What's in a location? A lot, as it turns out!

Summary

A review of the spatial distribution of poleis, region-by-region, illuminated outliers from their assigned region that warrented closer inspection. In addition, a search for, and analysis of, precisily co-located poleis was conducted. The exercise resulted in relocating some poleis in the catalog which is documented in detail in this article.

In summary, 69 (~6.7%) of the 1035 poleis contained in the catalog were reviewed, 28 were relocated to more accurate positions while an additional 22 were reclassifed as location "unknown" within their assigned region. Taken together, the relocated poleis represent ~4.6% of the catalog. The total number of unlocated poleis rose to 144 or ~13.9% of the catalog. The remaining 19 poleis were not relocated.

The latest version of the Polis-Inventory dataset is freely available from the app and can be obtained from a link found on the About page.

Background and Approach

With the advent of easy access to satellite maps we have become use to and expect data placed upon them to be accuratley located. Such is not the case with the database of Archaic and Classical Poleis. Location accuracy spans the spectrum from high accuracy, pinning a polis on top of its archeological site, to low accuracy, poleis that cannot be located more than assigning them to a region.

Extreme effort by multiple teams has been made to provide the best locations possible and the location data is continuously being improved. But the reality is more often polis locations are narrowed by the literary record alone and less often by the archeological record.

Placing the polis dataset on high accuracy base maps can amplify the visual inconsistency of some locations, but in some cases allows us to modify them to be extremely accurate.

The data presented in the Polis-Inventory app originates from the website polis.stanford.edu, which is managed by the classical department at Stanford University. For the purpose of this article this database is referred to as the S-Inventory.

The S-Inventory itself is derived from the work by Hansen & Neisen "An Inventory of Archaic and Classical Poleis", which is referred to as the Inventory.

In turn, the Inventory provides the location of each polis, where such data is available, by referencing a map and position in the Barrington Atlas by Talbert (Barr.). Refer to Appendix 1 below for a complete description of the treatment of location in the Inventory.

Not all poleis are located by the Inventory. As of this writing 144 (~14%) poleis are classified as "unlocated" in the Polis-Inventory app dataset. All of these poleis are associated with a region, and in this report these poleis are referred to as "unlocated" within their region. The approach for plotting these poleis in the Polis-Inventory app, being that they are unlocated, is described in the blog post The Lost Poleis.

The Inventory reports latitude and longitude for a location in Barr. to two decimal places, often rounded (they give an example of rounding a latitude of 38.29 to 38.30). The Barr. maps can be used to recover this precision, but location precision in Barr. would remain limited to ~1 kilometer at best (a hundredth of a degree).

Considering that the Barr. base maps do not coenside 1:1 with modern satillite maps, positions will shift further. This does not diminish the value of the dataset in its current state, but it does explain why a particular location might seem unexpected, for example, a coastal city located slightly to sea.

The S-Inventory curators made the first stride I know of towards improving the location data found in the Inventory. I have not spoken directly with the Stanford University team, but it seems clear that an attempt was made to improve the precision and accuracy of the Inventory locations and to add elevation data in the process. Some notes regarding the location data found in the S-Inventory can be found in Appendix 2.

A recent review of the S-Inventory locations led to a re-examination of a subset of the polis catalog. The re-examination resulted in a change in location for some of the poleis analyzed. The purpose of this article is to document the location changes made, which are included in the current version of the Polis-Inventory app (v1.6.2).

By design, the review was limited to a visual inspection of poleis plotted, region-by-region, looking for outliers. In addition, co-located poleis were reviewed. This approach led to a review of 69 of 1035 polis locations, just under ~7% of the catalog.

Most often a clerical error was the reason for a change in location. Other times, it was clear a change was warranted, but a unique answer was not in reach and a judgement, based on the various data sources available, was required.

Modifications to the S-Inventory

The following describes in detail the modifications to the S-Inventory database as of V1.6.2 of the Polis-Inventory app.

There are two types of modifications recorded, (1) polies repositioned to more accurate locations and (2) polies that were reassigned a location of "unlocated" within their region. Appended at the end of each report you will find the exact location data as reported in the Inventory.

Repositioned Poleis

The following poleis were relocated from the S-Inventory, mostly as a result of error spotting when plotting groups of poleis region-by-region.

Aitna, Sikelia (#8): Incorrectly possesses the same coordinates as Nakone, Sikelia, and is also incorrectly located in modern Messina, Sicily; repositioned to Aitna I, modern Katane, per the Inventory (Inventory: Map 47. Location of Aitna I is Katane … Aitna II is unknown "but is located somewhere on the slopes of Mount Etna")
Selinous, Sikelia (#44): The S-Inventory incorrectly located this polis in western Turkey. Its location was moved to coinside with the archeological ruins in south-western Sicily (Inventory: Map. 47 Lat. 37.35 Long. 12.50)
Poseidonia, Italia & Kampania (#66): Incorrectly possesses the same coordinates as Pithekoussai, Italia & Kampania; repositioned to the archeological site of Paestum (Inventory: Map. 45 Lat. 40.25 Long.15.00)
Pylos, Elis (#263): Not to be confused with Myceanean Pylos, both the S-Inventory and Inventory locate the polis over sea. However, the Inventory suggest a location near the modern city of Agrapidochori. Britannica suggest a location near the Pineios River which empties into the reservoir on the opposite side from Agrapidochori; repositioned near to Agrapidochori (Inventory: Map 58. Lat. 21.30 Long. 37.55)
Orneai, Argolis (#354) The S-Inventory co-locates Orneai with Mykenai, Argolis. The Inventory and web-search are definitive the location is near modern Lyrkeia; repositioned near modern Lyrkeia (Inventory: Map. 58 Lat. 37.45 Long. 22.35)
Belbina, Saronic Gulf (#359): Incorrectly possessed the same coordinates as Belbina, Lakedaimon; repositioned to the small island of Agios Georgios per notes in the Inventory and a web-search (Inventory: Map. 58 Lat. 37.30 Long. 23.55)
Athenai, Attica (#361): At my personal whim I moved the location directly on top of the Tholos in the Athenian Agora (e.g. to the archeological site of the Agora) (Inventory: Map. 59 Lat. 38.00 Long. 23.45)
Alponos, East Locris (#379) The Inventory co-locates Alponos with Nikaia, East Locris. The S-Inventory positions Alponos east of Thermopylai, as described by the Inventory, albeit slightly off shore, also co-located with Nikaia; repositioned just slightly onshore without changing the intention of the S-Inventory locations (Inventory: Map. 55 Lat. 38.50 Long. 22.35)
Halai, East Locris (#380): Both the S-Inventory and the Inventory locate Halai over sea southwest of the island of Potamos. However, the Inventory and a web-search place Halai at an archeological site near the modern city of Theologos located in East Locris; repositioned to this archeological site near Theologos (Inventory: Map 55. Lat. 35.40 Long. 23.10)
Nikaia, East Locris (#385) See Alponos above. It can be arugued both poleis should be restated unlocated, especially Nikaia; repositioned just to the side of Alponos so one polis does not overlap the other. (Inventory: Map. 55 Lat. 38.50 Long. 22.35)
Boion, Doris (#390): Its coordinates were missing in the S-Inventory, but the Inventory did list coordinates, which were adopted verbatim (Inventory: Map 55. Lat. 38.40 Long. 22.25)
Methylion, Thessaly & Adjacent (#402): Incorrectly possessed the same coordinates as Chen, Lakedaimon; repositioned based on a web-search to nearby the modern town of Mirina, Thessaly (Inventory: Map 55. Unlocated but see infra)
Therma, Aegean (#418) The S-Inventory has Terma incorrectly mapped to Ikos; repositioned to the island of Ikaria (Inventory: Map. 61 Lat. 37.40 Long. 26.10)
Herakleia, Thessaly & Adjacent (#430) Correctly co-located with Trachis, Thessaly & Adjacent. The S-Inventory location is confirmed by a web-search; Modified position slightly to coincide with the web-search and to offset slightly from Trachis (#432), a later name for this Herakleia (Inventory: Map. 55 Lat. 38.50 Long. 22.25)
Ikos, Aegean (#482) The S-Inventory has Ikos incorrectly mapped to Imbros; repositioned to the island of Alonissos (Inventory: Map. Lat. 39.10 Long. 23.55)
Imbros, Aegean (#483) The S-Inventory co-locates Imbros incorrectly with Ios, Aegean; modified position to the island of Gokceada (Inventory: Map. 51 Lat. 40.10 Long. 25.45)
Sindos, Thrace: Axios - Strymon (#551): The S-Inventory location does not coinside with its assigned region; repositioned based on a web-search to the suburb a Thessonaliki of the same name, a location also mentioned in the Inventory as a possible one (Inventory: Map 50. Lat. 40.40 Long. 22.45)
Dion, Thrace: Axios – Strymon (#569) Incorrectly co-located with Dion, Makedonia; moved position to near Mt. Athos based on the Inventory and a web-search (Inventory: Map. 51 Lat. 40.20 Long. 24.10)
Neapolis, Thrace: Axios - Strymon (#586): There are 4 poleis in the Inventory with the name of Neapolis. The S-Inventory locates number 586 over the modern city of Kavala, which suggests a confusion with Neapolis, Thrace: Strymon - Nestos (#634). The Inventory locates number 586 over sea in the Thermatic Gulf, which is more centered in the Thrace: Axios - Strymon region. The Inventory mentions the modern city of Polychrono as one possible location; repositioned the location between the Inventory location and Polychrono, on land, following the same latitude (splitting the difference, so to speak) (Inventory: Map 50. Lat. 40.00 Long. 23.30)
Skithai, Thrace: Axios - Strymon (#610) Co-located with Kithas, Thrace: Axios - Strymon. The Inventory proposes its position is the same as Kithas. The S-Inventory follows this suggestion yet updates both locations together; repositioned Skithai only slightly from Kithas to visually separate the two for plotting (Inventory: Not in Barr. Unlocated)
Neapolois, Thrace: Strymon - Nestos (#634): There are 4 poleis in the Inventory with the name of Neapolis. The S-Inventory locates number 634 over sea just south of the city of Kavala; repositioned to a location on land over the modern city of Kavala (Inventory: Map 51. Lat. 40.55 Long. 24.25)
Drys, Thrace: Nestos - Hebros (#644): Both the S-Inventory and the Inventory locate Drys over sea southwest of the island of Samothrake. However, the Inventory describes the location as being on the coast opposite (north) of Samothrake between Orthagoria and Mesambrie (confirmed by a web-search); repositioned just to the west of Mesambrie on the coast (Inventory: Map 51. Lat. 40.05 Long. 25.40)
Maroneia, Thrace: Nestor - Hebros (#646): The S-Inventory location does not coinside with its assigned region; repositioned based on a web-search to nearby the modern town of Maroneia (Inventory: Map 51. Lat. 40.50 Long. 25.30)
Kolonai, Propontic Asia Minor (#746): Incorrectly possesses the same coordinates as Kolonai, Troas; repositioned to the location provided by the Inventory verbatim (Inventory: Map 51. Lat. 26.55 Long. 40.20, but not securely located, see infra)
Prokonnesos, Propontic Asia Minor (#759): The S-Inventory has the Latitude / Longitude reversed from the Inventory; Reversed the coordinates according to the Inventory choosing Lat. 40.35 Long. 27.30 (Inventory: Map 52. Lat. 40.35-40 Long. 27.30-45)
Naulochon, Ionia (#857) The S-Inventory has Maulochon co-located with Itanos, Crete and Naulochon, Ionia. The Inventory states its location is the port of Priene; repositioned to the Priene archeological site (Inventory: Map. 61 Lat. 37.40 Long. 27.15)
Passanda, Karia (#922) Correctly co-located with Karbasyanda, Karia, offset slightly to visually separate the two for plotting (Inventory: Map. 56 Lat. 36.50 Long. 28.35)
Kydonia, Crete (#968) The S-Inventory has Kydonia incorrectly co-located with Itanos; repositioned to modern Chania (Inventory: Map. 60 Lat. 35.30 Long. 24.00)

Poleis Reassigned To Location "Unknown"

The following poleis in the S-Inventory were reclassified to location "unknown" within their region, mostly as a result of error spotting when plotting groups of poleis region-by-region. All but two are listed as unknown in location in the Inventory.

Nakone, Sikelia (#40): Shares the same location as Aitna, Sikelia (which is incorrectly located in modern Messina, Sicily); changed location to "unlocated" within its region (Inventory: Map. 47, Unlocated)
Phara, Akarnania & Ajacent (#133) The S-Inventory location has no relationship at all to the city’s description in the Inventory (and should probably be unlocated, per the Inventory notes); changed location to "unlocated" within its region (Inventory: Map 54. Lat. 38.40 Long. 21.35)
Akripos, Aitolia (#144): The S-Inventory location does not coincide with its assigned region nor the description in the Inventory; changed location to "unlocated" within its region (Inventory: Map 55. Unlocated, not in Barr.)
Messapioi, West Lokris (#163): The S-Inventory location does not coincide with its assigned region nor the description in the Inventory; changed location to "unlocated" within its region (Inventory: Map 55. Unlocated)
Trachis, Phokis (#195) Shares (erroneously) the same location as Herakleia and Trachis, Thessaly & Adjacent (which are supposed to be co-located). The Inventory gives a position over sea in the Gulf of Corinth and states the location is unknown (even though a coordinate is provided) and discusses 3 possible locations. The Inventory and a web search suggest the western tip of Euboea as the best candidate location; changed location to “unlocated” within its region (which will end up being positioned close to that given in the Inventory). (Inventory: Map. 55 Lat. 38.20 Long. 22.45, but see infra for a discussion of the unresolved location)
Ascheion, Achaia (#232): The S-Inventory location does not coincide with its assigned region nor the description in the Inventory; changed location to "unlocated" within its region (Inventory: Unlocated)
Lenos, Elis (#257): Shares the same location with 8 poleis** in the S-Inventory; changed location to "unlocated" within its region (Inventory: Map 58. Unlocated)
Opous, Elis (#261): Incorrectly possesses the same coordinates as Opous, East Locri; changed location to "unlocated" within its region (Inventory: Map 58. Unlocated)
Thraistos, Elis (#264): Shares the same location with 8 poleis** in the S-Inventory; changed location to "unlocated" within its region (Inventory: Map 58. Unlocated)
Euaimon, Arkadia (#269): Shares the same location with 8 poleis** in the S-Inventory; changed location to "unlocated" within its region (Inventory: Unlocated)
Koila, Arkadia (#277): The S-Inventory location does not coincide with its assigned region nor the description in the Inventory; changed location to "unlocated" within its region (Inventory: Unlocated (Meyer (1957a) 86))
Phoriera, Arkadia (#293): Shares the same location with 8 poleis** in the S-Inventory; changed location to "unlocated" within its region (Inventory: Unlocated)
Noudion, Tripylia (#308): Shares the same location with 8 poleis** in the S-Inventory; changed location to "unlocated" within its region (Inventory: Unlocated)
Aithaia, Messina (#312): Shares the same location with 8 poleis** in the S-Inventory; changed location to "unlocated" within its region (Inventory: Unlocated)
Phakion, Thessaly & Adjacent (#410): The S-Inventory location does not coincide with its assigned region nor the description in the Inventory; changed location to "unlocated" within its region (Inventory: Map 55. Unlocated (cf. Decourt (1990) 155-58))
Chen, Thessaly & Adjacent (#425): The S-Inventory location does not coincide with its assigned region nor the description in the Inventory; changed location to "unlocated" within its region (Inventory: Map 55. Unlocated)
Chedrolioi, Thace: Axios – Strymon (#566): Shares the same location as Smila, Thace: Axios – Strymon. The S-Inventory location does not coincide with its assigned region nor the description in the Inventory; changed location to "unlocated within its region (Inventory: Map. 56 Unlocated)
Smila, Thace: Axios – Strymon (#566): Shares the same location as Chedrolioi, Thace: Axios – Strymon. The S-Inventory location does not coincide with its assigned region nor the description in the Inventory; changed location to "unlocated within its region (Inventory: Map. 56 Unlocated)
Eion, Thrace: Strymon - Nestos (#570) The S-Inventory co-locates Eion with Eion, Thrace: Axios - Strymon. However the Inventory specifically states little is known of Eion, Thrace: Strymon - Nestos and it is not to be confused with the other Eion, Thrace: Axios - Strymon; changed location to “unlocated” in its region (Location: Not in Barr. Maps 50-51, Unlocated)
Polichnitai, Thrace: Axios - Strymon (#596) Possibly related to Stolos, the position the S-Inventory chose, another suggestion by the Inventory is modern Smixi, which would be in a different region altogether; setting to unlocated (Inventory: Unlocated, not in Barr. 51)
Iolla, Aiolla & SW Mysia (#812): The S-Inventory location does not coincide with its assigned region nor the description in the Inventory. The Inventory offers several options; however, changed location to "unlocated" within its region (Inventory: Map 56, Unlocated, but probably near Adramyttion)
Oiai, Rhodos (#998): The S-Inventory location does not coincide with its assigned region nor the description in the Inventory; changed location to "unlocated" within its region (Inventory: Map 60. Unlocated)

** The following eight (8) Poleis were all located at the same location in the S-Inventory (therefore plotted visually on top of each other): Ascheion in Achaia, Lenos and Thraistos in Elis, Euaimon and Phorleia in Arkadia, Noudion in Triphia, Aithaia in Messenia and Oios in Lakedaimon.

Polis Locations Not Modified

As mentioned, analyzing the distribution of poleis region-by-region and co-located poleis triggered a review of their locations.

The following is a list of 21 poleis inspected but not modified in location with respect to the S-Inventory.

Pithekoussai, Italia & Kampania (#65) Co-located with Poseidonia, Italia & Kampania; not repositioned (Inventory: Map. 44 Lat. 40.45 Long. 13.55)
Aigition, Aitolia (#143) Position appears correct; not repositioned (Inventory: Map 55. Lat. 38.30 Long. 22.10)
Kallion / Kallipolis, Aitolia (#147) Possition appears correct; not repositioned (Inventory: Map 55. Lat. 38.35 Long. 22.10)
Phara, Arkadia (#290) Torn, it probably should be reclassified as unlocated, but will keep the S-Inventory location as it plots in western Arkadia, per the Inventory (Inventory: Unlocated, but presumably in western Arkadia, …)
Belbina, Lakedaimon (#326) Co-located with Belbina, Saronic Gulf; not repositioned (Inventory: Map 58. Lat. 37.15 Long. 22.15)
Oios, Lakedaimon (#339) Shares the same location as the 8 co-located Poleis mentioned (see ** above). Keeping the S-Inventory location as the given Inventory location is more approximate and yet nearby. This is the only polis of the 8 that retained this location (Inventory: Map 58. Lat. 37.20 (?) Long. 22.25)
Mykenai, Argolis (#353) Co-located with Orneai, Argolis; not repositioned (Inventory: Map 58. Lat. 37.45 Long. 22.45)
Erineos, Doris (#391) The Inventory provides a location that makes much more sense than that in the S-Inventory. However, the location is the same as for Boion, Doris, so one city would plot on top of the other. It seems to me the Inventory writers meant to provide approximate locations in the region for both cities and so one logical solution would be to mark them both “unlocated" in Doris. Another would be to spatially offset one from the other slightly. As I’m not convinced to do either, I will use the S-Inventory location (Inventory: Map 55. Lat. 38.40 Long. 22.25)
Trachis, Thessaly & Adjacent (#432) Co-located with Herakleia, Thessaly & Adjacent; not repositioned (Inventory: Map 55, location as Herakleia (q.v.))
Ios, Aegean (#484) Co-located with Ikos, Aegean and Imbros, Aegean; not repositioned (Inventory: Map 51. Lat. 36.45 Long. 25.15)
Dion, Makedonia (#534) Co-located with Dion, Makedonia; not repositioned (Inventory: Map 49-50. Lat. 40.10 Long. 22.30)
Eion, Thrace: Axios - Strymon (#570) Co-located with Eion, Thrace: Strymon - Nestos; not repositioned (Inventory: Not in Barr. Map 50-51. Unlocated)
Kithas, Thrace: Axios - Strymon (#579) Co-located with Skithai, Thrace: Axios - Strymon; not repositioned (Inventory: Map 50. Lat. 40.25 Long. 23.05)
Stolos, Thrace: Axios - Strymon (#614) Co-located with Polichnitai, Thrace: Axios - Strymon; not repositioned (Inventory: Map 51. Lat. 40.20 Long. 23.40)
Kolonai, Propontic Asia Minor (#746) Co-located with Kolonai, Troas; not repositioned (Inventory: Map 51. Lat. 40.20 Long. 26.55)
Polichna, Troas (#789) The Inventory states “unlocated” and the current location does seem out of place, but Wikipedia references a location near to modern Balikliona, which is where the S-Inventory has it placed, so no change (Inventory: Map. 56. Unlocated)
Smyrna, Ionia (#867) Co-located with Naulochon, Ionia; not repositioned (Inventory: Map 56. Lat. 38.25 Long. 27.10)
Karbasyanda, Karis (#895) Co-located with Passanda, Karia; not repositioned (Inventory: Map 65. Lat. 36.50 Long. 28.35)
Itanos, Crete (#965) Co-located with Naulochon, Ionia and Kydonia, Crete; not repositioned (Inventory: Map 60. Lat. 35.15 Long. 26.15)

Appendix 1: Polis Location in the Inventory

Quoting the Inventory for location data we have:

"Map reference is to R. J. A. Talbert (ed.), The Barrington Atlas of the Greek and Roman World (Princeton, 2000), in this inventory abbreviated Barr. We record the map number,latitude and longitude, e.g.Delphoi: “Map 55. Lat. 38.30, long. 22.30.” The exact latitude is 38.29, but we record the coordinates with an exactitude of 5 minutes of arc only. If the polis in question is unlocated, we write for, e.g., Kallipolis: “Map 52. Unlocated.” If an unlocated polis is not recorded in Barr.’s map-by-map directory, we write for, e.g., Sombia: “Map 52. Unlocated. Not in Barr.” If the author agrees with the Barr. location, there is usually no further discussion of this issue. If the author disagrees with the Barr. location, a
discussion of the issue follows either here or in a following section (indicated by “see infra”)."

Parsing this further:

If the polis is referenced in Barr. and is identified on a particular map, the coordinates provided by the Inventory are a rounded reference to make it easier for you to find the location in the referenced Barr. map.
If the polis is referenced in Barr. but is not identified on a particular map, then the Inventory states the polis is "Unlocated".
If the polis is not referenced in Barr. at all, then the Inventory states the polis is "Unlocated. Not in Barr." In this case all we can really reference with respect to location is the assigned region (and the Inventory possesses a chapter on poleis that are also unassigned to a region, which we treat as region "unlocated" and position it in the open Aegean).

How the Polis-Inventory app plots all "unlocated" poleis is described in detail in the article The Lost Poleis.

Appendix 2: Polis Location in the S-Inventory

The S-Inventory website states the following:

"Source of Latitude / Longitude information

H = Hansen & Neisen, N= Pleiades Names, P = Pleiades Places, V = Van Alfen’s Nomisma site (drawing from Pleiades Places), G= Google Maps/other believed-reliable internet resources."

The S-Inventory presents more data than just the poleis contained in the Inventory and so the above description is broader than needed, but it is included as written for completeness. When considering just the 1035 poleis referenced in the S-Inventory, the Latitude / Longitude source is always listed "H".

As mentioned, I have not spoken directly to the Stanford University team, but it's clear they did more than just use "H" as described above. They attempted to improve on the Inventory locations, where possible, and then referred to Google Earth or similar to determine an elevation. At least this is the current assumption.

Their method for dealing with unlocated poleis was undoubtably different than that followed in the Polis-Inventory app and this could be a source of some of the discrepencies enumerated above regarding unlocated poleis.