How did the Romans build straight roads? If they were building a road a couple of hundred miles long, say from London to Chester, how did they know which direction to start?
They did roads as we did back in the 80’s before you had full maps of the world.
You want your road to minimize cost of transportation, so you wanted for the road to:
- have as small a distance as possible
- be as level as possible and
- cost as little as possible.
That meant you wanted a straight road, a level road and a road that crossed few rivers and creeks.
With that in mind you actually walked (or rode in a horse) the freaking 100 miles firstly, taking note of the forced points, or points you had to use because they had no alternatives.
This meant that you used your reason to look for major river crossings, mountain passes and even major creeks and swamps.
Thus, you divided the road in smaller segments, for example, to reach the next creek or river or the next hostal.
When you could not observe the river or creek, you used literally a tower that you had to build or you selected a peak or tree in the general direction and marked a constant slope line.
Real life modern surveying tower built to survey lines across Lake Superior in 1902. They were dismountable and were carried along the lake to create survey lines across tens of miles. Romans could use similar towers when needed in extremely flat terrain, like in Belgium
Notice that the previous tower is two towers in one, an inner tower to support the instruments and an external tower for surveyors so your movements would not affect the instruments.
Using towers or distant peaks as references you could trace extremely long straight lines by hammering stakes into the soil marking a center line for the road or aqueduct.
To align the stakes you used a groma, which is the grandfather of theodolites and total stations we use today, with no telescope but with plumbs to mark a straight line, like this one.
You observed the strings, like when you use the sights of a gun.
You worked with at least two (or three) guys helping you.
One of those helpers held the start of a chain (grasping it by the brass handle in the following image) made of metal links right where the base of the groma was and the other guy, in front, at the end of the chain, hammered the stake at 10 to 50 meter intervals, depending on the length of your chain.
This is an ancient surveying chain that is around 200 years old, but in the end it uses a very slightly modified version of Roman measures.
Ten links measure a rod (around 5 meters) marked with the small bronze label with one spike, two spikes (10 meters), three spikes (15 meters) and the full chain (20 meters, give or take).
Eighty full chains, thus, measured a bit over 1.600 meters, equal to an ancient Roman mile, which is one thousand double steps of 1.6 meters each.
Mille means one thousand (steps) in Latin, hence the name.
The guy in front stretched the chain and then hammered into the ground the stake where you said to him aligning him with the groma (“right, right, right, a bit to the left, there!!”), using the strings as a visual.
The third helper use a machete or similar to cut vegetation ahead of the team and make a path for the other two “chain men” which is the name we still use for the people that help a surveyor.
Tracing that preliminary line, you did an also preliminary survey, producing a map of a strip of ground around the intended line you annotated in your first walk.
In that preliminary survey, when you are in a completely level plain and there is no water (a true rarity), you can trace a line of stakes in the general direction you already knew you had to follow because you walked or ride a horse first, using a peak or group of trees or a tall tower as reference to trace long straight lines.
However, to build a road in perfectly level ground in continuous straight lines is impossible, not only because that kind of ground does not exist as water runs to some place and slowly but surely creates a ground that slopes towards some draining place, but is also not desirable.
You want the road itself and its lateral drain ditches to drain towards some place, so we do a lot of work to give the road at least a 1 percent, hopefully 2 percent slope, so water can run off the road and not make puddles that in a short time will make the road impassable or freeze in winter, with the end result that such level, straight road will be destroyed by water and vehicles in a few winters.
So, the line of stakes you trace on the ground tries to follow a constant slope, and the axis or center of road you trace starts to meander when you do such thing.
For that, you use a constant slope level, like an Abney spirit level.
Romans had similar instruments that you carry on your pocket.
This particular Abney level is shown at zero percent (top image) and at 42 percent of slope (bottom image)
Now, you must have notice that the groma had four strings.
Two of the strings were to create a line of stakes that marks the preliminary axis of the road, at a constant slope when going up or down.
The other two strings allow you to trace perpendiculars at each stake and measure the transversal slope.
While one person keeps the instrument aligned with the stakes that show the preliminary line, another person can look at perpendiculars and annotate how high is the terrain with respect to the center stake.
With that operation, you end with a center line, with angles measured every time you change your straight line, with stakes measured every 50 meters or 10 rods and knowing how steep is the transversal slope along your survey line.
This strip that you measured along the constant slope axis had to be wider in places where you thought you needed deviations from the preliminary line.
You could do this strip narrower where you thought that the path was more restricted or where the transversal slope was minimal.
Anyone who has dug a ditch with his or her own hands knows you want to excavate as little as possible.
This means that, once you have measured and staked the center of your preliminary line, you have to design cuts and fills so they are balanced, because you don’t want to carry a lot of dirt around.
You try to use the dirt from the cuts to be compacted and used in fills.
For this you need to show in the field to your crew precisely how much you want to cut or fill along the sides of the road, which you do using what we call “slope stakes”, to mark the chamfer at each point.
Knowing the height of ground at the center line of the road (marked CL) you calculate the distance “d” at which you have to locate the slope stake (black circles) for a given lateral slope of the embankment marked as 1 vertical and S level and a depth of embankment or a depth of cut “h”
Of course your slaves (or your workers or your motor grader operator) expect from you to be able to compensate the material excavated (marked with a red circle) with the material used to fill (marked with black circles), so you become an artist of taking the middle ground, literally, so if you have to create an embankment in a 200 meter stretch, there is a cut where the fill material will be excavated in the next 200 meter stretch and not five kilometers away wich forces your crew to move literally thousands of tons of dirt for kilometers and they will look weirdly at you.
For example, in the next picture, you have a real slope stake, similar to the one marked with a red circle in the previous image.
In this particular stake you have to dig 10 feet down (it says -10 at the top, almost hidden behind the cloth) to reach the road level.
The stake must stay and you have to dig to the right of the stake only.
You have to give the cut you are making a slope of 3 meters down (vertically) for each 1 meter you dig towards the left (horizontally) to the top of embankment (at the bottom of the stake it says 3:1 TOE).
So, you need not only to walk the road but to design on a map, establishing a center line for the road or aqueduct, and the location along the axis and lateral location of every slope stake.
Besides that you had to design every bridge, drain, culvert and gutter you found the road will need.
Here you have an example of a small segment I picked from the Map of the Hudson River Rail Road from New York to Albany.
If you enlarge the image, you can observe to the left of the town a survey preliminary line marked as “not adopted” around Poughkeepsie
A survey line includes all I said: a staked centerline, that at each turn of the road has circular curves (I did not explain how to trace and space the stakes on the curves, you can check any book on road design), with measured transversal slopes along the line and with slope stakes to mark how much to dig and fill and with a design for every culvert, drainage and bridge.
Once you did all this job you could plan how much resources in money, people, machinery and materials you needed, so the owner of the road could know how much he had to pay, how much people were needed and where they will work, be in a quarry or on the road itself along the 100 miles you mention.
Of course, nobody did 100 miles in one stretch, by the way, as you can logically conclude from the effort I have just described, so usually you divided the problem in smaller parts.
So, to answer your question, there is a lot of work involved in several stages of refinement and you do not start just saying to yourself: “this must be the direction to Chester, so let’s make the road toward this direction because intuitively I think this is it”.
A road builder can see in his mind the road well before it is built, because he has walked every inch of the terrain, not only from London to Chester but has seen every river and rivulet, probed every soil along the line and tested several alternatives first on a map.
In ancient times you could use, besides the groma, a drawing table on top of the groma, where you drawn on the field, as you were tracing the lines of the road, the road itself and the curves it made, by drawing lines on paper on top of the table that represented the lines of the road at a scale, so after the field work you ended with a drawn representation of the road and the lateral slopes that allowed you to design the slope stakes.
Here you can see some Renaissance instruments very similar to those used by Romans before the invention of telescope.
In the blue circle you see how you use a level to measure the lateral slopes of the terrain before the road is made, to be able to design the lateral slope stakes
In the green circle you see a drawing table to the right and the result plot traced on the table by aligning the instrument on top of the paper with the stake lines
In the yellow circle you see how you could use a measured base CD and the drawing table to measure at scale the distances to the building across the river without chaining
In the red circle you see a “perambulator” or wheel that measures distances by measuring rotations of the wheel
Today is easier with electronic maps but nothing substitutes the work of a farmer going out to his field to see where to plant a crop and the farmer knows his field well before planting.
In the same way nothing substitutes the work of a road engineer walking the road to check every detail, tree, and trace of old floods well before anyone can even see the start of the works.