. 14, 1 would seem to belong better with Class E, whereas pl. 14, 2 is perhaps rather happier in Class C from which, incidentally, pl. 13, 4 probably needs to be exiled (Dunbabin, , Western Greeks 277, n. 4Google Scholar). As not all this material has been accessible to me I can make only tentative suggestions at this stage.

¹⁵ See p. 221.

¹⁶ It is, of course, impossible now to estimate the volume of production, but it seems likely that the number of pieces derived directly or indirectly from a single archetype not infrequently amounted to several thousands.

¹⁷ E.g. by Martha, Cat. des fig. en terre cuite du Mus. de la Soc. Arch. d'Athènes, Introd. xxiii.

¹⁸ ‘Abformung und Typenwandel in der antiken Tonplastik’, Opuscula Archaeologica Inst. Rom. Reg. Suec. II 1 ff. For earlier references to the problem see especially 1 n. 1, 13 n. 3, 21 n. 4 in her study. Terracotta figurines, as opposed to reliefs, are treated more particularly on 21 ff.

¹⁹ Cf. Knoblauch, , Studien zur archaisch-griechischen Tonbildnerei in Kreta, Rhodos, Athen und Böotien 105Google Scholar; Kleiner, op. cit. 4 ff.; Neutsch, op. cit. 3 ff.

²⁰ See p. 222.

²¹ Jastrow, op. cit. 2 ff.; Neutsch, op. cit. 6. The degree of shrinkage varies according to the clay, its preparation, and the technique (e.g. thickness of walls) of the figurines themselves, but this variation is not sufficient to influence a division into generations as here made. A greater difficulty in this regard arises from the diversity presented by certain ‘parallel moulds’ (on which see p. 223). Most of the shrinkage occurs when the piece is drying in the air, only a little in the firing. Thus, of the two small protomai shown on Plate 44 (b1), only the right-hand one has been fired, the other having merely been left to dry in the air for a few days. The difference in size is only slight.

²² Because of the totally different meaning it has acquired (see p. 217) ‘type’ can no longer be used to express this mechanical relationship. ‘Series’ is employed by Knoblauch (op. cit. 105) with something approaching our meaning, while Kleiner's use of ‘Reihe’ (op. cit. 3 and frequently) also to some extent adumbrates our term.

²³ I wish here to record my gratitude to Mrs. I. Triandafillidhis of Amarousi, Athens, for her kindness in affording me the facilities for producing and firing the modern pieces illustrated on Plate 44. They are in Attic clay and have been baked in a wood-fired kiln at an estimated temperature of 850–900° C. The fact that some of the pieces have a texture approximating to actual archaic ones whereas others appear to have been fired at a slightly higher temperature than their ancient counterparts (e.g. the right-hand one of the two protomai on Plate 44 (b1) has been baked to the hard red commonly found among Attic terracottas of the Roman period) suggests that in archaic times a lower temperature was used, though perhaps not one very much lower. The shrinkage of the seated statuette shown in Plate 44 (a) is as follows. The measurements are in metres.

²⁴ Jastrow, op. cit. 21 ff.

²⁵ Frickenhaus, , Tiryns I 86Google Scholar; BSA XXXII 33 ff.

²⁶ E.g. MA XXXII 210; Van Ufford, Terres cuites siciliennes 62 ff.

²⁷ Winter, op. cit.; unfortunately his estimate of identity frequently requires checking, especially in the case of the archaic pieces. Likewise the very useful bibliographical and other information in the catalogue to Knoblauch, op. cit., is occasionally vitiated by a somewhat irresponsible use of the terms ‘identisch’ and ‘ähnlich’.

²⁸ E.g. Hesperia XI 380 ff., no. IV. a.5.; Hesperia Suppl. VIII 353 ff. (to which should be added Danish National Museum, no. 7769, Breitenstein, Cat. of Terracottas, no. 351 and pl. 40).

²⁹ ActArch VIII 1 ff.

³⁰ Cf. Neutsch, op. cit. 6.

³¹ Very simple work of this kind has been done on the mould producing the small protomai in Plate 44 (b1). Fineline incision is sometimes also employed on the archetype, but much less frequently and with much less striking results. It is usually readily distinguishable from incision on the mould and, as it normally appears on all moulds from the same archetype, has no relevance to the question of identifying the different parallel moulds discussed on p. 223. Where both processes are employed to the full one may propose the following general distinction: intaglio-work on the mould results in sharp raised masses on the finished article, whereas archetype-work produces sharp sunken ones as well as laying out the broader more even areas.

³² Except, of course, for the possibility of much later re-issues from an archetype that had been preserved. Moreover, it seems likely that in some cases at least the archetype and first generation moulds were produced by a professional artist who also worked in other materials, and under such circumstances we may regard the actual creative process as being somewhat sundered from the mass-production that next took place in the coroplast's workshop. This is a question that will be considered more fully in my subsequent study. Here let it suffice to draw attention to the definite lack of sympathy between the modeller of the original and the mechanic who did the painting on, e.g., many archaic Attic terracottas, and the clumsy adaptations and even actual misuse of good moulds observed by Wolters, (Corolla Curtius 96 ff.).Google Scholar

³³ Hesperia IV 370, fig. 26; VIII 285, fig. 1.

³⁴ Hesperia VIII 285 ff.

³⁵ This is not, of course, to imply that all fine-line incision on ordinary bronzework was necessarily done after casting; for example, one may reasonably suspect that the hair on the Acropolis bronze no. 6445 (De Ridder, no. 740; Langlotz, Bildhauerschulen, pl. 35, middle) may have been rendered by a series of overlapping groups of wavy lines cut with a fine comb-like instrument on the wax model from which the bronze was cast. But in such cases, too, the incision is done, so to speak, on the positive and not by intaglio-working of the negative.

³⁶ Hesperia VIII 287. And so, most recently, Züchner, in JdI LXV–LXVI 199.Google Scholar As to Züchner's revival (op. cit. 202) of Brunn's, suggestion (Kleine Schriften II 102)Google Scholar that Melian reliefs were made from moulds, perhaps of wood, that had been cut out in intaglio, here it need only be stressed that intaglio-working on the mould plays a very important part in the production of these reliefs, but that it is surely done on moulds of clay, themselves taken from clay archetypes.

³⁷ Cf. the pieces shown on Plate 45. It is to be hoped that my Attic article will make many other examples known.

³⁸ An error in the extent of the ear-blank has not been concealed as it was felt that this, too, might be not without its technical interest. It might incidentally be mentioned that the use of green fuel in firing caused this archetype's Attic clay to turn a green colour not unlike that of Corinthian!

³⁹ Jastrow, op. cit. 5, 26.

⁴⁰ The pieces shown all belong to the Acropolis Museumxs collection of terracottas, and the photographs used have been taken with the kind help and permission of the ephor of the Acropolis. They are: (a) no. 462; (b) no. 978; (c) no. 498; (d) no. 94; (e) no. 450; (f) no number; (g) no. 494; (h) no number.

⁴¹ Perhaps a little further incision was done on one or two of the figurines after moulding, but its rarity is shown by a consideration of the numerous exact replicas.

⁴² In the case of these deep archaic Attic single moulds in particular, not only do both these operations need to be carried out before shrinkage has had time to set in and produce distortion, but also the clay must still be very pliant when taken out, since otherwise the stresses exerted on it in its withdrawal tend to cause it to crack. A little practical experimentation shows the need for some modification of the common assumption (e.g. Schneider-Lengyel, , Griechishe Terrakotten 11Google Scholar; Neutsch, op. cit. 7) that the clay was left to dry in the mould before withdrawal.

⁴³ The diagram restricts itself on the whole to the material illustrated, and under such circumstances cannot show the whole of either dimension of the series. I know of no first generation example of F.2.a. and assume provisionally that it is the product of a second generation parallel mould. As an example of the third generation (F.3.a.) Agora T. 2210, might be quoted.

⁴⁴ p. 217.

⁴⁵ p. 225.

⁴⁶ At times, however, individual series appear to stand quite by themselves with no readily identifiable stylistic ‘relatives’. In such cases we may regard the limits of the group and the series as coinciding.

⁴⁷ Consistency, however, is essential. One typical example should suffice. In dealing with archaic terracotta heads two measurements have been found to be particularly valuable. Both are taken with dividers (with wooden instead of metal points where necessary!). They are: (a) the distance from the centre of the mouth to the inside corners of the eyes (a mean where there is variation); (b) the distance from under the chin at the point where the underside of the jaw joins the column of the neck to the top centre of the forehead where the hair or veil begins.

Both these measurements show a small degree of non-significant variation due to distortion of the soft clay either of the mould or of the figurine, while, in the case of (a), different intaglio renderings of the eyes on parallel moulds may cause further small variations. One has also to allow for differences in the shrinkage of the clay, though these seem usually to be insignificant. However, despite these small elements of variation, these two measurements used with discretion provide a valuable key for checking mechanical attributions and for identifying the different generations involved.

It will be noticed that both these measurements are more or less vertical ones. Horizontal measurements of heads are of much less general use because of the extent of the distortion and non-significant variation one gets in this plane, particularly in the case of thin-walled protomai.

⁴⁸ p. 217.

⁴⁹ Different ‘adventures’ of this kind will be examined in my archaic Attic study.

⁵⁰ An analogy, though only a partial one, may be drawn between the relationship of type to both group and series and that of vase-shapes to the stylistic criteria of vase-painting.

⁵¹ Cf. the breadth of our definition of the group, p. 218 ff.

⁵² With a few interesting exceptions, until Hellenistic times mechanical evidence of this nature is not extensive enough to enable the systematic consideration of the pattern of the employment of derivative moulds by specific workshops.

⁵³ I.e. apart from the limited evidence from ‘signed’ pieces, and, where it applies, vastly supplementing it.

⁵⁴ p. 221.

⁵⁵ p. 221.

⁵⁶ Jastrow, op. cit. 10, 26.

⁵⁷ Cf. particularly Hesperia XI 365 ff.