"The Zama Campaign Reconstructed" Topic

The Zama Campaign Reconstructed: Forgotten Battles and Massinissa's Triumphs

By the autumn of 202 BC the fate of Carthage hung in the balance. Hannibal had returned to Africa to defend his mother-country against Scipio`s army and he was aided by Massinissa and his Numidian light cavalry. We are well aware that history has it recorded that all was decided in one great, dramatic clash of arms at a place that became known as Zama. But the truth about the so-called Zama campaign however, was that it was brought to a conclusion by a succession of ambushes and ruses that employed forces which were largely comprised of Massinissa`s Numidian light horsemen. These stratagems and deceptions were directed according to Scipio`s grand strategy, whose Roman Consular army by itself, was far too small to defeat Hannibal in open battle and at the same time maintain his blockade of Carthage and siege of Utica…

New evidence and astro-archaeological data from three monuments in Tunisia will form the basis of this Zama campaign reconstruction.

Planned Publication: Summer 2026.

Hmm, interesting.

I'll be watching for this

Here`s an update and a preview of two of the book`s illustrations.

The first is a modified Stellarium 25 screenshot from the position of the most recently excavated Numidian monument at Djebel el Berouag, Tunisia.

It has been known to archaeologists for over ten years, but was only recently excavated in 2023-24 by the Tunisian INP and the Catalonian ICAC and the UB (Institut Català d'Arqueologia Clàssica and the Univeristat de Barcelona) – see the report: Kallala, N., Torchani, M., Cherif, S. & Jnen, M. Africa XXVII Institut National du Patrimoine, (2024).
This triumphal monument is the third to be identified as belonging to a single state building programme in the reign of Micipsa. I have attempted to date this and the monuments at Kbor Klib and Chemtou by their orientations. Chemtou and Djebel Berouag align with sunrises and Kbor Klib with a sunset. All these alignments occur within one month of each other.
The Stellarium 25 application allows us to 'rewind' the sky and match the monument's measured alignment to a past sunrise/sunsets and the date upon which the alignment matches the orientation of the monument at Djebel el Berouag is the 3rd September 120 BC. This places its construction within the last few years of Micipsa`s reign.

I propose in my book that this is the site of Q. Minucius Thermus`s ambush that appears in Appian`s Punica:

"… learning that Hannibal was very short of supplies and was expecting a convoy, sent the military tribune, Thermus, [Quintus Minucius Thermus.] by night to attack the supply train. Thermus took a position on the crest of a hill at a narrow pass, where he killed 4,000 Africans, took as many more prisoners, and brought the supplies to Scipio." (Appian Punica 8:36.5-7)

Image 1: Modified Stellarium 25.2 image of the data at Djebel el Berouag for the 3rd of September 120 BC – the proposed construction date for the monument that stood there. (Red lines struck through to deter FB`s use of the original image.)

link

Image 2: A detail of a French map from 1925, The cyan arrow shows the orientation of the monument at an azimuth of 78.5°, overlooking the proposed battlefield where reinforcements bound for Hannibal`s army were trapped between the hill in the west and marshy ground to the east. Tunisia Flle. No. LII-B1-C33, Ebba Ksour Paris: National Geographic Institute, (1925) 07-08-02-011 1886

link

The site of the battle was not far from the city of Obba.

The site of Djebel el Berouag is 32 kilometres west of Seba Biar and the real battlefield of "Zama".

But the truth about the so-called Zama campaign however …

I know this has been an ongoing hypothesis on your part and we've debated this in the past on TMP so I am not going go over the material again but I find it very unlikely.

Marcus,
It`s all a question of what you choose to believe. After considering all the sources and rejecting the elements in ALL of them that appear incredible, or implausible this hypothesis is what I`m left with.
I know some will favour Polybius over and above all others and excuse his bias when Hannibal is taken in by Scipio`s magnanamity and as a result feels compelled to parley with him after recycling the three spies` story for a third time. The material has greater contextual cohesion now – especially with the new finds at the Djebel el Berouag site. The reconstruction is based upon the dating of the monuments, and I go in to that in more detail in the book than you may have seen before in discussions here. Historians may have been unaware or dismissive of the archaeology before now. Only from 2005 has Kbor Klib been discussed as a potential battlefield monument. I have since made claims for both Chemtou and Djebel el Berouag – the latter only being made known to us after 2014 and as I mentioned above, only recently excavated in 2024. If you need more to convince you perhaps you should read my papers on the orientations of the monuments and their frieze shield devices on academia.edu ?

@ Dave Crowell and Grattan54 thanks for your signs of encouragement, it is much appreciated. For more information you`d be welcome to join my fb page @ link

Regards, Mike.

For me, the Djebel Berouag monument filled in the missing link in the puzzle of the Zama campaign chronology last year. Its location provides valuable context and points to the reason why it was that Hannibal marched from Zama to be drawn into a fight with Scipio close to Seba Biar (15k as the crow flies from Zama Regia). That battlefield site is now overlooked by the remains of the Kbor Klib monument, which in 2005, Duncan Ross proposed was the site of the battle Zama. The details of this, the dating of the monuments and the calendar dates of the battles that they point to will be explained in the book.

My update to the book is that I've just edited the footnotes and bibliography, laid out the contents, and tidied up the illustration list.
Although I planned for late summer, I may well be ahead of schedule and able to publish in the next two months!

The final proofreading is going well, and so I will probably publish at the end of this month or the beginning of June.

Delays and more delays! As I intend to use KDP and IngramSpark, this needs to be coordinated with simultaneous release dates (too technical). And so, I will be waiting for the slowest printers to catch up; ISBN`s to become activated, previews to update, printing queues, shipping times (posting proof copies), and so publication is probably a few weeks away now.

The Zama Campaign Reconstructed is now available from Amazon! It took me by surprise, but as per their shipments, they delivered earlier than they had advised !

Here`s the Amazon link to the book:

link

If you`d like to read a sample of the first three chapters, of my book, then this is available on academia.edu: link
or you can find the same sample on my amazon author`s page: link

Nebeur vs. Djebel el Berouag: The Logistics That Redrew a Campaign

For a long time while researching my book, The Zama Campaign Reconstructed, I was convinced Hannibal's movements and Thermus's ambush spot lay somewhere along the hill passes of the modern P17 route. The geography seemed plausible, but the strict chronological timelines from Polybius and Livy always felt slightly strained.

When I first attempted to locate the site of Q. Minucius Thermus's night ambush against Hannibal's supply train (Appian, Punica 8:36.5-7), the hill passes near Nebeur along the modern P17 route were an incredibly tempting candidate. The narrow passes fit the ideal tactical profile for an ambush site. Located 50 kilometres north-west of Kbor Klib, Nebeur, Le Kef, Tunisia was also conceivably situated on the route that reinforcements from Vermina would have taken.

But then ancient military history isn't just about finding a good hill; it's about the maths of march distances, timeline constraints and ancient logistics.1

On hearing about excavations at Djebel el Berouag, situated just 32 kilometres west of Seba Biar, I calculated that Thermus with Masinissa attached might be able to march there in time and the site fell into place perfectly and my earlier Nebeur hypothesis failed to match the dates that the solar orientations of the monuments suggested.

And so with the addition of the new piece in the puzzle, the marching times matched the distance from Partha (Henchir Lorbeus) to the el Berouag site. The terrain at Djebel el Berouag reveals a narrow pass between hills and some marshy ground to the east of the modern route C17, providing the perfect geographical anvil for Thermus's hammer. The recent archaeological excavations of the Numidian triumphal monument anchor the entire action.2

The P17 route near Nebeur looked like a very good candidate for the site of the battle on a modern map, but the physical, archaeological evidence and logistical tracking suggest quite strongly that Djebel el Berouag was where Thermus`s trap was actually sprung.

1. You can read about this in post 5 on 01-19-2026, 03:52 PM link

2. Nabil Kallala et al., "Nouvelle découverte d'un monument Numide Hellénistique à Djebel El Berouag," Africa 27 (2024) academia.edu/128472897

The Zama Campaign Reconstructed
Part of the reconstructed Zama campaign chronology (August–September 202 BC)
Extract From: The Zama Campaign Reconstructed (Appendix B, p. 145)
Author: Michael Collins

Abstract & Research Context

Uses marching rates from Vegetius's De Re Militari to analyse five critical routes between Tunes, Naraggara, and Killa. The routes and data align with recent excavations and surveys in Tunisia, at Kbor Klib and Djebel El Berouag.

Scipio's Routes of March from Tunes to Naraggara, Killa and back

This appendix provides a detailed breakdown of the distances of five of Scipio's marches: to Naraggara from Tunes at the start of the campaign, to Killa via Partha and As Sars (and the battle of Zama), and then his return to Tunes/Castra Cornelia.
Distances covered could be greater dependent upon conditions that may vary due to terrain, weather and road conditions, but the following calculations are based upon an average of the Roman legionary "ordinary step" or "military pace" as detailed in Flavius Vegetius Renatus' De Re Militari marching at a rate of 20 Roman miles per day which equates to 18.6 English miles, or approximately 30 kilometres. This is a conservative estimate for mounted forces, but as these were often accompanied by light infantry, the foot marching times would be appropriate. Numidian forces under Massinissa's command may probably moved much further and faster than these proposed distances below, as even an ordinary horse walking pace is faster than that of infantry. Massinissa's troopers would have been able to cover these route distances with ease. And so, even though the campaign chronology seems somewhat compressed, the Numidian light horse would have managed longer and faster marches.

Scipio's Routes of March from Tunes to Naraggara, Killa and back

Tunes – Naraggara (modern Sakiet Sidi Youssef)
Kilometres/English miles/Roman miles/Length of march in days
217/135/146.5/7.5

Naraggara – Partha/Lares (modern Henchir Lorbeus)

Kilometres/English miles/Roman miles/Length of march in days
52.58/32/38/2

Partha/Lares – (modern As Sars)

Kilometres/English miles/Roman miles/Length of march in days
17.49/10/10.87/0.5

As Sars – Killa/Margaron (modern Seba Biar)

Kilometres/English miles/Roman miles/Length of march in days
15.79/15.81/10/0.5

Killa/Margaron (modern Seba Biar) – Tunes

Kilometres/English miles/Roman miles/Length of march in days
150/93.2/101.36/5

The Zama Campaign Reconstructed. Appendix 3
Author: Michael Collins
Edited Extracts From: The Zama Campaign Reconstructed (Appendix C, pp. 147–149)
Target Focus: Methods of Astro-Archaeological Dating and Alternative interpretations

In order to fix the exact calendar dates for the monuments and the battles that they commemorate, the first step was to establish their azimuths (bearings in relation to the celestial north pole). This was achieved by using the Google Earth Pro ruler function and measuring the strongest lines produced by the imagery.
In the case of Kbor Klib the measurements of its azimuth were taken from the interior sides of the two door entrances and the average of these measurements were then compared to orientation of Kbor Klib found in the 2001 work of César Esteban, Juan A. Belmonte, and others: Orientations of Pre-Islamic Temples of Northwest Africa published in Archaeoastronomy no. 26.
Firstly one needed to narrow the search for the date. As the monument faces west north-west and so a solar orientation was to be expected rather than a stellar one.
Knowing that the events commemorated were sometime in the autumn of the year of 202 BC is of great help of course, but a solar orientation there are usually two times in a year at which the structure would have been aligned with a sunset.
Using the Stellarium25 application the aim was to select a year in which the azimuth was best aligned with the average of measurements taken on the Google Earth Pro rule function and this also matched the lowest altitude measurement for the Sun at sunset.
After the work by Naïdé Ferchiou on Kbor Klib and that by Frederick Rakob on the stylistic dating of the Chemtou monument, I narrowed the search down to the last decade or so of King Micipsa`s reign. The year of 120 BC was found to provide the best matches for the Kbor Klib azimuth.
For the astronomical year of -121 on Stellarium25 (there is no year zero), this azimuth of ~280° produces sunsets aligned with the monument; on the 13th April at 18:48:35 and the 4th of September at 18:46:35.
The April date was rejected because it was too early given the historical circumstances. It would produce a time for the battle that would predate the Roman senate`s annual appointments of officials and it would also precede the delayed departure of Tiberius Claudius Nero and his fleet sailing for Africa.
The monument and battle therefore is dated to the 4th of September on the Julian calendar.

Astroarchaeological Methodology and the Tunisian Monuments

Gerald Hawkins' ground-breaking work demonstrated in Stonehenge Decoded1 that ancient monuments often marked culturally significant dates through celestial alignments. Another noteworthy example of a solar alignment was the older Parthenon on the Acropolis, Athens which was aligned so that the temple's interior and the statue of Athena were illuminated on specific days as the sun rose. This astronomical event represented the sacred presence of the goddess Athena and functioned as a time-keeping marker for ceremonies and festivals.2
With a similar intent, the three Numidian monuments' orientations on the 30th and the 31st of August and the 28th of September, 121 BC (on the proleptic Julian calendar) were aligned to commemorate Massinissa's victories over Vermina and Hannibal. Each monument overlooked its battlefield from an elevated position on the sunset or sunrise of a precise calendar date.
I believe we can confidently have faith in the Numidians' ability to record and keep pace with these dates of commemoration. In the 120`s BC, the Numidians likely used the Egyptian calendar, which was reformed by the Ptolemaic decree of Canopus in 238 BC, adding a leap day to a year of 365 days every fourth year. This would have enabled a precise tracking of dates over the 81-year gap, from 202 to 121 BC and it supports the monuments' accurate alignments. The precision of Numidian time-keeping probably surpassed that of the Roman pre-Julian calendar and its irregular intercalations.
The measurements taken in my study and the use of applications were in accordance with Clive Ruggles' guideline that:
"The main necessity is to determine the azimuth, horizon altitude, and declination in the direction "indicated" by any structural orientation…".3
We can be confident that the alignments with the sun's movements for all three monuments, as shown in the Stellarium application, are adjusted for precession and confirm the narrow calendar time period of August 30th to September 28th, 121 BC.
Although the Earth's precession may affect solar and celestial alignments over centuries (as noted by Ruggles, 2014) the effects on the dating of the monuments in Tunisia would be minimal. As an example of the degree of error in historical alignments, the Temple at Karnak demonstrates how the alignment of the winter solstice sunrise has changed over time. The alignment has moved by 0.54° in 4,000 years.4 The difference in degrees at the temple between 202 and 121 BC would just over 0.27°. But more importantly, the measurements are relative, and so the time-frame would be unchanged, with less than one month separating our points of reference in the campaign chronology. So, if we take the movement of the sun in 4,000 years at the Temple at Karnak as a guide, the difference in the position of the sunrise in the 81 years between 202 and 121 BC would be 0.01107° and is therefore not a significant error:
0.54°/4,000 (years) = 0.000135° per year x (82 years) = 0.01107°
It should be noted that whilst Google Earth Pro is a useful application, it has its limitations and far more accurate measurements could be gathered by employing GPS and astrogeodetic surveying techniques. Hopefully, a combination of these methods would confirm my rough estimations.
Ruggles also states that: "…the credibility of any interpretation needs to be assessed in terms of social theory, the strength of the material evidence in its support, and the quality of the corroborating evidence from history and/or ethnography, as available."5
Belmonte et al.6 demonstrate that pre-Islamic North African mausolea share eastward solar orientations and were common for commemorative purposes in North-West Africa.
With this local tradition, shared architectural and artistic elements within these monuments suggest a grand and highly coordinated building project dedicated to Massinissa's victories rather than a collection of constructions with arbitrary alignments in the landscape. The astroarchaeological data and the architectural and art-historical evidence would favour construction in the last years of Micipsa's reign.

Summary of Calculation Details:

1. Multiple azimuth measurements were taken using the Google Earth Pro Ruler Line "heading" function. From these multiple measurements, mean averages were calculated.
When the images were of lower quality, more measurements of the monuments' azimuths were taken.
2. The coordinates and elevation of sites were also recorded using Google Earth Pro.
3. The coordinates and elevation were entered into the Stellarium 25's location function, and the date was set for -120 (there being no year "zero"). For this calculation, I had previously used the Planet.calc site.
4. The sun was selected on the screen, and the time was adjusted by days, hours, minutes, and seconds until the Sun's azimuth equalled the "heading" measurements taken on Google Earth, and its altitude was as close as possible to 00° 0' 0".
5. For the cases of Chemtou and Djebel El Berouag, I used Grok3 xAI to calculate the adjusted time of sunrise due to the apparent horizon (the level from which the sun was seen to rise) being closer than it would have been at sea level.

Alternative interpretations of the monuments' purpose and function
Several alternative functions have been proposed for the three monuments. Most popular,(since the evidence for their having been used for specific religious purposes is lacking, and there being no inscribed dedications to Punic deities, nor tombs discovered in association with them), is that these monuments were frontier markers. That may apply to two of the monuments which face eastward and in the direction of Carthaginian territory (Djebel El Berouag and Chemtou), but this is not the case for Kbor Klib, which, facing west, looks inwardly into Numidian-held territory and away from the enemy, to whom a statement of "under newownership" would normally be directed.
The proposed dating of the Chemtou7 and Djebel el Berouag8 monuments is approximated to the last quarter of the second century, and this dating does not align with their positioning on the Numidian-Carthaginian border.
In the case of Chemtou, Massinissa had made Bulla Regia his capital in 156 BC. Bulla Regia was situated 17 kilometres to the east of the Chemtou monument, and so a frontier marker in 121 BC would have been placed further to the east of that city, not close to the town of Simutthus.
All of our monuments, however, are too far to the west and too far removed from the Carthaginian borderlands of the early 150's BC to suggest that they were triumphs erected to celebrate Massinissa's victory over Hasdrubal at a battle or the siege of Oroscopa in the country surrounding the town in 151 BC. This campaign ended in total disaster for Hasdrubal's army.9
In 1990, Naïdé Ferchiou, suggested that:
"…the toponym Rechga was a contraction of Oroscopa, or of the Punic-Numidian name, which the Greeks themselves may have distorted into Oroscopa?"
She proposed that Oroscopa could be Kef Rechga or Mount Kaf Rashqa,10 which is situated approximately 45 kilometres to the north-east of Chemtou's Numidian monument. This being a disputed ground in the 150's would place Chemtou's monument well within Micipsa's territory in the 120's. It was therefore too far from Massinissa's battleground and frontier of 151 BC.
The reasons for not being a boundary marker would also apply to the monuments at Kbor Klib and Djebel El Berouag.

Endnotes: Astroarchaeological Methodology and the Tunisian Monuments

1. G. S. Hawkins and J. B. White, Stonehenge Decoded (New York: Doubleday, 1965), 20. Hawkins notes that in 1740, Dr. William Stukeley first observed the alignment of the Heel Stone as "the principle line of the whole work, [which points to] the northeast, where abouts the sun rises, when the days are longest." On page 103, Hawkins cites the earliest observations by John Smith (1771) and "Warltire" (1792), the latter describing the site as "a vast theodolite for observing the motions of the heavenly bodies." In the 1840`s, the Rev. Edward Duke noticed that station stone positions 91–92 and 93–94 align to midsummer sunrise and midwinter sunset. Hawkins's work was first published in Nature in 1963.
2. George Pantazis, "The Symmetric Placing and the Dating of Parthenon and Hephaisteion in Athens (Greece)," Mediterranean Archaeology and Archaeometry 14, no. 3 (2014): 278, table 2.
3. Clive L. N. Ruggles, ed., Handbook of Archaeoastronomy and Ethnoastronomy (New York: Springer, 2014), chap. 27, "Analyzing Orientations Part II," abstract.
4. Amelia Carolina Sparavigna, "The Karnak Temple and the Motion of the Earth's Axis" (Polytechnic University of Turin, 2016), 3. "For Karnak temple, the sunrise azimuth on winter solstice is 116.20 degrees today and was 116.74 degrees, in about 2000 BC."
5. Ruggles, Handbook of Archaeoastronomy and Ethnoastronomy, chap. 24, "Nature and Analysis of Material Evidence Relevant to Archaeoastronomy," abstract.
6. J. A. Belmonte et al., "On the Orientation of Pre-Islamic Temples of North Africa: A Re-appraisal (New Data in Africa Proconsularis)," Mediterranean Archaeology and Archaeometry 7, no. 1 (2007).
7. Friedrich Rakob, Simitthus II: Der Tempelberg und das römische Lager (Mainz: Philipp von Zabern, 1994), 1–38.
8. Nabil Kallala et al., "Nouvelle découverte d'un monument Numide Hellénistique à Djebel El Berouag: Mausolée, Mausolée-Santuaire ou Sanctuaire-Autel?" Africa 27 (2024), academia.edu/128472897/.
9. Oroscopa appears in Appian`s Punica, 14.70.6–14.74.1.
10. Naïdé Ferchiou, "Habitats fortifiés pré-impériaux en Tunisie antique," Antiquités Africaines 26 (1990): 47, link

The Zama Campaign Reconstructed is available now. If you want to dive deeper into the tactical maps, troop movements, and the other appendices, you can find the book at
link