(PDF) Identification of potential structural coherence based on the ceramic material from provincial necropolis II at Ancient Ge

P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 Identification of potential structural coherence based on the ceramic material from provincial necropolis II at Ancient Gerulata (SR) using quantitative analyses Identifikace potenciální strukturální koherence na základě keramického materiálu z provinciálního pohřebiště II v antické Gerulatě (SR) za pomoci kvantitativních analýz – Alina Szabová*, Zuzana Porubčanová – 1. Site introduction KEY WORDS Ancient Gerulata is (as part of Limes Romanus) the only ar- chaeological site of its kind in Slovakia. It lies in the territory Gerulata – graves – Roman Period – pottery – volume – shape – statisti- of the present-day village of Rusovce, which is now part of the cal analysis – structural coherence capital, Bratislava. In the Roman period, it was a borderline mil- itary camp (later a fort) with its own civil hinterland (Schmid- tová, Mathédesz 2018, 58), which consisted of vici, villa rustica ABSTR ACT (Varsik 1999, 215) and the autochthonous agricultural settle- ment (Hlavatá, Varsik 2019, 429–430). Its immediate vicinity to The ceramic vessels from necropolis No. II were redocumented during the the border of the northern Danube and its location on the road ongoing revision of archaeological material from cemeteries belonging to the connecting Carnuntum (Petronell – Bad Deutsch-Altenburg, Roman military fort with civil hinterland – Gerulata. This necropolis con- Austria) and Ad Flexum (Mosonmagyaróvár, Hungary) meant tains the highest informative value for both archaeological and anthropolog- that Gerulata had not only a strategic and military significance ical material. Therefore, for the needs of this study, a set of pottery was cho- but played an important trading role between the Roman Em- sen from that particular site, which consisted of approximately 120 complete pire and the barbarian societies. Due to the presence of the ford various-shaped artefacts to be used for volume calculation. The vessels pri- across the river, the probability of a harbour placed somewhere marily served for offerings but were also partly used as urns. Drawing and around the camp is extremely high. Unfortunately, constant photo documentation was used to create 3D models of the individual arte- changes in the river flow across the centuries leave little hope facts. These visualisations were a data source for the quantitative measure- for finding its closer location (Schmidtová, Mathédesz 2018, ment of the volume variations. Models and volume calculations were created 57–58). Roman army units occupied Gerulata from the second in Blender open-source software before identifying a correlation between half of the 1st century AD (Varsik 1996, 22) until the end of the vessel shape and volume and the graves in which they were found. The main 4th century or even the beginning of the 5th century AD. Both subject of this research was to identify potential structural coherence within the civilian settlement and the cemeteries belong to the same the funerary habits using the univariate statistical analysis produced in time frame (Schmidtová, Mathédesz 2018, 61). A mild hiatus PAST 3 open-source software. during or shortly after the Marcomannic Wars can be observed inside and outside the camp in the form of a destructive burnt layer (Pichlerová 1983b, 74). * Corresponding author – E-mail address:

https://doi.org/10.47382/pv0621-04 Received 6 February 2021; received in revised form 8 April 2021. Available online 30 June 2021. Fig. 1. Topography of ancient Gerulata based on the present state of archaeological Copyright © 2021 Czech Acad Sci, Inst Archaeology Brno, and the authors. excavation. Based on Schmidtová, Mathédesz 2018, 57; Varsik 1996, 131, Abb. 8. This is an open access article under the CC BY-NC-ND 4.0 license (https://creativecommons.org/licenses/by-nc-nd/4.0/). Obr. 1. Topografie antické Gerulaty založená na současném stavu výzkumu. Podle Competing interests: The authors have declared that no competing interests exist. Schmidtová, Mathédesz 2018, 57; Varsik 1996, 131, Abb. 8. 91 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 The military camp had four main building phases through- (Hlavenková et al. 2015, 140–143). There is a definite possibility out the Roman period. The first, the earth-and-timber phase, that some of the burial grounds (especially II and III) are only dated between the second half of the 1st and the second half fractions of one large long-lasting necropolis (Varsik 1996, 36). of the 2nd century AD, was characterised by a typical dou- However, this will have to be proved by further excavations. Ci- ble-pointed ditch and a rampart with a wooden palisade vilians and soldiers were probably buried together without any (Varsik 1996, 24). This was possibly established by the Cohors differences (Pichlerová 1983b, 77). V Callaecorum Lucensium, according to the finds of stamped In this study, we used archaeological material from necropo- bricks (Schmidtová, Mathédesz 2016, 126–127). However, at lis II, which is situated next to the modern cemetery (Fig. 2). The the beginning of the 2nd century AD, Ala I Cannanefatium ar- necropolis has the highest number of well-documented graves rived from the Lower Rhine region, which caused construction (258) with 88 inhumation and 170 cremation burials discovered changes in the inner area of the camp. This unit remained in there to date (Hlavenková et al. 2015, 142). The inhumation bur- Gerulata until the beginning of the 4th century AD. During the ials were dominated by simple pits although brick tombs and Marcomannic Wars in the second half of the 2nd century AD, alcove or shaft graves were also present. The cremation graves a large stone castell was built in its place (Varsik 1996, 26). In consisted of simple pits, urn graves, pits with or without cre- this period, temporary camp No. II also existed about 300 m mation residuals and busta (Pichlerová 1981, 21–26). Burying from the permanent camp (Varsik 1996, 31). The third phase began there under the rule of the Flavian dynasty and ended lasted from the end of the 3rd to the first half of the 4th century some time at the beginning of the 3rd century AD (Pichlerová AD. The castell was partially reduced, and the fortification wall 1981, 17–18). The main excavation took place between 1968 and was added. The final phase is distinguished by the late antique 1973 although a couple of graves were already discovered in 1949 fort, which is dated from the second half of the 4th to the be- and earlier (Pichlerová 1981, 13). Several more graves were later ginning of the 5th century AD (Varsik 1996, 28–29). By then, identified in the 1980s (Pichlerová 1983a, 209) and the 2000s Gerulata was inhabited by a new unit called Equites Sagittarii – by rescue excavations (Schmidtová, Vorlíčková 2006, 175). The equestrian archers (Varsik 1996, 37). During the gradual decay unusual biritual character of this cemetery could have been hy- of Limes Romanus, these troops completely abandoned Gerulata pothetically caused in this time period by the presence of the au- (Schmidtová, Mathédesz 2018, 61). tochthonous population although a different explanation is also Several burial places were created during the Roman occu- possible. For example, the influence of eastern religious cults, pation, possibly along the roads leading from the camp (Fig. 1). the tradition of certain social groups, insufficient exploration of Five to six cemeteries (Hlavenková et al. 2015, 139–143) and similar sites or a local anomaly (Krekovič 1992, 73–75). several isolated graves have been discovered up to the pres- ent day in residential and industrial zones (Pichlerová 1983a, 2. Data set and formulation of hypotheses 208). Cemeteries No. Ia, II and V are from the Early Roman pe- A collection of 121 ceramic vessels (Tab. 1) deposited in the riod; the later phase includes No. Ib, III, IV and eventually VI Archaeological museum of Slovak National Museum (hereinaf- ter referred to as SNM) was chosen for the volume calculation. Most of the vessels were complete, but several were preserved only fragmentary. In some cases (e.g., 5, 23, 25), we were able to reconstruct the missing part of the original proportion due to the fieldwork documentation (archive of the Archaeological mu- seum SNM) and calculate the absolute volumes. However, some- times (e.g., 18, 67, 98) it was possible to only reconstruct a major part of the vessel, thereby only the minimum volume was calcu- lated. This collection contains typical fine yellow ware and both fine and coarse-grain grey ware (Krekovič 1998, 42–44). Terra sigillata, mostly from Lezoux, Rheinzabern and Westerndorf, was also present. Types such as Drag. 18/31 (No. 30), Drag. 33 (No. 36), Drag. 36 (No. 41) and Drag. 37 (No. 34) appeared in the spectrum (Gabler, Pichlerová 1996). Nevertheless, it was necessary to adequately classify the mate- rial and formulation of the key hypotheses before the volume cal- culation and statistical analysis (Neustupný 2007, 136). The five- shape group categorisation proved to be the most suitable and was also optimal for the following analyses. The first and quite homo- geneous group involved 26 jugs and pitchers (Fig. 3, 4) with heights between 15.8–40 cm, a maximum bulge of 10–20 cm, a mouth diameter of 5–8.33 cm and a base diameter of 5.9–8.98 cm. The second group consisted of 28 bowls and plates (Fig. 5, 6) with heights between 3.2–14.5 cm and diameters of 7.2–26.5 cm. Even though the bowls are much higher than the plates, their volumes are similar because of the generally wider diameter of the plates. All the representatives of terra sigillata belong to this shape group. Fig. 2. Grave distribution in the main part of the researched territory of necropolis II (excavation in 1949 and 1968–1973). The exact location of the later The third group is the largest (36 pieces) and also a quite uniform discovered graves was not published. Based on the original distribution plan from group, which contains convex forms, such as ovoid and pot-shaped Pichlerová 1981. vessels (Fig. 7, 8). The height of these vessels ranges between Obr. 2. Distribuce hrobů v hlavní prozkoumané části pohřebiště II (výzkumy v letech 1949 a 1968–1973). Přesná lokace později objevených hrobů není publikována. 9–22 cm, with a maximum bulge of 8–18 cm, a mouth diameter of Založeno na originálním distribučním plánu podle Pichlerová 1981. 8.2–15.5 cm and a base diameter of 3.6–8.1 cm. 92 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 No. Inv. No. Vessel Absolute Absolute Grave No. Burial type Grave type Gender Age Vessel location Grave shape height volume (Grave part) dating (AD) (cm) (ml) 1 21099 Jug 43 7208.14 63 Cremation Pit with cremation Unknown Adult ? 100–200 residuals 2 21052 Jug 39 5353.4 LXIII Inhumation Simple pit Male Adult Right lower p. 50–200 3 21209 Jug 37.5 5517.4 127 Cremation Simple pit Unknown Adult Middle p. 50–200 4 21237 Jug 36 4058.1 97 Cremation Bustum Unknown Adult Left lower p. 50–100 5 21283 Jug 34 4398.78 104 Cremation Pit with cremation Unknown Adult Left upper p. 100–150 residuals 6 21282 Jug 32 3060.71 104 Cremation Pit with cremation Unknown Adult Upper p. 100–150 residuals 7 20986 Jug 30 2610.4 XXVII Inhumation Simple pit Unknown Child Middle upper p. 50–150 8 14878 Jug 30 2180 III Inhumation Simple pit Male Adult Left lower p. 50–100 9 21073 Jug 29 2346.73 LXXXVI Inhumation Simple pit Male Adult Middle lower p. 150–250 10 21450 Jug 28.8 2113.03 159 Cremation Simple pit Unknown Unknown ? 50–200 11 15381 Jug 28 2156.2 43 Cremation Bustum Unknown Adult Right lower p. 50–150 12 15326 Jug 27 2248 29 Cremation Bustum Unknown Adult Middle p. 50–200 13 14870 Jug 27 2198 I Inhumation Simple pit Unknown Child Right lower p. 100–200 14 14918 Jug 26.5 1003 XV Inhumation Simple pit Unknown Child Right upper p. 50–150 15 21102 Jug 26 1885.63 64 Cremation Pit with cremation Unknown Adult ? 50–200 residuals 16 24227 Jug 26 2178.54 171 Cremation Pit with cremation Unknown Unknown ? 50–200 residuals 17 21449 Jug 24.4 1481.49 158 Cremation Pit with cremation Unknown Adult Middle p. 50–200 residuals 18 24230 Jug 24 2816.46 172 Cremation Pit without Unknown Unknown ? 50–200 cremation residuals 19 21016 Jug 24 1299 XLII Inhumation Simple pit Male Adult Right upper p. 50–150 20 15380 Jug 21.5 1011.96 43 Cremation Bustum Unknown Adult Middle p. 50–150 21 21044 Jug 21 988.08 LVIII Inhumation Simple pit Unknown Child Middle lower p. 100–00 22 24215 Pitcher 20.3 1014.85 168 Cremation Pit without Unknown Unknown ? 100–200 cremation residuals 23 20908 Jug 20 1473 49 Cremation Urn grave Unknown Adult ? 50–200 24 21020 Jug 20 709.05 XLIII Inhumation Simple pit Unknown Child Middle upper p. 100–150 25 21251 Jug 19.5 1176.5 97 Cremation Bustum Unknown Adult Left lower p. 50–100 26 14911 Pitcher 15.8 617.8 XIII Inhumation Brick chamber Unknown Child Right upper p. 50–150 27 15346 Bowl 14.5 3513.7 33 Cremation Pit with cremation Female Adult Upper p. 100–200 residuals 28 21458 Bowl 10.4 1485.43 155 Cremation Pit with cremation Unknown Adult ? 100–200 residuals 29 24229 Bowl 8.5 1453.52 172 Cremation Pit without Unknown Unknown ? 50–200 cremation residuals 30 21217 Bowl 8.5 1682.64 94 Cremation Pit with cremation Unknown Adult Middle p. 150–250 residuals 31 21308 Bowl 8.3 851.5 111 Cremation Simple pit Unknown Adult ? 100–150 32 21308 Bowl 8.2 923.87 111 Cremation Simple pit Unknown Adult ? 100–150 33 14882 Bowl 6.88 82.12 IV Inhumation Simple pit Unknown Child Right upper p. 50–150 34 21103 Bowl 6.5 517.05 66 Cremation Pit with cremation Unknown Adult Upper p. 100–150 residuals 35 20925 Bowl 6.15 256.89 51 Cremation Pit with cremation Female Adult ? 100–150 residuals 36 21131 Bowl 5.7 105.93 75 Cremation Pit with cremation Unknown Adult ? 100–200 residuals 37 15383 Bowl 5.4 863 43 Cremation Bustum Unknown Adult Left upper p. 50–150 38 15384 Bowl 5.2 354.4 43 Cremation Bustum Unknown Adult Middle p. 50–150 39 15385 Bowl 5 276.1 43 Cremation Bustum Unknown Adult Left upper p. 50–150 40 15386 Bowl 4.8 298.4 43 Cremation Bustum Unknown Adult Left upper p. 50–150 41 21225 Bowl 4.5 446.36 95 Cremation Simple pit Unknown Adult ? 50–150 42 21032 Bowl 4.5 471.07 LI Inhumation Simple pit Unknown Child Left middle p. 100–150 93 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 No. Inv. No. Vessel Absolute Absolute Grave No. Burial type Grave type Gender Age Vessel location Grave shape height volume (Grave part) dating (AD) (cm) (ml) 43 21135 Bowl 4.1 144.06 75 Cremation Pit with cremation Unknown Adult ? 100–200 residuals 44 15387 Bowl 4.1 108.7 43 Cremation Bustum Unknown Adult Middle p. 50–150 45 21277 Bowl 3.7 392 104 Cremation Pit with cremation Unknown Adult Middle lower p. 100–150 residuals 46 21344 Bowl 3.7 72.64 124 Cremation Bustum Unknown Adult Upper p. 50–200 47 21262 Bowl 3.2 71.71 99 Cremation Pit with cremation Unknown Adult ? 100–150 residuals 48 21430 Plate 4.8 506.72 149 Cremation Pit without Unknown Adult Middle p. 100–200 cremation residuals 49 20910 Plate 4.6 824.6 50 Cremation Pit with cremation Unknown Adult Upper p. 100–200 residuals 50 15305 Plate 4.3 941.2 17 Cremation Pit with cremation Unknown Unknown ? 100–150 residuals 51 15321 Plate 3.75 761 27 Cremation Urn grave Unknown Unknown Lower p. 100–150 52 21320 Plate 3.6 241.46 114 Cremation Pit with cremation Unknown Adult Upper p. 100–200 residuals 53 21215 Plate 3.5 316.39 94 Cremation Pit with cremation Unknown Adult Upper p. 150–250 residuals 54 21045 Plate 3.32 421 LVIII Inhumation Simple pit Unknown Child Middle lower p. 100–200 55 20935 Ovoid-shape 22 2384.3 53 Cremation Simple pit Unknown Unknown ? 150–250 56 21318 Ovoid-shape 21 2138.02 114 Cremation Pit with cremation Unknown Adult Upper p. 100–200 residuals 57 21049 Pot-shape 19 1837.49 LXII Inhumation Simple pit Unknown Adult ? 100–150 58 21454 Pot-shape 17.9 1960.61 154 Cremation Pit with cremation Unknown Adult Middle p. 100–150 residuals 59 21184 Ovoid-shape 17 1058.55 89 Cremation It with cremation Unknown Unknown ? 100–200 residuals 60 21469 Pot-shape 16.7 1419.01 161 Cremation Simple pit Unknown Unknown ? 100–200 61 24223 Pot-shape 16 1125.46 171 Cremation Pit with cremation Unknown Unknown ? 50–200 residuals 62 15329 Ovoid-shape 14.69 889.69 30 Cremation Pit with cremation Unknown Adult ? 100–200 residuals 63 21455 Ovoid-shape 13.9 975.51 154 Cremation Pit with cremation Unknown Adult ? 100–150 residuals 64 21329 Pot-shape 13.5 575.23 117 Cremation Pit with cremation Unknown Adult ? 50–150 residuals 65 14887 Pot-shape 13.2 1160.9 VII Inhumation Simple pit Unknown Child Middle lower p. 50–150 66 15308 Pot-shape 12.98 610.8 19 Cremation Pit with cremation Unknown Adult Upper p. 50–150 residuals 67 14919 Ovoid-shape 12.8 892 XV Inhumation Simple pit Unknown Child Right upper p. 50–150 68 20985 Ovoid-shape 12.8 595.62 XXVI Inhumation Simple pit Unknown Child Right lower p. 150–200 69 20965 Pot-shape 12.7 537.4 XX Inhumation Simple pit Female Adult Left upper p. 100–150 70 21003 Pot-shape 12.6 548.27 XXXV Inhumation Simple pit Unknown Child Left middle p. 100–150 71 24228 Pot-shape 12.3 575.25 172 Cremation Pit without Unknown Unknown ? 50–200 cremation residuals 72 21017 Pot-shape 12 492.75 XLII Inhumation Simple pit Male Adult Right upper p. 50–150 73 15316 Ovoid-shape 11.79 444.5 27 Cremation Urn grave Unknown Unknown Middle p. 100–150 74 14884 Pot-shape 11.62 481.3 IV Inhumation Simple pit Unknown Child Left upper p. 50–150 75 21031 Pot-shape 11.2 381.56 LI Inhumation Simple pit Unknown Child Left middle p. 100–150 76 21101 Pot-shape 11.1 413.11 64 Cremation Pit with cremation Unknown Adult ? 50–200 residuals 77 21459 Ovoid-shape 9.5 371.74 155 Cremation Pit with cremation Unknown Adult ? 100–200 residuals 78 14903 Ovoid-shape 9.5 309.2 XI Inhumation Simple pit Unknown Child Middle lower p. 50–150 79 21059 Ovoid-shape 9.4 284.02 LXXIV Inhumation Simple pit Male Adult Middle p. 100–150 80 21306 Ovoid-shape 9.4 337.44 111 Cremation Simple pit Unknown Adult ? 100–150 81 20987 Ovoid-shape 9.3 289.7 XXVII Inhumation Simple pit Unknown Child Lower p. 50–150 94 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 No. Inv. No. Vessel Absolute Absolute Grave No. Burial type Grave type Gender Age Vessel location Grave shape height volume (Grave part) dating (AD) (cm) (ml) 82 14880 Ovoid-shape 9.05 283 III Inhumation Simple pit Male Adult Left lower p. 50–100 83 21439 Ovoid-shape 9.05 288.71 156 Cremation Pit without Unknown Adult ? 50–200 cremation residuals 84 21270 Ovoid-shape 9.05 214.27 101 Cremation Urn grave Unknown Child ? 150–200 85 15317 Ovoid-shape 9 284.2 27 Cremation Urn grave Unknown Unknown Lower p. 100–150 86 21274 Ovoid-shape 9 293.33 102 Cremation Pit with cremation Unknown Adult Upper p. 50–100 residuals 87 14879 Ovoid-shape 8.9 280 III Inhumation Simple pit Male Adult Middle p. 50–100 88 21074 Ovoid-shape 8.6 323.61 LXXXVII Inhumation Simple pit Unknown Child ? 50–200 89 15287 Ovoid-shape 7.9 208 10 Cremation Simple pit Unknown Unknown ? 50–100 90 15320 Ovoid-shape 6.9 115.3 27 Cremation Urn grave Unknown Unknown Lower p. 100–150 91 21216 Vase 33.5 6345.82 94 Cremation Pit with cremation Unknown Adult Middle p. 150–250 residuals 92 15296 Vase 19 968.4 14 Cremation Pit without Unknown Adult ? 150–250 cremation residuals 93 24224 Vase 17 1334.62 171 Cremation Pit with cremation Unknown Unknown ? 50–200 residuals 94 14883 Vase 11.9 128.72 IV Inhumation Simple pit Unknown Child Left upper p. 50–150 95 14925 Amphora- 10.7 270 II Inhumation Simple pit Unknown Child ? 150–250 shape 96 15301 Cup 24 1233.8 15 Cremation Pit with cremation Unknown Adult ? 150–250 residuals 97 20912 Cup 22.5 1703.4 50 Cremation Pit with cremation Unknown Adult Lower p. 100–200 residuals 98 15395 Cup 12.99 782 13 Cremation Simple pit Unknown Unknown ? 100–200 99 20992 Cup 16.5 775.4 XXVIII Inhumation Alcove grave Male Adult Left upper p. 100–200 100 14871 Cup 9.25 206.8 I Inhumation Simple pit Unknown Child Left middle p. 100–200 101 15319 Cup 8.6 143.7 27 Cremation Urn grave Unknown Unknown Upper p. 100–150 102 21307 Cup 4.1 89.59 111 Cremation Simple pit Unknown Adult ? 100–150 103 15351 Urn 29 5705.9 34 Cremation Urn grave Male Adult ? 50–150 104 20907 Urn 27 6427 49 Cremation Urn grave Unknown Adult ? 50–200 105 21276 Urn 24 3244.59 104 Cremation Urn grave Unknown Adult Left middle p. 100–150 106 21147 Urn 23 2398.39 78 Cremation Urn grave Unknown Child ? 100–150 107 21295 Urn 22.4 3757.8 108 Cremation Urn grave Unknown Adult ? 100–150 108 21106 Urn 22 2846.27 68 Cremation Urn grave Unknown Unknown ? 100–150 109 21290 Urn 21.55 2848.51 107 Cremation Urn grave Unknown Child ? 50–150 110 21434 Urn 21.49 2674.63 150 Cremation Urn grave Unknown Child ? 100–150 111 15379 Urn 21 3570.6 43 Cremation Urn grave Unknown Adult Middle p. 50–150 112 21325 Urn 20.91 2414.63 115 Cremation Urn grave Unknown Unknown ? 100–150 113 20930 Urn 19 3508.39 52 Cremation Urn grave Unknown Adult ? 50–200 114 21268 Urn 18.75 1865.71 101 Cremation Urn grave Unknown Child ? 150–200 115 21288 Urn 18.6 2082.57 106 Cremation Urn grave Unknown Child ? 50–200 116 21129 Urn 18 1414.34 74 Cremation Urn grave Unknown Adult ? 50–200 117 21195 Urn 17 2452.74 92 Cremation Urn grave Unknown Adult ? 50–200 118 24222 Urn 16.2 1165.37 170 Cremation Urn grave Unknown Unknown ? 50–200 119 21158 Urn 15.23 814.19 84 Cremation Urn grave Unknown Child ? 150–200 120 24221 Urn 14 857.99 169 Cremation Urn grave Unknown Unknown ? 50–200 121 15318 Urn 8.5 197.5 27 Cremation Urn grave Unknown Unknown Middle p. 100–150 Tab. 1. Pottery data set with calculated absolute volumes (No. = vessel number in the photo/model tables; Inv. No. = vessel inventory number in the Archaeological museum of Slovak National Museum, hereinafter referred to as SNM). Tab. 1. Keramický datový soubor s vypočtenými absolutními objemy (No. = číslo nádoby ve fotografických/modelových tabulkách; Inv. No. = číslo nádoby v inventáři Archeologického múzea Slovenského národného múzea, dále jako SNM. 95 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 Fig. 5. Pottery set with calculated volumes. Group No. 2: Bowls and plates. Deposited in the Archaeological museum SNM. Photo by A. Szabová. Obr. 5. Keramický soubor s vypočtenými absolutními objemy. Skupina č. 2: Misky a talíře. Uloženo v Archeologickém múzeu SNM. Foto A. Szabová. Fig. 3. Pottery set with calculated volumes. Group No. 1: Jugs and pitchers. Deposited in the Archaeological museum SNM. Photo by A. Szabová. Obr. 3. Keramický soubor s vypočtenými absolutními objemy. Skupina č. 1: Džbány a konvice. Uloženo v Archeologickém múzeu SNM. Foto A. Szabová. Fig. 6. Generated models of the pottery set with calculated volumes. Group No. 2: Bowls and plates. Deposited in the Archaeological museum SNM. Models by Z. Porubčanová. Obr. 6. Vygenerované modely keramického souboru s vypočtenými absolutními objemy. Skupina č. 2: Misky a talíře. Uloženo v Archeologickém múzeu SNM. Modely Z. Porubčanová. Fig. 4. Generated models of the pottery set with calculated volumes. Group No. 1: Jugs and pitchers. Deposited in the Archaeological museum SNM. Models by Z. Porubčanová. Fig. 7. Pottery set with calculated volumes. Group No. 3: Convex vessels. Obr. 4. Vygenerované modely keramického souboru s vypočtenými absolutními Deposited in the Archaeological museum SNM. Photo by A. Szabová. objemy. Skupina č. 1: Džbány a konvice. Uloženo v Archeologickém múzeu SNM. Obr. 7. Keramický soubor s vypočtenými absolutními objemy. Skupina č. 3: Baňaté Modely Z. Porubčanová. nádoby. Uloženo v Archeologickém múzeu SNM. Foto A. Szabová. 96 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 Fig. 8. Generated models of the pottery set with calculated volumes. Fig. 11. Pottery set with calculated volumes. Group No. 5: Urns. Deposited in the Group No. 3: Convex vessels. Deposited in the Archaeological museum SNM. Archaeological museum SNM. Photo by A. Szabová. Models by Z. Porubčanová. Obr. 11. Keramický soubor s vypočtenými absolutními objemy. Skupina č. 5: Urny. Obr. 8. Vygenerované modely keramického souboru s vypočtenými absolutními Uloženo v Archeologickém múzeu SNM. Foto: A. Szabová. objemy. Skupina č. 3: Baňaté nádoby. Uloženo v Archeologickém múzeu SNM. Modely Z. Porubčanová. Fig. 9. Pottery set with calculated volumes. Group No. 4: Others – vases, amphora-shape vessels, cups. Deposited in the Archaeological museum SNM. Photo by A. Szabová. Obr. 9. Keramický soubor s vypočtenými absolutními objemy. Skupina č. 4: Ostatní (vázy, amfory a poháry). Uloženo v Archeologickém múzeu SNM. Foto A. Szabová. Fig. 12. Generated models of the pottery set with calculated volumes. Group No. 5: Urns. Deposited in the Archaeological museum SNM. Models by Z. Porubčanová. Obr. 12. Vygenerované modely keramického souboru s vypočtenými absolutními objemy. Skupina č. 5: Urny. Uloženo v Archeologickém múzeu SNM. Modely Z. Porubčanová. The fourth group (Fig. 9, 10) was the most diverse one due to the lack of representatives, and, therefore, has the lowest informa­tive value. This had to be taken into consideration in the following steps (particularly in statistics). The first part of this group was formed by five vases and amphora-shape vessels with a height range between 10.7–33.5 cm, a mouth diameter of 6.21–15.22 cm and a base diameter of 2.9–8.89 cm. The second part was formed by various types of cups (seven), although four Fig. 10. Generated models of the pottery set with calculated volumes. Group No. 4: were folded beakers (ger. der Faltenbecher). The height ranges Others – vases, amphora-shape vessels, cups. Deposited in the Archaeological between 4.1–24 cm, the mouth diameter between 5.63–10.75 cm museum SNM. Models by Z. Porubčanová. and the base diameter between 2.9–6.5 cm. The calculated vol- Obr. 10. Vygenerované modely keramického souboru s vypočtenými absolutními objemy. Skupina č. 4: Ostatní (vázy, amfory a poháry). Uloženo v Archeologickém umes of the folded beakers are slightly higher than the real ones múzeu SNM. Modely Z. Porubčanová. because of the uneven folding of the cup wall, which could not be 97 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 accurately measured. The last (fifth) group (Fig. 11, 12) was com- The third section has already addressed all the shape groups; prised of urns with a height between 4.9–29 cm, a maximum bulge otherwise, it was the same as the last one and followed the vol- of 6.5–15 cm, a mouth diameter of 5.85–19.5 cm and a base diam- ume differences in the various grave categories. The next cate- eter of 3.25–10.8 cm. Even though most of the urns had a convex gory formed a different type of question – whether a correlation shape similar to the vessels in the third group, we decided to create between the volume and the other variables exists (e.g., Is there an individual category for them because of their different function a correlation between the vessel volume and the age of the de- in the grave inventory. The wall thickness was also important for ceased?). The final three sections used the whole set of pottery the volume calculations and oscillated mostly between 0.3–1.0 cm, (not only the part with the calculated volumes). The fifth cat- depending on the type of vessel and the body area. egory examined the significant differences in the frequency of The total number of vessels identified from the necropolis is the vessels within the referred variables. The following category currently 409, although most of them are obviously preserved in also traced the differences within the variables but instead of fragments. For potential structural coherence identification, we frequency, we worked with the shape spectrum. The last analy­ worked not only with the vessel volumes but also with the shape sis attempted to find a correlation between the shape and the groups. For this reason, when working with the hypotheses that other variables. By creating the two question types (difference/ dealt with the shape, we used the whole set of pottery (including correlation), we ensured partial mutual control of the results. fragments), not only the measurable part. Concerning the funer- ary practices, primary variables such as the burial type (inhuma- 3. Methods tion/cremation), grave type, gender or age of the person buried Three-dimensional models of each vessel from the collection and the chronological interval and location of the artefact in were created in Blender v2.79 open-source software. The same the grave were chosen for testing. Gender and age identification software was also used to calculate the approximate absolute comes from the two individual anthropological studies (Pichle­ volume of the vessels. For the 3D reconstruction, a drawing and rová, Stloukal 1977, 1981). In addition to elementary questions photo documentation of the vessel profile and its dimensions such as the average volume of the different shapes, the amount (height, throat diameter, maximum bulge, base diameter and of pottery in the graves etc., we created several question sections thickness of the walls) was needed. In most cases, drawing doc- where the null and the alternative hypothesis were formed in umentation with a relevant scale was sufficient for the model each section. The first category was defined by the question of generation. Still, the dimensions acquired by documentation can whether the volumes statistically significantly differ in the vari- slightly differ from the real dimensions, especially the thickness ous shape groups. Because the urns are, unlike any other ceramic of the vessel, which is often not constant. The lower part of the vessels, the functional element of the burial practice (not the vessel usually has thicker walls than the rest of it. The wall thick- offering), they were treated separately first. Therefore, we asked ness of the different parts of the vessel was measured where pos- if there was a significant difference in the urn volumes within the sible (e.g., on bowls, plates and other wider vessels). However, aforementioned variables (grave type, gender, age etc.). this could not be done with some shapes with a narrower throat Graph 1. Normality of data distribution presented by the normal probability plots on the left and by the tests for normal distribution on the right. The normal data distribution of volumes in the fifth shape group (urns) in the upper part and the abnormal distribution of volumes in the third shape group (convex vessels) in the lower part. Graf 1. Normalita datové distribuce, která je nalevo zobrazena na normálních pravděpodobnostních grafech a napravo pomocí testů normálního rozdělení. V horní části jsou prezentovány objemy 5. tvarové skupiny (uren), které mají normální distribuci dat, a ve spodní části jsou objemy 3. tvarové skupiny (baňatých nádob), které nemají normální distribuci dat. 98 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 (e.g., amphora-shaped vessels or jugs). Therefore, there could be can also be seen (except the histogram) in the normal proba- a minor difference between the real and the calculated volumes. bility plot, which also shows a correlation coefficient for every First, the shape of the inner profile of each vessel was created variable (Graph 1). Nonetheless, various other graphic visualis- using Bezier’s curves. Afterwards, the inner body of the vessel ations were also useful (Graph 2, 3, 5, 7–10). was created by function “spin”, where the inner profile spun 360° around the Z-axis. Bezier’s curves were transformed to vertices 4. Results for better manipulation with the shape of the model. Because of As mentioned above, 121 3D vessel models were created. the vertices, the individual mesh sections of the model could be Eight vessels were preserved in the fragmented condition. There- adjusted to the shape of the real vessel. Before calculating the fore, the calculation of the absolute volumes was impossible. In absolute volume, the model needed to be closed with polygons at these cases, the program could only calculate the minimum vol- the top and bottom. Finally, the normals were calculated to cor- umes. Half of these belong to the category of the convex vessels, rect vertices, which could deform the morphology of the model. two to the jugs, one to the urns and one to the cups. However, The program could then calculate the absolute volume of the excluding these vessels (as well as the folded beakers) made no vessel in cm 3 or m 3, which was then converted to ml or l. This way significant difference in the results, so we decided to keep the of estimating vessel volumes in Blender was inspired by Köster’s collection complete on account of the higher variability. paper (2015, 1–8). A similar method was used in the study by The highest absolute volume of all the models measured in Emmitt (2020, 4–6), where it also proved to be effective. The the jugs’ group was 7,208 ml. The lowest absolute volume of the technique for calculating vessel volume from profile drawings jugs was 617 ml, which is the highest amongst the lowest vol- was also discussed by Senior and Birnie (1995, 319–334). umes of all vessels. This category also had the second-highest The veracity of each hypothesis was then validated or dis- average absolute volume of all the shape groups with 2,450 ml. proved by univariate tests in PAST 3 open-source software for First were the urns with an average absolute volume of 2,596 ml. scientific data analysis (Hammer et al. 2001, 2–3). The testing This group also had the second-highest absolute volume with is basically a comparison of the observations that are predicted 6,427 ml. The jugs and the urns had higher volumes than other in the hypothesis and the observations of the evidence in reality categories in general. The lowest absolute volumes were meas- (Demján 2020). The first step was to create the null hypothesis, ured within the bowls and plates group and the convex vessels. which states that observations are just a product of coincidence These categories also had the lowest average absolute and maxi- (Neustupný 1973, 192). For example, the prediction that there mum absolute volumes. This means a 656 ml average volume for is no statistically significant difference between the urn volume bowls and plates and 737 ml for convex vessels. The highest vol- of the adult and the child. We could assume that the alternative ume for bowls and plates was 3,513 ml and 2,384 ml for convex hypothesis we defined (e.g., there is a statistically significant vessels. The lowest absolute volume (about 71 ml) was measured difference between the urn volume of the adult and the child) in the bowls and plates group. In the fourth group, the lowest was valid if the null hypothesis were disproved (Otárola-Cas- volume was 89 ml. The shape variety of this group is visible on tillo, Torquato 2018, 437). For choosing the adequate type of the volume calculation results. The highest absolute volume in statistical analysis, it was important to first check the data dis- this category was 6,345 ml, the third-highest of all the volumes. tribution (Graph 1), which could be observed on several plots The average absolute volume was 1,116 ml. Finally, the average (the normal probability plot in particular) or by using tests for volume of all modelled vessels was 1,511 ml. normal distribution; e.g., the Shapiro-Wilk test (Weber 1997, The regular frequency of the vessels in the grave ranged be- 206–207). According to the results of each data set, paramet- tween 0 and 4. There were no vessels in 84 graves, one in 63 graves, ric or non-parametric tests were used to validate the hypothesis two in 47 graves, three in 22 graves and four in 13 graves to (Neustupný 1973, 228). be specific. The higher amount of the vessels is less common From the two-sample parametric tests, the T-test for equal (7× 5 pieces, 5× 6 or 7 pieces, 4× 8 pieces and only 1× 9 pieces). means was applied and the F-test for equal variances was less Within the shape variability, the most frequent were jugs (or frequently used. From the non-parametric tests, the Mann-Whit- pitchers) with 116, then bowls (or plates) with 99, and a convex ney U test for equal medians was applied and Mood’s median and the Kolmogorov-Smirnov test for equal distributions was less frequently used (Weber 1997, 206–209). The level of signifi- cance (Drennan 2010, 29–31) was 5% in all cases. From several sample tests, on the one hand, One-way ANOVA and Tukey’s range test for equal means were used, and on the other hand, Mann-Whitney pairwise and Kruskal-Wallis were used (Weber 1997, 206–209). For testing the presence of the association be- tween the variables, Chi-squared and Fisher’s exact test (Weber 1997, 208) were applied. However, the expected frequencies us- ing the totals of the variables had to be calculated beforehand (Neustupný 1973, 193). Data visualisation also formed (besides the testing of indi- vidual hypotheses) a significant part of the statistical analysis. For this purpose, a histogram (Graph 8) was often used where both the frequency and our data distribution combined with normal distribution (expressed as a Gaussian curve) are visible. Box plot (Graph 4) expressing median, 25th and 75th percentile, Graph 2. Pie chart picturing the total numbers of various shapes of the vessels standard error and deviation, outliers or the maximum and min- within our data set (including fragments). imum digits, were also effective (Drennan 2010, 11–13, 37–41). Graf 2. Koláčový graf zobrazující celkové počty různých tvarů nádob v našem The comparison of our data distribution to normal distribution datovém souboru (včetně fragmentů). 99 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 large variability and the low number of individuals. The prob- ability of different results after the increase of the artefacts amount is rather high. The set of urn volumes had a normal distribution (Graph 1), so in this case, we used the parametric tests (T-test, F-test). We could not search for the differences in the case of gender because there were too many unidentifiable individuals although the age analysis provided interesting results. The adult urn volumes were significantly different from the children’s (t-test, p[same mean] = 0.04054) and the unknown’s as well (t-test, p[same mean] = 0.019724). However, children’s and unknown’s urns had statistically similar volumes (t-test, p[same mean] = 0.2987). The volumes of the adult urns had a significantly higher mean value (Graph 4). No difference was recognised within the time Graph 3. Percentile graph picturing the vessel volume distribution in the different intervals (t-test, p[same mean] = 0.10677 – 0.78207). The same shape groups. applies to the vessel locations as the majority of the urns were Graf 3. Percentilový graf zobrazující distribuci objemů nádob v různých tvarových located in the middle part of the grave (other locations occurred skupinách. only sporadically). After the analysis of the whole set of calculated volumes (urns + offerings), the higher mean of the vessel volumes in the cremation graves was noticeable (Mann-Whitney, p[same med.] = vessel with 78 was the third most frequent. There were 31 vessels 0.020481) although after excluding the urns, there was no sig- identified as urns with 48 vessels in the “other” category (vases, nificant difference (Mann-Whitney, p[same med.] = 0.20797). amphora-shaped vessels and cups). Thirty-seven of the pottery Within the individual shape groups, only the fourth group had fragments were unidentifiable (Graph 2). considerably higher volumes in the cremation than in the inhu- Most of the data sets used did not have normal distribution mation graves (Mood median, p[same med.] = 0.0083153; Kol- (Shapiro-Wilk, p[normal] for jugs = 0.03588; bowls = 5.655 × 10 -6; mogorov-Smirnov, p[same dist.] = 0.029391). Among the grave convex vessels = 1.67 × 10 -5; other = 5.872 × 10 -5; urns = 0.2312) types, there was a difference documented only between the sim- and many of them contained a couple of outliers. For this reason, ple pit inhumation graves and the urn graves – the urn graves had we used the non-parametric tests more often than the paramet- a higher mean volume (Mann-Whitney pairwise, p = 0.03803). ric ones throughout the analysis. First, we attempted to detect No difference was registered between the genders (Mann-Whit- significant differences or similarities between the volumes in ney pairwise, p = 0.6156 – 0.9123). Adult graves had a mildly the various shape categories. As a result, we found that there is but not significantly higher mean volume (Graph 5) than the no significant difference between the volumes of jugs and urns children’s (Mann-Whitney pairwise, p = 0.1001) and unspec- (Mann-Whitney, p[same med.] = 0.30649). These two categories ified graves (Mann-Whitney pairwise, p = 0.6355). The jugs also had the highest mean volume (Graph 3). Other statistically (t-test, p[same mean] = 0.03541) and the others (Mann-Whit- significant similarities could be observed between bowls/plates ney, p[same med.] = 0.095836) had a particularly higher mean and convex vessels (Mann-Whitney, p[same med.] = 1.8672). On volume in the adult graves. On the time scale, there was no the one hand, the convex vessels were generally higher but on significant difference between most of the intervals (Graph 6). the other hand, the bowls were mostly wider, which caused their parallel volume range. The fourth category (vases, cups etc.) is statistically similar to bowls (Mann-Whitney, p[same med.] = 0.50659) and also to convex vessels (Mann-Whitney, p[same med.] = 0.79857). However, in this case, we had to consider the Graph 4. Box plot picturing the urn volume distribution within the different age groups. Graph 5. Line chart picturing the vessel volume distribution in the different age Graf 4. Krabicový diagram groups. zobrazující distribuci objemů uren v různých věkových Graf 5. Liniový graf zobrazující distribuci objemů nádob v různých věkových skupinách. skupinách. 100 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 Graph 8. Histogram picturing the distribution of ceramic vessels in the inhumation and in the cremation graves. Graf 8. Histogram zobrazující distribuci keramických nádob v kostrových Graph 6. Box plot picturing the vessel volume distribution within the various time a kremačních hrobech. intervals (+ outliers). Graf 6. Krabicový diagram zobrazující distribuci objemů nádob v různých časových intervalech (+ extrémní hodnoty). Graph 9. Stacked bar graph picturing the shape distribution of ceramic vessels in the inhumation and cremation graves. Graf 9. Skládaný sloupcový graf zobrazující distribuci tvarů keramických nádob v kostrových a kremačních hrobech. frequently used locations were the lower-middle and upper-right parts (Graph 7a). The only significant difference appeared be- tween the grave centre and the left part. The vessels in the cen- tre had a higher mean volume (Mann-Whitney, p[same med.] = 0.04157). The majority of the vessels in the cremation graves (except the urns) were located in the lower part (Graph 7b) with the least in the left part of the grave. No significant difference was found there (Mann-Whitney pairwise, p = 0.1449 – 0.7018). No correlation between the volumes and the other variables was proven (Chi-squared, p[no assoc.] = 0.10869 – 0.2447). Un- Graph 7 a, b. Percentile graph picturing the vessel volume distribution in the fortunately, three of the variables (gender, time and location) different parts of inhumation (a) and cremation (b) graves. could not be tested, even after the partial unification of the op- Graf 7 a, b. Percentilový graf zobrazující distribuci objemů nádob v různých částech kostrového (a) a kremačního (b) hrobu. tions, because of the expected low frequencies. As previously mentioned, in the last three analyses we engaged the whole data set of 409 vessels and fragments. In the cremation graves, there were significantly higher amounts of vessels (Graph 8) Only the wide interval of 50–200 AD differed (had a higher than in the inhumation ones (Mann-Whitney, p[same med.] = mean value) from three of the others (Mann-Whitney pair- 3.1108 × 10 -15). There was a difference between the simple pit in- wise, p for 50–150 AD = 0.00991; for 100–150 AD = 0.0318; for humation graves and almost all types of cremation (Mann-Whit- 100–200 AD = 0.02813). In the inhumation graves, there was no ney pairwise, p for simple pit = 2.597 × 10 -5; p for graves with- vessel located in the right-middle part of the grave. The most out the cremation residuals = 0.0001093; p for graves with the 101 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 cremation residuals =  1.578 × 10 -12; p for urn graves = 7.548 × 10 -9). between the shape and location as well as the shape and gender All the cremation types had a higher number of artefacts. Busta could not be tested (because of the expected low frequencies). and graves with cremation residuals were distinguished within A middle intense correlation (Contingency C = 0.32077) was the cremation, which had a higher number of finds (Tukey’s pair- proven between the shape and the chronology (Chi-squared, wise, p = 0.03179). p[no assoc.] = 0.0031372) and none within the age of the de- The graves with an unidentified gender had a higher mean ceased (Chi-squared, p[no assoc.] = 0.30752). value than males (Mann-Whitney pairwise, p = 0.001316) and according to some of the tests (e.g., Mann-Whitney pairwise, 5. Interpretation and discussion p = 0.03186), also females. However, the difference was lower for The standard number of pottery in graves in necropolis II in the females. In terms of the age of the deceased, the unidenti- Gerulata was 0–4. On average, the cremation burials had more fiable individuals were also the most numerous (Mann-Whitney vessels than the inhumation burials, which also corresponds pairwise, p for adults = 0.006016; p for children = 0.02595). In with other items from the grave inventory. This could be because terms of chronology, the interval of 50–200 AD again differed (had the cremated population was the wealthier one, as stated in other a lower mean value) from many others (Mann-Whitney pairwise, studies (Krekovič 2011, 102). The highest number (9 pieces) of p for 50–150 AD = 5.912 × 10 -6; p for 100–150 AD = 3.708 × 10 -8; vessels was found in grave No. 95. However, besides the vessels, p for 100–200 AD = 5.714 × 10 -8; p for 150–250 AD = 0.0005247). there were only two other items present – a ceramic lamp and However much fewer of the graves belonged to this time interval. a wooden box with iron fittings (Pichlerová 1981, 145–146). The shape variation of the whole data set showed no signif- There were between two and six other artefacts in the graves icant difference (Mann-Whitney, p[same med.] = 0.54676) be- rich with pottery. Nonetheless, we could identify some of the tween the inhumation and cremation burials (Graph 9). Within graves (in the context of this necropolis) as relatively wealthy. the grave types, only the urn graves differed (Mann-Whitney For example, in female grave No. 1, two ceramic lamps, a bone pairwise, p = 5.69 × 10 -8 – 1.289 × 10 -6) because of the obvious fact pin, a bronze tweezer, three bronze circular rivets, a bronze coin that there was a high number of urns there (Graph 10). Analysis from the reign of Emperor Domitian (81–96 AD), a small glass of the gender and age showed no statistically significant differ- vase, an iron knife, a wooden box with iron fittings, iron rivets ence (Mann-Whitney pairwise, p for gender = 0.4329 – 0.6623; and nails, and the bones of a dog and a small ruminant (goat/ p for age = 0.1143 – 0.5177). In terms of chronology, most of sheep) were found. The pottery spectrum consisted of two jugs, the time intervals showed similar results (Mann-Whitney pair- a plate, an ovoid vessel, a terra sigillata bowl (Drag. 33) and wise, p = 0.06187 – 0.9808). However, there was a difference of fragments of four other unidentifiable vessels (Pichlerová 1981, the interval 150–200 AD from all of the others (Mann-Whitney 83–85; Gabler, Pichlerová 1996, 85). pairwise, p = 0.004872 – 0.01507). Testing of the vessel location Various combinations of jugs, bowls, plates and convex in the cremation graves showed similar results in all parts of the vessels were the most common. The analysis showed the use grave (only the urns were mostly located in the centre as men- of larger vessels (particularly jugs and urns) in the graves of tioned above). In the inhumation graves, the right part differed adult individuals. This was already indicated by the study of the (Mann-Whitney, p[same med.] = 0.018881) from the left (this urn shape indexes by Professor Krekovič, who recognised the was caused by the fact that the jugs dominated the right). The last possibility of correlation between the urn size and the age and analysis revealed a correlation of medium intensity (Contingency gender of the person buried (Krekovič 2014, 816). Our analysis C = 0.20114) between the vessel shape and the type of burial (Chi- confirmed this between the age and the urn size and showed the squared, p[no assoc.] = 0.00347337), as well as a high intensity same correlation in the case of jugs. Unfortunately, the coher- correlation (Contingency C = 0.62922) within the grave type var- ence between vessel size and the gender of the deceased was not iable (Chi-squared, p[no assoc.] = 1.8625 × 10 -27). A correlation confirmed, although that was probably caused by the low num- ber of gender-identified graves in our data set. The parametric tests also discovered significant parallels between the volumes of children’s urns and the urns of the unidentifiable individuals. This could indicate that a certain amount of the unidentified de- ceased might be children or (according to the theory mentioned above) women. Another argument for the possibility of uniden- tified graves being partially children’s graves is that these graves contained a significantly higher number of vessels in comparison with the graves of adult men and women. A volume similarity between certain shapes such as jugs, urns, bowls and convex vessels is, within a limited range of provincial pottery, typical for the area and time period and is an understandable result. We could have obtained a more di- verse outcome if there was a higher number of artefacts in the fourth shape group. The analysis also showed the larger vessel volumes in the cremation than in the inhumation burials, which was mainly caused by the presence of the urns and by the higher variability of the pottery in the cremation graves. This also trans- ferred to the correlation between vessel shape and the type of Graph 10. Stacked bar graph picturing the shape distribution of ceramic vessels in burial (and grave type). Within the chronology differences, we the different grave types. 1 – simple pit, 2 – brick tomb, 3 – others (mostly alcove- found no significant changes in most of the time intervals. Only or shaft graves). a wide interval of 50–200 AD had a higher vessel volume than Graf 10. Skládaný sloupcový graf zobrazující distribuci tvarů keramických nádob v různých typech hrobů. 1 – jednoduchý jámový hrob; 2 – cihlová hrobka; 3 – další others, but also the lowest amount of the pottery. Next, the in- (především výklenkové či šachtové hroby). terval 150–200 AD had a different shape variety than the others. 102 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 Unfortunately, both these results were caused by the low number Gabler, D., Pichlerová, M. 1996: Terra sigillata von Gerulata. of individuals in those intervals and do not indicate any chron- In: K. Kuzmová, J. Rajtár (eds.): Gerulata I. Nitra: Archeologický ological changes in the pottery spectrum. In the inhumation ústav Slovenskej akadémie vied v Nitre, 45–143. graves, vessels were mostly cumulated by the feet or next to the Hammer, Ø., Harper, D. A. T., Ryan, P. D. 2001: PAST: skull of the buried, which were standard locations in many Pan- Paleontological statistics software package for education and data nonian cemeteries. In the cremation graves, the central area pre- analysis. Palaeontologia Electronica 4(1), 1–9. vailed in the case of urns and the lower grave part in the case of Hlavatá, J., Varsik, V. 2019: Sídlisko autochtónnej panónskej populácie offerings. However, a considerable amount of the fragmentarily v Rusovciach: Prvé výsledky archeobotanických analýz. In: Sedem preserved vessels were concentrated in more than one grave part. kruhov Josefa Bujnu. Studia Historica Nitriensia 23, Supplemen- tum 2. Nitra: Univerzita Konštantína Filozofa v Nitre, 427–448. 6. Conclusion Hlavenková, L., Schmidtová, J., Zeman, T. 2015: Paleodemografia A combination of several multidisciplinary methods (3D model rímskej populácie Gerulaty v 1. až 4. storočí n. l. Historická generation, volume calculation, application of statistical analy­ demografie 39(2), 137–179. sis) helped us to detect certain indications of the structural co- Köster, J. 2015: Calculate vessel volume with Blender. Academia herence of the archaeological material (in this case, the pottery [online] ©2021. 03-02-2015 [cit. 2021-01-11.]. Availlable from: ware) and some of the burial factors in necropolis No. II in an- https://www.academia.edu/11200820/Calculate_vessel_volume_ cient Gerulata. Specifically, a relation between the size of the with_Blender. vessel and the age of the dead, or between the amount of the Krekovič, E. 1992: Skelettgräber der älteren römischen Kaiserzeit im vessels and the burial type, was discovered. Naturally, we cannot Licht der Funde aus Gerulata. Carnuntum Jahrbuch 1991, 71–79. generalise these results to all of the data in this particular col- Krekovič, E. 1998: Römische Keramik aus Gerulata. Bratislava: lection (e.g., not all of the unidentified individuals with smaller Filozofická fakulta Univerzity Komenského v Bratislave. urns must necessarily be children). Nevertheless, it could be Krekovič, E. 2011: Deti v Gerulate. Musaica XXVII, 101–106. partially valid (Neustupný 2007, 165–167). Krekovič, E. 2014: Gender identity and pottery: the size of urns in Next to the validation of some of the hypotheses, the dis- cremation cemeteries. Rei Cretariae Romanae Fautorum Acta 43, proving of others was also important, for example, the negative 815–817. evidence in the matter of changes in the volume or shape of the Neustupný, E. 1973: Jednoduchá metoda archeologické analýzy. vessels during the time intervals with a sufficient quantity of Památky archeologické LXIV(2), 169–234. subjects (apart from the intervals 50–200 and 150–200 AD). The Neustupný, E. 2007: Metoda archeologie. Plzeň: Katedra archeologie application of different methods throughout the process of cre- Západočeské univerzity v Plzni. ating this study brought a more diverse perspective on the long Otárola-Castillo, E., Torquato, M. G. 2018: Bayesian Statistics in known, but to date, mostly conventionally treated site. Archaeology. Annual Review of Anthropology 2018, 435–453. Pichlerová, M. 1981: Gerulata Rusovce. Rímske pohrebisko II. Bratislava: Acknowledgements Archeologický ústav Slovenského národného múzea v Bratislave. We would like to express our gratitude to the staff of the Ar- Pichlerová, M. 1983a: Pokračovanie výskumu v Bratislave- chaeological museum SNM, especially to the director PhDr. Ju- -Rusovciach. Fortsetzung der Grabung in Bratislava-Rusovce. raj Bartík, Ph.D., for allowing the realisation of revision, then Archeologické výskumy a nálezy na Slovensku v roku 1982, 208–210. to PhDr. Klára Füryová, Mgr. Igor Choma, Sandra Csabiová, Pichlerová, M. 1983b: Postavenie Gerulaty na stredodunajskom and particularly to Daniela Lacová, for their assistance during Limes Romanus. Zborník Slovenského národného múzea LXXVII. the documentation and archive investigation. Revision is part História 23, 63–94. of the ongoing dissertation – A. Szabová: Funerary practices Pichlerová, M., Stloukal, M. 1977: Kostrové hroby z římského in Pannonia and the examination possibilities of protohistoric pohřebiště II v Gerulatě. Časopis Národního muzea v Praze. Řada populations based on the example of cemeteries from ancient přírodovědná 146, 29–40. Gerulata. Supervisor: Mgr. Balázs Komoróczy, Ph.D. Institution: Pichlerová, M., Stloukal, M. 1981: Žárové hroby z římského Department of Archaeology and Museology, Masaryk University. pohřebiště II v Gerulatě. Časopis Národního muzea v Praze. Řada This study was created during the long-term cooperation of both přírodovědná 150, 184–200. authors with the Institute of Archaeology of the Czech Acad- Schmidtová, J., Mathédesz, Ľ. 2016: Nové nálezy jednotky piatej emy of Sciences, Brno. Our thanks go namely to Mgr. Balázs Ko- kohorty Lucensium. In: I. Bazovský (ed.): Zborník na pamiatku moróczy, Ph.D., and Mgr. Marek Vlach, Ph.D., for their guidance Magdy Pichlerovej. Štúdie. Zborník Slovenského národného múzea. during the creation of this study. Last but not least, we would Archeológia, Supplementum 11. Bratislava: Slovenské národné like to thank Mgr. Peter Tóth, PhD., for the introduction to the múzeum v Bratislave – Archeologické múzeum, 125–130. statistical analysis, and to MA. Joshua Emmitt, PhD., for the Schmidtová, J., Mathédesz, Ľ. 2018: Rímsky kastel Gerulata. consultation on model generation. In: R. Šenkirik, L. Gembešová, H. Bakaljarová, M. Škrovina (eds.): Dunajský Limes a odkaz rímskej antiky na ľavom brehu Dunaja. References Bratislava: Bratislavský samosprávny kraj, Academia Istropolitana Demján, P. 2020: Štatistická analýza v archeológii. In: You Tube Nova, 52–66. [online]. 2. 4. 2020 [cit. 2021-01-11]. Available from: Schmidtová, J., Vorlíčková, B. 2006: Pokračovanie výskumu na https://www.youtube.com/watch?v=kMoTy23eYaQ. stavbe Odkanalizovanie Rusoviec. Archeologické výskumy a nálezy Drennan, R. D. 2010: Statistics for Archaeologists. A Commonsense na Slovensku v roku 2004, 175–177. Approach. Second edition. New York: Springer. Senior, L. M., Birnie, D. P. 1995: Accurately estimating vessel volume Emmitt, J. 2020: Formation and function: Middle Holocene pottery from profile illustrations. American Antiquity 60(2), 319–334. from Kom W, Fayum, Egypt. Quaternary International 555, 126–134. DOI: 10.2307/282143. Availlable also from: www.jstor.org/stable/282143. [cit. 2021-01-11] DOI: https://doi.org/10.1016/j.quaint.2019.09.032. Varsik, V. 1996: Archäologische Topographie des antiken Gerulata. Available from: https://www.sciencedirect.com/science/article/pii/ In: K. Kuzmová, J. Rajtár (eds.): Gerulata I. Nitra: Archeologický S1040618219307827?via%3Dihub. ústav Slovenskej akadémie vied v Nitre, 7–43. 103 I d e n t i f i c a t i o n o f p o t e n t i a l s t r u c t u r a l c o h e r e n c e b a s e d o n t h e c e r a m i c m a t e r i a l f r o m p r o v i n c i a l n e c r o p o l i s I I a t A n c i e n t G e r u l a t a (S R) u s i n g q u a n t i t a t i v e a n a l y s e s X S z a b o v á , A ., P o r u b č a n o v á , Z . X P ř e h l e d v ý z k u m ů 6 2 / 1, 2 0 21 X 91–10 4 Varsik, V. 1999: Die Notgrabung einer römischen Siedlung des pohlaví jako proměnné však bylo výrazně ovlivněno nízkým po- 2.–3. Jahrhunderts im Hinterland des Kastells Gerulata (vorläufige čtem identifikovaných jedinců z této nekropole. Ergebnisse 1995–1997). In: J. Tejral (Hrsg.): Das mitteleuropäische V rámci chronologie pohřebiště nebyly zjištěny žádné vý- Barbaricum und die Krise des römischen Weltreiches im 3. Jahrhun- znamné rozdíly ve tvaru či objemu nádob. Nepatrné odlišnosti dert. Spisy Archeologického ústavu AV ČR Brno 12. Brno: Archäo- způsobila nepravidelná distribuce dat v jednotlivých časových logisches Institut der Akademie der Wissenschaften der Tschechi- intervalech. V poloze nálezů nebyly taktéž zaznamenány signi- schen Republik Brno, 215–233. fikantní rozdíly. V kostrových hrobech se nádoby koncentrovaly Weber, Z. 1997: Statistická analýza archeologických dat. především v blízkosti lebky nebo dolních končetin zesnulého, In: J. Macháček (ed.): Počítačová podpora v archeologii. Brno: Ústav což je standardní umístění na mnohých provinciálních po- archeologie a muzeologie Filozofické fakulty Masarykovy hřebištích. V kremačních hrobech se urna vyskytovala hlavně univerzity, 19–214. uprostřed hrobu a milodary v jeho dolní části (v případě, že ji bylo možné určit na základě orientace). Pokud by se v tvarové Resumé skupině „ostatní“ dochovalo v lepším stavu větší množství ná- Antická Gerulata (Bratislava-Rusovce, Slovensko) byla vojen- lezů, dosáhly bychom v rámci této analýzy spíše heterogenních ským táborem (později pevností) s civilním zázemím na severo- výsledků. I přesto jsme pomocí kombinace více metod získaly panonském Limes Romanus. Existovala od 2. poloviny 1. století n. l. určité náznaky strukturální koherence mezi keramickými nálezy a zanikla na počátku 5. století (Schmidtová, Mathédesz 2018). a pohřebním ritem. Některé naše hypotézy se podařilo valido- Místní obyvatelé pohřbívali své mrtvé na několika pohřebištích vat, důležité je však i to, že jiné ne (např. žádné změny v rámci (obr. 1), avšak pro tuto studii je podstatné pohřebiště č. II, kde jednotlivých chronologických stupňů). Ačkoliv nemůžeme inter- byl prozatím prozkoumaný nejvyšší počet hrobů (obr. 2). Dato- pretaci výsledků generalizovat na celý datový soubor, na určitém váno je mezi 2. polovinu 1. století a počátek 3. století (Pichlerová stupni pravděpodobnosti ji lze považovat za autentickou. 1981). Keramický soubor z této nekropole sestává ze 409 nádob (většinou se jedná o fragmenty). Pro výpočet objemů bylo pou- Contacts žito 121 nádob (tab. 1), které se dochovaly kompletně, případně Alina Szabová alespoň z větší části. Všechny nálezy se nacházejí v depozitáři Ústav archeologie a muzeologie Archeologického múzea Slovenského národného múzea (SNM). Filozofická fakulta, Masarykova univerzita Nádoby byly na základě tvaru klasifikovány do pěti skupin Arna Nováka 1 (obr. 3–12): džbány, misky a talíře, baňaté nádoby, urny a ostatní CZ-602 00 Brno (vázy, amfory a poháry). Po vygenerování 3D modelů a výpo-

čtu objemů následovalo testování hypotéz, které se zabývaly buď statisticky významnými rozdíly nebo korelací mezi objemy Zuzana Porubčanová (či tvarem nádob) a různými okolnostmi pohřebního ritu (typ Ústav archeologie a muzeologie pohřbu, typ hrobu, věk a pohlaví pohřbeného, datování hrobu Filozofická fakulta, Masarykova univerzita a poloha nálezu). Tvorba modelů a výpočet objemů byly reali- Arna Nováka 1 zovány ve volně dostupném software Blender v2.79 za pomoci CZ-602 00 Brno kresebné a fotografické dokumentace a rozměrů jednotlivých

nádob. Hypotézy se testovaly za použití univariačních statis- tických metod ve volně dostupném software Past3 pomocí pa- rametrických a neparametrických testů. Vizualizace frekvence a distribuce dat jsou prezentovány prostřednictvím různých grafů a diagramů (graf 1–10). Nejvyšší průměr měly objemy džbánů a uren, nejnižší pak zase misky a baňaté nádoby. Obě dvojice keramických tvarů měly zároveň podobnou objemovou škálu. V hrobech se vyskytovaly obvykle nejvýše čtyři nádoby, přičemž nejčastější byly různé kombinace džbánu, misky/talíře a baňaté nádoby. V žárových hrobech byla zjištěna nejen výrazně vyšší frekvence nádob, ale i signifikantně vyšší průměr jejich objemů. Druhá uvedená sku- tečnost je však částečně způsobena přítomností uren, které se samozřejmě v kostrových hrobech neobjevily. Vyšší počet ke- ramiky v kremačních hrobech koreluje i s ostatními hrobovými nálezy na pohřebišti. Je proto možné, že žárově byla pohřbívána ekonomicky silnější sociální vrstva populace (Krekovič 2011). Objemy uren dospělých jedinců byly signifikantně vyšší než ob- jemy uren dětí, které však korelovaly s objemy uren jedinců, je- jichž věk nebylo možné identifikovat. Z tohoto důvodu můžeme předpokládat, že minimálně některé z věkově neurčených hrobů jsou dětské. Výskyt objemnějších nádob v hrobech dospělých je- dinců byl prokázán i v případě džbánů. Zdá se, že nádoby s nej- vyšším objemem dominují v hrobech dospělých mužů. Po nich následují ženské hroby a nakonec hroby dětí. Frekvence všech nádob byla naopak vyšší v hrobech žen, a zejména neurčených jedinců, přičemž značné procento zde s jistotou tvořily děti, je- jichž ostatky je však složité pohlavně identifikovat. Testování 104