Daily Life in the Limesgebiet : Archaeozoological Evidence on Animal Resource Exploitation in Lower Danubian Sites of 2nd–6th Centuries AD

: Roman Limesgebiet was both a border between the Greco–Roman world and barbaricum, but also the contact area between these two parts of the ancient oicumene. In Moesia Inferior, this area was established after the Roman conquest in the 1st century AD, in order to defend the Danube border. This article analyses segments of the cultural landscape development in the Roman Danube limes: e.g., the animal resources of subsistence as a paleoeconomy component and reconstruction of the environmental context in the area. The settlements under archaeozoological study are: Sacidava, Capidava, Dinogetia, Noviodunum, Aegyssus, and Halmyris. The settlements have exploited a relatively large faunal spectrum, with taxa of 19 mammals being identiﬁed. Animal husbandry had a major importance in paleoeconomy and was focused on cattle, sheep/goat, and pig; other domestic species are horse, donkey, dog, and cat. The hunting is of small importance, according to the frequency of animal remains, and the forest species are dominant. Red deer and wild boar are present in all assemblages, and they are the dominant wild species in terms of frequency; roe deer, hare, wolf, aurochs, fox, beaver, badger, Eurasian otter, and pine marten have been also identiﬁed. Aquatic resources are represented by molluscs and ﬁsh.


Introduction 1.Study Area: The Roman Limes in Dobrudja
The Roman conquest and occupation of the Lower Danube Region resulted in a system of fortifications of the Danube border known as the limes-a system composed of three main elements: military roads, different types of fortifications and sites, and a specific environment, which over time has become a cultural landscape in itself.
Special interest in the history and archaeology of Roman borders (Limesforschung), manifested since the end of the 19th century, makes the concept of limes evolve and refine over time.The term "limes" emerges to define the demarcation between the Roman and non-Roman areas [1]; "limes" also depicts, in its classical period of evolution (the 2nd-4th centuries) a space under Roman military protection, considered to be an area of assertion of the supremacy of a superior civilization.On the other hand, it is rather a huge area of contact between the two parts of the ancient oicumene that gradually becomes a specific area, called in the German literature, Limesgebiet ("limes area").The Roman borders were rather key interaction areas as the economy of the empire depended to a considerable extent on trade relations with the neighbouring populations, thus terms like "separation areas" or "connection areas" are more suitable to define the realities of the limes [2].
Although the Roman limes area has been intensively researched since the end of the 19th century, for the Dobrudjan segment no reliable data can be attributed to their foundation, being much more realistic to accept some stages.Perceived as a geopolitical Diversity 2022, 14, 640 2 of 16 unit since antiquity, named by Strabo as ή µικρά Σκυθία (Strabo, VII, 4, 5), the territory nowadays known as Dobrudja was gradually military and administratively organized over the course of the 1st century AD, according to the Roman criterion.However, being permanently a constituent part of larger administrative entities, the creation of limes structures in Dobrudja is inextricably linked to the creation of limes structures in the larger entity, namely Moesia, and the administrative entities derived from it.The constitution of the Moesic limes, bordering the bank of the Danube River, began with the appearance of the province of Moesia at the end of the Augustus reign (27 BC-14 AD) and the beginning of the reign of Tiberius (14-37 AD) [3].From the Flavian times there seem to have been merely outposts in Dobrudja [4], the strongest control structure of the Danube course was classis Flavia Moesica [5], probably formed around 45-46 AD [6].In the second half of the 1st century this area became strategically more valuable, Vespasian (69-79 AD) significantly strengthened this stretch of the frontier while reorganizing the entire Empire's defence [7].The foundations of the limes are now attributed to Moesian governor Rubrius Gallus, in response to the Daco-Sarmatian invasion of 69-70 AD, by establishing a series of garrisons in the area [3].Trajan (98-117 AD) completed the engagement of this space within the limits of the Roman administrative and fiscal system [8], a situation consolidated under emperors Hadrian (117-138 AD) and Antoninus Pius (138-161 AD).After the division of Moesia and creation of the two new provinces by Domitian (81-96 AD) (in 86 AD, Moesia Superior and Moesia Inferior), the limes on the Lower Danube acquired its definite form which remained basically unaltered until the end of the Principate (284 AD).Due to the administrative reforms started by Diocletian (284-305 AD), Moesia Inferior, from which most of its maritime territories were separated, became the new province Scythia Minor [9].The inscription in which is mentioned C. Aurelius Firminianus, vir perfectissimus, dux limitis Scythiae [10] (p.155) is the earliest proof of the existing new province, dated between 286-293 AD [11] or 285-292 AD [12].During the 6th century, but prior to 535 AD, according to Hierocles (Synecdemos), 15 urban centres were mentioned in this province, of which 12 were on the territory of Dobrudja, including Capidava, Noviodunum, Aegyssus and Halmyris [13], which are settlements relevant to the present study.
After a short period of economic recovery and revival, as a result of the measures taken by Justinian (527-565 AD), the Dobrudjan limes were abandoned sometime in the second decade of the 7th century by Heraclius (610-641 AD), under Avar-Slavic pressure and the strategic losses suffered by the Empire in the East.Before that, however, the limes was partially destroyed in 598 AD-the mapping of the last coins found in fortresses show that only the eastern part of the limes survived after 598 AD [14].The gradual and constant diminution of Roman military power and economic activity as a whole led to the cessation of life in many localities in Dobrudja and the final disappearance of what was the Roman limes.The real cause was the general economic decline that affected almost all the provinces in the outskirts in the second half of the 6th century [14].
In the present study, we take into consideration only settlements located in Romanian Dobrudja, a territory delimited south by the border with Bulgaria and north by the Danube River, which separates it from Ukraine, with an area of about 15,485 km 2 .The specification is necessary because, over time, the name Dobrudja has also been used for territories currently in the territorial composition of the Bulgarian state.

Settlements under Archaeozoological Study: Archaeological Context
In the following, we present a brief description of the archaeological situation of the settlements on the limes from which the faunal material analysed in this study was collected; the order of presentation is from south to north of the territory (Figure 1).
Sacidava.Systematic archaeological research took place between 1969 and 1982, and has been resumed in recent years by a team of archaeologists from the Museum of National History and Archaeology in Constant , a [15].We identified a Roman fortress (2nd-3rd centuries) and a late Roman and Roman-Byzantine fortification (4th-7th centuries).The first fortification seems to date from the time of Trajan and is related to bellum Dacicum.
During the Principality, the auxiliary camp was occupied by the cohors IV Gallorum, and after the year 114 AD, from cohors I Cilicum milliaria.Stamps of legio I Italica and legio XI Claudia were also discovered [16].In the late Roman era the fortification was rebuilt, attesting the stationing of a cuneus equitum scutariorum [17].Archaeologically, 12 construction phases have been identified, 10 of which belong to the Dominion era, including from the time of Justinian until the first decades of the 7th century [18].
Capidava.This early Roman and late Roman fortification was built in the first phase at the beginning of Trajan's reign and then rebuilt or repaired in the late era.The vestiges of the early era are less known, but it is assumed that the fortification had approximate dimensions of 105 × 125 m, and based on the epigraphic material we know that the fortification was successively occupied by cohors I Ubiorum [19] (p.24) and then by cohors I Germanorum [19] (pp.16,36) starting with Hadrian's reign.Stamps of the legio XI Claudia [19] (p.53) and legio V Macedonica [19] (p.54) were also discovered as were stamps of the legio I Italica (and of the beneficiarius consularis [19] (p.41), probably in connection with statio portorii, certified at Capidava).In the late period, the fortification was rebuilt after the Gothic destruction in the middle of the 3rd century and was occupied by new troops: cuneus equitum Solensium and vexillatio Capidavensium.Recent research in military baths related to the first castellum revealed the constructive contribution of legio XI Claudia pia fidelis to the edification of the ensemble in the first decade of the 2nd century, with expansion and repairs in the following decades [20].
Dinogetia.The early fortification is completely superimposed by the late fortification.Inscriptions and tegular material from legio V Macedonica, legio I Italica, cohors II Mattiacorum [19] (pp.260, 267), cohors I Cilicum [19] (p.264), as well as a statio of the classis Flavia Moesica [19] (p.263) were discovered there.The late Roman fortification has 1.2 ha and a trapezoidal shape, with 14 towers (horseshoe-shaped and fan-shaped), a monumental gate to the east and secondary gates to the north and west.Streets, housing estates (including a 28 × 19 m domus), and warehouses were identified, as was a late Roman praetorium/principia, a basilica from the 5th century [21].The baths, from the 3rd to the 4th centuries (25 × 15 m), located 100 m south of the fortress, have been researched and preserved [20].
Noviodunum.This was described as the main statio of the fleet, classis Flavia Moesica [19] (pp.273, 281).After 168 AD, traces of some detachments from legio I Italica [19] (p.271) and from legio XI Claudia pia fidelis [19] (p.276) were archaeologically attested.The civil settlement, probably a civitas or a vicus Noviodunum [19] (p.233), became municipium Noviodunum, sometime in the early 3rd century or even earlier, during the joint reign of Marcus Aurelius and Commodus (177-180 AD) [16].In the late period, the fortress became the headquarters of legio I Iovia Scythica and a shelter for milites primi Constantiani [19] (pp.284, 271, 276).There is archaeological attestation of the late Roman fortification, trapezoidal in form, surrounded by several defence ditches.The northern enclosure, for the most part, is today under the waters of the Danube River.It is underwater for a length of about 200 m.The walls are 3 m thick, and seven semi-circular towers are evident.The southern enclosure is the most visible as three towers have been unearthed, one corner (fan-shaped), one intermediate (horseshoe-shaped), and the middle, a rectangular bastion (over 35 m long and four supporting pillars of the interior superstructure) [20].
Aegyssus.Originally an indigenous fort, Aegyssus was garrisoned by the Romans quite early, during the 1st century [19] (p.286).It served as statio of the Danube fleet.During the Principality, an auxiliary camp probably operated there.During the archaeological research, seven towers were discovered (two presumed), outlining constructions made in the 3rd, 4th and 6th centuries, and also stamps of cohors II Flavia Brittonum and classis Flavia Moesica.In the 4th-5th centuries, Aegyssus became the garrison of one cuneus equitum armigerorum and especially the headquarters of a praefectus ripae legionis I Ioviae cohortis V pedaturae inferioris.It was rebuilt during Justinian's time, following a construction process initiated under Anastasius (491-518 AD).Three construction phases were identified at the enclosure wall and the related towers [20].
Halmyris sive Thalamonium (Salmorus?).A fortress and statio of the Danubian fleet-many former sailors of the fleet founded near Halmyris, including even a sailors' village, vicus classicorum [22], inhabited by veterans [16].Due to archaeological research from the auxiliary camp, two phases for the early period were highlighted: the wooden camp, hypothetically dated in the time of Trajan, and the stone camp, a trapezoidal form, probably erected at the beginning of Hadrian's reign.Repairs were made in the time of Caracalla (198-217 AD); stamps and inscriptions of legio I Italica and legio XI Claudia were identified.The Late Roman fortification was provided with 12 "U"-shaped towers and was surrounded by two waves of defence to the west and three to the south.It began to be rebuilt during the reigns of Aurelian and Probus, and was completed at the beginning of the Tetrarchic era (a refoundation inscription).It was repaired during the time of Theodosius I and Anastasius, when it appeared in sources as a civitas, and was rebuilt in the time of Justinian [20].

Materials and Methods
Archaeozoological studies, giving information on animal resources for food, husbandry strategies, or consumption practices, can contribute to the knowledge of the common elements in the Roman world, but also of the geographic and cultural diversifications by region.In this context, the present study can contribute with archaeozoological data regarding the above mentioned six settlements.
The faunal samples considered in this study contain household waste that reflects the practice of mollusc harvesting, fishing, and animal husbandry.We mention that the 5 of 16 remains of fish and birds should be considered undervalued, since the sediments of the sites from which they come have not been sieved.All the samples were hand collected, and this can cause an overrepresentation of large size animals (e.g., cattle, red deer), but also an underrepresentation of small ones (e.g., small mammal carnivores, fish, and birds).The quantity of faunal material varied by archaeological site (e.g., small samples of 178 remains at Capidava and 180 remains at Dinogetia; large samples of 2207 and 3553 remains at Halmyris).Bone preservation was variable from one site to another, but generally it was in good condition.
The archaeozoological sample was analysed according to standard procedure [23].The faunal remains, anatomically and taxonomically identified, are described in terms of their frequency based on the number of identified specimens (NISP) and the minimum number of individuals (MNI).

Results and Discussion
Archaeozoological information for the sites of Capidava, Dinogetia, Noviodunum, and partially Halmyris is taken from the literature, as shown in Table 1; the information for the sites of Sacidava, Aegyssus, and partially Halmyris are our own unpublished data.Molluscs.Molluscs are often used as a food source in various human communities, past and present.In the analysed faunal samples, the presence of molluscs is reduced, as shown in Table 1.Thus, mollusc shells were reported only for the sites of Dinogetia, Aegyssus, and Halmyris.In the Dinogetia faunal sample, 16 remains of snails (Helix sp.) were identified [27]; in the Aegyssus sample, only two remains of freshwater mussels (Unio sp.); in the case of Halmyris sample, El Susi [24,25] mentions just nine remains of molluscs, but as unpublished data freshwater and marine mussels are identified (i.e., Unio sp.-11 remains, and Mytilus sp.-one, respectively).
Fish.Although we would have expected the frequency of fish remains to be high in the samples presented as the sites are located on the banks of the Danube, the archaeozoological data indicate an underestimation of them, as mentioned above, mainly due to the collection method-by hand and not by sieving the sediments.Fish remains were identified at all sites considered in the present study (Table 1).Sturgeon and Teleost fish have been identified.These species are the same as species present today in the area where the sites are located.Except for sturgeons, which are anadromous, the rest of the species mentioned in Table 2 are freshwater fish.Catfish (Silurus glanis) and carp (Cyprinus carpio) are more commonly mentioned in the previous studies.Reptiles.Only three dermal plaques of turtle (Testudo sp.) have been identified from a sample of Halmyris (Table 1).
Birds.The presence of bird remains in the studied samples is low; bird remains are likely underestimated, as in the case of fish.Domestic species were identified (Table 3), and many of the unidentified remains at the genus/species level belonged to wild birds.Mammals.Mammal remains, domestic and wild, predominate in all samples considered in this paper (Table 1).The proportion of unidentified mammal remains by species is relatively low, varying between 14% in Sacidava and about 32% in Noviodunum.
As Figure 2 and Tables 4 and 5 show, among the identified mammal remains, species of domestic mammals predominate in all samples both as NISP (number of identified specimens) and MNI (minimal number of estimated individuals).
Wild mammals.The list of identified wild mammal species can be followed in Figure 3 and Tables 4 and 5. Red deer (Cervus elaphus) and wild boar (Sus scrofa) appear in all samples and are the predominant remains of the wild mammal group (Figure 3).Roe deer (Capreolus capreolus) follow in order of frequency of remains identified in the four sites, then the aurochs (Bos primigenius), which are present in five sites.Smaller wild mammals are less represented; they appear better represented in the site of Halmyris.Hunting was mainly practiced for local economic purposes (e.g., procurement of meat and raw materials such as deer antlers, hides, and furs), but also for internal or external trade needs.
In order to evaluate the biotopes exploited by hunting, the identified wild mammals were grouped according to the ecological affinities in forest species (i.e., red deer, wild boar, and pine marten), forest edge/open field species (i.e., roe deer, aurochs, and hare), semiaquatic species (i.e., otter and beaver), and eurytopic species (i.e., wolf, fox, and badger).As Figure 4 shows, in all the analysed sites the forest species predominate, followed by those of the forest edge/open field area (excepting the site of Dinogetia, where the semiaquatic species are in the second place for frequency).Mammals.Mammal remains, domestic and wild, predominate in all samples considered in this paper (Table 1).The proportion of unidentified mammal remains by species is relatively low, varying between 14% in Sacidava and about 32% in Noviodunum.
As Figure 2 and Tables 4 and 5 show, among the identified mammal remains, species of domestic mammals predominate in all samples both as NISP (number of identified specimens) and MNI (minimal number of estimated individuals).and Tables 4 and 5. Red deer (Cervus elaphus) and wild boar (Sus scrofa) appear in all samples and are the predominant remains of the wild mammal group (Figure 3).Roe deer (Capreolus capreolus) follow in order of frequency of remains identified in the four sites, then the aurochs (Bos primigenius), which are present in five sites.Smaller wild mammals are less represented; they appear better represented in the site of Halmyris.Hunting was mainly practiced for local economic purposes (e.g., procurement of meat and raw materials such as deer antlers, hides, and furs), but also for internal or external trade needs.In order to evaluate the biotopes exploited by hunting, the identified wild mammals were grouped according to the ecological affinities in forest species (i.e., red deer, wild boar, and pine marten), forest edge/open field species (i.e., roe deer, aurochs, and hare), semiaquatic species (i.e., otter and beaver), and eurytopic species (i.e., wolf, fox, and badger).As Figure 4 shows, in all the analysed sites the forest species predominate, followed by those of the forest edge/open field area (excepting the site of Dinogetia, where the semiaquatic species are in the second place for frequency).In order to evaluate the biotopes exploited by hunting, the identified wild mammals were grouped according to the ecological affinities in forest species (i.e., red deer, wild boar, and pine marten), forest edge/open field species (i.e., roe deer, aurochs, and hare), semiaquatic species (i.e., otter and beaver), and eurytopic species (i.e., wolf, fox, and badger).As Figure 4 shows, in all the analysed sites the forest species predominate, followed by those of the forest edge/open field area (excepting the site of Dinogetia, where the semiaquatic species are in the second place for frequency).Auroch, beaver, and red deer are identified in the settlement under study, but today they are not found in the area, and the first species, auroch, is now extinct.Domestic mammals.The frequency of domestic mammals varies slightly from one sample to another both as NISP and as NMI (Tables 4 and 5).The following species have been identified: cattle (Bos taurus), sheep (Ovis aries), goat (Capra hircus), pig (Sus domesticus), horse (Equus caballus), donkey (Equus asinus), dog (Canis familiaris), and cat (Felis domesticus).Cattle, sheep/goat, pig, horse, and dog are present in all the analysed sites; donkey does not appear at Noviodunum; cat is rare and only reported at Halmyris.
Cattle dominate in all samples, both as NISP and as MNI (Tables 4 and 5; Figure 5).Sheep/goat are in second place in frequency, both as NISP and as MNI, in Sacidava, Capidava, and Noviodunum, followed by pig.The positions are reversed at Dinogetia and Aegyssus, where the pig frequency exceeds that of sheep/goat.However, the frequency differences between sheep/goat and pig are not high.
Cattle, sheep/goat, pig, horse, and dog are present in all the analysed sites; donkey does not appear at Noviodunum; cat is rare and only reported at Halmyris.
Cattle dominate in all samples, both as NISP and as MNI (Tables 4 and 5; Figure 5).Sheep/goat are in second place in frequency, both as NISP and as MNI, in Sacidava, Capidava, and Noviodunum, followed by pig.The positions are reversed at Dinogetia and Aegyssus, where the pig frequency exceeds that of sheep/goat.However, the frequency differences between sheep/goat and pig are not high.Estimation of slaughter ages focused on dental characteristics, but data are not precise for all the analysed samples.For some samples (e.g., Capidava, Dinogetia, and Halmyris) [24,26,27], only vague estimates of slaughter ages were published.Based on the samples that we analysed (Noviodunum, Aegyssus, Sacidava, and Halmyris) (Figure 6), the ratio by dental age categories in cattle and sheep/goat is in favour of mature specimens, more pronounced in the case of cattle; in pig, immature specimens slightly predominate (Table 6).In the main domestic mammal species (Bos taurus, Ovis aries/Capra hircus, Sus domesticus), the age of slaughter was estimated based on dentition (the replacement of deciduous dentition, and M3 tooth wear) [23].The estimated individuals were grouped into two classes: immature (deciduous dentition; not fully erupted M3 molar) and mature (fully erupted M3 molar; M3 in wear stages).These results suggest that the exploitation of cattle was oriented to obtaining by-products (e.g., milk, traction force, and reproduction), especially in Noviodunum and Halmyris settlements.The exploitation of sheep/goat Estimation of slaughter ages focused on dental characteristics, but data are not precise for all the analysed samples.For some samples (e.g., Capidava, Dinogetia, and Halmyris) [24,26,27], only vague estimates of slaughter ages were published.Based on the samples that we analysed (Noviodunum, Aegyssus, Sacidava, and Halmyris) (Figure 6), the ratio by dental age categories in cattle and sheep/goat is in favour of mature specimens, more pronounced in the case of cattle; in pig, immature specimens slightly predominate (Table 6).In the main domestic mammal species (Bos taurus, Ovis aries/Capra hircus, Sus domesticus), the age of slaughter was estimated based on dentition (the replacement of deciduous dentition, and M3 tooth wear) [23].The estimated individuals were grouped into two classes: immature (deciduous dentition; not fully erupted M3 molar) and mature (fully erupted M3 molar; M3 in wear stages).These results suggest that the exploitation of cattle was oriented to obtaining by-products (e.g., milk, traction force, and reproduction), especially in Noviodunum and Halmyris settlements.The exploitation of sheep/goat seems to have been of the mixed type, both for primary products (e.g., meat) and secondary ones (e.g., milk and wool).In pig, the interest was in primary products, obtained by slaughtering animals at an optimal age.
In other provinces on the limes, a similarity is found in the ratio between the animal species raised or hunted, the differences being given by local specifics (e.g., growing conditions, habits of the local population, and the military, etc.).
Diversity 2022, 14, x FOR PEER REVIEW 12 of 17 seems to have been of the mixed type, both for primary products (e.g., meat) and secondary ones (e.g., milk and wool).In pig, the interest was in primary products, obtained by slaughtering animals at an optimal age.In Roman Switzerland, a borderland with one legionary camp (Vindonissa) and some military camps until the end of the first century AD, but a heavily militarized area starting from the second century AD, cattle was the most important animal species used.Thus, the ratio of cattle, pig, and sheep in the civilian town of Augusta Raurica is 45:38:17 on average.At Vindonissa, in military contexts, the food habits of the three successive legions living there were quite different: while legions XIII and XI preferred pig, legion XXI consumed mainly sheep/goat, but in the nearby canabae mostly cattle bones were found.Other examples are given by the results of food customs in three military fortresses of the 4th century AD, as following: in Kaiseraugst (Castra Rauracense), cattle clearly dominate (60%); in Schaan (Liechtenstein), cattle and pig have equal frequencies (around 30% each), and sheep/goat is relatively well represented (21%); in Pfyn (Ad Fines), pig and cattle are also equal represented (around 40% each), but the frequency of wild animal remains is rather low (6%) [32].
In Gamzigrad-Felix Romuliana, Serbia, according to archaeozoological results, domestic species were the main source of meat and secondary products (e.g., milk, physical strength, wool, and hair) [33].
In Bulgaria, Novae, the consumption of domestic animals reached an average of 93-94%.During the 2nd-6th centuries, cattle registered 35-55%, pigs 30-45%, and sheep/goat 22-29% [34].Recent archaeozoological analyses confirm that the meat eaten in Novae had been principally supplied by pig and cattle, and less by sheep/goat [35].More, the comparison between archaeozoological samples from Novae, Iatrus, and Nikopolis-ad-Istrum shows a difference between Novae and Iatrus, but there is a similarity between Novae and Nikopolis.In Iatrus, the contribution of cattle was definitely higher than that of pig, while in the other two towns, there was a domination of pig over cattle.ln all settlements, the remains of sheep/goat were in the third or second position [36].At Nikopolis-ad-Istrum, in the 2nd-6th centuries, the situation of the three main domestic animals can be presented as follows: cattle about 20.6%, pig about 42%, and sheep/goat about 32.2%.During the early Roman period (100-175 AD), pig and sheep/goat show similar frequencies, followed by cattle; but from the middle Roman period (175-250 AD) until the early Byzantine period (450-600 AD), pig seems to have dominated, followed by sheep/goat and then cattle.Sites such as Iatrus (late 4th-late 5th centuries) have high amounts of cattle (67-72%), with less pig (15-20%) and sheep/goat (13%).Other sites such as Béla Voda, Novae and the earliest and latest phases at Iatrus (early 4th century and late 7th century) still have a predominance of cattle but in lower percentages (39-48%), with higher amounts of pig (20-38%) and sheep/goat (20-32%).Nikopolis-ad-Istrum, in contrast, has higher frequencies of pig (37-53%) and sheep/goat (30-37%) but less cattle (17-27%) [37].
On the Dacian limes, as a result of the archaeozoological analyses (samples of Micia-Vet , el, Bologa-Resculum, Românaşi-Largiana, Brâncoveneşti, Răcari, Buridava-Stolniceni, Apulum, and Porolissum), it was found that animal husbandry was an important source for meat, milk, work force, etc.The Roman army favored the spread of large sized animals with increased muscle mass.In the settlements where cattle predominated, there was a small number of pig and vice versa; in the rural sites the pig was preferred more than in the urban ones [38].
For a better understanding of the presented data on dental age of slaughter in the main domestic species, a comparison with the archaeozoological data from other settlements of the Roman Empire, on the Germanic limes (Netherlands) and the Danubian one (Novae, Nikopolis-ad-Istrum) is useful.Thus, the research carried out for the 1st-2nd century AD dwelling in the Houten-Castellum area (Netherlands) regarding cattle mortality profile indicates the highest slaughter peak occurred at 18-30 months, while mortality of young calves (0-1 month) was low (5-8%); 12-17% were killed at 8-18 months [39].The archaeozoological analysis of three samples from Tiel-Passewaaij (administratively belonging to civitas Batavorum, province of Germania Inferior) indicates that, in the 1st century to the beginning of the 2nd century, the mortality profile for cattle (39.1%) peaked at 18-30 months and 43% of all cattle survived to adulthood; in the 2nd century until the beginning of the 3rd century, the mortality profile for cattle (55.9%) shows a peak at the prime ages for beef (18-36 months) and an overrepresentation of adult and older cattle (24%) is evident, but more than two-thirds of cattle survived to over four years; in the end of the 3rd century to the 4th century, the mortality profile for cattle (47.1%) shows peaks at 8-18 months and 18-30 months.Also, pig (22.1%) was mainly killed at 14-21 months and 21-27 months [40].
Closer to the studied area in this work, in Bulgaria, on the Danube Limes (Novae), the obtained data indicate that cattle intended for consumption was aged from about 3 months to over 10 years.However, the greatest amount of meat was supplied by fully grown individuals with fully developed muscle mass, aged 3.5-7 years.In the case of pig, the inhabitants of Novae consumed meat and fat from animals slaughtered at the ages of 2-3 months and 8 years.Most of the meat came from individuals slaughtered at the age of 2-3.5 years.Sheep/goat meat came from individuals slaughtered at an age of 3 months to 4-5 years [34].At Nikopolis-ad-Istrum, pig generally appear to have been killed at different ages ranging from juveniles to mature adults.Most animals, however, were slaughtered before attaining adulthood [37].

Conclusions
The archaeozoological study describes animal resources exploited in Lower Danubian settlements of 2nd-6th centuries AD, contributing to the daily life reconstructions of the Dobrudja area of the Limesgebiet.
The Dobrudjan segment of the Danubian limes proved to be a dynamic structure, which contributed both to the Romanization of the territory and to the dissemination of Roman culture and civilization north of the Danube River in barbaricum.A substantial contribution was made by Roman army troops relocated there in large numbers, attested archaeologically as the headquarters of the Danube fleet, four legionary camps, six auxiliary camps, eleven camps and a network of small and medium fortifications.One of the basic needs of the garrisons stationed in these military units was the supply of food.Regarding the supply of products of animal origin, specific and continuous activities have been carried out in this respect, both in terms of livestock and wildlife.Thus, it is possible to document the constitution of a new society specific to limes, but also the identification of some Roman cultural patterns of everyday life, as it appears so far from the research of the sites presented in this study.
Animal husbandry was a major component in the food economy in all the studied settlements.The remains of domestic mammals are predominant in all samples, compared to wild mammals; they range from 78.8% at Halmyris [24,25] to 97.8% at Noviodunum [29,30].Eight domestic species have been identified.Communities mainly consumed meat of cattle, sheep/goat, and pig, which appear with significant frequencies in all samples, with the predominance of the former.Cattle remains are dominant in all samples (except in one of Halmyris), the proportion ranging between 41.5% at Dinogetia [27] and 60.2% at Noviodunum [28].

Figure 2 .Figure 2 .
Figure 2. Proportions of wild and domestic species in the total identified mammal remains: (a) NISP-number of identified specimens; (b) MNI-minimal number of individuals.

Figure 3 .
Figure 3. Proportions of species in the wild mammal group.

Figure
Figure 4. Proportions of wild mammal species by ecological group.

Figure 3 .
Figure 3. Proportions of species in the wild mammal group.

Figure 3 .
Figure 3. Proportions of species in the wild mammal group.

Figure
Figure 4. Proportions of wild mammal species by ecological group.

Table 1 .
Distribution of animal remains by taxonomic group.

Table 2 .
Identified fish in Lower Danubian archaeological sites.

Table 3 .
Identified birds in Lower Danubian archaeological sites.

Table 4 .
Quantification of mammal remains by NISP (number of identified specimens) in the assemblages.

Table 5 .
[24,25]ication of mammal remains by MNI (minimal number of individuals) in the assemblages.*Quantification of mammal remains by MNI is not complete in the publications of El Susi[24,25].

4 .
Proportions of wild mammal species by ecological group.

4 .
Proportions of wild mammal species by ecological group. ).