North Africa has received considerable attention as a possible location for Atlantis since the beginning of the 19th century. Gattefosse and Butavand are names associated with early 20th century North African theorists. They, along with Borchardt, Herrmann and others have proposed locations as far west as Lixus on the Atlantic coast of Morocco, on through Tunisia and Libya and even as far east as the Nile delta.
One of the earliest writers was Ali Bey El Abbassi who discussed Atlantis and an ancient inland sea in the Sahara. The concept of such an inland sea, usually linked with Lake Tritonis, has persisted with the Chotts of Tunisia and Algeria as prime suspects. There is acceptance that a seismic/tectonic convulsion in the vicinity of the Gulf of Gabés cut off this inland sea from the Mediterranean. Diodorus Siculus records this event in his third book dating it to around 1250 BC. If such an event did not occur, how do we explain the salt laden chotts? However, proving a connection with Atlantis is another matter.
Whether this particular geological upheaval was related to the episode that destroyed parts of ancient Malta is questionable as the Maltese event was one of massive subsidence.
It should be kept in mind that Plato described the southern part of the Atlantean confederation as occupying North Africa as far eastward as Egypt (Tim.25b & Crit.114c). This of course conflicts with the idea of the Atlanteans invading from beyond ‘Pillars of Heracles’ situated at Gibraltar since they apparently already controlled at least part of the Western Mediterranean as far as Italy and Egypt.
*One of the principal arguments against Atlantis being located in North Africa is that Plato clearly referred to Atlantis as an island. However, as Papamarinopoulos has pointed out that regarding the Greek word for island, ‘nesos’ “a literary differentiation between ‘island’ and ‘peninsula’ did not exist in alphabetic Greek before Herodotus’ in the 5th century BC. Similarly, there was not any distinction between a coast and an island in Egyptian writing systems, up to the 5th century BC.” In conversation with Mark Adams[1070.198] Papamarinopoulos explains that in the sixth century BC, when Solon lived, nesos had five geographic meanings. “One, an island as we know it. Two, a promontory. Three, a peninsula. Four, a coast. Five, a land within a continent, surrounded by lakes, rivers or springs.”
Personally, from the context, I am quite happy to accept that the principal city of the Atlantean alliance existed on an island as we understand the word. This was probably north of Tunisia, where a number of possible candidates exist. However, it may be unwise to rule out a North African city just yet!
Another argument put forward that appears to exclude at least part of North Africa is that Plato, according to many translations, he refers to Atlantis as ‘greater’ than ‘Libya’ and ‘Asia‘ combined, using the Greek word, ‘meizon‘, which had a primary meaning of ‘more powerful’ not greater in size. Atlantis could not have been situated in either Libya or Asia because ‘a part cannot be greater than the whole’. However, if Plato was referring to military might rather than geographical extent, as seems quite likely, North Africa may indeed have been part of the Atlantean alliance, particularly as Plato describes the control of Atlantis in the Mediterranean as far Tyrrhenia and Egypt.*