Theophanes dated the earthquake to the third regnal year of Emperor Constantine V, who ruled from 741 to 775. His account is spare but vivid: a sign in the north, dust descending on various places, and destruction at the Caspian Gates. The 11th-century chronicler George Kedrenos repeated the account nearly word for word, drawing directly from Theophanes as his source. That the earthquake was recorded at all in Byzantine sources speaks to its magnitude. Constantinople's historians did not typically note seismic events in distant provinces unless they were catastrophic enough to generate reports that traveled the full length of trade and military communication networks back to the capital.

The name is the puzzle. "Caspian Gates" was a geographic term ancient and medieval writers applied to several different mountain passes across the Caucasus -- the narrow corridors linking the Mediterranean world to the Iranian plateau and Central Asia. By the Byzantine era, the term most commonly referred to the pass at Derbent, the fortified gap between the Caucasus mountains and the Caspian Sea in what is now Dagestan, Russia. Derbent was known as "the Gate" or "the Gate of Gates," a choke point that armies had contested for centuries. But an Armenian translation of Pseudo-Callisthenes' Alexander Romance identifies the Caspian Gates with the territory of Talis, in what is now northern Iran. A 1982 seismic catalogue by Ambraseys and Melville placed the earthquake east of Ray, Iran, in the valley of Tang-e Sar-e Darreh.

In the 19th century, the German geographer Karl Ernst Adolf von Hoff and the Irish engineer Robert Mallet proposed yet another location: the pass of Dariel, a gorge through the central Caucasus on the Georgia-Russia border. The proliferation of candidates reflects the ambiguity built into ancient geographic terminology. Greco-Roman authors used "Caspian Gates" loosely, and by the time Theophanes wrote his chronicle, the term had accumulated multiple referents across centuries of shifting political geography. Modern seismologists studying the earthquake have noted that the region between northern Iran and the Caucasus is one of the most seismically active zones on Earth, where the Arabian plate grinds northward into the Eurasian plate. A large-magnitude event anywhere along this collision zone could have generated the effects Theophanes described.

What survives is an echo: a few lines in a Byzantine chronicle, a mention of dust in the sky, and a geographic name that points in several directions at once. The earthquake itself has been absorbed into the ground, its rupture healed by thirteen centuries of tectonic motion. But the Caspian Gates remain -- not as a single location but as an idea, the narrow places where mountains pinch the roads between civilizations into corridors that shake and crack. Whether the ground broke at Derbent, at Talesh, or in some valley east of Ray, the lesson is the same: the mountains that guard the gates between worlds are not as solid as they appear.