Limitatio

vol. XIII p.672-701

Limitatio

1. Sources and literature

The main sources are the Roman agrimensores. In the following article, they will be cited according to Lachmann’s Gromatici veteres (Blume, Lachmann and Rudorff Schriften der röm. Feldmesser I 1848), despite the fact that Thulin’s edition in the Corpus agrimensorum Romanor. I 1, 1913, which unfortunately has remained incomplete, has been a great step forward: in that edition, he regularly provides helpful cross-references. <60> Here, I have only provided page numbers for Thulin’s edition where it would be difficult to locate it from Lachmann’s numbering, if for instance their ordering is different, or if their readings are fundamentally different. <page break 672/673> As well as the articles on agrimensores (see vol. I p.894) and gromatici (see vol. VII p.1886), and Thulin’s edition itself, see his works on the manuscripts of the agrimensores Abh. Akad. Berl. 1911 Anh. II and Rh. Mus. LXVI 1911, 417ff., and on the history of the transmission of the corpus agrimensorum, see Goeteborg’s Handlingar 1911/12. In addition to inscriptional evidence, the remains of Roman limitationes in today’s road-networks can be used as a source (see below sect. 20-24). <10> Out of the modern literature, the most important works are Rudorff Gromatische Institutionen, Feldmesser II 227ff. and Mommsen Die Libri coloniarum ibid. 143ff.; Ges. Schr. V 146ff. (on the last, see Pais Memorie Accad. dei Lincei XVI 1920, 55ff.), as well as Nissen Die L., Templum 1869, 1ff. Max Weber Röm. Agrargesch. 1891. Mommsen Zum röm. Bodenrecht, Herm. XXVII 1892, 79ff.; Ges. Schr. V 85ff. Meitzen Die röm. Landmessungen und Flureinteilungen, Siedelung und Agrarwesen I 1895, 284ff.; <20> other works by Schulten, especially Die röm. Flurteilung und ihre Reste, Abh. Ges. d. Wiss. Gött. N. F. II nr. 7, and in particular Barthel Röm. L. in der Provinz Africa, Bonn. Jahrb. ĈX 1911, 39-117. Cf. also the short overviews in Kornemann’s article Coloniae vol. IV p.573f., as well as the articles Centuria nr. 5 (Kubitschek), Cardo and Decumanus (Schulten), and Limes (Fabricius). <30>

2. History

Limitatio refers to the uniquely Roman practice of partitioning the ground into equally-sized pieces by drawing strictly vertical and strictly horizontal limites, which are parallel to each other and which cross each other at right-angles, and which are equally spaced (Hyg. Gr. 181, 14 = 145, 19 Th.: omnis limitum connexio rectis angulis continetur). According to the tradition which we can trace back to Varro, limitatio was considered to have had Etruscan origins (Front. 27, 13. Hyg. Gr. 166, 10. Nissen 10ff. On the potential etymology of groma as being derived from the Greek γνώμων via Etruscan, see Schulze S.-Ber. Akad. Berl. 1905, 709). <40> We encounter limites in the Oscan writings on the Cippus Abellanus (v. 19 liímíto[m], cf. Bücheler Comment. in hon. Theod. Mommseni 233ff.), and decumani on the bronze inscription from Agnone (v. 47 dekmanniúís): this could be because of Etruscan influence, but could also be because of Roman influence, or some other Italic community. <50> In Roman tradition, limitatio first appears at the very founding of the city, in the ager privatus, when Romulus split the land up into pieces of two iugera (heredium) and distributed it among the citizens by lot, with one hundred of those pieces making up a centuria (Fest. p.53: centuriatus ager in ducena iugera definitus, quia Romulus centenis civibus ducena iugera tribuit, see vol. III p.1960, 48ff.). Moreover, the annalistic historians writing about the distribution of land at the founding of the earliest colonies assume it was done by lot, and they also write about land being distributed one portion each to every man when discussing the earliest agrarian laws: both of these things would be impossible without a limitatio of the relevant land (see vol. XII p.1153ff.). <60> The first reference we have to limitatio as an item of general knowledge is in Plautus’ Poenulus, where the poet v. 48f. has the speaker of the prologue say the following about his lot: eius nunc regiones, limites, confinia determinabo, ei rei ego finitor factus sum. All four terms that appear here belong to the technical language to do with surveying. Furthermore, even in the republic, limitationes took place on a large scale outside of Italy. <page break 673/674> The province which the decemviri introduced e lege Livia immediately after the fall of Carthage in 146 (Lex agrar. from the year 111 CIL I² 585 v. 81), in all probability, immediately underwent limitatio across all of its land that had become ager Romanus (see below p.700 sect. 24). <10> The same thing happened to the province Cyrenaica introduced in 74 BCE, where the kingdom which Ptolemy Apion had bequeathed to the Roman people was split into square sections via limites 6000 ptolemaic feet long (Hyg. 121, 25ff. who however believes that this happened under Trajan; the use of local units of measurements for limitationes in the provinces was also rather common. Hyg. loc. cit. Rudorff 282 and 421).

The Gracchan agrarian laws must have contained precise specifications on limitatio. <20> Of the boundary-stones which were used for these purposes, a great deal are extant (CIL I² 639-645), and details from the lex Sempronia are cited in the writings of the agrimensores (169, 1 on keeping the limites clear; 242, 8 on the form and size of the lapides Graccani; and on the form and size of the centuriae and on the orientation of the limites in the distributions made under the lex Sempronia cf. the libri coloniarum passim). <30> Now, after the agrarian reform bill in 111 BCE CIL I² 585, of which fragments are extant, had been passed, and after both the land making up the ager publicus had been declared to either remain in possession of the existing landowners or to be exchanged for ager privatus, and the land distributed by the three members of the Gracchan agrarian committee had been declared ager privatus, the limitatio of these plots of land was overseen by the populares in 109. <40> A large fragment of the relevant plebiscitum from this year can be found in the manuscripts of the agrimensores, where in ch. 53-55 a lex Mamilia Roscia Peducaea Alliena Fabia has been preserved verbatim, though not in the archaic spoken form (p. 263-266, cf. Bruns Fontes⁷ 95ff.; the law has been often misunderstood as the Lex Iulia agraria of 59, see above vol. XII p. 1185. Frontin. writes about the sermo antiquus of the law, cf. Agenn. 37, 24. 66, 15 = 27, 5 Th.). This same law was referred to by Cicero and by the agrimensores by the shorter term lex Manilia (de leg. I 55. Front. 11, 5. 12, 12 = 61, 3 Th. 43, 20. 74, 17 = 33, 14 Th. Sic. 144, 19. Hyg. Gr. 169, 7). <50> Of its authors, C. Mamilius, who gained the cognomen Limetanus from it (Cichorius Röm. Stud. 124), was well known to be a leading popularis, as well as Sex. Peducaeus, whose tribunate is commonly misplaced in the year 113. Like Peducaeus, the three other men who proposed this law will have been ancestors of the Caesarians with the same names, which people have up until now confused with the authors of this particular law. <60> A L. R(oscius) and a C. F(abius) also appear as men in charge of minting coin around the time of 109 BCE (Babelon Monn. républ. Rom. I 485. Grueber Coins of Rom. Republ. II 256). The clauses of this law which refer to keeping the limites clear (c. 54) were transferred into the Lex Ursonensis (c. 104), and the specifications about protecting the termini from c. 55 were transferred into a lex agraria of C. Caesar (Caligula) Dig. XLVII 21, 3, with edited stipulations about the relevant criminal proceedings. <page break 674/675> As far as it is possible to tell, the Lex Mamilia would have contained thorough specifications on limitatio and on all other matters to do with agrarian law relevant to the deductio of colonies and the establishment of municipalities etc. (Fabricius S.-Ber. Akad. Heidelb. 1924/5 Abh. 1, and then Kroll vol. XII p.2397. For opposing views, see Hardy Class. Quarterly XIX 1925, 185ff.). <10>

Additionally, all later agrarian bills, the imperial statutes relating to agrarian matters, as well as the local bylaws in the colonies and municipalities, contained specifications on the size of the centuriae; the width of the limites; the size, form, inscription, and position of the termini; and other things relevant to the limitatio. <20> The works of the agrimensores and the so-called libri coloniarum offer a great deal of evidence for this (on the credibility of the libri coloniarum, which Mommsen considered too unfavourably, see Pais loc. cit. 55ff.). Finally, as well as accounts of limitatio in general, and details about more ancient systems, the writings of the agrimensores also contain recommendations and instructions from the authors on how to carry out new limitationes, or how to verify and correct prior ones. <30>

3. The basic structure of a limitatio

The foundation of every limitatio are the two main limites, perpendicular to each other: the decumanus (maximus) and the kardo (maximus). For an explanation of the term decumanus (a mensura denum actuum, cf. Sic. 152, 26), and the other etymologies brought up by Varro and by others, see vol. IV p.2314f. Rönsch’s attempt to show the term to be Greek in origin (from δαίω, δαίζω, divido and κόμμα incisura), in Jahrb. f. Philol. 1880, 501, have been appropriately rejected by Walde² 222. <page break 675/676> The ancients themselves, theorising that the east-to-west orientation of the decumanus and the south-to-north orientation of the kardo were the original ones, derived the term kardo from a cardine caeli (Front. 28, 15 = 12, 2 Th. and the references there). <10>

The decumanus splits the entire area to be limited into a right half and a left half (see below, fig. 1), while the kardo splits it into a nearer and a further one (fig. 2), each from the viewpoint of the person carrying out the limitatio, as per the system (Front. 28, 9: decimanus dividebat agrum dextra et sinistra, cardo citra et ultra).

These are the terms used regardless of how the system is oriented, regardless of whether it is being used in the sky, or for more practical purposes: they always remain the same. <20> Accordingly, citra cardinem refers to the nearer half - the pars postica, which was also called citrata - while ultra cardinem refers to the further half - pars antica or ultrata. The two halves formed by the decumanus are also called the pars dextrata and sinistra (Front. 28, 3: ultra antica, citra postica nominaverunt). <30> By splitting an area by both the kardo and the decumanus at once, we produce four regiones (Fig. 3), and these were routinely referred to by four letters, forming an abbreviation, as per the following: dextra decumanum citra kardinem (DDKK or CK, from the viewpoint of the person carrying out the limitatio this was the region on the right and close by), dextra decumanum ultra kardinem (DDVK, on the right and further away), sinistra decumanum citra cardinem (SDKK, on the left and close by, as eg. on the Gracchan terminus CIL I² 640, see below p.681) and sinistra decumanum ultra cardinem (SDVK, on the left and further away). <page break 676/675> <50>

Nissen 15 has previously presented this system, and explained it with diagrams that match with fig. 3 here, while Rudorff 341 and 346 and those following him have got ultra = antica and citra = postica flipped the other way around, entirely without evidence to justify it, and entirely in spite of the very clear ancient explanations we have (cf. Barthel 40ff., where the correct terms for the regions are once again emphatically determined. In the same vein, the article in vol. III p.1587 needs correcting. <60> This mistake has thrown all scholarship on the matter into an unfixable confusion. The arguments in Toutain Mém. des ant. de France 1910, 84ff. are also completely wrong, because they start from the assumption that the regiones ultrata and citrata were determined by the viewpoint of the person conducting the limitatio, while the regiones dextrata and sinistrata were determined by the course of the sun. <page break 675/676> <50> The terms sinistra and dextra are, of course, determined for the decumanus in exactly the same way as the citra and the ultra are for the kardo, cf. also Barthel 125 postscript).

Occasionally, the line described by the kardo was also called the postica, presumably because it bordered the area that lay behind the observer, and the line described by the decumanus which extended outwards from the observer and divided the pars antica was also accordingly called the antica: Fest. 233 posticam lineam in agris dividendis Ser. Sulpicius appellavit ab exori[ente sole ….], Paul. Fest. 232 postica linea in agris dividendis ad occasum spectat, cf. Gloss. II 20, 39 antica linea: διάμετρος. (On orienting the system according to the compass points, see below sect. 13). <60> These terms for the two main perpendicular lines, which were also used in augury for the purpose of setting up a templum, are behind the following custom purported by the gromaticus Dolabella: <page break 676/677> the founders of military colonies on conquered territory would, secundum auspicium, fix a cross with the words antica and postica on the doorframe of the temple to their gods (Grom. 303. 22ff.). <10> This claim has been confirmed by a recent [as of 1926] find in Lambaesis in the temple to Aescula: a block of limestone, on the top side of which is a cross in bronze foil, with writing on it as depicted below (Monceaux Comptes de l’Acad. 1920, 179ff., Rev. archéol. 1921, 459).

4. Further division

In order to divide the grid further, equally spaced limites are added parallel to the two main lines, in all four directions. <20> These are also referred to as decumani and kardines, so the main lines are called maximi in order to distinguish them. In earlier time periods, once again going from the perspective of the person carrying out the limitatio, the lines going ‘straight out’, parallel to the decumanus, were called limites prorsi, and those ‘running across’, parallel to the kardo, were called limites transversi (Front. 29, 10: haec vocabula in lege quae est in agro Uritano in Gallia [at Ravenna], item in quibusdam locis adhuc permanere dicuntur. <30> These terms are also used in Hyg. Gr. 206, 10ff.; cf. 167, 16, where once again the system is imagined to be oriented towards the west, hence Barthel 47, 3). In the Lex Mamilia c. 54 and 55, as well as in the article of the Lex Ursonensis c. 104 taken from it, the phrase limites decumanique refers to all of the limites. Mommsen’s view that limes simply refers to the kardo here does not only contradict the Lex Agraria of 111 v. 89 and the Lex Ursonis c. 78 itself, where limites is also used specifically for the decumani, but also every other use of the term limes: moreover, this view contradicts the fundamental meaning of the word limes itself. <40> The use of the two separate words limites decumanique will likely just be pleonasm (see above p.572). <50> Finally, in certain areas people also named the limites a caeli regione (orientales etc. 247, 22) or a loci natura, like eg. at Fanum Fortunae where the limites running along the sealine were called maritimos, and those directed towards the mountains were called montanos, or also Gallicos because they were also directed towards Cisalpina (Front. 30, 1. Sic. 153, 16; other examples Lib. col. 215, 4. 225, 5. 227, 12 etc. Other additional practically illogical terminology can be found in an interpolated section in Liber colon. I 225, 6, on which cf. Mommsen Feldm. II 165f.; Ges. Schr. V 161f.).

5. The widths of the limites

The widths of the different limites were carefully bounded by fixed surveying lines, called rigores, and were carefully maintained by ditches, called sulci, in land that had been ploughed (Balb. 98, 10. Hyg. Gr. 192, 15). <60> The specific widths differed according to the needs of traffic (Agenn. 89, 14. Hyg. Gr. 169, 9. 175, 1). The widths for the different kinds of limites were specified in the Lex Mamilia (Hyg. Gr. 169, 6), and were also prescribed in the Leges datae for the coloniae (Hyg. 111, 9. Hyg. Gr. 194, 3. Sic. 158, 18), and could have had very different measurements. <page break 677/678> The decumanus maximus and the kardo maximus, as the main lines of travel, had the broadest widths: according to a building contract, a lex agris limitandis from the triumvirs, and secundum legem et constitutionem divi Augusti, this was 40 and 20 feet. Every fifth limes (the quintarii, which also served traffic, and were therefore also called actuarii) had a width of 12 feet. The rest, which only served as paths across fields and were only kept clear using hoes, and which were accordingly called subruncivi, were 8 feet wide (Hyg. 111, 12. Hyg. Gr. 168, 9. 194, 9. Lib. col. 212, 4). <10> The elder Hyginus prescribes 30, 15, or 12 feet for the decumanus maximus and the kardo; 8 feet for the subruncivi (111, 12). The limites that lay in between the quintarii were sometimes only set out as simple lines, and were therefore called linearii (Sic. 158, 14). However, according to Hyginus, it was the case for all limites, including the subruncivi and linearii, that their width should be non minus quam qua vehiculo iter agi possit (120, 21). <20> The paths taken up by the limites could also contain water ditches, or fossae limitales. The Lex Mamilia c. 54 and the Lex Ursonensis c.104 after it contain the stipulation that, in any plot of land governed by those laws, limites and decumani could not be obstructed, built on, or ploughed over, nor could the flow of water in the fossae limitales be interrupted. <30> Furthermore, these coloniae laws also stipulated that, if a limes crossed a farm or a building, the owner was required to provide a route around their land or a path across their farm (Hyg. 121, 1. Sic. 158, 22). Likewise, if a limes passed through terrain that couldn’t be traversed, the owners would also have to provide a route around it (Front. 24, 9. 58, 13).

Although the limites had to be kept open, it is unclear whether that simply means that the relevant pieces of farmland had to remain accessible, or whether it meant that they had to be open to all kinds of passing traffic. <40> According to the Lex Sempronia, the Lex Cornelia (of Sulla), and a Lex Iulia, all limites barring the subruncivi had to be open to public use (Hyg. Gr. 168, 16: iter populo sicut per viam publicam debetur). All agrarian bills contained similar stipulations (Front. 24, 6. Hyg. 120, 18. Lex. agrar. v. 89. Lex. Urson. c. 78; cf. the numerous details in the Libri colon.). <50> The ground cut through by the limites was either removed from the pool of land to be distributed (exceptus), or it was included. In the first case, the area of the section of land was measured from the edge of the limites (Balb. 98, 11), and in the latter case it was measured from a linea mensuralis per limitem (Sic. 158, 8; cf. Hyg. 120, 23). <60>

6. Centuriae

The plots of land bordered by the intersecting limites were called centuriae (Paul. Fest. 116: limitatus ager est in centurias dimensus), and the plots bordered by the quintarii - made up of 25 centuriae each - were called saltus (Sic. 158, 21; though Varro r. r. I 10, 2 has it differently, stating that a 2x2 square of 4 centuriae total was called a saltus, see vol. III p.1961, 55. On the ancient explanations of the term centuria and its origins, ibid. p.1960, 5). <page break 678/679> When the centuriae had an area of 200 iugera, which wasn’t only the theoretical area or the one considered most ancient, but also the area most often used, then these centuriae formed squares with sides 20 actus = 2400 feet (710.4m) in length (Front. 30, 17. Nipsus 293, 9). This was the size prescribed for limitationes in the Lex Sempronia. <10> The details given in the Libri colon. are also confirmed by the extant Gracchan termini marked by the boundaries of the praefecturae Atinas and Tegianensis in Lucania, and of the ager Compsanus in the land inhabited by the Hirpini, and are also confirmed to some extent by the accounts of the colony Neptunia founded by C. Gracchus (209, 6. 21. 210, 7. 261, 1 with CIL I² 639. 642-644, and cf. Pais 56. 60). For the military colonies founded in Italy by Sulla, Caesar, and Augustus, these same sources consistently show us square centuriae of 200 iugera in area. <20> Furthermore, we also find rectangular centuriae, eg. in Beneventum, Velia, and Vibo, all 16 x 25 actus = 200 iugera (Sic. 159, 22. Lib. col. 210, 1. 209, 10. 19), or coloniae with other areas, like an area of 210 in Cremona, which presumably measured 20 x 21 actus (Front. 30, 19; Hyg. Gr. 170, 19), or an area of 24 x 20 actus = 240 iugera (Sic. 159, 14 without mentioning the specific location), or 80 x 16 actus = 640 iugera in Luceria (Lib. col. 210, 15), or an area of 40 x 20 actus = 400 iugera in Emerita, the Augustan colony for veterans in Spain (Hyg. Gr. 171, 1; see vol. V p.2494). <30> The centuria triumviralis of only 50 iugera mentioned by the agrimensores, ie. a 10x10 actus square (Front. 30, 20. Hyg. Gr. 170, 18), may relate to ager quaestorius, or may have arisen from mixing things up with the actual land distributed to individual recipients (Mommsen seems to have assumed the latter in Herm. XXVII 81, 5; Ges. Schr. V 87, 5). <40> The centuriae of 210 or 240 iugera in size were perhaps measured out for specific recipients, given that their sizes must have been intentional. During the republic, the size of the centuriae must have been even more flexible, if all the wildly different areas for the individual portions of land given out at the founding of colonies (those, at least, that have been handed down to us) are to add up to make the right number of iugera to form a centuria (Mommsen Herm. XXVIII 81; Ges. Schr. V 87). It is also of note that we occasionally find portions of land 66 ⅔ iugera large, ie. a third of a normal centuria (see below sec. 11, and the discussion of the different sizes prescribed in vol. IV p.575). <50>

7. Notation for the centuriae

The centuriae were referred to by which of the four regions they were in, and by numbers that specify their distance from the decumanus and the cardo maximus. For example, >DD III KK IIII refers to the centuria in the third column to the right of the decumanus maximus (ie. between the 2nd and 3rd limes prorsus), and in the 4th row on ‘this side’ of the cardo maximus (ie. between the 3rd and 4th limes transversus measured from the perspective of the person conducting the limitatio: here, this is directly behind them) (Front. 14, 2. Hyg. Gr. 201, 1; see the diagram below). <60>

Likewise, the actual limites bordering the centuriae were numbered in the same way, measuring from the main lines outwards. <page break 679/680> In the agrimensores, there was a difference of opinion as to whether the decumanus and cardo maximus were counted as the first, or whether the numbering began from the actual regions themselves, even though the latter numbering system was universally employed in the agrarian laws of the classical period and was the only proper system to use (Hyg. Gr. 173, 18ff.; cf. Hyg. 112, 13ff. and cf. Schulten Flurt. 31f. and Barthel 44f., who traces the agrimensores’ false opinion that the decumanus and cardo maximus should be counted as number one back to an error made by the specialists writing the works, because the practice was foreign to them. Either way, the younger Hyg. 207, 10 uses this system of measurement in his proposal of a law concerning the ager vectigalis in the provinces). <10>

8. Placing of the stones

The grid created by limitatio was marked and preserved using boundary stones, called termini or lapides. They were specifically and legally forbidden from being moved (Lex Mamilia c. 53 and 55). Their inscriptions made it easier for people to find their way through the plots of land that had been subject to limitatio (cf. Nipsus 290, 17-295, 4). <20> Their shape, material, size, placement in and on top of the ground, spacing apart, and inscriptions were precisely prescribed, whether in the agrarian or colonial laws themselves, or in the contracts drawn up for the limitationes which were based on those laws. The agrimensores therefore differentiated between termini or lapides Gracchiani, triumvirales, Augustei, Tiberiani, Claudiani, etc. (Lib. col. 212, 10. 242, 7-243, 17; cf. 348, 1), and give a large range of instructions that differ on these details. <30> Given this, the set of rules prescribed for limitationes carried out under Augustus were taken to be the baseline / exemplary set. In his contracts, Augustus had requested that stones marked with the numbers of the limites should be placed on all four corners of the centuriae (Hyg. Gr. 172, 2). They therefore stood in the middle of the crossroads between the limites, in mediis tetrantibus (ibid. 194, 12 = 157, 13 Th.). <40> At the very top of the boundary stones, which in older time periods were usually round, they inscribed two lines that intersecting each other at right angles, called the decussis, which showed you the direction of both of the limites (Nips. 286, 16: decussati in capitibus lapides). The agrimensores provide a very large variety of prescriptions for each of these properties (in addition to the sources already mentioned, cf. Hyg. 126, 21 = 89, 17 Th. and 281, 17 = 75, 16 Th.; cf. Rudorff 271ff.). <50>

9. The inscriptions on the stones

The following patterns of inscriptions on the boundary stones for 36 centuriae are good examples for the system for creating the inscriptions to go on the boundary stones (main sources: Hyg. 111, 16-112, 8 = 71, 10-72, 13 Th. Hyg. Gr. 194, 17-196, 14 = 157, 18-159, 17 Th. - Mommsen Die Bezeichn. der Grenzsteine, Herm. XXVII 90ff. = Ges. Schr. V 95ff. is made less helpful because of the confusion between citra and ultra, and because of numbering the decumanus and cardo maximus the dec. and c. primus contrary to what was the usual practice. In this regard too, Barthel 42ff. has presented the material in essentially the correct way). <60>

To begin with, all stones along the decumanus maximus bear the marking DM, and all stones along the cardo maximus bear the marking KM. The stone in the centre of the entire system bears the marking DM KM. All stones in the regio dextrata bear DD, all in the sinistra bear SD, while all stones in the regio ultrata bear VK and all in the regio citrata bear KK. <page break 680/681> These are followed by the numbers associated with the limites (172, 3), which are counted outwards from the decumanus maximus and the cardo maximus in all four directions. The stones that lie on the diagonals, on the outer corners of the centuriae, in angulis clusaribus, also always bear the name of the relevant centuria. As well as letters and numbers, there is also a note detailing the origin of the particular limitatio, like on the Gracchan termini which have the names of the IIIviri a(gris) i(udicandis) a(dsignandis) (CIL I² 639-644). <10>

These written marks were either found on the apex of the stone, in vertice, in capitibus, or they were found on the main body, in lateribus. <50> In the first instance, the letters which refer to the limites are placed next to each other at right angles, making the direction of the limites clear as well.

The direction is then usually also given by the decussis in the middle, or by lines on the edge of the sides of the stone. Accordingly, on the Gracchan terminus from the ager Campanus nr. 640, on its upper side is written s(inistra) d(ecumanum) I, k(itra) k(ardinem) XI. <60> This tells us that the stone stood at the intersection between the first decumanus on the regio sinistra and the 11th cardo of the regio citrata. At the same time, it also describes the centuria bordered by it and the DM KK X stone. <page break 681/682> If the inscription was made on the main body of the stone, where there was more space for writing numbers, given that the larger limitations could go up to the hundreds (Hyg. Gr. 173, 9), a distinction was made between in fronte and in lateribus, and the writing was either inscribed sursum versus or ad terram deorsum versus (Hyg. 111, 24. Hyg. Gr. 195, 12). <10> In this case, the terms DM and KM should appear in fronte, ie. on the side of the stone that faced the midpoint of the grid, the terms DD and SD should always be on the side of the stone facing the decumanus maximus, and the terms VK and KK should face the cardo maximus (Hyg. Gr. 194, 20. 195, 11; cf. also Hyg. 111, 25. 112, 2. 5. 6). <20>

10. Differences in inscriptions

There are a great deal of examples which deviate from the example system described above. On the Gracchan boundary stone 639 from Atina in Lucania, there is a decussis on the top which has inscribed the symbols ⦙כ and ⦙＜. But on the round body of the stone, next to the names of the triumvirs, the symbols K VII are written from top to bottom. <30> The stone must therefore have stood at the intersection of the decumanus maximus and the 7th cardo, and its position would have made it clear whether it was citra or ultra kardinem. <40> On the 18 extant quadrilateral stones for a limitatio which was carried out by the Legio III Augusta in 29/60 CE in Africa (see below p.699), alongside the inscription regarding who carried out the limitatio, on that same side four of them bear full references to two limites: going by the numbers, these also appear to be the quintarii, eg. DD LXX and VK CCLXXX (CIL VIII c. IV p.2307 nr. 22786f. Barthel 61ff. nr. 1). <50> Other stones from the same limitatio bear only the reference to the limites on two adjacent sides, eg. DD LXXX … and VK CCLXIIII (CIL i. Barthel nr. 11). Following the rule stated above (p. 681), these inscriptions also bore the names of the centuriae which the stones stood at the angulus clusaris of. <60> In a third group, this was marked by the symbol 〉 written before the numbers, eg. V·K|〉CC·LXV and S·D·〉·XL·V (CIL p.2309 nr. 22789. Barthel nr. 18). Many other stones belonging to the same limitatio have no writing on them at all, and only bear the decussis on the top, the inscribed cross which shows the direction of the limites. Also, according to the agrimensores, often only the termini on the decumanus and cardo maximus bore written inscriptions, and the others were left with no writing at all (muti, Hyg. Gr. 171, 18). <page break 682/683> In areas lacking in stone, sometimes they only placed stones on the quintarii, and just set up wooden posts on the other limites (Hyg. 112, 11).

The agrimensores had also differentiated between other systems for making the inscriptions and explained them using illustrations. <10> Unfortunately, as we have them now, the texts and the figures have become almost incomprehensible (cf. esp. Hyg. Gr. 172, 7-173, 15 = 137, 3-138, 13 with fig. 80-84 Th., where a system for making the inscriptions is described in which the inscriptions should be ordered along three one-quarter sections of the round lower side of the stone such that the quarter describing the centuria, that being the quarta lapidis portio clusaris, is left blank and without writing. On this, cf. Barthel 43, 2, where there is a translation of some sections which others have misunderstood). The details shared by Nipsus about how to find particular centuriae using the boundary stones are particularly informative. <20> He first supposes a grid oriented towards the west (290, 24 and 294, 3, see below p.685/6 diagram 1), and supposes that the stones only bear the numbers of the limites: for this situation, he explains how it is possible to determine the regiones by looking at whether the numbers increase or decrease when travelling in various directions (290, 26-291, 18 and 294, 5-16). <30> Then, he does the same for a grid oriented towards the north (292, 5 and 294, 16, see below p.685/6 diagram 4), where the method is the same (291, 18-192, 17 and 294, 16-295, 4). <30> With the help of our diagrams, these explanations are easy to understand in their entirety.

All of this is only relevant to the termini that stood at the intersections between limites. Occasionally, however, lapides medii or epipedonici were set up in between them, though they probably got their names from the fact that they were shorter (Lib. col. 213, 9). <40> Augustus stipulated that the borders of the individual portions of land - the sortes or acceptae (see below, l.57) - should be marked by termini roborei (Hyg. Gr. 172, 6), and following an especially thorough practice recommended by the younger Hyginus, even the individual iugera should be marked out by pali actuarii, to make it easier to create the acceptae (192, 9). <50>

11. Distributing the land post-limitatio

Once the limitatio was complete, its distribution amongst the colonists or other recipients was done iugum by iugum, such that they received a particular modus of farmland within the centuriae, whether those were individual portions when the land was handed out to each man individually, in nominibus, by lot, or whether the land was distributed to groups to then divide among themselves. When the land was distributed by lot, the portions were called sortes, though generally they were also called acceptae. Their borders were not marked by limites, but by rigores. <60> In contrast to the outer borders of the centuriae, they were often called interiectivi (Agenn. 41, 8. 72, 16; the limites intercisivi however refer to sections of land that was not part of a centuria, cf. the explanation on Frontin. 2, 3 = 52, 14 Th. Only in Lib. col. I 213, 11 is the term intercisivi used to refer to the lines dividing the individual plots of land distributed to veterans, and this only happens once). <page break 693/684> The sizes of the acceptae were not only different between the different instances of land distribution, but were also different within one and the same colony, with regard to the military rank of the recipient and the quality of the land (Sic. 156, 9. 15. Hyg. Gr. 169, 10. 176, 12. Lib. col. 222, 13. 224, 12; cf. Weber 20ff.). During the republic, the centurions received twice as many iugera as the average soldier, and the cavalry also received a larger area (Liv. XXXV 40. XL 34; on the different sizes of acceptae see Mommsen Herm. XXVII 81, 4. 5; Ges. Schr. V 87, 4. 5. Cf. above p.679). <10> If the area of the centuriae wasn’t divisible by the number of acceptae, the acceptae were also split between adjacent neighbouring centuriae (Front. 14, 1. Hyg. 113, 4. Hyg. Gr. 204, 7). The agrimensores also make thorough recommendations for how to distribute the land as properly as possible (Hyg. 113, 1. Hyg. Gr. 199, 18. Rudorff 367ff.). <20>

12. Expanding a limitatio

Expansions to the limitationes were done in very different ways. To begin with, they may only encompass one small section of territory, then the entire area of a particular municipality, then multiple municipalities or even entire provinces. Within an area that had undergone limitatio, only arable land qua falx et aerator ierit (Hyg. 112, 24 secundum legem divi Augusti, cf. Hyg. Gr. 201, 8. 203, 16) was really limited. <30> The leftover arable land that lay outside the centuriae and the grid of limites, but between this grid and the outer borders of the relevant area, was called ager extra clusus (Front. 8, 7. 22, 2. Agenn. 87, 4 = 47, 9 Th.). Where this land was due to be distributed, but where there wasn’t enough of it to make an entire centuria, it was called subsecivum, ab subsecante linea (cf. Lex agrar. v. 66: ager locus in ea centuria supsicivove). The same name was given to land in the middle of the centuriae bordered by limites that was left over after the land had been distributed (Front. 6, 5. 20, 3. Hyg. 132, 25. Sic. 155. 27). <40> It would also sometimes happen that one part of the land within an area, particularly in a municipality, did not undergo limitatio, even if it had been ploughed or was arable. These sections of land were called agri soluti or in absoluto relicti, in contrast to the limitibus et terminis publice obligati. This land, and any other land that had not undergone limitatio, was only sporadically bordered more arcifinio (Hyg. Gr. 179, 15. Rudorff 252f. 393f.). <50> Uneven land, or land that had been put to one side by the founder of a colony, was exempt from being distributed, and hence called exceptum, and did not have any limites put upon it. These pieces were called loca relicta, and in legal matters were handled like the subseciva (Front. 21, 8. Agenn. 79, 19. 86, 28. 87, 4). The loca excepta also included public roads and paths, streams of water, and aqueducts that were already present before the limitatio took place, as well as shrines, graves, and any individual structures which the auctores divisionis were required to preserve in accordance with the lex data for the particular colony (Lex agrar. v. 89. Lex. Urson. c. 78. Hygin. 120, 12. 134, 7. Sic. 157, 11). <60> Also exempt from the limitatio were the flood plains for rivers: this was a large point of contention when it came to how they were handled during the technical land surveying process, as well as how they were handled when it came to legal agrarian matters (Front. 51, 22. 52, 11. Hyg. 120, 7. Sic. 157, 18; cf. Rudorff 399). <page break 684/685> Finally, even the decumanus and cardo maximus were sometimes interrupted by mountains within land subject to limitatio. The methods which were used to get around such problems with surveying the land are given in particular detail by Frontinus (33, 14ff. Hyg. Gr. 180, 17). <10>

The terrain within an area which was distributed was referred to in its entirety as a pertica (Front. 26, 6). If there was not enough land, then plots from neighbouring areas were added in, called praefecturae, implying that they formed their own secondary jurisdiction. They were also subject to their own limitatio (Front. 26, 6. Agenn. 80, 3 and especially Sic. 159, 26. Hyg. Gr. 171, 3).

The choice of the centre point, or origin, for the limitatio for an area of land was also dependent on the topography. <20> In the best case scenario, the grid could begin ex ipsa colonia, as in Admedera in Africa, where the decumanus and cardo maximus pass through the four city gates like the main streets inside a camp (Hyg. Gr. 178, 10. 180, 2. 191, 12. 194, 4; other examples are provided by the diagrams of limitationes for certain areas in the manuscripts of the agrimensores, see below p.692). Where that wasn’t possible, the origin should at least be put near to the city, or even in optimo solo, that being where the area of a colony had been extended when the costs of the neighbouring cities had been deduced (178, 12. 179, 8). <30>

13. Orientation of the limitatio

Up to here, the specific orientation of the limitatio grid has been mostly put to one side, even though it does play a large role in the prescriptions given by the agrimensores, and even though they almost always require the decumani and cardines to be oriented according to the main compass points. <page break 685/686> The limitationes which have left a measurable impact on today’s road network, on the other hand, seem to have always (with one single apparent exception, see below p.687) given not the slightest bit of attention to the cardinal directions, and instead appear to only have been oriented according to practical concerns. <10> These sorts of practical concerns are also mentioned by the agrimensores, but with evident disdain (see below p.687). In order for somebody to find their way around an area that has undergone limitatio, they do need to have an approximate idea about which cardinal directions the decumanus and cardo are roughly pointed towards, and in the detailed prescriptions given in the libri coloniarum this was probably specified for each individual place (cf. 209, 9. 17. 19. 22 etc.). <20> Given all of this, when the agrimensores speak about limitationes in terms of compass directions, this does not completely line up with reality, though for the most part they still use the compass directions as reference points because of the general desire for engineers to come up with academic-sounding proposals and theoretical situations and so on: something which is still the case today. In our ancient sources, and especially in modern works on the process of limitatio, these theoretical situations have lead to some large misunderstandings: among others, the constant confusion between citra and ultra (see above p.675). <30>

In order to make this completely clear, we will first consider the evidence for limitationes oriented by the compass points both in the theory and in practice, and demonstrate the four possible cases by the use of diagrams. Here, the arrow-heads mark out the directions of the limitationes. <page break 686/685> <50>

1) With reference to Varro, the agrimensores follow the discipline of Etruscan haruspices and the orientation of Greek temples by assuming that the decumanus originally went westwards (diagram 1), and recommend this as the optima ac rationalis agrorum constitutio (Front. 27, 13. 31, 1. Hyg. Gr. 166, 10; cf. Paul. Fest. 71 s. v. Decimanus). <60> However, we do not find a decumanus oriented towards the west in the libri coloniarum, nor do we find it in any of the diverse limitationes which we have material evidence for. Even the younger Hyginus, who agrees with Frontinus in taking this to be the original way of doing things, does not seem to be aware of any actual examples, and he explains that the much more common easterly orientation was a later inversion of that system (169, 18). According to Barthel, this entire idea stems from Varro, who seems to have taken the Hellenic tradition of orienting their temples towards the west, and transferred it onto Etruscan religion and the Roman limitatio (114ff.). <page break 685/686> And yet, Nipsus, who unlike the other agrimensores likes to keep away from anything too theoretical, chooses to assume a westerly orientation in his first example for how to orient yourself in ager assignatus (290, 23-291, 18 and 294, 5-16; his detailed and self-consistent instructions are illustrated by our first diagram). <60>

2) The elder Hyginus (3, 23. 30 = 54, 7. 12 Th.) recommended the decumanus to be oriented towards the east (diagram 2), and according to the libri coloniarum this was common practice in Italy (209, 9. 210, 5. 13 lege Sempronia et Iulia. 217, 18). In Aeclanum, the Gracchan termini CIL I² 643-645 demonstrate that this practice can be traced back to the Lex Sempronia. <page break 686/687> For the Spanish Augustan colony Emerita, both the limitatio and the praefecturae were oriented towards the east (Hyg. Gr. 171 3 and 7). According to the agrimensores, they determined which exact direction was east by observing where the sun set on the actual day on which the surveying took place (Front. 31, 4. Hyg. Gr. 170, 3. 182, 8. 183, 13, where potential mistakes to be made there are pointed out, and various methods to determine the meridian and thereby also determine the proper east-to-west line are set out. Orientation using the sun is also in Lib. col. 210, 15 and 223, 14, on which cf. Barthel 99. Isodor. Etym. XV 4, 7 mentions a templum oriented towards the east: cuius partes quattuor erant: antica ad ortum, postica ad occasum, sinistra ad septentrionem, dextra ad meridiem spectans). <10>

3) The agrimensores, proceeding only from their theory, state that it is incorrect to orient the limitatio towards the south (diagram 3) (Front. 29, 4: contra sanam rationem, ut in agro Campano qui est circa Capuam, ubi est kardo in orientem et decimanus in meridianum; cf. Hyg. Gr. 170, 14). This note about Capua can actually be verified by the Gracchan terminus CIL I² 640 and the remaining impact of the limitatio on today’s road network, though it also makes sense given practical concerns (see below p.697). <30> Varro himself prescribes a southerly orientation for a Temple of the Heavens (de l. l. VII 7: eius templi partes quattuor dicuntur, sinistra ab oriente, dextra ab occasu, antica ad meridiem, postica ad septentrionem), which Varro’s contemporary, the lawyer Serv. Sulpicius Rufus, also prescribes for limitatio in the source cited above p. 676, 59, and which Serv. Georg. I 126 prescribes as well. A southerly direction for the decumanus is attested in the Liber coloniarum I for Consentia, Vibo, Clampetia, and Beneventum (209f.). Moreover, when Hyginus is describing how the termini are inscribed, he supposes this kind of southerly orientation (111, 16 = 71, 10 Th.). <40>

4) Finally, the agrimensores can’t have been unaware of the possibility of a decumanus oriented towards the north, because Nipsus uses it in one of his hypothetical scenarios with which he describes how to find your way in an area that has undergone limitatio (291, 18-292, 17 and 294, 16-295, 4, with regard to our diagram 4). In his derogatory critique on auspices and traditions regarding these orientations, the orator Arnobius nat. IV 5 takes all four cardinal directions as possible, and other sources allow us to conclude that for every augurium, they had to decide which direction to orient it towards (see Wissowa above vol. II p.2340ff., where the relevant sources are listed). <50>

When it comes to deciding the orientation of limitationes purely from practical concerns - relicta caeli ratione - the agrimensores give the following examples:

1) The decumanus is aligned such that it extends as far as possible along the land undergoing the limitatio (Front. 29, 1: non nulli aliud secuti, ut quidam agri magnitudinem, qui qua longior erat, fecerunt decumanum; cf. Hyg. Gr. 170, 10). <60> This method of orientation can be attested in the limitationes for both of the African provinces: the older one from 146 BCE, and the one under Caesar. In the latter, where we know precisely where the main limites lay, not only does the decumanus extend along the length of the entire province, but the cardo also follows the line where the width is greatest (see below p.69ff. sect. 23-24). <page break 687/688> This was also the method used in the ager Campanus, where the decumanus also lines up with the direction of the meridian (see below p.697).

2) The limitatio is superimposed on top of existing streets (Hyg. Gr. 179, 11: quibusdam coloniis decumanum maximum ita constituerunt, ut viam consularem transeuntem per coloniam contineret, sicut in Campania coloniae Axurnati; decimanus maximus per viam Appiam observatur, cf. 169, 3). <10> As well as those Italian limitationes whose impact can still be seen in today’s road network, we can add those of Parma, Regium Lepidum, Bononia, Forum Cornelii on the Via Aemilia, and a north-westerly one from Tarvisium on the Via Postumia (see below p.698, cf. Barthel 96).

3) In order to prevent a new limitatio from falling on top of a pre-existing one from a neighbouring colony, they were often given a different direction (Front. 31, 7: multi, ne proximae coloniae limitibus ordinatos limites mitterent, exacta conversione discreverunt. Hyg. Gr. 170, 9). <20> All of the limitationes in Galla cisalpina that we are still able to recognise followed this rule. Accordingly, the limitationes from Brixia and Cremona near Oglio, and those of Patavium and Tarvisium near Musone, are at an angle to each other (Schulten Flurt. panels I and V). It also wasn’t too uncommon for land that had already undergone limitatio to undergo it again, but in a different direction (Hyg. Gr. 178, 2. Sic. 165, 13). <30>

All of the evidence considered here about orienting the limitationes according to the compass points and according to practical concerns allow us to conclude that the theory, as Varro writes it, did not line up with actual practice. Unlike the orientations of temples and cities, limitatio had no ties to religious considerations: it was very secular in its nature. <40> Though we do find a note on determining the main limites using the groma “like when taking auspices, perhaps in the presence of the founder of the colony” (Hyg. Gr. 170, 5: posita auspicaliter groma, ipso forte conditore praestente), this hardly implies that any universal religious ritual always took place at the start of each limitatio: in fact, this note would actually show that if there were such a religious ritual, it would have been done incorrectly. The Lex Sempronia never once stipulated that the limitationes should be oriented according to the cardinal directions: near Aeclanum, the decumanus maximus heads towards the east, while in the Ager Campanus it heads towards the south (see below p.686f.). <50>

To conclude: in effect, this all boils down to the fact that the direction that the limitatio pointed in was chosen, and overwhelmingly so, with regard to practical purposes. <60> And even if it isn’t wrong to say that the compass points would certainly appeal as the natural directions to opt for, it is nevertheless right to doubt whether it actually was much of a common practice to orient the limitationes according to those compass points, or whether it was the work of engineers coming up with theoretical scenarios in their academic writings that first gave rise to such techniques (cf. Barthel 114ff., whose fascinating arguments would lead to rejecting these theoretical methods of orienting the limitationes according to the compass points, even if he himself does not take them that far). <page break 688/689>

14. Origins and implementation

The sets of terminology used to describe the limites were differentiated from each other according to specific technical considerations which we have already mentioned above, and they were named after the people who had first come up with them and first put them to use. In the libri coloniarum, there is frequent mention of the use of limitibus Gracchanis to assign land, though that system had also been used under Augustus (209, 16). In other cases, new limitationes were carried out (232, 14: Cales, ager eius limitibus Gracchanis antea fuerat adsignatus, postea iussu Caesaris Augusti limitibus nominis sui est renormatus; cf. 233, 15). <10> There have been a great many instances of partitioning land limitibus or centuriis Augusteis. Even Claudius and Nero made use of them (238, 20. 237, 15). Moreover, we find land being partitioned and assigned limitibus Sullanis or mensura Syllana (236, 4. 237, 5 with the additional comment lege Syllana. 238, 11), limitibus Iulianis (236, 1 lege Augustiana. 238, 15) and limitibus triumviralibus (226, 9. 237, 22). These examples allow us to conclude that the phrase limitibus Sullanis, or the like, need not imply that the limitatio in question happened under Sulla (or anybody else) if there is no mention of a lege Syllana or similar (Mommsen Feldm. II 188, 55; Ges. Schr. V 178, 1). <20>

Furthermore, these examples are enough to show us what sorts of men were considered to be the auctores limitationum. All limitationes that took place either at the founding of a colony or for the purpose of land redistribution were always carried out by one or more ordinary or extraordinary officials, who were granted the power to do so by a situational law, a law that universally applied, or a decree from the senate. <30> “The decree to undergo limitatio is, for land, the same thing as the bestowal of citizenship is for a person: that decree always found its basis in a decision made by an auctor who had been granted either general authority, or specific authority, by the nation” (Mommsen loc. cit.). Up until Caesar’s agrarian reform law in 59, these would have been commissions of 3, 5, 10, or 20 men, who were dubbed curatores (Cic. leg. agr. II 17; cf. Lex Mamilia c. 55). <40> In Rullus’ agrarian bill, he demanded that instead of decemviri, finitores ex equestri loco ducentos should be granted with propraetorial imperium: these men, of course, would hardly have been the expert engineers that the architecti employed by the decemviri were. Instead, they would have been attendants or agents for the colonising commission (Cic. loc. cit. 45: Xviri cum inperio …, cum illa delecta finitorum iuventute; cf. 32. 34. 53). Caesar’s military colonies were later implemented by legati pro praetore, like Q. Valerius Orca and P. Alfenus Varus, whom the dictator had passed this negotium on to (Cic. fam. XIII 4. 5). <50> Augustus, as it happens, seems to have tasked regular magistrates with the process, like Octavius Musa who carried out limitatio at his request on a section of the Ager Mantuanus (Schol. Serv. Ecl. IX 7, where he is referred to as limitator ab Augusto datus, cf. Prosop. II 425, 14; the condiscipulus Vergilii refers to Alfenus). <60>

During the republic, as well as under Augustus, the technical jobs associated with limitatio were awarded to people on a contractual basis. From one such contract, the lex agris limitandis metiundis under the triumviri Antonius, Caesar (Octavianus) and Lepidus, a fragment has been preserved in liber coloniarum I: this being specifications for the width of the limites and the properties of the termini (211, 24). A similar thing happened for a locatio operis under Augustus (Hyg. Gr. 172, 4). <page break 689/690> According to the earlier documents, the price was calculated per centuria, which was itself calculated as being a subsecivum of over 100 iugera: any subsecivum that was less than 100 iugera but more than 50 counted as half a centuria (213, 1). In earlier time periods, the professionals could also be called finitores (Plaut. Poen. 49, see above p.673), later mensores, agrimensores, gromatici (see vol. I p.894. vol. VI p.2329, though there, the finitores demanded by Rullus are incorrectly understood to be professionals of this kind. vol. VII p.1886. Rudorff 320ff.). <10> During the empire, limitations were carried out by people in the army, or even by the legions themselves. On one of the termini from the year 29/30 mentioned above, in the territory of Numidia near Chott el Fejej around Africa nova, it says: legio III Augusta leimitavit C. Vibio Marso proconsule III (CIL VIII 22786a. f. k. Barthel 61. 64; see above p.682). <20> Equally, under Trajan an evocatus Augusti, professionis nostrae (sc. agrimensorum) capacissimus lead the limitatio for a colony for veterans in Pannonia (Hyg. 121, 7), and in the year 126 (for the date, see Mommsen Feldm. II 178, 45) a miles datus a Metello Nepote managed it, in 141 a praetorian miles, and in 149 a centurio cohortis XXVII mensoribus intervenientibus carried out some land surveying in Italy (Lib. col. 244. 252; cf. 251 on the inscription of a terminus under Antoninus Pius which says determinante … mil. coh. VI pr. mesore agrario). <30>

The optimi mensores would have had to lay down the decumanus and cardo maximus, as well as the limites quintarii, to make sure that no mistakes were made (Hyg. 112, 9. Hyg. Gr. 191, 14). Here, as in any instance where they would have needed to find a right angle, the groma was the main tool used by the surveyors, which they colloquially referred to as a ferramentum = ‘device’. <40> In 1912, an intact metal piece from a groma was discovered, which has allowed us to make a complete reconstruction of this very delicately designed tool (Della Corte Mon. dei Lincei XXVIII 29ff., then Novotny Germania VII 1923, 22ff. with figures. This find has also shown that the reconstruction I provided in Schulten’s article Groma in vol. VII p.1884 was broadly correct). The most detailed instructions on how to use the ferramentum are found in Hyginus the younger and in Nipsus (193, 3. 285ff.). We may also add here the building markers, metae, cannae, signa, and the measuring sticks, decempeda (Rudorff 335ff. Della Corte 83ff.). <50> Like the intersection between the via praetoria and the via principalis in a camp, the centre point of the limitatio was often referred to as the groma (Non. 63. Hyg. Gr. 180, 8; cf. de mun. castr. c. 12).

15. Maps of the land

A map of the land, a forma, was created as a document to record the completed limitatio. It was either engraved onto bronze tablets (aes, tabulae aeris), or drawn on top of marble, wooden slats, parchment, canvas, or similar (Sic. 154, 14, where other terms like typus, pertica, etc. are used, and 160, 3). <60> According to the agrimensores (the main source being Hyg. Gr. 202, 11, and on marking out the forma 196, 17, cf. Mommsen Feldm. II 152; Ges. Schr. V 151. Rudorff 404ff. Weber 57f. and especially Schulten Herm. XXXIII 561ff. XLI 38f., as well as Barthel 45ff.), the forma covered only the part of the territory that had undergone limitatio: only the pertica. <page break 690/691> Individual formae were also produced for the praefecturae that had undergone their own separate limitationes (Sic. 160, 3). For the ager arcifinius, there is only a private plan of the land without any official evidential value (Sic. 138, 15). The formae extra limitationem for the assignationes sine divisione of a few portions of land are exceptions (Sic. 160, 18). <10> In a forma, the entire limitatio would be presented, and the decumanus, the cardo, and in earlier times the quintarii too, would be drawn on with thicker lines to make them stand out (Hyg. Gr. 175, 3), though all centuriae would at least have their borders drawn on (ibid. 202, 15. Balb. 98, 13). Inside the centuriae, which would have had to have retained their default names, they would also mark the names of the people who received the acceptae, their sizes, and what specifically the farmland was used for (Agenn. according to Front. 46, 11 = 36, 16 Th.: habere debet aes primum locum, deinde modum, deinde speciem; Hyg. 124, 7: in divisa et adsignata regione … formis per centurias certus cuique modus adscriptus est, cf. 121, 11. Sic. 156, 4. Agenn. 77, 3 = 36, 16 Th.). <20> The same was the case for the subseciva (Hyg. 121, 17; cf. Lex agrar v. 7. 78. 80, which, given v. 44. 66. 89, concerns land that has undergone proper limitatio). <30> Inside the centuriae, the borders surrounding the plots of land owned by different people were only rarely drawn on, which we are able to conclude from the evocatus Augusti under Trajan, who had carried out limitationes on the land for the veterans in Pannonia (Hyg. 121, 10 = 84, 12 Th.: in aere, id est in formis, non tantum modum quem adsignabat adscripsit aut notavit, sed et extrema linea unius cuiusque modum comprehendit: uti acta est mensura adsignationes, ita inscripsit longitudinis et latitudinis modum). As well as the data, adsignata, the formae also contained precise significationes for the concessa, excepta, reddita veteri possessori, commutata pro suo, for the loca publica, sacra, sepulcra, etc. (Hyg. Gr. 197, 5. 202, 11). <40> The waterways, too, where they overlapped with the land undergoing limitatio, appeared on the formae with the modus of the land they flowed over, as well as the land that made up their flood plains (Hyg. 120, 7. 125, 6. Nipsus 293, 6).

These instructions given by the agrimensores can be confirmed by the fragments of a land map from Arausio, which Schulten has recently [as of 1926] handled in detail (Herm. XLI 25ff.; CIL XII 1244 and Add. p. 824; cf. also Mommsen Herm. XXVII 103; Ges. Schr. V 108 and Weber 35. 279f. with reconstructions in append. 1). <50> They contain a diagram showing the division of the land into right-angled centuriae, and the proper names of those centurae, eg. S·D·X·C·K·X or DD XIII CK IIII. However, considering the lists of records that we find alongside the diagram, these fragments appear to come from a graphical land registry used for tax purposes. <60> The map was oriented such that the decumanus maximus ran from bottom to top (as in diagram 4 on p.686 above), meaning that the regions SD and DD also lie to the left and right of the viewer respectively, and CK is below while VK is above. Since the centuriae are drawn on 14cm tall and 11.6cm wide, the whole map that the three fragments are from must have been at least 2.78m wide and 1.26m tall (24 centuriae across, and 9 high). The whole thing, then, even if the map itself was made up of separate pieces, must have been placed on the wall of a very large building. <page break 691/692>

In addition to these special plans describing the land that had undergone limitatio, there must have also been maps showing a simple overview of the whole territory, including the loca inculta, silvae, pascua, compascua (Agenn. 85, 24. Hyg. 112, 25), as well as the mountains and their names, if that made sense for a given location, at least (Nipsus 293, 7), and finally the borders of the territory and the names of the neighbouring territories. <10> In all likelihood, these are the sorts of plans which the much smaller diagrams of land maps found in the manuscripts of the corpus agrimensorum, and especially in Hygin. the younger, trace back to. In these diagrams, the limitationes are only roughly sketched out. As Schulten Herm. XXXIII 534ff. has shown, these diagrams do not always match up with the text that they belong with, implying that the diagrams were not constructed from the text, but rather are based at least in part on some actual example of a map. <20> They also contained the territories, mountains, rivers, streets, and city buildings themselves (cf. the photographic reproductions in Thulin Corp. agr. I).

The formae for territories that had undergone state-required limitationes were taken to the tabularium or sanctuarium principis or Caesaris in Rome, along with the appropriate land registers and documents (on these see Schulten Herm. XLI 39ff.). Additional copies were stored in the relevant locations themselves (Sic. 154, 24. Hyg. Gr. 202, 17. 203, 3). <30> They were authoritative in official circumstances, and during legal disputes they served as vital pieces of evidence (Sic. 138, 11. 154, 23. Rudorff 284f.). For this reason, the formae also bore the names of the agrimensores who had officially carried out the limitatio (examples in Lib. col. 244). Furthermore, they also served more practical uses, eg. for finding your way around an area if there weren’t any termini or other landmarks (Nips. 293, 4). <40> In the year 77 or 78, Vespasian had the borders of the land which Sulla had dedicated to Diana Tifata re-established ex forma divi Augusti (Dessau 251). Under Trajan, some borders were established in Africa, secundum formam ab imperatore missam, next to the area near Chott el Fejej that had undergone limitatio under Tiberius: for this project, the copy of the forma stored in Rome must have been used at least to some extent when the new limitatio was carried out (CIL III 22787. 22788 and cf. Barthel 89ff.). <50>

16. Ager per strigas et scamna divisus

There is a second type of land division, to be differentiated from ager limitatus and centuriatus: this is ager per strigas et scamna divisus. Strigae and scamna were, as it happens, also sometimes used during limitationes, and we do come across hybrid systems, which we may also refer to as a mixed limitatio-system. <60> The sources from the agrimensores considered here were first gathered and discussed by Schulten Bonn. Jahr. CIII 13ff.

Strigae and scamna, ‘strips’ and ‘banks’, refer to rectangular plots and strips of land: these were individual sub-sections of a larger unit of land, and the difference between strigae or scamna was nothing to do with any properties of the land itself, but rather their general position relative to the larger unit of land. <page break 692/693> When the subsections of a larger unit of land ran along its length, they were called strigae, but when they were stacked along the width, they were called scamna (Front. 3, 3; cf. Hyg. Gr. 207, 1 quod in latitudinem longius fuerit, scamnum est, quod in longitudinem, striga). Accordingly, when describing a plot of land, the vertical strips which run parallel to the length of the land were called strigae, and the horizontal strips were called scamna, and compass directions had nothing to do with it. <10> This terminology was therefore always dependent on what the whole plot of land was like, and there could be a lot of variation in it: it could be one individual piece of farmland that formed part of some arbitrary extension, it could be a centuria, or a saltus, or an entire municipality. In any case, these subsections of land, those being the scamna and strigae, were equal to each other, as should be self-evident.

An area of land was called scamnatus when it was split into subsections horizontally, ie. when the overall plot of land was longer than it was wide, as eg. in the individual centuriae that were 240 iugera in total, being 24 actus long and 20 actus wide (Nipsus 293, 11), meaning that each of the 24 scamna was 10 iugera in area. <20> On the other hand, this area of land would be called strigatus if it had been divided along its length, into 20 strigae that were each 12 iugera in area. A third situation is also possible: where some of the land inside a centuria is divided along the its length, but some is split horizontally. All of these three situations do appear inside the libri coloniarum (the sections below, however, have been disregarded by Mommsen Herm. XXVII 85, 2; Ges. Schr. V 90, 4 as later interpolations, though I don’t think that’s right): <30>

1. Ager per centurias et scamna adsignatus in Bovianum, Aufidena, Histonium (231, 8. 259, 17. 260, 10).

2. Ager per centurias et strigas adsignatus in Aletrium (230, 9).

3. Ager per strigas et per scamna in centuriis adsignatus in Aequiculi, Nursia, Reate (255, 17. 257, 5. 26; which shows us that in other places, too, scamna and strigae were often subdivisions of centuriae.) <40>

17. Hyginus the Younger’s proposed system for land division

We can see just how important the use of scamna and strigae were in farmland that had undergone limitatio in Hyginus the younger’s own proposition, found at the end of his work de limitibus constituendis, for a new division of the ager vectigalis in the provinces (206, 3 - 208, 4 = 169, 4 - 171, 4 Th. The text is briefly and ingeniously explained by Oxé in Bonn. Jahrb. CXXVIII 20ff., but its main gist has been completely misunderstood. Barthel 46f. has done a better job of interpreting it; cf. also Weber 22ff. and Mommsen Herm. XXVII 86. 89f.; Ges. Schr. 91. 94f.). <50> In this situation, the scamna and strigae describe the centuriae of the ager limitatus. They were joined together in groups of eight, four in each direction, in right-angled fields called quadrae, and slotted into a proper limitatio grid (interstitione limitari mensuras per strigas et scamna agemus). <60> The dividing lines were in this case called rigores, though they had the usual width given to limites (omnium rigorum latitudines velut limitum observabimus), and Hyginus also makes use of other terminology borrowed from limitatio: he sets down a decimanus maximus and a cardo each 20ft wide; the limites actuarii transversi and prorsi (which equate to the quintarii in a limitatio) are 12ft wide; while the scamna and strigae, like the centuriae, are separated from one another by 8ft wide rigores linearii. <page break 693/694> In the sections bordered by the actuarii, which would be called the saltus in a limitatio proper, the four scamna lay parallel to the cardo and the four strigae lay parallel to the decumanus such that two scamna and a row of strigae lie between the transversi (see the figure below, which Oxé has reconstructed following the diagrams in the Corpus agrim. Th. fig. 137. 137a. 140). <10>

As with a field that has undergone limitatio, termini would be set up all across it, and those on both the decumanus and cardo maximus would be inscribed as usual with DM KM, and then DM LIMES II, KM LIMES II, etc. And where the actuarii intersected each other, ie. on the four corners of the quadrae, as well as the symbols referring to its specific region the termini would also have the numbers for which row of strigae and scamna it was in, eg. DD VL STRIGA I SCAMNO II. <20> Hyginus does not specify how large the spaces between the limites should be, or the areas of the scamna and strigae. Though Oxé seems to make the strange conclusion from Hyginus’s remark about calculating tax ad modum ubertatis per singula iugera (205, 15) that the scamna and strigae would only have been one iugus in area, this clearly cannot have been the case. <30> If it was, then the 12ft wide limites actuarii here, on the ager vectigalis, would have only enclosed 8 iugera, when usually in ager assignatus they enclose 5000! Rather, the only difference between ager vectigalis divided according to Hyginus’ proposed method versus ager assignatus that has undergone limitatio proper is the following: inside the saltus, instead of there being 25 square centuriae divided by 8ft wide paths, there would be 8 scamna and strigae. <40> His proposal, which won’t have been used in practice, was a middle ground, halfway between limitatio and scamnatio, which the ager per strigas et scamna in centuriis adsignatus seems to have provided him with a good example for (on attempts to demonstrate Hyginus’s system actually being used, see Barthel 48f.). Hyginus wanted to make the legal difference between ager vectigalis and ager immunis apparent through his proposed system. <50>

18. The main type of ager per strigas et per scamna divisus

This leads us on to the main type of ager per strigas et per scamna divisus, which is contrasted with the limitatio. For this method, the land is divided not by limites, but only by the borders around plots of land owned by different people: per proximos possessionum rigores (Front. 3, 1; cf. Weber 26; on the other hand, Mommsen Herm. XXVII 82, 6; Ges. Schr. V 88, 2 refers to Balbus 98, 9 as he interprets rigores here as the pairs of parallel lines bordering the paths, though this can hardly be correct). <60> Given the above, unlike the ager limitatus, the ager scamnatus does not have a standard system for laying down paths, and the sizes of individual pieces must have varied rather a lot. When they ran along the length of the whole plot of land, they were called strigae, and when they ran horizontally, they were called scamna, which should prompt us to consider the shape of the entire area of land. The ancient commentaries on Frontinus make use of compass points, claim that the length referred to the dimension going from north to south, and that the width referred to the west-to-east dimension (Agenn. 3, 12 = Comment. 53, 21 Th.). <page break 694/695> As an example, Frontinus bring up Suessa Aurunca (3, 2). In the liber col. I 237, we read the following about this location: iter populo non debetur. ager eius pro parte limitibus intercisivis et in lacineis est adsignatus. Similar details are given for Atina (230, 4: ager pro parte in lacineis et per strigas adsignatus), Ostia (236, 7: in praecisuris, in lacineis et per strigas), Terventum (238, 14: in praecisuras et strigas), and Anagnia (230, 15: per strigas). <10> In the same sources we also find notes about areas of land whose ager is entirely or pro parte, lociis variis divided intervisivis limitibus, with or without the clarification in laciniis. In such situations, a large variation in distances between the limites are given - anywhere from 1400ft to 2500 ft - and they seem to have been oriented depending on the properties of the location in question (240, 11. 252, 8). <20> This sort of haphazard division seems to have taken place in earlier limitationes (229, 6. 235, 17. 236, 5; in contrast 259, 20; cf. Agenn. 1, 28 = Comment. 52, 13 Th.). Rudorff uses this as the basis for his theory that the limites were symbolically named ‘cutting’ or ‘destructive’, because they ‘destroyed’ the original non-standardised land-boundaries after Roman conquest (296ff., and cf. Schulten Bonn. Jahrb. CIII 25). <30>

The difference between dividing up land into scamna and strigae versus applying limitatio is, therefore, the fact that in the first case, there are no paths with public rights of way - the limites themselves - and the plots of lands are only separated by rigores or by arbitrary tracks cutting through the overall area of land, as well as the fact that the plots of lands themselves consisted of ‘strips’ and ‘banks’, or just ‘snippets’ and ‘subsections’. Frontinus makes the distinction all the more clear by using diagrams: ager ergo limitatus hac similitudine (fig.) decimanis et cardinibus continetur, ager per strigas et per scamna divisus et adsignatus est more antiquo in hanc similitudinem (fig.), qua in provinciis arva publica coluntur (3, 5). <40> Accordingly, ager scamnatus was originally the usual way of dividing land for the purpose of distribution, which Suessa Aurunca serves as an example for, and then it later became the usual method for dividing ager publicus in the provinces. This difference also highlights the link between the division of the land and its legal quality (on this, cf. especially Weber 12ff. and Mommsen Zum r. Bodenrecht, Herm. XXVII 79ff.; Ges. Schr. V 85ff.). <50> The following list will serve to justify the approaches I have described here.

19. The uses of limitatio proper

Limitatio was used in the following situations:

1. The establishment of citizen- and Latin- colonies during the republic (cf. Mommsen Feldm. II 154f.; Ges. Schr. V 153). <60>

2. The founding of municipia, praefectura, fora, and conciliabula in Italy (Lex Mamil. c. 55; this was presumably to do with the individual territories within the larger areas of land in which plots of land given out following the lex agraria of 111 that used to be ager publicus had become ager privatus, Fabricius S.-Ber. Ak. Heidelberg 1924/25 1. art. 22ff.); see above p.675). <page break 695/696>

3. Handing land over into full Roman property by distributing it one plot each. (Sic. 154, 9 may imply that this land did not undergo limitatio. However, the terminus of the Gracchan III vir(i) a(gris) i(udicandis) a(dsignandis) or a(dtribuendis) CIL I² 639 from Atina in Lucania was on land that had undergone limitatio (see above p.682), which could not have been ager colonicus, since Atina was a municipium. For the most part, besides, it would be impossible to conceive of distributing land without having already performed a limitatio). <10>

4. The distribution of land to veterans by founding colonies in Italy and in the provinces since the time of Sulla. The citizen and military colonies are such a key type of ager divisus et assignatus that the agrimensores often use these terms and simply ‘colonial land’ interchangeably (Mommsen loc. cit. 154 = 153, where the examples are listed). <20>

5. Surveying ager publicus in Italy, which was exempt from occupation or distribution, and was rented out for specific purposes, like the ager Campanus, which the Gracchan terminus CIL I² 640 comes from. On the legal state of the ager Campanus up to Caesar’s second agrarian law in 59 BCE, cf. Mommsen Ges. Schriften I 108 (see below sect. 20). <30>

6. Surveying conquered provincial land, or provincial land that had come into possession of the state by other means. This includes in particular the huge limitationes undertaken in Africa after 146 BCE and under Augustus, as well as the limitationes in what used to be the kingdoms in Cyrenaica after 74 BCE (see above p.674 and below sect. 23).

7. Surveying and dividing ager vectigalis in the provinces (Hyg. Gr. 205, 1: multi huius modi agrum (sc. ager arcifinium vectigalem) more colonico (see note in 4) decimanis et kardinibus diviserunt, hoc est per centurias, sicut in Pannonia. <40> The land near Chott el Fejej in Africa that underwent limitatio in 29/30 will also have been ager vectigalis.

8. Surveying ager quaestorius due to be sold in individual pieces (laterculi, plinthides) of 10 actus squares of 50 iugera (Hyg. 115, 15. Sic. 136, 17. 152, 24. 154, 1; cf. Rudorff 285ff.). <50>

9. Surveying imperial territories in Africa (cf. the Lex Manciana which was passed before Trajan, Bruns Font.⁷ 295 nr. 114 v. 7, where subseciva are mentioned, which are only found in land that has undergone limitatio, and the inscription on the ara leg. Hadr. Bruns nr. 115, 2, 1 and 116, 2, 9 with the reference to centuriae). Here, the land was divided into centuriae of 200 iugera all the way into the 5th century (Cod. Theodos. XI 28, 13 from the year 422; cf. Barthel 49ff.). The fact that we find an imperial saltus that has undergone limitatio in Africa, which is otherwise entirely unattested, may have something to do with the older limitationes which were extended to cover the entire province. <60>

10. Surveying boundaries of foreign territories in the provinces or in some plot of ager privatus, which only actually needed to be undertaken per extremitatem, by imposing a forma (Front. 5, 3: hunc agrum multis locis mensores, quamvis extremum mensura comprehenderint, in formam in modum limitati condiderunt). <page break 696/697>

11. Pliny mentions a purely private form of limitatio that took place in vineyards, with an 18ft decumanus which two wagons could pass alongside each other in, as well as a cardo of the same width, and 10ft limites transversi (n. h. XVII 169). <10>

When it comes to the division land into scamna and strigae, all the evidence we have has already been handled in sect. 16. Frontinus says nothing about which distributions this more ancient method of division would occur in versus which distributions would take place under limitatio proper. <20> The comments in the libri coloniarum do not allow us to answer this question along chronological or other lines, either: Suessa was an ancient Latin colony, a municipium in Cicero’s time, and then a veteran colony under the triumvirs, and Bovianum was an Augustan veteran colony. The other places whose territory was divided into scamna and strigae are only known as municipia or praefecturae. <30> Hyginus the younger confirms that it was mainly used for ager publicus in the provinces, in that he complains about the limitatio of the ager vectigalis (205, 1) and proposes his own method of dividing the land into scamna and strigae, as well as privatas limitum observationes, instead of the usual rigores (ibid. 8; cf. otherwise Mommsen Herm. XXVII 85f.; Ges. Schr. V 91f.).

20. Remains of Roman limitationes

Today’s road networks in various places, especially in Campania, in Gallia cisalpina, and in Africa, preserve some Roman limitationes. <40> The most important of these remains are in the Ager Campanus, because the Gracchan terminus from 132/31 BCE (CIL I² 640) with the inscription SD I KK XI (see above p.681) and Frontinus’s note that the cardo was oriented to the east and the decumanus to the south (29, 4) allow us to reconstruct it. The decumanus maximus begins in the north, where the volturnus, the northern border of the limitatio grid, emerges from the mountains, and then travels to the west across ancient Capua and Atella, and ends in the southern corner of the plain not far from Napoli. <50> The line would have continued through the centre of Napoli itself along the Toledo (Via Roma) towards the Castell dell’ Ovo. It lines up almost perfectly with the meridian, but cuts through the plain when extended as far as possible. When the map was recorded in Beloch Campanien Pl. XII, that being 1817-1819, the decumanus maximus north of Capua was still preserved as a 3.6km long path through a field. <60> The cardo began at the foot of Monte St. Michele on the eastern end of the plain at Maddaloni, and passed through it along the most northern bend of the Clanis where it was most wide. The rest of the decumani, up to the 16th, are almost entirely preserved in the pars sinistra (eastern) south of Maddaloni; and in the pars dextra they are partially preserved up to the 16th south-east of Vico; the cardines in the pars citrata (northern) have left behind significant remains up to the 8th to the north of Capua Vetere and up to the 20th 7km south of Aversa, occasionally in pieces up to 10km long. <page break 697/698> The grid of square centuriae with sides 710m = 20 actus in length, ie. the usual size of 200 iugera, are especially clear to the south east of Capua (cf. Beloch Campanien² 310. Schulten Flurteil. 30ff. with panel VI, though the regiones and the site of the decumanus maximus have not been correctly determined, and cf. especially also Barthel 97f.). <10>

21. Gallia cisalpina

The remains of limitationes of part of a Roman colony in Gallia cisalpina have been even more fully preserved. Schulter has collected around 20 examples from the Po valley and Venetia, including some which even today offer an excellent picture of the Roman limitationes (loc. cit. 15ff. with reductions of the relevant parts of the carta topografica to a 1 : 150000 scale). <20> At Parma, the remains of at least 22 centuriae stretch out in one direction, and 33 in the other direction, at around 450km², and the magnificent limitatio to the north east of Patavium’s grid of roughly 250 centuriae extended accordingly is still almost entirely apparent. The centuriae measured on average 710m squares, ie. 200 iugera, and only in a limitatio to the south west of Treviso (Tarvisium) at Musone do they seem to have had a different area. <30> They are all oriented in different directions, but none of them follow compass directions. At Fidentia (Borgo San Donnino), Parma, Regium Lepidum, Bononia (lmitatio to the west of Reno), Forum Cornelii (Imola), the systems may have been linked to the Via Aemilia built in 187 BCE, though the street could not have been built along land that had already undergone limitatio. <40> In contrast, the limitationes from Tannetum, Mutina, and Bononia to the east of Reno, are at about a 3-4° angle to the Via Aemilia. These limitationes must therefore have been older. The limitatio to the north west of Treviso seems to have used the Via Postumia as its decumanus maximus. Wherever two limitationes meet each other, they always do so at an angle. The limites were often marked by water ditches, the fossae limitales of the Lex Mamilia (see below p. 778). The limitationes did not extend long enough to meet the larger rivers, like the Po. <50>

22. Germania

In Germania we are also able to find traces of Roman limitationes. The inscription CIL XIII 8254 on an altar in Cologne, which the possesores dedicated ex vico Lucretio scamno primo, seems to refer to a region of the city, and not to a plot of land. <60> The dedicatory inscription in Obrigheim around Neckar ibid. 6488, in contrast to Zangemeister’s reading of aed(em) si(gnum) a[g]r(um) (centuriarum) IIII, states that the four centuriae, which if they had the usual area would comprise of 800 iugera, would be an absolutely ridiculously large dedication (cf. Schulten Bonn. Jahrb. CIII 12 and 37ff. and Barthel 48f., who in the latter inscription reads arg[enteum with a specification of weight. In a rough copy of the inscription, which I am grateful to Gropengießer for, Drexel does seem to have found a small I behind ∧̣CR; however, he supposes that the symbol ⵢ before the numeral IIII did not refer to centurias, but sicilicos (cf. Haug in Wagner Fundstätten in Baden II 395f.). <page break 698/699> Four sicilici were 2400 square feet, or 210m²: an entirely appropriate size for a small donation. Regardless, it has still not been ruled out that the land in the Agri Decumati had undergone limitatio. However, the traces of ancient land division which Meitzen III 157 is inclined to see in a few country paths in the area Friedberg in Oberhessen are too insignificant to allow us to say with certainty whether they were Roman or not. <10> Accordingly, we cannot be at all certain about Dopsch’s claims that rest on this (Grundlagen der europ. Kulturentw. I 339f.). The same goes for the conjectures about remains of Roman land division in the area around the Rhine and Moselle, which Philippi has referred to (Gött. Gel.-Anz. 1920, 49). <20> The traces of a Roman limitatio which Schumacher has observed in Rhenish Hesse (Mainz. Ztschr. XVI 14f. Siedlugns- u. Kulturgesch. der Rheinlande II 221f.) are somewhat more likely to be correct, though they are still not surefire enough. Finally, I will also direct attention towards the outstanding work on the issue of land-division in Augusta Raurica near Basel which Burckhardt-Biedermann has carried out with as much detailed knowledge as caution (Die Kolonie Augusta Raurica 51ff.; see below vol. I A p.293). <30>

23. Africa

In the province Africa, the remains of huge limitationes are very clearly preserved in today’s road network, which have been completely considered in detail by Barthel (1911, 52ff.) following on from the work by C. T. Falbe (1833) and Schulten (L’arpentage rom. en Tunisie, Bull. arch. du Comité 1902, 129ff.). <40> Near Carthage, there are still many traces of a grid of 710m square centuriae with an area of 200 iugera, whose north-westerly limites form a 29° angle with the meridian. To the north, these remains extent up to Bagradas, and in towards the mainland they stretch 70km long almost reaching the fossa regia (see above p.661): most notable are the remains of a Roman military road over 6km wide preserved in pieces, stretching from Carthage to Theveste, and going in the same direction as the limites. <50> The limitatio passes over the peninsula of Cape Bon to the south east, which it presumably covered completely, and to the south it stretched at least 90km from Carthage up to the 36th degree of latitude (Barthel loc. cit. and panels II and III).

At this point, this grid encounters a second one, whose limites are not only preserved in today’s road network, but even some of the ancient stonework still survives: this second limitatio contained square 200-iugera centuriae, and we are also sometimes able to make out subdivisions within them. <60> These remains stretch 100km to the south across Hadrumetum and Thapsus all the way down to near Thysdrus. Even in the area of the modern city Sfax, ie. in the area of Thenae, we can still make out traces of this limitatio (Barthel 57ff. with panels III-V). This second system was angled 39° to the west of the meridian. <page break 699/700> This exactly matches the orientation of the limitatio carried out near Chott el Fejej in 29/30 CE by the legio III Augusta (see above p.682 and p.687). Barthel has been able to reconstruct the limitatio grid, basing his work on the stones found there and the regiones and numbers marked on them, though Toutain, who first published the termini, interpreted them wrongly (Toutain Mém. présent. à l’Acad. XII 1907, 341ff. and cf. Barthel Wochenschr. kl. Philol. 1909, 1257. Toutain Mem. Soc. des ant. de France 1910, 79ff. Barthel Bonn. Jahrb. CXX 1911, 60ff. and 125 for sources). <10> Accordingly, the decumanus maximus began somewhere along the north coast of Africa, slightly to the west of Hippo Regius, and continued down to Syrte Minor, which it cut through around 12km to the east of Tacape. The cardo maximus went from Promuntorium Mercurii (Cape Bon) down to Theveste. <20> Both of the main limites were therefore placed in such a way that they cut through the widest parts of the province when reckoned from north-west to south-east, and from north-east to south-west.

24. Dating the limitationes in Africa

Given the above, we are able to conclude the following about the dates of the limitationes that took place in Africa: the second limitatio, in one self-consistent grid, spanned over the entire limitatio carried out by Caesar and left to the senate as the province Africa by Augustus in 27 BCE. <30> The absolutely huge amount work of land-surveying, which the lmitatio done in 29/30 CE near Chott el Fejej was attached on to, presumably traces back to Augustus. The other limitatio, which has left traces around Carthage, must have therefore already existed at that time, since the military road going from Carthage to the camp in Theveste can only have been built after it. It never stretches beyond the borders of the ancient province as it was introduced in 146 BCE, which means that it must have come before Caesar. <40> Presumably, we are looking at the remnants of the limites which the relevant part of the Lex agraria of 111 BCE refers to (CIL I² 585 v. 89). The same law (v. 80) tells us of the presence of formae publicae, which therefore point us back towards the time at which the province was first introduced. That, then, with reference to the X viri ex lege Livia, was when the limitatio took place, and C. Gracchus’s land distributions will have happened following it. <50> Like the Augustan limitatio in the expanded province, it was oriented lengthwise across the area up to the Fossa Regia which was taken over in 146 (Barthel 73ff. with panel I). <60> Both limitationes, the one during the republic and the one under Augustus and extended by Tiberius, are marvellous works of Roman land surveying. They lasted throughout the entire empire, and were still apparent in the 4th and 5th centuries in the African provinces including Numidia, both in the imperial territories as well as in land that lay outside it (Barthel 49ff. Though in 117ff. he aims to show that the peculiar misplacement of the eastern coast of Africa in Ptolemy’s map from Hippo Diarrhytus to Syrte Minor was the result of a misunderstanding in the orientation of the second limitatio grid, and in so doing presumes that he was using a provincial forma for geographical purposes, we are unable to comment here on how correct this is). <page break 700/701>

[Fabricius.]

This article is referenced by: Augures, C. Mamilius Limetanus (7), Spectio