1 The project was funded by the Deutsche Forschungsgemeinschaft and generously supported by the Soprintendenza Speziale per il Colosseo, il Museo Nazionale Romano e l’area archeologica di Roma and the Materialprüfungsanstalt in Bremen.

2 The samples originate from different wings of the complex, mainly the Domus Augustana, the Stadium and the Domus Severiana, and were taken in the years 2005, 2008 and 2010. For the provenance and a list of the samples see Schmölder-Veit et al. 2016: 332–337, Tab. 1, Fig. 1-5.

3 DTA = differential thermal analysis; ESEM = Environmental Scanning Electron Microscope; TG = thermogravimetry; XRD = X-ray diffraction; EDX = energy dispersive X-ray spectroscopy.

4 Because of the condition of the break, it was impossible to take samples from these mortars. Schmölder-Veit 2016: 315–320, Fig. 29–32 with further reading; Schmölder-Veit et al. 2016: 355.

5 Possibly because of its very unusual position in the center of the basin, the drain did not function very well and had to be relocated to the outer rim.

6 The courtyards were built in the early Flavian period, but because of the position of the basin’s different drains, it is evident that the bottom layer of mortar belongs to the first basin dating from the reign of Domitian; for details of the archeological record see Schmölder-Veit 2012: 190, 196-198 (basin 337); for the exact position of the cavity see D’Elia, Le Pera Buranelli 1985: 177, Fig. 1.

7 Schmölder-Veit 2012: Fig. 16; Schmölder-Veit, Kilian 2012: 223, Fig. 1. 4; see below.

8 In other fountains the raised floor can be extrapolated from the replacement of the drain, see Schmölder-Veit 2012: 196-197, 199, Fig. 14, 17,18.

9 It might be coeval with modifications in the post-Hadrianic period, when shallow basins were incorporated into the floors of all four porticos of the sunken peristyle (Schmölder-Veit 2012: 201-212, Fig. 23, 24).

10 Schmölder-Veit 2016: 296–305 with further reading; Schmölder-Veit et al. 2016: 334, Fig. 2.

11 Schmölder-Veit 2016: 298–301; Schmölder-Veit et al. 2016: 334 f. 352. 356, Fig. 2, sample 2010_27.

12 At first glance, this mortar appeared to be very similar to the first layer in the basin in the sunken peristyle (see above). But as it was not possible to take a sample in the Domus Augustana, the similarity was not verified by microscopy analysis.

13 The middle layer might stem from a renovation in the Antonine period (thickness 8 cm; Schmölder-Veit 2016: 301–303 and Schmölder-Veit et al. 2016: 334, Fig. 2 sample 2010_06), while the bottom layer is of unknown date and shows at the moment no similarity to other alterations (thickness 12 cm; Schmölder-Veit 2016: 301 and Schmölder-Veit et al. 2016: 334, Fig. 2 sample 2010_07; interpreted as an intermediate phase of unknown date).

14 Schmölder-Veit 2016: 304 and Schmölder-Veit et al. 2016: 334, Fig. 2 sample 2010_02.

15 Lancaster 2005: 13–16, 58 f., 211 f., 246 no. 12, Fig. 43.

16 Bipedales in these basins were not used as a waterproofing layer, but to facilitate the accurate slant of the basin floors (see above).

17 Augustus (Domus Tiberiana: Tomei 2012: 304, Fig. 3); Nero (Domus Aurea: Fabbrini 1982: 8, Fig. 7), Domitian (Domus Augustana: Schmölder-Veit, Kilian 2012: 220-223, Fig. 1. 2); Hadrian (Stadium: Schmölder-Veit et al. 2016: 334, Fig. 2 sample 2010_09 and 2010_10), Septimius Severus (Domus Severiana, see below and Schmölder-Veit 2016: 321) or under Diocletian in his monumental public baths (Friggeri, Magnani Cianetti 2014: 135, Fig. 4).

18 Schmölder-Veit 2016: 306-315 and Schmölder-Veit et al. 2016: 335-336, Fig. 5 sample 2010_22, 2010_23).

19 White marble plaques were only attached to the second layer (Schmölder-Veit 2012: 208-210). A similar case might be preserved in the residence fountains (= Ninfeo del Palazzo) in the Villa Hadriana with a blue colored basin, and a second top layer as foundation for the marble slabs (Salza Prina Ricotti 1998: 386–389).

20 Schmölder-Veit 2012: 194-196, 198-199, 202-203, 208; Schmölder-Veit, Kilian 2012: 219-227.

21 Schmölder-Veit 2012: 198-199; Schmölder-Veit, Kilian 2012: 223.

22 Bitterer 2012: 234-235, Fig. 3.

23 On the Palatine Hill there might be only one exception: The reservoir in the eastern part of the Palatine was proofed with two layers of quite identical mortar. Because of this similarity and the building’s function as a reservoir the two layers might have been applied at the same time (Schmölder-Veit 2016: 320-322 and Schmölder-Veit et al. 2016: 355 samples 2010_18, 2010_19).

24 Laura Thiemann provided the analysis of the cross sections. Frank Schlütter was dedicated to further analysis on thin sections, at the ESEM and chemical analysis at the Amtliche Materialprüfungsanstalt Bremen. Marie D. Jackson, University of Arizona, attended to the research on the pozzolane.

25 The number of samples for further analysis was reduced due to partly small sample sizes. Besides, the preceding grouping allowed to choose representative pieces. The smaller amount of samples could thus be more thoroughly analysed.

26 The following samples were selected: 2008_05; 2008_06; 2010_02; 2010_03; 2010_04; 2010_05; 2010_06; 2010_08; 2010_09; 2010_10; 2010_11; 2010_14; 2010_16; 2010_18; 2010_19; 2010_20; 2010_26; 2010_27.

27 XRD/ DTA, chemical analysis and ESEM exemplary on samples 2010_03 and 2010_18.

28 Ringbom, Hohlfelder 2011; Lindroos et al. 2011; Langley et al. 2011.

29 First research was done by Heinz-Otto Lamprecht (1984: 41-67). More recent analysis is presented for example by Heinrichs, Gerdes, Schön 2014:123-134 and Jackson et al. 2009: 2481-2492.

30 Cyclododecane is a alicylic hydrocarbon which evaporates at room temperature.

31 For characterization of the properties of the mortars a terminology used in geology and petrography was applied (see also Lenz 2000: 5-18). Descripiton of the colour of the samples was always done on dry samples and based on the Munsell Rock-Color Chart, 8^th edition, 1995.

32 Schmölder-Veit et al. 2016: 352, Tab. 3, 354 Tab. 4.

33 Sample 2010_03; Sample 2010_18.

34 Binding medium: Aggregate Sample 2010_03 (1 : 1); Sample 2010_18 (1 : 0,8).

35 The exact identification of the hydraulic binding medium was not possible, but the presence of lime lumps and the almost exclusively use of volcanic additives leads to the conclusion, that volcanic ash might have played a crucial role concerning the hydraulicity of the binding medium.

36 See also Jackson et al. 2010: 47: Munsell-scale 10YR 4/4 and 10R 2/2 - 5R 3/4.

37 See also Jackson et al. 2010: 47: Munsell-scale 10YR 4/2 and 5YR 5/6.

38 Fig. 1 and Schmölder-Veit 2016: 320-322, Fig.34-35.

39 Fig. 1 and Schmölder-Veit 2016: 317-319 Fig. 30-32. For the whole group see Schmölder-Veit et al. 2016: 352-353, 355, Tab.3

40 See also: Jackson et al. 2009: 2481–2492; Jackson et al. 2010: 36–74; Jackson et al. 2011: 728–742.

41 Carbonates, Mica and other finer clay minerals in the matrix of the crushed ceramics are still intact, which proves that they can not have been heated over 800 °C. On the other hand there was no trace of Kaolinite in the matrix, which was surely in the primary material. Kaolinite is decomposed at a temperature around 570 °C.

42 Siddall 2011: 163, 165 f.; Poblome, Schlitz, Degryse 1998; Peña 2007: 263, 267 f.