# Primary roadside and rupestral evidence expansion

17 October 2026. Current scope: 42 physical-object dossiers, 51 inscription phases, 84 object-history events, six alternative inscription assignments and 122 typed location assertions, 69 mapped. This supersedes the 39-case hierarchy release for current counts. It preserves the existing private audience and source geometry.

## Result and admission logic

T2a expands from zero to three reviewed primary roadside settings. T1 remains one: Donnas was reassessed and qualified, while no additional strict rupestral milestone passed the inspected evidence. T2b remains the single qualified Casa de Campo association. T3 7, T4 17, T5 9, T6 3 and X 1 are unchanged. Unreviewed bulk catalogue points remain U.

| Admission | Original setting | Chronology | Location and source limits |
|---|---|---|---|
| Smith’s Shield, RIB 2308 = CIL VII 1187 | Surviving base explicitly reported still beside the Stanegate; upper shaft later split and moved | Primary text unresolved; secondary formula assigned to the fourth century in RIB; Hadrianic identification remains conjectural | No object survey or excavated socket claimed. CIL concordance checked through RIB; original CIL page not directly collated. Catalogue point remains T6. |
| Taucheira–Ptolemais, FS12 | Field report identifies an in-situ socket, adjacent fallen inscribed shaft and physical coastal road | Undated inscription; survey in 2009 does not date the stone | Previously unpublished in the inspected 2016 report; no invented CIL number or second object. Published map lacks a usable coordinate grid. Proposed museum transfer is not completed custody. |
| Čatež ob Savi, ILJug 1176 / ILSl 1, 198 | Investigator records the ancient seating recorded in a highway-cut exposure beside a 6.40m road; modern blasting fragmented the stone | Antoninus Pius, 139–141 CE from a qualified fragmentary tribunician numeral | Discovery-year reports conflict, 1958/1959; neither is silently preferred. Six fragments form one object. CIL entries cited by Petru are comparanda, not this object’s identifiers. |

The first admission uses the criteria’s equivalent documented primary-base setting rather than a newly claimed socket excavation. FS12 is a field survey, not a stratigraphic excavation. Čatež has a published road section and seating description. Each assessment has four separately sourced admission gates. These sources establish a local relation; source independence from individual Itiner-e segments and matching road-surface chronology remain unassessed.

All three exact original-position coordinates remain null. Scholarly certainty about a primary context does not produce a survey coordinate. Original stations remain directly readable in the atlas locator and inspector. Repository and inherited catalogue positions cannot inherit T2a. The new filter distinguishes T2a from T2b while retaining the T2 family.

## Publication checks and corrections

[RIB 2308, Support, Edition and Apparatus](https://romaninscriptionsofbritain.org/inscriptions/2308) supplies the base/shaft distinction and historical readings; [Historic England 1014820, Details](https://historicengland.org.uk/listing/the-list/list-entry/1014820) identifies the local Stanegate remains. Birley’s full 2014 research note was inspected; the 2017 article was accessible only as an abstract. Possible Hadrianic/Tungrian interpretation is not presented as a named workforce or established construction event.

[Weber-Karyotakis and Berger 2016, pp.24–26, figs.1–3 and note4](https://doi.org/10.19272/201610601002) directly describes FS12. Text and relevant figures were inspected in the university-hosted PDF. The abstract’s plural milestones does not justify inventing a second separately identified case. An earlier Tetrarchic milestone discussed nearby is a different object and does not date FS12.

[Petru 1960–1961, pp.27–29, figs.1–3](https://ojs.zrc-sazu.si/av/article/download/10041/9187/27506) supplies Čatež’s primary setting, fragments and qualified date. [Lovenjak 1997, p.89](https://ojs.zrc-sazu.si/av/article/download/8887/8021) explicitly reports 1958, while EDH and Lupa metadata report 1959. This is one disputed discovery event with source alternatives, not two discoveries or a continuous 1958–1959 event. Petru’s regional reconstruction interpretation is distinguished from the fragment’s lack of a complete roadwork verb.

Čatež catalogue ID **EDCS-10001181** uses `[14.9588038,45.9705365]`, conflicting with the published discovery west of Čatež near Brežice. It is the fifth quarantined catalogue coordinate. The raw point is preserved, withheld from display and distance comparisons, and not silently replaced by an estimated point. The museum location and historical original station remain separate.

Donnas remains a fixed local horizontal passage marker. Schneider identifies XXXVI but argues that the visible rutted surface need not be contemporary with it, inferring a higher original level and later reworking. This is a scholarly reconstruction, not a measured original pavement level. Nearby 1474 lettering does not date the Roman numeral. The source is entered in Object, Timeline and the claim ledger with unknown dates for the relative reworking sequence. [Schneider’s section in Schiedt–Schneider–Herzig 2007, pp.142–145, especially pp.143–144 and table2 p.145](https://journals.ub.uni-heidelberg.de/index.php/vuf/article/view/17795/11606). Full relevant text was inspected; remote rendering failed, so no visual inspection of that PDF is claimed.

## CIL priority and withheld candidates

Direct  CIL checks included **II Supplementum 6211**, printed p.993, and **III 346**, printed p.61. Rendered page images were inspected and added to the in-atlas candidate review. Paiva has disputed mileage-marker function and a large rock does not itself prove continuous bedrock. Sarikaya is a rock-cut roadworks text without demonstrated mileage function. Gospodjin Vir remains a road-repair inscription lead, without a newly inspected original CIL page or verified current emplacement. They are not added as T1 milestone cases.

The local EDCS screen and its hash, multilingual search scope and access limits are recorded in [rupestral expansion audit](primary-expansion-rupestral.md). No comprehensive global rupestral corpus is claimed. The official CIL/MuseoTorino links to V/2 failed through Arachne; the Donnas concordance remains explicitly publication-reported, without a false direct-collation claim.

Twelve candidates are retained in the separate admissions review rather than inflating the milestone count. These include Codley Gate’s contradictory relocation history, Volta di Rote’s insufficiently established physical road setting, Sangusín’s modern return, Rantis road spolia and Bet Neṭofa’s inaccessible full excavation report. Full source scopes and decisions: [western audit](primary-expansion-west.md), [eastern audit](primary-expansion-east.md), the `data/primary-review.json` download in Sources & method.

## Validation and release

73 semantic/source checks and 70 browser checks (50 regression and 20 primary-context checks) passed, with no runtime exceptions. Both JavaScript files passed syntax checks. Desktop and mobile renders, including the CIL control scans, were inspected. Final results are retained in `qa/data-audit.json`, `qa/browser-results.json` and `qa/primary-context-browser.json` in the portable package. Source PDFs used in new checks are hashed and retained locally outside the static Site and redistribution archive. Eleven public-domain CIL page renders are bundled; modern source PDFs and modern publication figures are linked, not bundled.

The publication receipt is `deployment.json`. Local checks do not establish field autopsy, present custody, independent archaeological certification or comprehensive CIL collation.