The "dread" of the headword phrase (ai)anh/s) is a poetic adjective of uncertain etymology. The connection implied by the second gloss here between this word and the adverb ai)ei/ ("always") has been accepted by some scholars (cf. LSJ s.v.) and has left its mark on translation and interpretation (see, for instance Jebb's translation of the word in the Sophocles passage quoted below: "weary" (i.e. unrelenting)). Chantraine, however, disputes this etymology and sees instances where the term has unequivocal connotations of "everlasting", "persistent" or "wearisome" as secondary usages based on folk etymology. Fairly early examples occur, however, e.g. Aeschylus, Eumenides 672.
[1] Sophocles, Ajax 672 (text at web address 1), with Lloyd-Jones' translation from the Loeb edition. The Suda's glosses are derived from scholia on this verse. The same verse is quoted at iota 296 and lambda 322 with a different accentuation: ai)anh=s rather than ai)anh/s (see the second part of this entry). With that accentuation the adjective "dread" applies to "night" rather than "circle". Most mss. of Sophocles actually exhibit this alternative accentuation, but some Sophoclean mss. and most modern editors present the form recorded in the headword phrase here: ai)anh/s. See Jebb and Kamerbeek ad loc. The belief that the adjective modifies "night" here may have inspired the frequent glossing of the word as "dark."
[2] Probably derived from a scholion on Sophocles, Electra 505, where the headword occurs (web address 2), though the etymological connection between the word and ai)a/zw (on which see alphaiota 4) is also reflected in scholia and commentary elsewhere (e.g. scholia to Pindar, Pythian 4.236, Isthmian 1.49). No modern scholar seems to take this etymology seriously. On the accentuation issue, see previous note. There is probably no significant semantic distinction between the two forms. cf. alphaiota 4.