THE ANTHROPOLOGICAL ASSESSMENT OF THE LATE ROMAN CEMETERY AT SOMOGYSZIL-DÖGKÚTI DŰLŐ

The results of a general anthropological examination of 140 individuals from a late Roman period cemetery at Somogyszil-Dögkút site are presented in this paper. The population had a more or less balanced sex ratio, lived a fundamentally peaceful life suggested by the low frequencey of bone injuries, and according to their morphoscopic traits, they all belonged to the Caucasoid group. Based on the biological distances calculated from selected linear measurements of male crania, the population of Somogyszil-Dögkút proved to be quite similar to several other late Roman period cemeteries in Transdanubia, as well as to some local Avar period series. This raises the possibility of a significant local continuity between the late Roman and late Avar period on this territory, however other potential explanations cannot be ruled out. Some anthropological characteristics of the human skeletal material unearthed from graves oriented differently than the cemetery’s norm suggest the presence of immigrants in the community. Their biological background cannot be traced from the present data, however a few skeletal evidence proposes the probability of a Sarmatian origin.


INTRODUCTION
The first finds of a late Roman period cemetery at Somogyszil-Dögkút was found in 1964, and until the end of 1968 altogether 148 graves had been excavated by the archaeologist Balázs Draveczky. 1 A full archaeological publication of the cemetery was given first by Alice Sz. Burger 2 with some basic anthropological information about the human skeletal remains provided by Tibor Tóth. However, recently both the archaeological and anthropological material was put to a re-assessment. 3 In this paper a more detailed version of the results of the anthropological examination is presented.

MATERIAL AND METHODS
The human skeletal material of Somogyszil-Dögkúti dűlő site is housed in the Department of Anthropology of the Hungarian Natural History Museum under the inventory numbers 68.126.1. -68.126.124. and 68.150.1. -68.150.17. The general preservation of the bones are quite mediocre on average with many of them being strongly incomplete and/or badly preserved.
For scoring morphological sex, altogether 21 anatomical characteristics indicating sexual dimorphism were used. 4   varied. The presence of maxillary torus is quite rare, whereas no shovel-shaped incisors occurs in the studied material. There are differences in the frequencies of palatinus torus and fossa praenasalis as well as in the occurrence of the curved and straight foreheads between the two sexes.
The metric data and indices of the skulls are presented in Tables 5-6, while Table 7 shows the summary statistics of the skulls.
On the basis of the cranial indices, the majority of the late Roman community of Somogyszil-Dögkút can be characterized by a medium long/long (length-width index), low/moderately high (length-height index), and moderately high/high (height-width index) skull. Male foreheads are most often moderately wide or wide (transversal-frontoparietal index), whereas this index shows greater variability among females. Cranial capacity falls into the moderately large/large categories in both sexes. The face and upper face are medium wide/medium high in around half of the cases. Male orbits are usually low, whereas female orbits are low/moderately high. Nasal width varies in both sexes (Table 8).
For the calculation of the mean sigma ratio only the sigma ratios of those cranial measurements and indices were used that were composed of at least seven individual data. The mean sigma ratio of the cranial measurements is 103.53 for male skulls, and 102.08 for female skulls. The value for the cranial indices is 112.18 for males, and 137.21 for females respectively. Thus, the variability of the linear measurements (bearing information mainly about cranial size) indicates a natural population (the theoretical value for an average level of heterogeneity is 100.00), however the sigma ratios of the indices (bearing information about cranial shape) show a more mixed population, particularly for females, with a greater level of heterogeneity than that is typical for a natural population. This may suggest that the community was composed not just by local people, but some extralocal gene flow (immigrants that differed from the locals in their cranial shapes) were also present in it.
It is important to point out that due to the combined effect of the relatively small sample size and the poor preservation of the skeletal material the number of recorded cranial measurements and morphological attributes are rather small. Thus, every above mentioned result, and the conclusions drawn from it should be handled with caution as the small sample size goes along with an increased possibility of random effects that may influence the results.
Based on morphoscopic traits, every examinable skull belongs to the Caucasoid group. With respect to the traditional (and quite subjective, thus, in our opinion, "semi-scientific") racial typology, it can be stated that Nordoid/Mediterranoid, Cromagnoid and archaic chamaecran types occurred in the population (Figs 2-6).

Metric characteristics of the postcranial skeleton, estimated stature
The individual postcranial measurements and the estimated stature are presented in Tables 9-10. The mean height of males is 165.94 cm, while it is 156.86 cm for females. Taking a look at the N-S/S-N oriented burials, it can be seen that three males from those graves were suitable for stature calculation (Grave no. 18: 171.33 cm; Grave no. 41: 166.05 cm; Grave no. 50: 174.33 cm). If they are excluded from the stature mean calculation, the average height of males lowers to 165.31 cm. Two of these three males (Grave no. 18 and Grave no. 50) had a stature significantly exceeding this mean value. Although far-reaching conclusions can hardly be drawn owing to the extremely low number of cases, if this record is viewed together with the data on skulls, it indirectly underpins the assumption that the N-S/S-N oriented graves contained the burials of an immigrant group that in terms of certain anthropological traits differed substantially from the overwhelming majority of the community using the cemetery.
The average stature (calculated from the lower limb bones) of the population of the Carpathian Basin during the Roman Age was 166.09 cm for males, and 156.91 cm for females. 27 The average height of the Somogyszil-Dögkúti dűlő population fits nicely into this picture.

Traumas and injuries
The frequency of fractures and other injuries caused by accidents or interpersonal violence is fairly low in the Somogyszil-Dögkúti dűlő population. Fracture of the lower limb was identified in one case: a healed fracture was found in the proximal third of the right fibula of an adult woman interred in Grave no. 16. Healed fractures of the upper limb was noted in three cases: a woman buried in Grave no. 20 had broken her right ulna near the distal end, a mid diaphyseal fracture healed with an angulation was observed on the left radius of the man interred in Grave no. 23 (Fig. 8), and a mature man in Grave no. 41 had broken his left radius in the distal third. In addition, one other bone trauma was found: an ~ 11 mm long healed depressed fracture probably caused by a moderately sharp implement was identified on the left parietal bone of an elderly woman from Grave no. 19 (Fig. 9).

Biological distance between Somogyszil-Dögkúti dűlő and other archaeological populations
The pairwise Penrose distances (C R 2 : "size" and "shape" combined) between the Somogyszil-Dögkúti dűlő and other male cranial series are presented in Table 11.
From the comparative male cranial series below the 1% error band (C R 2 ≤0.196) are the late Roman period sites of Esztergom-Bánomi dűlő (C R The biological distance-based relations of the close analogies of Somogyszil-Dögkúti dűlő male cranial series is visualized by a dendrogram (Fig. 10). Accordingly, Somogyszil-Dögkúti dűlő clusters together with the Avar period populations of Kaposvár-Road 61, Site 26, Toponár 40-es őrház [Toponár watchman's house No. 40] and Kereki-Homokbánya. This joins together still within the 1% error band threshold value (0.196) with an other cluster composed of the Avar period series of Kaposvár-Fészerlakpuszta and Zalakomár-Lesvári dűlő II, and the  Table 9. Table 10. Table 11. Penrose distance between the male cranial series of Somogyszil-Dögkút and other comparative series
32 The original metric data providing publication : Bottyán 1975. Roman period cemeteries of Esztergom-Bánomi dűlő and Tác-Margittelep. The other series fall outside of this threshold, and thus, connect to Somogyszil-Dögkúti dűlő more loosely.
With regard to the geographical distribution, the close analogies of the late Roman period population of Somogyszil-Dögkúti dűlő are all from Transdanubia, with a single exception (Tiszaderzs). With respect to the dating of these series, parallels can be found both in the Roman and in the Avar period. Viewed in a broader perspective, this indicates a continuity in the selected cranial measurements from the Roman into the Avar period on the territory of Transdanubia. On a theoretical level, this result may suggest a significant continuity of the local populations, or it may indicate that a large part of those groups that probably immigrated and settled down in the territory during the 5 th -8 th centuries possessed similar craniometric features as the locals had, or the combination of the two. However, with the present analysis, this question cannot be answered. Table 11. Penrose distance between the male cranial series of Somogyszil-Dögkút and other comparative series (cont'd)