unfulfilled the promise about Ephialtes; he mentions twice in language very similar, but not identical (i.175; viii. 104), the fact, not worthy of such signal prominence, that when any untoward event threatened the city of Pedasus, the priestess of Athena there was liable to grow a beard. More remarkable still, he refers in two places to what he will say in his 'Assyrian Logoi' (i.106; i.184), which are not to be found. The actual end of the work is hotly fought over. Can it, a mere anecdote about Cyrus, tacked on to an unimpressive miracle of Protesilaus's tomb, be the close of the great life-work of an artist in language? It is a question of taste. A love for episodes and anecdotes is Herodotus's chief weakness, and Greek literary art liked to loosen the tension at the end of a work, rather than to finish in a climax.
As to the 'Assyrian Logoi,', the most notable fact is that Aristotle seems to have read them. In the Natural History (viii.18) he says that "crook-clawed birds do not drink. Herodotus [Some MSS. Ἡσίοδος which is hardly possible.] did not know this, for he has fabled his ominous eagle drinking in his account of the siege of Nineveh." That must be in the 'Assyrian Logoi.'
This clue helps us to a rough theory of the composition of the whole work, which may throw some light on ancient writings in general. If Herodotus was telling and writing his 'Historiai' most of his life, he must have had far more material than he has given us, and parts of that material doubtless in different forms. It is "against nature" to suppose that a 'Logographos' would only utilise a particular 'Logos' once, or never alter the form of it. The treatment of the Pedasus story shows how the anecdote unintentionally varies and gets inserted in [142]
