GEOMETRY AND ASTRONOMY

IN THE PROJECT OF ANCIENT STRAIGHT ROADS:

THE CASE OF REGINA VIARUM
MAGLI Giulio (I), REALINI Eugenio (I),
REGUZZONI Mirko (I), SAMPIETRO Daniele (I)

Abstract. Many ancient cultures planned and constructed long, straight roads,
sometimes taking an uttermost care of their “going straight”. In many cases, interesting
geometrical and astronomical methods were applied to gain an astonishing accuracy. As
it turns out, this “desire of going straight” was not always due to practical purposes but
rather to symbolical, cosmological motivations. Sometimes, the very existence of the
roads was connected mainly to such motivations. We discuss here some key examples
and present the recent results we have obtained about the project of the masterpiece of
the Roman roads, Via Appia.

Keywords: Ancient geometry, Roman astronomy, Roman roads

Mathematics Subject Classification: Primary 85A03; Secondary 01A20

1 Introduction

Straight roads are the fastest way to connect two points on a plane surface. Of course, since the
earth is round, no “straight lines” exist at all on it. However, provided that the two ends of a road
are not too far apart, the line which can be traced starting from one end and pointing to the final
destination does not differ, in practice, from the true shortest path between the same points (which
of course is the arc of great circle which passes through them). Minimization of the distances was
thus a key reason in the project of the majority of the straight roads built in ancient times, for
instance by the Romans and the Incas. However, this is not the end of the story. Indeed, in many
cases ancient straight roads do not bend even when reasonableness would require a local change of
direction; in other cases, the width of the roads or even their very existence is due to symbolic – as
opposed to utilitarian – reasons.
From the technical point of view, the problem of tracing a road composed by subsequent straight
segments is readily solved. Indeed, it suffices to use fire signals during the night, or mirror signals
during daylight hours, day by day, during the construction. Though, this procedure works well only
if the length of the segment is of the order of some kilometres. Indeed, again, the earth is round. As
a consequence, the visible horizon (“how far it is possible to see, due to the earth roundness’’) is
limited, and it is an easy exercise of trigonometry to show that the distance in kilometres at which
an object of “zero” height can be seen from an height of H metres approximately equals the square

1
root of 13H kilometres. Thus, for a person 1.70 metres high, the visible horizon is about 5
kilometres; from a tower 20 metres high it is of the order of 16 kilometres. In case the sighted point
is at non-zero height, the two horizons sum up; thus, for instance, the top of a tower 20 metres high
can be seen at a distance of 21 kilometres by a person 1.70 metres high. Practically, such figures
imply that, if the two ends of a straight road are at a greater distance, it becomes extremely difficult
to fix the direction of the road by a single measure; of course, the use of subsequent aligned
segments is always possible (each based, for instance, on three or more inter-visible fire signals) but
the possibility of introducing errors increases considerably.
For this reason, suitable methods allowing for the required accuracy were devised and adopted by
many ancient cultures. These methods involved the use of geometry and astronomy, in some cases
in a rather sophisticated way. As a result, the exercise of “reverse engineering” of the project of
ancient roads is in many cases a quite interesting and difficult challenge, which is far from having a
general solution. We discuss here a few known cases and then present briefly the results we have
recently obtained in the case of the most famous of the Roman roads, Via Appia.

Fig. 1. Map of the earthworks at Newark. The first section of the Great Hopewell road is shown to the left.
(James and Charles Salisbury 1862).

2 Examples of ancient straight roads

Straight roads having a ceremonial (as opposed to utilitarian) character are a characteristic of many
pre-Columbian cultures. Perhaps the most famous example is the network of roads constructed by
the Anasazi in their heartland, Chaco Canyon (Sofaer, Marshall and Sinclair 1989). Most of these
roads do not exhibit utilitarian features and suggest cosmological purposes: actually, many of them
do not connect Chaco with other inhabited communities, and lead only to topographic features such
as pinnacles, springs, or lakes. In particular, there exists a long ceremonial route, having a width of

2
9 metres, constructed around 1100 AD. This “Great North Road” starts near the north rim of Chaco
Canyon and ends north-west at Kutz Canyon following three straight segments; the second one (that
leading to the so-called Pierre's Complex) runs with less than half of a degree of error to true north
for 16 kilometres. Precise methods for the astronomical determination of geographic north were
thus known to the Anasazi, as it is made clear also by the orientation to the cardinal points of many
buildings in Chaco Canyon, such as the famous Pueblo Bonito.
Other pre-columbian examples come from the Maya's Sacbeob (sacred roads). For instance, a
causeway connected Yaxuna and Cobá, in the Yucatan. The road, about 100 kilometres length, runs
along subsequent straight segments, the first one, starting at Yaxuná, is reported to be nearly 60
kilometres long (Villa Rojas 1934). Recent archaeological investigation has argued for the existence
of an even longer, peninsular-wide Maya road running east-west for approximately 300 kilometres
from the ancient city of Tihó (modern Mérida) to Puerto Morelos, on the east coast (Mathews
1999).
The Anasazi and the Maya roads are not the most ancient straight roads of America, though. Indeed,
the most ancient example pertains to the Hopewell culture flourished between 100 BC and 400 AD
in a wide area of the United States. Among the most characterizing features of Hopewell culture are
the monumental earthworks which characterize their ceremonial centres (Charles and Buikstra
2006). Many such earthworks are lost today, but they were documented in the 19th century by the
survey carried out by Ephraim Squier and Edwin Davis (1848). In addition, traces of Hopewell
works can be found through aerial photography and other non-intrusive techniques. Most of the
Hopewell earthworks have a regular, geometrical form, such as circles and squares but also more
complex figures, like octagons (Romain 2000). They were certainly connected with the religious
world of this people; in many cases the absence of extended human settlements nearby as well as
the analysis of the archaeological findings point to an interpretation of the concentrations of
earthworks as pilgrimage centres (Lepper 1996). This holds true, in particular, for the enigmatic
“circle+octagon” structures located at Newark and at High Bank respectively. These earthworks are
separated by a distance of about 90 kilometres; the High Bank complex, near Chillicothe, is barely
visible today, while the Newark one is preserved, although unfortunately used as a Golf court, They
show striking similarities in design. Indeed, they are composed of a huge circle – around 320 metres
of diameter, identical in dimensions in both sites - connected with an octagon. The High Bank
octagon is smaller with respect to Newark’s, but the disposition in plan is identical and both are
orientated to similar astronomical targets; however, the plans are skewed 90° with each other.
Examining ancient maps of the Newark earthworks, especially the one drawn by James and Charles
Salisbury in 1862 (Fig. 1), it can be noted that an ancient road, composed of two parallel earthen
walls, runs to the south west. During the course of the years, various attempts have been made to
find this road; finally, archaeologist Brad Lepper succeeded in recovering (both on the ground and
on aerial photographs) various parts of it (see Lepper 1995, 2006). In particular, a well documented
segment is located 26 kilometres south of Newark, and another one runs near the terminus at
Chillicothe. Lepper was thus led to the conclusion that this “Great Hopewell Road”, composed of
two parallel earthen walls at least one meter high and separated by 60 metres, once connected
straight away the two sites, a conclusion further strengthened by a recent LIDAR analysis (Romain
and Burks 2008).
The road is very likely to be interpreted as a ceremonial pathway. The two ends are too far apart to
be inter-visible and therefore the construction involved serious difficulties, probably solved with
astronomy (Magli 2009). In fact, the direction of the road runs along azimuth 31° 25’ from High
Bank to Newark (211° 25’ from Newark to High Bank) and is orthogonal to the direction of the
setting sun at the summer solstice (respectively, to the rising sun at winter solstice). However, how

3
can such an alignment – measurable only for the few days bridging the solstice, when the sun
moves extremely slowly from day to day at the horizon – be carried on with precision during the
construction of the road? A possible solution is the use of the rising and/or setting positions of
bright stars. Due to precession, such positions vary year by year (indeed, day by day), but of course
the variation is very small. Given the azimuth of the road, it is therefore possible to inspect the
period within which the road was constructed. At azimuth 211° 25’, an observer looking south-west
with a flat horizon would have seen the bright star Fomalhaut, of the constellation Piscis Austrinus,
setting in alignment with the road. The alignment is optimal around 250 BC. At this date,
Fomalhaut was setting at an azimuth of 211° 25’ with an altitude of 1° while at azimuth 31° 25’, an
observer looking north-east would have seen the bright star Capella, of the constellation Auriga,
rising in alignment with the road (the latter alignment is optimal at a slightly later date around AD
100, which seems to fit better the period of construction).
The above mentioned facts open the possibility that the Great Hopewell Road was constructed in
such a way to be oriented 90° with respect to the summer solstice sunset, and that this alignment
was maintained during the construction using the observation of bright stars at the horizon.

Fig. 2. The first section of Via Appia, from Rome to Terracina

The above discussed, pre-columbian examples are all based on astronomy. Straight roads having a
symbolic character are not unique to the Americas, though. Many examples are known from the
Romans. Perhaps the best studied are those of Britain (Ferrar and Richardson 2003). One example
is Fosse Way that linked Exeter in South West England to Lincoln in the East Midlands, with a
sequence of straight sections. Built in the second half of the first century AD, its length is 292
kilometres, and its errors of alignment are such that the whole road remains inside a strip of ten
4
kilometres in width. The azimuth of the road is close within 1% to the arc whose tangent is the
rational fraction 3/5. The method which was used is, thus, very probably the following. First, a grid
oriented to the cardinal points through astronomical observations was constructed point by point.
Second, the road was traced as the hypotenuse of a right triangle in which the two short legs were
integer numbers (in this case 3 and 5; a tendency to the use of Pythagorean triangles – i.e. right
triangles with all the legs integers, as (3,4,5) – is documented as well in other examples). This
simple but very clever method, if applied rigorously, is very precise and, further, it has a very
important feature: the engineers did not need the use of dimensional units, such as measures of
length, or angles. Indeed, given an arbitrary unit, the triangle which generates the orientation of the
road with respect to the cardinal points can be reproduced in any desired scale.
Recently, we have conducted a wide survey and analysis of the most famous of the Roman straight
roads, the section of Via Appia between Rome and Terracina. As we shall see in the next section, the
incredible complexity of the project of this road can be explained only if an equally complex
interplay between both geometry and astronomy is taken into account.

3 The Regina Viarum between Rome and Terracina

In his poem Silvae (end of the 1 century AD) the Roman poet Statius writes Appia longarum teritur
regina viarum that means “Appia is said to be the queen among the longest roads”. Effectively, the
road - constructed by the Roman consul Appio Claudio around the year 312 BC to connect Rome
with Capua - is an astonishing work of engineering (Quilici 1990). In particular, the first section of
the road, leading to Terracina, is composed by two perfectly straight segments connected by a short
zig-zag section aimed to cross the Alban hills at Colle Pardo. The first of such segments runs for 36
kilometres, the second crosses the Pontine Marshes going straight for as long as 62 kilometres
(Fig. 2).
To our knowledge, no one has ever tackled seriously the problem of the project of this road and, in
particular, of that of its longest and spectacular segment, although it is well known – and actually
obvious to anyone still today – that the path proceeds with impressive straightness. Further, it is
usually given for granted in the literature that the segment crossing the Pontine marshes runs with
azimuth 135°, that is along the north-west/south-east direction. Such a feature – if true – clearly
simplifies the project of the road. Indeed, astronomy can be used to determine the north celestial
pole and then an instrument like the Roman groma (i.e. a coeval surveying instrument) can be
efficiently employed first to find the meridian and then to bisect two times, obtaining the north-
west/south-east direction. However, taking some sample measures along the road with a precision
magnetic compass, as well as relatively precise measures using satellite images, we realized that
this azimuth is wrong: our measures in fact repeatedly gave azimuth ~136°. A deviation of one
degree south of south-east may appear small, but if it was kept for the whole length of the road then
it is definitively too much to think about an error of the Roman mensores (i.e. land surveyors). So
motivated, we finally decided to test ideas about the construction and the meaning of the road by
performing a high-precision GPS survey of the Colle Pardo-Terracina segment. Details on our
measurement campaign, which involved the solution of several non-trivial issues, as well as
a complete analysys both from the technical and the historical point of view can be found in (Magli
et al., 2012). Here, we report on our main findings.
1) The azimuth of the road is 135° 57', with a maximum error in the GPS measurement of ±10'. This
means that, if the Roman mensores wanted to orientate the road along the inter-cardinal direction,
then they really committed an error around one degree, and certainly not less than 47'. This error is

5
too much to be acceptable (also due to the point 3 below). Therefore, whatever strange it may seem,
we must conclude that Via Appia was deliberately orientated along an azimuth very close to 136°.
2) The road was almost certainly constructed together with the planning of a centuriation of the
Pontine marshes. This centuriation was discovered by Cancellieri (1990). She noticed the existence
of fossil traces of a regular division orientated to the cardinal points on aerial photographs. We were
able to confirm the existence of the grid by finding additional traces (Fig. 3). In particular, we have
found a full square of the original centuriation, located at Codarda (Latina).
3) Performing a computer reconstruction of the centuriation grid and superimposing the projected
grid on the satellite map of the Pontine area, we discovered that the road and the grid intersect very
close to the beginning of the straight segment under Torre Elena. It is, therefore, very likely – if not
certain – that the road was traced starting from this node used as the main survey point and,
therefore, with the key help furnished by the centuriation grid. Since indeed the grid was orientated
to the cardinal points, it allowed for a precise determination and collimation of the desired azimuth
section after section. To confirm this we have estimated also the original accuracy of the grid using
a priori single lines measured on satellite images, and it turns out to be very good, the grid being
orientated about 10' east of north. In order to keep the desired direction of the road with the
astonishing accuracy we can see still today, it is very probable that the construction proceeded along
the grid with a single building site (instead of two at the two ends or even more than two distributed
along the path), the one that started at the south easternmost node.

Fig. 3. Reconstruction of the centuriation of the Pontine Marshes (solid lines denote existing traces).

To summarize, the grid, and the existence of only one direction of progression towards Rome, are in
our view the keys that were needed to explain the great precision of the builders of Via Appia
between Colle Pardo and Terracina. However, having at disposal the cardinally oriented grid, the
obvious solution for the road was to trace it along the diagonal of the grid itself. It should be noted
that no topographical explanation can be found for this riddle: since the road and the diagonal differ

6
of only one degree in azimuth, the deviation that Appia accumulates on the east-west direction with
respect to the diagonal which starts at the same node is of less than 1.5 kilometres at the end of the
segment. On the other hand, as already mentioned, clear examples of the use of a method based on
a cardinally orientated grid have already been documented in Roman contexts.
Therefore, we analyzed the data in order to understand if ractional fractions were used.
Interestingly, this is almost certainly the case, but only for the first straight segment which runs
straight from Rome along the solidified lava plateau called Capo di Bove. Here once the general
direction was fixed (perhaps using the Aventine Hill as a distant foresight, Quilici Gigli 1990)
a rational fraction of 6/5 was probably adopted. Our estimate for the azimuth is in fact 140° 8';
subtracting 90° gives 50° 8', which is extremely close to 50° 12', the arc whose tangent is 6/5. But
what about the other segment? To obtain an angle close to 45° 57' one should use a ratio like 31/30.
Clearly, this would be pretty illogical, having at full disposal the simplest ratio of all, 1/1. The
solution must therefore be elsewhere, and effectively there is another, striking possibility.
Looking along azimuth 315° 57', that is, from Terracina to Colle Pardo, namely the direction along
which the road was constructed according to our hypothesis, the bright star Castor, the brightest star
of the constellation Gemini, was setting around 312 BC with an azimuth which is in excellent
agreement with that of the road. The azimuth of Castor at an altitude of one degree (appropriate for
a first magnitude star; recall that the horizon is practically flat) was indeed 315° 50'. Castor's setting
was visible approximately from the first days of January to the first days of June (proleptic Julian
dates).

4 Discussion

To get more insight into the symbolic meaning of straight roads it is necessary to merge technical
(topographical and astronomical) information with historical, cultural and archaeological
information. Such an inter-disciplinary approach is actually fundamental in any
archaeoastronomical research (Magli 2009). An example is the Anasazi case, where hints at the
cosmological meaning of straight roads can be obtained from the ethnographic background, in
particular from Historic Pueblo cosmology and ceremonies. There are indeed many symbolic uses
of roads in Pueblo ritual and myth; in particular, in the cosmological myth straight roads represent
the paths travelled by the first people when they emerged from the “world below”; similar paths are
travelled in the opposite direction by the souls of the dead (for a complete discussion and references
see Sofaer, Marshall and Sinclair 1989).
As far as the Romans are concerned, we are still very far from a complete understanding of the
presence of astronomical and cosmological concepts in their art and architecture, which anyway
appear to be quite more complex than previously thought (Hannah 2009, Hannah and Magli 2011).
Our results on the Via Appia show that the original project of the road arose from a complex
interplay of geometrical and astronomical techniques. It is usually said that Appius conceived Via
Appia as a main route (with final destination Capua) for military supplies during the Samnite wars.
The connection with Capua (and therefore with the war operational zone) was however already
assured by the Via Latina, running along the basis of the Lepini mountains on the border of the
Pontine marshes. It is usually given for granted that Via Appia is much shorter than Via Latina, but
a careful analysis via satellite images shows that the difference is actually negligible, precisely due
to the fact that Via Appia heads towards Terracina first. Thus, the saving of time does not explain
the construction of the road. As a consequence, some authors tried to frame Via Appia in a more
general cultural horizon, that of the Hellenistic influxes on Rome in the same period (Humm 1996).

7
The Via Appia might indeed be viewed as a “Hellenistic road”, celebrating the successes of its
builder. In such a context, it seems reasonable to think that the “ideology” connected to the
construction of the Appia road was strengthened by an astronomical orientation related with the
“military” environment within which the project was conceived. As it is well known indeed, the
Dioscures (Castor and Pollux) were protectors of the Roman army and, according to the Roman
historians, they even fought as allies of the Romans at the Battle of Lake Regillus, at the beginning
of the fifth century. Lake Regillus is not identified with certainty, but was certainly one of the (today
dried) basins located in the short strip of Albani hills which overlooks the Pontine marshes (Pareti
1959). Therefore, Via Appia was running from Terracina in the direction of the setting of Castor and
also towards the place of the first appearance of the divine twins as allies of the Romans.
To conclude, in view of the above reported results the huge, seemingly exaggerated engineering
project entertained by the Romans for the construction of Via Appia can be now better understood.
The road served as a military connection during the Samnite wars, but represents also a synthesis
between the Roman “practical” mentality and a system of religious symbols and beliefs which were
of mandatory importance for the temper of the army during a key moment of the Roman history.

References

[1] Cancellieri, M. 1990: Il territorio Pontino e la Via Appia. In S. Quilici Gigli (ed.) Quaderni
del Centro di studio per l'archeologia etrusco-italica 18 (Rome), 61-71.
[2] Charles, D. and Buikstra, J. (eds.), Recreating Hopewell. University Press of Florida. (2006).
[3] Ferrar, M.J. & Richardson, A., (2003) The Roman Survey of Britain, BAR, Series 359.
[4] Hannah, R. (2009) Time in Antiquity Rutledge, NY.
[5] Hannah , R. and Magli, G. 2011: The role of the sun in the Pantheon's design and meaning.
Numen 58(4), 486-513.
[6] Hively, R. and R. Horn (1982) Geometry and Astronomy in Prehistoric Ohio. Archaeoastr.
4:S1-S20.
[7] Hively, R. and R. Horn (1984) Hopewellian Geometry and Astronomy at High Bank,
Archaeoastr. 7:S85-S100.
[8] Hively, R. and R. Horn (2006) A Statistical Study of Lunar Alignments at the Newark
Earthworks, Midcontinental Journal of Archaeology.
[9] Humm, M. 1996: Appius Claudius Caecus et la construction de la via Appia. Mélanges de
l'École Française de Rome, Antiquité 108(2), 693-746.
[10] Lepper, B.T. (1995). Tracing Ohio’s Great Hopewell Road. Archaeology, vol. 45, no. 6, 52 –
56.
[11] Lepper, B. T. (1996) The Newark Earthworks and the geometric enclosures of the Scioto
valley: connections and conjectures. In A View from the Core: a synthesis of Ohio Hopewell
Archaeology, edited by P. J. Pacheco, pp. 224-241, Ohio Archaeological Council,
Columbus.
[12] Lepper, B. T. (2005) (Ed.) Ohio Archaeology: An Illustrated Chronicle Of Ohio's Ancient
American Indian Cultures. Orange Frazer Press.
[13] Lepper, B. T. (2006) The Great Hopewell Road and the role of pilgrimage in the Hopewell
interaction sphere. In Douglas K. Charles and Jane E. Buikstra, eds., Recreating Hopewell.
University Press of Florida.
[14] Lekson, S. H. (1999) The Chaco Meridian: Centers of Political Power in the Ancient
Southwest AltaMira Press, N.Y.

8
 [15] Magli, G. (2009) Mysteries and Discoveries of Archaeoastronomy, Springer-Verlag, NY.
[16] Magli G., Realini E., Reguzzoni M., Sampietro D., 2013, Uncovering a masterpiece of
Roman engineering: the project of Via Appia between Colle Pardo and Terracina. Subm. to
Oxford Journal of Archaeology.
[17] Mathews, J (1999) The Long and Winding Road: the Regional Maya Sacbe FAMSI Reports
Quilici L. (1990) Il rettifilo della Via Appia tra Roma e Terracina - La tecnica costruttiva in
10° Incontro di Studi del Comitato per l'Archeologia laziale CNR,Roma,pp. 41-60
[18] Romain,W. (2000). Mysteries of the Hopewell: Astronomers, Geometers, and Magicians of
the Eastern Woodlands, University of Akron Press, Akron.
[19] Romain, W.F., and Burks, J. (2008) LiDAR Imaging of the Great Hopewell Road. Current
Research in Ohio Archaeology
[20] Squier, E. and E. Davis (1848) Ancient Monuments of the Mississippi Valley. Smithsonian
Contributions to Knowledge 1.
[21] Sofaer, A., Marshall, P., and Sinclair, R. (1989) The Great North Road: A cosmographic
expression of the Chaco Culture of New Mexico, in World Archaeoastronomy, Edited by A.
F. Aveni, pp. 365-376. Cambridge: Cambridge University Press.
[22] Sofaer, A. and Sinclair, R. M. (1986b). Astronomic and related patterning in the
architecture of the prehistoric Chaco culture of New Mexico. Bulletin of the American
Astronomical Society 18, 1044-5.
[23] Villa Rojas, A. (1934) The Yaxuna-Coba Causeway. Carnegie Institution Publication, 436.
Washington: Carnegie Institution.
[24] Pareti, L. 1959: Sulla battaglia del Lago Regillo. Studi Romani 7, 18-30.

Current address

Giulio Magli
Faculty of Civil Architecture,
Politecnico di Milano, Milano, Italy

Eugenio Realini
Research Institute for Sustainable Humanosphere,
Kyoto University, Japan

Mirko Reguzzoni
Department of Environmental, Hydraulic, Infrastructures and Surveying Engineering, Politecnico di
Milano, Milano, Italy

Daniele Sampietro
Department of Environmental, Hydraulic, Infrastructures and Surveying Engineering
Politecnico di Milano, Como, Italy