Road certainty says how sure the atlas is about where a Roman road ran. Itiner-e sorts every segment into certain, conjectured or hypothetical, and on our map the three classes are drawn as a solid, a dashed and a dotted line.
Key takeaways
- Certain means well documented and traced to within about 50 m in mountains or 200 m on plains, and only 2.7% of our road length is in that class.
- Conjectured is the big group at 89.8%: the road is known, but the line is drawn with less precision.
- Hypothetical covers 7.4%: routes that are not pinned down, sit on shifting ground, or are only suspected.
- Certainty is about where the line is, not whether a road existed.
How road certainty works
Every piece of road in the Itiner-e atlas carries a certainty label. It is recorded for each segment, and it describes "spatial accuracy and confidence in the digitised location" of the line [1]. In plain words: how far could the real road be from the line you see?
There are three classes, and our map draws them differently so you can tell them apart at a glance.
- Certain (solid line). The segment is well documented in the sources and was digitised to within less than 50 m in mountainous terrain and less than 200 m on plains [1].
- Conjectured (dashed line). A road we know about, drawn with lower accuracy because the sources say less about its exact course [1].
- Hypothetical (dotted line). Used for roads that are identified but not located, for roads on ground where the track was less fixed or several parallel tracks may have existed (the paper gives deserts and flood plains as examples), and for roads speculated to have existed without enough evidence to call them certain or conjectured [1].
The label is a judgement made by the atlas team from the evidence behind each line. The paper says conjectured and hypothetical lines are informed by Roman remains and earlier research, then fitted to the geography of the area [1].
For the short version of this on the map, see how to read the map, and for how far to trust the whole dataset see where the data comes from.
An example from our data
Across the whole network, 8,181 km is certain (2.7%), 268,710 km is conjectured (89.8%) and 22,281 km is hypothetical (7.4%).
Britain shows how lopsided this is. The United Kingdom page records 0% certain and 98.8% conjectured. The Fosse Way is 99.1% conjectured and 1% hypothetical. That does not mean nobody knows where it runs. It means the line on our map is not tied to a surveyed Roman surface everywhere.
Italy does better. The Via Appia has 93 km of certain road, 16.1% of its length, and the rest is conjectured.
The thing people get wrong
Many people read "conjectured" or "hypothetical" as "probably not real". That is not what the labels mean. The Itiner-e authors state that all the included roads were used at some point in the Roman period, but that their precise locations are not certain [1]. So the question the label answers is "how exactly do we know the line?", never "was there a road?".
The map also shows roads as they stood around AD 150, and a conjectured line can be tens of metres or more off the true course. Use it to find the general route, and check the ground, local records or the Itiner-e atlas before you plan a walk along it.
Frequently asked questions
Does a conjectured road mean the road is made up?
No. Itiner-e says every road in the dataset was used at some point in Roman times. Conjectured describes how precisely the line is known, not whether the road existed.
Why is so little of the network certain?
Because long stretches of Roman road have rarely been excavated or survived to the present day, so most lines rest on older maps, place evidence and modelling. Only 2.7% of the length is certain.
Which Roman roads in Britain are certain on the map?
In this data, almost none: the United Kingdom total is 0% certain.
Sources
- Itiner-e: A high-resolution dataset of roads of the Roman Empire, Scientific Data (Brughmans et al., 2025).
- Itiner-e: the digital atlas of ancient roads, Itiner-e.
- Itiner-e dataset (version 2024 1.3), Zenodo.
- Every figure from our own map is worked out from the datasets on the data sources page (checked against Itiner-e version 2024 1.3).