Maps and archives

Terrain on an old map: relief, water, roads and crossings

How to read contours, gullies, marshes and old channels on an old map, tell a road from a forest ride or a stream, read bridges and fords, and test a hunch.

32 min read · 18 sections · free

An old map holds more than names and roads. Its most durable layer is the shape of the ground: heights, valleys, catchments, ridges, gullies, banks and low ground. A settlement can vanish, a road can be re-routed, a stream straightened, a marsh drained — but surface form changes slowest of all, and that is why it explains best why something stood where it stood.

A topographic map renders three-dimensional surface shape on a flat sheet using contour lines; the USGS describes the principle this way: alongside natural features the sheet carries roads, boundaries, major structures and other human-made elements. On a historical sheet all of that gives you a landscape hypothesis, not a finished answer about present-day access, passability, preservation or legal status.

This article reads the terrain as a whole. First the shape of the surface — contours, valleys, ridges and gullies. Then the water: rivers, ditches, marshes, dams and old channels. And finally the lines along which people moved across that ground — roads, forest rides, boundaries and crossings. The symbols and labels themselves are covered in How to read an old map; this is about how they add up to a landscape, and where the claims a sheet can make actually stop.

The governing rule: the map forms the question, it does not confirm the fact. A sheet states exactly one thing — at the moment of survey the landscape was described this way. What survived, what changed and why is checked against another sheet, an overlay, a document and observation on the ground.

Why landform beats a single point of interest

A lone symbol is misread more often than people expect: it may have been generalised in a later edition, shifted in printing, or simply confused with a similar sign. Relief and hydrography set a wider context — where the watershed runs, where water naturally drains, which valley a road could have followed, why a bridge stood at one point and not three hundred metres downstream. That does not make relief proof, but it lets you test a hypothesis against several independent signs at once.

Question to the map	What relief and water help you check
Why does the road take a detour?	It may skirt a steep slope, a marsh, a gully or a floodplain, or pick an easier saddle
Why is this settlement or farmstead here?	Access to water, a dry terrace, a workable slope, a crossing, protection from flooding, or a link to the road
Why does a line look like a road, a stream or a boundary?	Slope and catchment logic help separate a watercourse from a linear feature, but they do not replace the legend
Where should the confirming document be?	The landscape type suggests the question type: bridge, dam, drainage, land improvement, ferry, road plan, canal project
Why do the old and modern layers disagree?	The channel may have moved, the marsh may have been drained, the banks reworked — or the map simply generalised the line

The symbols and labels themselves are covered in How to read an old map. Relief does not replace the legend: it adds a third independent sign next to the legend and the labels.

Contours: the base language of surface shape

A contour line joins points of equal elevation. It is not a path, not a boundary and not a physical object on the ground — its job is to convey the shape of the surface. The closer contours sit to each other, the faster elevation changes over a short distance; the further apart, the gentler the ground. The number beside a line usually gives absolute height, but the units, the contour interval and the labelling rules have to be checked on that specific sheet rather than assumed from habit.

What you see on the map	What it usually says about the ground	What must not be concluded automatically
Contours packed close together	The slope is steeper than where contours are sparse	That the ground is impassable, or that no road ran across the slope
Contours widely spaced	Relatively gentle ground	That the ground is dry, safe or free of restrictions
Closed loops with elevations rising towards the centre	A rise or hill	That it is a barrow, an embankment or a man-made feature
Closed loops with elevations falling, or carrying hachures or a label	A depression or hollow — depending on the mapping tradition	That it is a pond, a quarry or an archaeological feature: without the legend and other sources no such conclusion is drawn
Lines forming a pronounced V shape	Often a valley or swale; the apex of the V points upslope, towards the head of the catchment	That a permanent stream necessarily runs there

Diagram of four landforms shown by contour lines: valley, ridge, saddle and depression — Discovera schematic

The core rule: one landform, several checks

Contours help you pose a question; on their own they do not answer it. Look at spot heights, neighbouring lines, hydrography, labels, the scale and the date of the sheet. Some old series show relief with contours, others with hachures, hill shading or a mixture of methods. The rules of a modern sheet cannot be carried over to a historical tradition without reading its legend: the same landform is drawn differently in two editions more often than you would like.

Valley, ridge, saddle and gully

Valley and swale

A valley is a low line between slopes that normally collects surface runoff. On the map it reads as a run of contour bends turned towards the head of the catchment. Inside it there may be a stream, a dry swale, a ditch, a road — or nothing marked at all: it depends on the scale and on when the survey was made.

Ridge and watershed

A ridge separates drainage directions. It often explains why a road follows the dry high line rather than the bottom land. But a historical road could equally follow the valley — for water, for settlements, or for an easier gradient. Compare the surface form with the other features on the sheet and do not turn a general principle into a strict law.

Saddle

A saddle is a low point on a ridge between two higher sections. On routes it sometimes explains where a crossing over high ground was chosen. On an old map, though, a saddle is a good question to test, not evidence of an old highway: you still need the legend, a link to the road network, the neighbouring sheets and sources from another period.

Gully and washout

Gullies, dry valleys, swales and washouts are tied to catchment and gradient. How they are drawn depends on the series: a line, hachures, contours, a label, or a combination. The point is to tell a natural landform from an artificial cutting, a ditch or a boundary — and for that you use the scale, the direction of slope, the bridges and roads nearby, and later editions of the map.

Landform	Question to the map	How to test it
Valley	Where is the head of the catchment and where is the bottom land?	Spot heights, stream direction, modern elevation models, the adjacent sheet
Ridge	Does it separate catchments, and does it steer roads and boundaries?	Contours on both flanks, labels, roads and forest compartments
Saddle	Is there a connected route through it, or is it just a landform?	The road network, the scale, an edition from another year, an archival route plan
Gully	Is this a natural form, a ditch, a cutting or an editorial convention?	The legend, hydrography, aerial imagery, different years of the map

Water on an old sheet: river, stream, canal, ditch, marsh and old channel

Water obeys relief, but a historical map shows more than a natural process. Canals, ditches, dams, ponds, drainage lines, bridges and fords make up an artificial hydrography. A blue or winding line therefore does not always mean the same thing.

Feature	What the map can tell you	What has to be checked separately
River or stream	Flow direction, the link to the valley, bridges, tributaries, marshes	Whether flow is permanent, the modern channel, straightening and bank changes
Canal or ditch	An artificial water line or drainage, sometimes tied to an economic feature	When it was built, what it was for, its present status and safety
Pond and dam	Water tied to a farm, a road, a mill or flow regulation	Whether the feature survives, whether water level changed, whether a regime or restriction applies
Marsh, boggy low ground	Ground and water conditions on the date of the sheet, and any drainage system	Drainage works, seasonality, present-day passability and protected status
Oxbow, old channel	Traces of earlier water movement, or a cartographic depiction of older hydrography	The exact date of the change and how it relates to modern relief
Ford, crossing, bridge	An organised water crossing at the time of survey	Whether it exists, whether it is accessible, and whether it is safe today

A historical sheet is most useful when you compare not one water feature but a sequence: river, floodplain, road, bridge, settlement, mill with its pond. Such a chain suggests which documents to look for next: a bridge design, land-improvement files, a plan for a hydraulic structure, mill accounts, correspondence about flooding.

How to read the lines that run beside water and cross it is covered below, in the sections on roads and crossings.

How the landscape changes over time

Even durable relief does not stay put. Water moves channels and banks; people build dams, straighten rivers, drain marshes, cut ditches, remove spoil, raise embankments and dig cuttings. On top of that, cartographers generalise the same landform differently at different scales — so part of the disagreement between two sheets comes from the editor, not from the ground.

Change	How it can look when maps are compared	The mistake to avoid
Channel straightening	Meanders on the old sheet, a straight cut on the new one	Treating every mismatch as an overlay error
Marsh drainage	A marsh on the old map, fields and a ditch network on the new one	Concluding that the old conditions still hold today
A dam built or removed	A pond appears, changes shape or disappears	Treating an old dam symbol as working infrastructure
Road embankment or cutting	The route changes how it relates to the contours	Confusing an engineered form with a natural ridge or gully
Extraction and quarrying	A pit appears; contours and roads shift	Reading any depression as a historical natural landform
Generalisation by scale	Detail disappears and the line is simplified on the smaller-scale sheet	Comparing the axis and minor bends across scales literally

Historical topographic collections are valuable for exactly this comparison: the USGS states plainly that the old series preserve a layer of observations from a particular period, rather than a permanent truth about the ground.

The practical consequence: a mismatch between the old and the new layer is not always a georeferencing error. Before you drag control points, separate overlay error from a real change on the ground; the method is in an old map over the satellite view.

Roads and other lines: why they mislead more often than points

On an old map a line is rarely just a line. The same thin stroke can be a post road, a field track, a footpath, an estate boundary, a forestry ride, a ditch, a stream, a railway, an earthwork bank, or a late editorial overprint. Getting it wrong is easy: the eye likes to continue a familiar shape and build a route where the surveyor drew something else entirely.

Linear symbols are never read from one feature alone. The meaning comes from the drawing of the line, its colour, its label, its junctions, the relief, the scale, the date and the legend of that particular series. The USGS sets out the basic principle: symbols show features as points, lines or areas depending on their size and extent. But exactly how a route, a watercourse or a boundary looks depends on the cartographic school and the period.

A map records how a feature was described at the moment of survey. It does not prove that the line survives on the ground, that it is passable, that access is open, or that it means the same thing today.

Post road, track, forest ride and crossing — four different lines on an old sheet — Discovera schematic

A point symbol usually stands for one identifiable thing — a church, a mill, a station, a single building. A line runs for kilometres and changes character along the way. In one stretch it is thin, in another doubled, in a third it carries a label or hachures. You cannot settle its meaning from its shape on a phone screen.

What changes the meaning of a line	The question to ask
Colour and weight	Is this water, a road, a boundary, a relief stroke, or an editorial correction?
Continuity	Does the line really carry on, does it die at the sheet edge, break in woodland, or change category?
Junctions	What happens where it meets a river, a railway, a boundary, another road?
Labelling	Is there a name, an abbreviation, a road number, a feature designation, or an owner or parcel note?
Relief	Does the line follow the logic of a descent, a bridge, a saddle, an embankment or a channel?
Scale and date	Can this map even separate road, path and forestry ride; was the feature already obsolete at publication?

The most useful habit here is to stop asking “what is this line?” and start asking: “which three features confirm that this is a road?” If there is no confirmation, what belongs in the journal is not a confident reading but a hypothesis with a stated confidence level.

Seven kinds of linear feature

Names and symbols differ between series, but it helps to sort a line into a broad family first and only then go to the legend of the specific sheet.

Family	What it can include	What to look at
Routes	Post road, metalled road, unmetalled track, country lane, footpath, street, railway	Width, doubled outline, labelling, bridges, junctions, links to settlements
Water	River, stream, ditch, canal, seasonal watercourse	Colour, direction toward a water body, sinuosity, labels, marsh and bridges
Boundaries	Administrative, estate, forestry, parish, state	Tick marks, symbols set along the line, labels, links to compartments and boundary markers
Forestry and estate lines	Ride, compartment edge, drain, baulk, fence, planting line	Rhythm, straightness, grid pattern, ties to woodland or fields
Engineered features	Embankment, cutting, dam, telegraph line, pipeline, power line on later maps	Tick symbols, paired outlines, labels and the character of junctions
Relief forms	Gully, ridge, scarp, bank, terrace edge	Hachures, contours, shading, spot heights, and the absence of transport links
Editorial elements	Grid, sheet boundaries, overprints, later corrections, georeferencing lines	Neat line, print colour, typeface, and no relationship to the ground

The family is not an answer; it narrows the search. The answer comes from the legend of the edition: the National Library of Scotland keeps Ordnance Survey characteristic sheets separately by scale and period precisely because the graphic language of a map only makes sense inside its own series. Labels, abbreviations and lettering are a layer of their own — see How to read an old map.

Observation — hypothesis — check

A route rarely exists on its own. It joins things, changes character at a settlement, crosses water at a structure, and picks the easier landform. A watercourse, by contrast, obeys the relief; a boundary tends to follow administrative or tenurial logic; a forestry ride usually sits inside a compartment grid.

Observation on the sheet	Hypothesis	How to check it
The line runs to a settlement, bridge, station or crossing and continues beyond it	A route	Legend, adjacent sheet, a larger-scale edition, modern orthophotos
The line winds along low ground, gathers tributaries and enters a water body	A watercourse	Colour, hydrographic symbols, direction of fall, labels
The line is almost straight and forms a regular grid inside woodland	A forestry ride or compartment edge	Woodland context, parallel lines, compartment numbering
The line follows field edges or an administrative outline, often dashed	A boundary or baulk	Legend, labels, cadastral or administrative layer
The line carries distinctive junction symbols and a made formation	A railway or engineered feature	Legend, stations, level crossings, embankments and cuttings
The line breaks on slopes and sits next to relief hachures	A gully, scarp or other relief form	Contours, spot heights, the shape of the catchment

No single feature works in isolation. A straight line can be a road or a ride; a winding one can be a path or a stream. Confidence appears only once the hypothesis has been tested against an independent source.

The most accessible independent source for linear features is imagery from above. Wartime aerial photography of 1939–1945 from the NARA holdings is collected on the Discovera aerial imagery page: it shows the ground before post-war field amalgamation and drainage schemes, and often answers whether a feature was physically on the ground at all, rather than only on paper.

Crossings: bridge, ford, dam and ferry

A water crossing is one of the most informative subjects on a map and one of the easiest to romanticise. A bridge, a ford, a dam and a weir are different kinds of infrastructure and may differ in age, accessibility and status. Some series give them their own symbols; others show only a break in the road, or a label.

Feature	What it tells you about the map	What it does not tell you
Bridge	At survey time an organised crossing existed; route and water were joined by a built structure	That the bridge survives, is accessible, or stands in the same spot today
Ford	The surveyor marked a possible crossing of the watercourse without a permanent bridge	That the crossing is safe or legally permissible now
Dam or weir	Water and working infrastructure were linked; channel and level may since have changed	That the structure is sound or bears weight; such places must never be treated as a route
Ferry crossing	A link between the banks was organised, sometimes seasonally	That it ran continuously or left a visible trace in the modern landscape

Reading a crossing correctly depends on its surroundings: the road up to it, the continuation on the far bank, the label, the nearest settlement, a mill, a railway, a change in relief. But historical information is not an invitation to visit or survey the site. Banks, bridges and water structures change condition, and access to them is regulated separately.

Water infrastructure is not a route. Dams, weirs, sluices and their approaches change condition, are usually in someone's care, and can be dangerous regardless of what a 19th-century sheet shows.

Relief as an independent check on a line

Relief rarely draws a road for you, but it tests a hypothesis well. A route usually seeks a sensible way across a slope, an easier way round an obstacle, or a convenient landform. A watercourse, on the contrary, obeys the direction of fall. This is not a mechanical rule: historical roads sometimes ran awkwardly by modern standards, and the surveyor may have generalised the line. But relief is a useful second, independent voice.

If the line…	Ask yourself
Crosses the contours	Is this a climb or a descent, a road on a slope, or a feature unrelated to movement?
Runs along a valley	Could this be water, a valley road, a boundary, or several features side by side?
Approaches a sharp scarp	Is there a bridge, a zigzag, a cutting, a label or any other sign of an engineering solution?
Barely responds to relief and holds a straight course	Might it be a forestry ride, a grid line or a boundary?

On old maps the relief may be shown not by modern contours but by hachures, wash or shading. So establish the cartographic tradition and the scale first, and only then draw a conclusion about the line. How to read contours and landforms is covered at the start of this article.

The routine: from observation to a tested hypothesis

Instead of "found an old road" or "there was a mill here", it pays to run a short check. It is the same for relief, water and lines, and it turns a pretty picture on the screen into a transparent working hypothesis.

Step	Question	What gets recorded
1. Map passport	Which series, scale, year, sheet number and scan quality?	The conditions under which interpretation is admissible at all
2. Legend	How exactly does this series show relief, water, marsh, roads, ditches and crossings?	Confirmed and assumed meanings of the symbols
3. Form and connections	What do the contours and spot heights support? Which two or three checkable features does the line lead to — a settlement, a bridge, a station?	A description of the observation, without a conclusion
4. Logic of water and movement	A natural catchment or an engineered feature? A route, a watercourse, a forest ride or a boundary?	A hypothesis with its arguments
5. Another sheet, another date	Does the form and the feature repeat on a neighbouring or different edition?	Independent confirmation or a discrepancy
6. Overlay	Where does the old sheet match the modern layer, and where not?	A zone of high, medium or low confidence
7. Archival question	Which document could explain the feature or the change?	A clear request topic: drainage plans, a bridge file, a road project
8. Legal status	What does the law say about this land — separately from the map?	The basis for any lawful visit

Order matters more than speed here. Skip the map passport and the legend, and every later step tests not the ground but your guess about what the surveyor meant.

Why the routine matters is clear from international practice. The USGS explicitly warns that symbols changed over time and that no single document covers every series and period. The same holds for the Shubert three-verst map, Red Army (RKKA) sheets, the German Messtischblatt and the Polish WIG series: each has its own vocabulary.

Step 6 is the bottleneck. An overlay of an old sheet on a modern layer is easy to nudge toward the answer you want, so check the georeferencing by the rules in an old map over the satellite view; the general context on verst scales and 19th-century sheets is in How to read an old map. Soviet series have their own quirks, covered in RKKA and General Staff maps.

Five questions to ask yourself

Before you write down “old road”, answer five questions. They take three minutes and remove most of the false confidence.

Which series am I reading? Year, scale, edition, sheet number — and have I seen the legend of that edition, not a similar one.
Which three independent features support the category? And do they really reduce to a single observation retold three ways.
Where does the line go and how does it end? Does it join checkable features, or simply die at the sheet edge.
What do relief, hydrography and the adjacent sheet say? Do they confirm, stay silent, or disagree.
Have I separated “does it survive” from “may I go there”? These are two different questions, and the second is not answered by a map.

If the answer to even one of them is “I don't know”, the confidence level is low. That is a normal working result, not a defeat: an honestly recorded low confidence costs far less than a high confidence built on a single stroke.

From the map to an archival question

A good archival question does not repeat the picture. It tests the hypothesis the map allowed you to state, and it names a place, a feature and a period — otherwise nobody can act on it.

Observation on the map	Weak question	Strong archival question
A pond and a dam beside a village	What was here?	Do plans, accounts or correspondence survive concerning the dam and mill at village N for 1880–1917?
A marsh that later became a field	Why did the marsh disappear?	Are there records of drainage, land improvement or ditch-cutting in area N for the stated period?
A road crossing a gully and water	Was there an old bridge?	Are there plans, estimates, accounts or correspondence about the bridge or crossing on road N for the stated period?
A different channel on the later sheet	Where did the river go?	Is there a straightening project, or records of channel regulation, floods or hydraulic works on river N?

The map creates the hypothesis; the archival document confirms, refines or refutes it. Do not bend the answer to fit an attractive landscape story: a negative result from a fonds is still a result, and it saves the next trip.

Where to address such a question, how it reads in correspondence, and how the whole route from an observation on the sheet to an institution's reply plays out on a real case, is in working with archives in Russia and Belarus.

A terrain log in Discovera

An observation that is not written down turns, a month later, into "I think there was a gully there" or "probably an old road". The terrain log is seven fields that keep the hypothesis separate from the fact.

Field	What to record
Map	Series, sheet, year, scale, link to the source
Observation	Contours form a valley; a thin watercourse line runs beside it; a bridge is marked at the crossing
Hypothesis	The route may have followed the valley and crossed the stream at an organised ford
Check	Two sheets compared; on the later one the channel is straightened; legend and archival plan still to be checked
Archival trail	Institution, fonds or catalogue, text of the request, reply, reference code
Confidence	High, medium or low — with the reason it is that and not higher
Legal status	Checked separately; a historical landscape is not permission to access or to work

An entry like that is more useful than the label "old road". It preserves your reasoning, lets you correct a mistake, and does not manufacture false certainty for your own benefit six months later.

The same separation of observation from conclusion works in the field: signal, context and interpretation are recorded apart — see from signal to context.

What the map does not prove, and does not give you

Most mistakes with linear features are not recognition errors. They are substitutions: the symbol was read correctly, and then a conclusion was drawn that the symbol never contained. It helps to keep a short list of what a map specifically does not prove.

It looks as if the map proved this	What it actually recorded
There was a road here, so there is one	That at the date of survey the compiler chose to show a linear feature of that category
There is a bridge, so there is a crossing	That an organised crossing existed at survey time, not necessarily in the same spot or the same form
The line runs to a village, so the village is inhabited	That the settlement existed in the form the surveyor recorded and generalised
A road is shown, so the place is open	Nothing at all about right of way, ownership, protected status or land designation
Two maps show the same thing, so it is confirmed	Possibly the second sheet was compiled from the first rather than surveyed afresh
The line looks old, so it is historical	Possibly it is an editorial element: a neat line, a grid, a join, an overprint, a georeferencing trace

Hence a simple rule: a map is evidence about a state of affairs on a date. It answers neither “what is there today” nor “what may be done there”. The second question is settled separately, and before you travel.

An old sheet tells you nothing about the landowner, nothing about protected status, and nothing about what is in the ground today. It permits nothing at all: the legal status of a parcel is checked separately and before you travel, not deduced from relief, a convenient saddle or a label a century old.

The map is not a permission. The legal status of the ground and the admissibility of any work are checked separately. If signs of an occupation layer, a burial or ordnance appear on site, work stops there and then, and the local notification procedure takes over from your own curiosity.

What the law says, and how to act when something turns up, is in metal detecting and the law.

Where this article sits in "Maps and archives"

The section is built as a route from the sheet to the document:

How to read an old map — symbols, labels and abbreviations: the alphabet of the sheet.
Terrain on an old map — this article: relief, water, roads and crossings.
Red Army and General Staff maps — another period and another cartographic tradition.
An old map over the satellite view — where the layers come from, how to georeference them, and how to check the overlay is not lying.
Working with archives in Russia and Belarus — where to send the question, and a case study from map to archive reply.

FAQ

Can contour lines cross each other?

In normal cartographic logic, no: a single point on the surface cannot have two different elevations at once. So a crossing on an old scan almost always means something else — damage to the sheet, an overprint, another symbol laid over the contour, a join in the scan, or a misreading. Rule those out first instead of rewriting the geometry of the ground.

Does a V shape in the contours always mean a stream?

No. The shape helps you spot a valley or a swale, but there may be no permanent watercourse in it either today or on the date of survey. Check the hydrographic symbols, the labels, the scale and the neighbouring sheets before you draw a stream on your own sketch.

Is an old marsh on the map still a marsh today?

No. It may have been drained, built over or flooded by a reservoir, or it may have changed naturally. The old symbol is a historical observation valid for the date of the sheet, not a description of the parcel today; present-day passability and protected status are checked separately.

I see a suspicious mound in the relief. What should I do?

Do not call it a feature. Record the observation, compare sheets from different years, read the legend, look at modern sources, and if it is warranted, write an archival question. Widening any intervention on the ground to get an answer is not an option: a mound can be of road, drainage or protected origin.

How accurately is relief shown on an old sheet at all?

Accuracy depends on the series, the scale, the survey method and the quality of the scan. At small scales the form is generalised, the contour interval is coarse, and minor swales vanish entirely. That is why the map passport — series, year, scale, condition of the sheet — is written down before any conclusion about the landscape.

Is every straight line in woodland an old road?

No. In woodland, straight regular lines are more often forestry rides, compartment edges or editorial elements of the sheet. Check the link to the road network, the legend of the edition, the adjacent sheets and modern sources.

If a road is visible on an old map, does it exist on the ground?

No. The road may have gone, shifted, been buried under modern infrastructure, or survive only as microrelief. A map records a state of affairs at the date of survey or publication, not today.

How do I tell a ford from a bridge?

Only from the legend of the series, the label and the cartographic context. A break in the road at a blue line is not enough on its own: it can mean a ford, a bridge below the generalisation threshold, or simply a simplified line.

Does a doubled line always mean a highway or post road?

No. Different series use a doubled outline for an improved road, an avenue, a canal, an embankment with a cutting, and a right-of-way strip. Pairing and weight indicate a category inside one legend; they are not a universal code.

The historical road looks interesting. Can I plan a search along it?

Only after checking the status of the land and the conditions of access separately. Historical interest in a route does not override protected designations, ownership or the law. Where the context looks complicated, intervention is stopped rather than widened.