Acknowledgements. The original idea for this article was suggested by Mr. M. R. Popham, and has grown further from discussion with him, Mr. M. S. F. Hood, and Dr. P. Warren. I should like to thank them all for their help and guidance both in exploring the complexities of the matter and in the actual presentation of the article. My thanks are also due to the Ashmolean staff for giving me the opportunity in the first place to study the Day Books of the Knossos excavations and Evans's collection of photographs, and allowing me to publish some here.

Abbreviations used in the article, in addition to those in common use in the BSA, are as follows:

RIBA Evans, A. J., ‘The Palace of Knossos and its Dependencies in the Light of Recent Discoveries and Reconstitutions’, Royal Institute of British Architects, xxxvi, 3 for 8 December 1928, 93–102.Google Scholar

OKT L. R. Palmer and J. Boardman, On the Knossos Tablets (Oxford, 1963).

PP Palmer, L. R., The Penultimate Palace of Knossos. Incunabula Graeca xxxiii (Rome, 1969).Google Scholar

DPK Popham, M. R., The Destruction of the Palace at Knossos. Studies in Mediterranean Archaeology xii (Göteborg, 1970).Google Scholar

¹ ‘Mycenaean Inscribed Vases III: the Consequences for Aegean History’, Kadmos xii (1973) 60–75.

² Boardman, , OKT 47 n. 1 and pls. xi, xii.Google Scholar Palmer raised objections to this point in PP 41–3, but his fullest treatment of the matter can be found in Kadmos xii 64—7 fig. 1.

One should also mention here an article by Woodard, W. S. in AJA lxxvi (1972) 113–25CrossRefGoogle Scholar, which deals generally with the North Entrance region. So far as the ‘baulk’ is concerned, his general views tally with the Boardman/ Popham approach. In one detail only does he significantly differ (118), in that he attributes the two boxes of J II 10 to the ‘baulk’, but is not so certain about J II 5. Despite Woodard's claims, Boardman (OKT 47) actually equates only J II 5 with the ‘baulk’, as does Popham, but neither of the J II 10 boxes. Apart from that point, there are only minor discrepancies, by and large not worthy of mention.

³ These photographs and some negatives are now housed in the Archive Room at the Ashmolean Museum, Oxford. They are in the process of being sorted and mounted in albums—having become disordered over the years. Together they provide, albeit somewhat patchily, a record of the dig at Knossos, and more briefly of other sites and finds from Crete, Mainland Greece, and even the Near East.

In considering their testimony and chronology, dates can seldom be fixed from publication elsewhere in known contexts. Rather one is compelled in part to compare the state of excavation with the written account in McKenzie's notebooks, BSA reports, and the various publications of Evans; and in part from the internal evidence of the photographs themselves.

⁴ One of the two photographs is reproduced by Boardman in OKT as pl. xi. A detail of the other, taken from the direction of the NE., is here reproduced as PLATE 9, and shows quite clearly the composition of the ‘baulk’. For an account of the excavation of the NEP, see Palmer, OKT 119–27, where relevant extracts from McKenzie's Day Books are quoted at length, and Boardman, OKT 45–9, for a more succinct account.

⁵ This is the plan compiled by Fyfe and first published in BSA vii (1901–2) 5 fig. 2, and later partly relabelled and reproduced in PM i, 397 fig. 286. A rough check can also be made by comparing PLATES 9 and 10b with the elevation in PM iii opposite 171 fig. 114, compiled by Piet de Jong. By measuring off the plates against the plan much the same dimensions are arrived at.

⁶ In some respects, PLATE 9 is the most important of all the photographs in that it shows clearly the nature of the ‘baulk’. It is not solid stone, but rather a homogeneous mixture of earth and a debris of small stone and sherds. Can it really be termed a foundation? And for what?

⁷ These photographs can be dated largely by indirect methods, though 1901 is well attested in McKenzie's Day Books as the year in which digging revealed the more northerly portions of the paved way of the NEP. PLATE 10a receives confirmatory evidence of its date, first in the unfinished state of the flat-roofed cover for the Throne Room, erected in 1901 (cf. BSA vii (1900–01) 2); and second in the absence of the support wall built round the Central Court in 1902. PLATE 10b and some other general shots not reproduced here not only show blocking walls still in position that are known to have been removed in 1902, cf. Fyfe's 1901 plan (BSA vii (1900–1) frontispiece), but also do not show signs of excavation at the northern end of the Eleven Pillar Hall and Threshing Floor Area beyond that, which were both explored in 1902 (BSA viii (1901–2) 3–8).

⁸ The stone-built drain was known to extend northwards some considerable distance, and had been explored from within in 1900 by some of the Cretan workmen, BSA vi (1899–1900) 50.

⁹ Of these photographs, a general shot is to be found in BSA viii (1901–2) 7 fig. 3, in which is just discernible the top of the ‘baulk’, the stone crossway, and a small part of the surrounding support wall of the Central Court, PLATE 11a is similarly a general shot, but being taken from a slightly different angle clearly shows the ‘baulk’ and the Central Court support wall, PLATE 11b is most likely to date to 1902—for the Threshing Floor Area is being investigated, and both the access way and, at the bottom corners of the photograph's vertical edges, the Central Court support wall are visible. But there is still a remote possibility that the photograph could date to very early in the 1903 season.

¹⁰ Lagrange, M-J., La Crète ancienne (Paris, 1908), opposite 8, pl. ii.Google Scholar Though the book was itself published in 1908, and written as a result of travels in Crete a year or two earlier, the photograph in question cannot really date later than 1903. A photograph in BSA ix (1902–3) 105 fig. 69, taken to illustrate the Theatral Area dug and reconstructed in 1903, shows the Theatral Area, the wood framework south of the NW. Portico, and the newly heightened Central Court support wall and the photographic tower near the Domestic Quarter—all features observable in Lagrange's plate. At present this view is not paralleled among Evans's photographs stored at the Ashmolean.

¹¹ This is the photograph reproduced by Boardman in OKT as pl. xii, which he dates to 1902 on account of the state of excavation. However, comparing it with pl. ii of Lagrange, it is clear that the two were most probably taken at about the same time in 1903—compare the state of the Machriteichos house in particular.

In the Ashmolean albums, there is another shot, similar in all respects, bar the presence of a small mobile photographic tower, to pl. xii of Boardman. This is labelled 01.81, and other similar identification marks are found on quite a few other photographs. They are either a simple two- or three-figure number (e.g. 63, 630, 637) or have a prefix 01 or 02 (e.g. 01.81, 01.44; 02.238). These are perhaps unlikely to represent 1900, 1901, 1902—even allowing that Evans may have added them much later when his memory of events and seasons had become blurred—for, as here, the difference in the state of the excavation between the 1901 and 1903 seasons is surely too great to be confused, especially with the adequate visual record of the photographs themselves before him. It is more likely then that they are an arbitrary series of numbers, perhaps for indexing or publication purposes. On the other hand, Mr. Popham has pointed out that some of the boxes of pottery stored in the Stratigraphical Museum have labelling that includes figures like 01 and 02, which in this case seems to indicate the year of excavation, cf. DPK 13.

¹² e.g. for 1903—cf. BSA ix (1902–3) 3; 105 fig. 69. The history of the photographic tower beside the Domestic Quarter is a complicated one: there seem to have been at least two, with various ‘sub-phases’ of use and disuse. The evidence for these is largely contained in the Ashmolean photographs, though a little can be found scattered throughout such publications as BSA, PM, and others.

¹³ Plans were ‘in hand’ by late 1928 for the work of reconstruction in the area (RIBA xxxvi, 3, 8 Dec. 1928, 99). Work of some sort was undertaken in that region in 1929 to judge from the remarks in Evans's Day Book (1929 Vol. i 3, 4, 7, 9). But by early 1930, or a trifle earlier, when PM iii was published, Piet de Jong had only ‘executed a series of restored plans and elevations’, which is borne out by the photographs of the NEP in that volume. Interest ingly, though, a photograph in the Ashmolean collection shows a canvas painting of the NEP by N. Lambrinides, dated 1930, in which the reconstruction is complete. The painting looks to have been copied from a photograph, of which a copy exists in the Ashmolean and part of which is reproduced in PM iv, i 15 fig. 7, and thus dates the completion of the work. There are plentiful photographs in PM iv, i of the finished state, e.g. 7–18 figs. 3–8.

¹⁴ This stage is represented by PM iii 159 fig. 105.

¹⁵ Kadmos xii 65–6: ‘The point (concerning the ‘baulk’) must be considered with all the greater care because the opposing sides are now within touching distance of complete agreement and the whole issue of the date of the tablets and the constructional history of the Last Palace can be decided here.

¹⁶ Kadmos xii 64–7.

¹⁷ There is one theoretical possibility that Palmer does not conceive of, namely that the layer above the stone NEP surface might have been an artificial dump of soil removed from elsewhere in the Palace site, and that therefore all the matter was intrusive. However, there seems no reason to assume that this had been the case.

¹⁸ Kadmos xii 66. In fact, this argument of Palmer's seems internally inconsistent. The West Bastion he puts as of LM II or later construction (Kadmos xii 66), and both the stone NEP surface and the later earth surface as of LM IIIB. It thus requires some mental gymnastics to reconcile these datings with the ‘foundations’ being both earlier than the West Bastion, which rests on or is cut into them, and yet later than the stone surface of the NEP, which he seems to imply underlay them. Further, it is not easy to appreciate why in cutting through the ‘foundations’ in the course of erecting the West Bastion such an unsightly mass should have been allowed to mar the effect of the fine ashlar masonry.

¹⁹ PP pl. 1 a and Kadmos xii 66. Actually Fyfe must have substantially completed at least the NEP region for this plan in 1902, as in BSA viii (1901–2) 5 fig. 2 there appears a detailed plan, which is obviously of the same basic format as that of 1903.

²⁰ There is just one possible instance in McKenzie's Revised Day Book of 1900, fig. 52 for Tuesday, 15 May, where the area of the deposit of the tablets, marked by xxx's, does not go right up to the Bastion in the south part of the sketch plan. However, it is difficult to tell whether or not a sufficient depth had been reached at this point to equate this with the ‘baulk’. Moreover, during the work of the next ten days or so, other plans which show either the spread of the tablets or the uncovering of the large drain do not respect or mark the limits of the area where the ‘baulk’ ought to be.

²¹ The relevant plans and text in PM come in four main passages: PM i 204–6 fig. 152, 393–400 figs. 284–8; PM iii 158–67 figs. 105–9, 114; PM iv, i 7–16 figs. 3–8. In the annual BSA reports, passages concerned with the NEP are: BSA vi (1899–1900) 48–51; BSA vii (1900–1) 68–72; BSA viii (1901–2) 3–8; BSA ix (1902–3) 23 fig. 11.

²² Palmer (OKT 237) in discussing Boardman's hypothesis, implies that Boardman attributed its existence to the Linear B tablets deposit:

‘The baulk … marks the edge of the area where the tablets were found. So far there can be no serious grounds for disagreement.’ This, as n. 20 above indicates, is difficult to uphold with certainty. Another viewpoint is to suppose it to be in some way connected with the discovery of the large drain running beneath the paved NEP surface (first reported in McKenzie's Revised Day Book on Wednesday, 23 May 1900, fig. 60). Another equally valid interpretation is that the ‘baulk’ was left as a sort of stratigraphical key. Whatever the truth may be, the fact that it was left need cause no surprise, since archaeology is not the sort of science necessarily to demand in all cases the total removal of one level before dealing with the next.

²³ Evans's Day Book 1913 Vol. i 110: ‘79. Outside passage in N. Entrance. Preponderantly MM III.’ This is elaborated in Pendlebury, Eccles, and Money-Coutts's ‘The Dating of the Pottery in the Stratigraphical Museum’ ii 8 as: ‘MM III; some LM I–LM III; one Neolithic’.

²⁴ Pendlebury, Guide to the Stratigraphical Museum 16—for the code (J II 5), location and year. Pendlebury, Eccles, and Money-Coutts, ‘The Dating of the Pottery in the Stratigraphical Museum’ ii 8—for details of the material. Pendlebury and Money-Coutts, ‘The Dating of the Pottery in the Stratigraphical Museum’ iii plan 14—for the position of die TP.

²⁵ The evidence for the equation is set out in detail in Popham DPK 43; and also by Palmer PP 39. Popham has suggested, DPK 13 and again in Kadmos xiii (1975) 123, postscript, the possibility that the codes on the labels of the storage boxes in the Stratigraphical Museum indicate their provenance above (A) or below (B) a floor level. In the case of TP 79 (J II 5), it carries the coding 98A. If Popham is correct, this should indicate that the source of the material was above floor level. Is it not a reasonable supposition that the floor level referred to is the paved surface of the NEP, and that accordingly the ‘baulk’ is meant, and a ‘baulk’ of earth and debris, not some stone foundation?

²⁶ Indeed, it has already been pointed out above that at its south end the ‘baulk’ does seem to continue below the level of the paved NEP surface by perhaps as much as one to one and a half ft., see PLATES 9 and 10b.

²⁷ Popham DPK 44. For Palmer's acceptance of Popham's criteria concerning his classification of the pottery see Kadmos xii 62.

²⁸ If one is correct in the assumption that part of the TP 79 material comes from above the NEP paved surface, and part from below, then it might be this test that Evans employed to give or confirm a date for the underlying drain. This would deal with Palmer's observation, PP 39: ‘This lower floor (the paved NEP surface) was dated by Evans to MM IIIA, but he offered no ceramic evidence from test pits below the pavement.’

²⁹ Cf. n. 26 above.