Mecca before Islam: 1) Diodorus and the Kaaba

The first in a series on pre-Islamic sources adduced for Mecca.
Next: 2) Macoraba.


Hieron d’ hagiōtaton hidrutai, timōmenon hupo pantōn Arabōn perittoteron.
“A very sacred temple has been established there which is highly revered by all the Arabs.”
—Diodorus Siculus, tr.  Stanley M. Burnstein

Ancient literary sources do not have much to say about the Red Sea coast of Arabia. A king of Egypt, Ptolemy II Philadelphus (r. 283–46 BCE), sent a fact-finding expedition down the coast to chart the region before he launched his campaigns into northern Arabia. The expedition, led by a certain Aristōn, produced a report for military use. The report has disappeared, but it is generally assumed to have been the main source of the historian and geographer Agatharchides (2nd century BCE) in his seminal work, On the Red Sea. This too has disappeared, but some of the text was preserved and reworked into three later studies: the Bibliotheca Historica of Diodorus Siculus (1st century BCE), the Geographica of Strabo (1st BCE–1st CE), and much later – but arguably the best preserved – the Bibliotheca of Photius (9th CE). These three sources, read side by side, give us some idea of what Agatharchides had to say about the western coast of Arabia.

According to Agatharchides, somewhere south of the Gulf of Aqaba is a bay or gulf (kolpos) that runs inland about 500 stadia—a very improbable number, roughly 90 km. The mouth is too craggy for ships to traverse. Around the bay live a hunting people called something like the Batmizomaneis or Banizomeneis. Diodorus uniquely includes the detail that a “very sacred temple has been established there which is highly revered by all the Arabs.” A few modern commentators have identified this temple with the Kaaba, which – according to Islamic tradition – the ‘Arabs’ had uniquely, universally revered long before Muhammad’s lifetime. The first to make this identification was probably Edward Gibbon, who was so confident in the connection, he admitted his surprise that no one had seen it before.

But Gibbon was mistaken. As our sources move on from the bay, southward down the coast, they encounter the lands of a people called the Thamūd (Thamoudēnoi). We have no hard and fast borders for the lands occupied by the Thamūd, but ancient and Islamic sources place them in the northwest of the Peninsula, not far from the Nabataeans and Arabia Petraea. Gibbon explicitly puts the temple between the Thamūd in the northwest and the Sabaeans in the Yemen – an area that may include Mecca – whereas a more careful reading of the Agatharchides tradition places the bay and its vaunted temple north of the Thamūd.

With these parameters in mind, more recent scholars have suggested the long, rocky bay near Wadi Aynuna. This fairly well fits the description, except that it is nowhere near 500 stadia long. Personally, I wonder if Agatharchides’ bay might run all the way to Aynuna from Tiran Island, where the Gulf of Aqaba opens onto the Red Sea coastline. The cluster of peninsulas, islands and reefs at that point may have presented a treacherous passage for ancient ships, resonating with his account. From there, a traveller would have followed the coast to Aynuna, going “inland” as Agatharchides sees it, before turning sharply southeast onto a relatively straight coast. Even this is more like 300 stadia than 500, and extremely speculative besides. We may never know for sure which place Agatharchides was talking about. The Batmizomaneis remain a mystery, as does their temple.

It may be tempting to keep hold of the temple’s description while discarding the geographical context. On this view, Agatharchides/Diodorus was half-right: there was a temple “very sacred” and “highly revered by all the Arabs”, but it was in Mecca, not the northwest. Unfortunately, that is to lean the heaviest on the source’s weakest point. It is entirely plausible that there was a successful temple in the northwest, whereas it stretches credulity that any temple anywhere should have been “revered by all the Arabs”.

The medieval Islamic historical tradition presents the Kaaba as the foremost temple of ancient Arabia, going back centuries. Whatever the merits and pitfalls of that tradition, the earliest sources do not obviously confirm this point. If there really was a centripetal temple throughout the millennium between Ptolemy II and Muhammad, then Agatharchides/Diodorus is the one and only ancient source that might explicitly describe it, and even that puts it in the wrong place. The literary sources for ancient Arabia are foreign and full of holes, but they are not completely ignorant. In my view, their failure to acknowledge a pan-Arabian temple would be a startling oversight.

We can of course choose to suppress the geographical context in order to salvage one convenient sentence. But it would be much easier – more parsimonious – to let the geography stand while moderating the temple’s significance. Agatharchides and his source(s) were outsiders who were more concerned with mapping territories than cataloguing religious practices. The statement about the temple should be read as a simple exaggeration. There was a successful temple near a bay in northwest Arabia. Diodorus does not mention the Kaaba.




Photius, Diodorus and Strabo (English): Stanley M. Burnstein (tr.), Agatharchides of Cnidus: On the Eritraean Sea (1989), 132–173, esp. 152–3 (§92).

Photius and Diodorus (Greek with Latin): Karl Müller (ed.), Geographi Graeci Minores, vol. 1 (1855), esp. 180–1 (§90).

Diodorus (Greek with English): C.H. Oldfather (tr.) Diodorus of Sicily, vol. 2 (Loeb 1935), esp. 216–7 [English].

Strabo (Greek with English): Horace Leonard Jones (tr.), The Geography of Strabo, vol. 7 (Loeb 1930), esp. 342–3 [English].


Further reading

Jan Retsö, The Arabs in Antiquity (2003), 295–300.

Edward Gibbon, Decline and Fall, vol. 5 (1788), 190–1 and n. 45.

Ian D. Morris

Ian D. Morris

Historian and occasional human, powered by tea and cat videos.


  1. “Diodorus does not mention the Kaaba.” End of story I think.Even J.Chabbi does not believe anymore to the story of Ibn Ishaq’s Mecca, capable of raising 30,000 strong to conquest Syria and a major step of in the caravans route.

  2. Way to dance away from the pretty obvious fact that it is talking about the Kaaba.

    The Asatir compiled no later than 250-300BC states that Ishmael and Abraham built Mecca. In the first translation of the Torah.

    Makoraba = Makkah + Rabba (Lord) = Lord of Makkah. Ibn Ishaq found an ancient inscription by the Kaaba saying ‘House of the Lord’.

    Even during the lifetime of Muhammad there was unanimity over the origin of the Kaaba amongst opposing parties.

    Mecca’s surrounding areas was very mountainous area so it’s going to be difficult for you to expect to historically find the right location.
    Historically, only the Kaaba has been seen as a Holy Temple revered by all Arabs. All evidence from the ancient time of the 600’s AD also affirm this.
    It’s EASY to just criticise and create ‘inflammatory’ content as cult followers put it.

    • Hi! It’s gratifying to know I made this look easy. Certainly a mission of this blog is to make these abstruse arguments a bit more accessible.

      I think you may be right that the Hijaz would have been tough for outsiders to navigate. Yet Agatharchides’ account correctly identifies the order of some known locations as you travel north-to-south along the coast. For him to misplace Mecca within the Hijaz would be one thing, but to place it further north than even the Thamud – and with such abundance of detail – would be a very strange mistake.

      I don’t assume that medieval sources are accurate when they describe what happened centuries before; I think it’s more rigorous to test those descriptions against earlier kinds of evidence. That’s a problem in this case, because Mecca and the Kaaba are not securely attested until after the Muslim conquests. That gives us a span of, say, nine centuries between the military report that described the northern temple and the appearance of Mecca and the Kaaba on the historical record. There is a big gap in both time and space; only a wilful interpretation can reconcile the two temples.

      I’ve written a much longer post about Makoraba, which explores the many attempts since the mid-seventeenth century to identify that name with Mecca. It’s worth a read as it unpacks some of the broader interpretive questions at stake when it comes to ancient Mecca. I also explain there that Greek k is unlikely to render Semitic k. I should add here that ‘House of the Lord’ is a different formula from ‘Mecca of the Lord’, which is what we should hope to see in the Arabic sources if that is what the name Makoraba is alluding to. Speculative etymologies are fun, but it’s better to root them in sources where possible.

      I’m glad you’ve brought up the Asatir, because it’s not very well known. It’s not a translation in fact, but a companion to the Torah in the form of a chronicle. A date in the 3rd century BCE was proposed by its modern editor Moses Gaster in 1927, but in 1943 Ze’ev Ben-Hayyim argued for a date from the mid-10th CE, which is now the consensus, from what I’ve seen. It would therefore post-date the rise of Islam. It says that Mecca was built, not by Abraham and Ishmael, but by Ishmael’s descendants. In the early Middle Ages the Ishmaelites were generally identified with the Muslims, so this is not at all surprising; some Jewish translations of the Torah even render biblical place names with the names of Muslim towns from their own time. I’ve tweeted a little about this before. If you follow the thread you’ll find a link to Gaster’s edition and translation, along with the relevant page numbers.

      What I think I’m demonstrating with this blog series is that it is obvious that ancient sources are talking about Mecca and the Kaaba, until we reflect on the assumptions we’re bringing to the evidence. It’s worth remembering that the claims we make about the past have histories all of their own. Before Gibbon, historians had access to Diodorus and they knew that Mecca existed, but they didn’t make the connection between Mecca and Diodorus’ temple, perhaps because they recognised that these were different places on the face of it. But once Gibbon had made that connection – and his Decline and Fall had become a smashing success – it then became obvious. It has been repeated uncritically, so that what we need today is a critical reevaluation of the evidence and our longstanding relationship to it.

      Thanks for stopping by; I hope this was useful.

Leave a Reply

Your email address will not be published. Required fields are marked *