1 Flow charts with further indications on stratigraphical relations are presented as annexes at the end of the chapter.

2 The mixed date attributed in preliminary pottery analyses is probably due to small worn out sherds deriving from the disintegration of mudbricks.

3 This also raises the question of the slope apparent in the area east of Building CD. There, it is clear that, nowadays, the level of the summit is gently sloping down towards the east. Is this slope the consequence of the progressive erosion and/or the result of disruptive agricultural practices or is it purely natural and already existed during the Bronze Age?

4 Joints were made between 15-04-0008, 15-04-0020, 15-04-0026, 15-04-0027, 15-04-0087; also between 15-04-0013 and 15-04- 0015.

5 Stratigraphically speaking, in both trenches, the situation is similar: the lowest level is characterised by MM II-III occupation directly on the bedrock over which a levelling fill – with many some traces of a destruction level with wall collapse just beneath – was laid out in preparation for a LM IB floor, particularly well preserved in trenches 1S and 1SE.

6 Stratigraphically speaking, in both trenches, the situation is similar: the lowest level is characterised by MM II-III occupation directly on the bedrock over which a levelling fill – with many some traces of a destruction level with wall collapse just beneath – was laid out in preparation for a LM IB floor, particularly well preserved in trenches 1S and 1SE.

7 The presence of such a feature is a recurring characteristic in many spaces of Zone 4 occupied during the pre- and Protopalatial periods (see below).

8 Together with deposits of a similar date found in Quartier Pi at Malia (Alberti et al. 2016), this MM III deposit – the first at Sissi – will certainly contribute to refine our understanding of this very important period in the region (see Macdonald & Knappett 2013).

9 The nature of the sediment suggests that this was a levelling fill/floor packing for the LM IB floor.

10 The following preliminary remarks concerning the ceramic material were made by C. Langohr and I. Caloi and slightly reworked by the author. I thank them warmly for their help and their contribution to this chapter.

11 The soil of this unit was fully sampled in the hope that information on the composition of the mudbricks may be collected. More specifically, the presence of small sea shells might be indicative of the use of sea grasses as temper in these mudbricks (Devolder 2005-2006: 67-69, fig. 2; Lorenzon 2016).

12 Representations of phalli are relatively unusual in Minoan Crete. A possible comparandum for 15-04-0012-OB003 was found in room 3 of Building B2 at Mochlos (Soles & Davaras 1994: 408; 1996: 189).

13 These slabs and sherds lying flat appeared to continue beneath the flimsy partition.

14 Nevertheless, the stone tools and the stone feature were resting a few centimetres above the reddish level, so the possibility that they were associated with another floor surface or a relaying of the lower one cannot be excluded.

15 Units 15-04-0011, 15-04-0019, 15-04-0024, and 15-04-0031 (see table 4.3 in Annexes).

16 The presence of pieces of charred wood and charcoal is a good indication of the presence of such wooden planks in the niche.

17 Some of these olive pits clearly seemed to come from the cups themselves.

18 This change in the soil matrix (which also became much softer and brown) may also be explained by the existence of a wooden floor on which the deposit from the niche would have fallen. This wooden floor would have burnt – leaving the clear traces of burning, ash and patches of charcoal – and protected the level on which it was laid out. Incidentally, it is only below unit 15-04-0031 that earlier sherds (EM II, MM I-III) were identified.

19 If there was indeed a wooden floor over the bedrock, the latter might only have been used as a surface during these earlier periods.

20 In all likelihood, at some point during the Neopalatial and before the final conflagration which sealed room 4.19, the staircase went out of use and refuse material was dumped in the stairs which spilled over in the basement space (see below).

21 See above for the wooden floor hypothesis.

22 Did an original wooden sottoscala burn and was then covered by the collapse of neighbouring walls that extinguished the fire? One element in favour of such a hypothesis is the fact that evidence for fire became more evident as we went deeper within the test. Furthermore, with wall D97, it is tricky to decide where the staircase actually started (see below).

23 Such a large amount of ash and fragments of charcoal could well indicate that at least part of the staircase was made of wood, whether it included actual wooden steps or was simply built on a wooden framework (Shaw 2009: 95).

24 The fact that several conical cups were found in stacks is in favour of such a hypothesis (see, for example, fig.4.69).

25 The slabs and the plastered steps found in other units could then come from the upper flight of the staircase, from built-in shelves, or, simply, from the sporadic use of stone elements in an otherwise wooden structure.

26 The fact that the courses of the west part of wall D55 seem to rest on soil rather than on bedrock just west of the gap contributes to corroborate this hypothesis.

27 Further indications of later activities in the area is the presence of post-Minoan ceramic material with joining fragments which dominate the poor pottery assemblage associated with unit 16-04-0252.

28 On figure 4.92, the dotted lines indicate the extent of each unit.

29 At some point, we wondered if this wall may not have been part of the missing east façade of Building CD, of which room 4.13 would then have extended further east than we originally thought and covered the street level and, presumably, the kerb. The type of masonry of this wall is reminiscent of some larger walls of Building CD and the presence of LM III material in the units opened in its vicinity are in favour of this hypothesis. This, of course, requires more detailed analysis.

30 As suggested in the previous note, this can tentatively be explained by the existence of a very poorly preserved LM III level (presumably associated with a Postpalatial phase of Building CD) mixed with building material (the earlier and eroded sherds come from mudbricks) which progressively gave way to a more homogeneous Neopalatial horizon (i.e. the fire destruction on top of the street level and the last use level of the latter).