None of the etymologies advanced is adequate on its own to explain the many diverse uses and translations of this adjective (see LSJ entry at web address 1; cf. alphaiota 244, alphaiota 245, alphaiota 247), with its related verb ai)o/llw (alphaiota 246), its compounds (alphaiota 249, alphaiota 250, alphaiota 251, kappa 2122), and the related but differently accentuated proper name Aiolos (alphaiota 252, cf. alphaiota 248). We can distinguish two different words, corresponding to the first two definitions here. A poetic phenomenon, exploiting ambiguity, intervenes in Homer and other poetry and greatly confuses the issue of which translation to use at any given point.
The first two definitions are also found in Hesychius (alpha2034) and the Synagoge le/cewn xrhsi/mwn (Anecdota Graeca, ed. Bekker, 1.356.21; ed. L. Bachmann 45.1; Photius, Lexicon alpha614, 52.19 Reitzenstein), scholia to Homer, Iliad 7.222, and to Odyssey 22.300; cf. Apion (de glossis Hom., fr. 74.214.16, but with poluki/nhton), and, with taxu/s for eu)ki/nhtos, Hesychius alpha2020, Eustathius on Odyssey 10.2 and many Homeric scholia (see Mette under *sx). Modern dictionaries such as LSJ assign these as the two primary meanings of a single word 'mobile, nimble, agile; multicoloured' (cf. Page in bibliography).
In the third century AD, Porphyry [OCD(4) s.v., pp.1190-1; pi 2098] had advanced the third definition. He rejected the meaning 'multicoloured' (hence the Suda's "or"), and explained all instances in Homer from the base of 'whirlwind' (see note 2 below) as 'swift, whirling' (Zet. 284 col. 1 - 286 col. 1.9; Liber 1; on Odyssey 20.27; cf. Etymologicum Magnum 37.3 and the exceptional scholion B to Iliad 22.509, eu)/strofoi kai\ poiki/loi). He correctly took a)/ella from the same root of rolling or turning as ei(le/w. Modern etymologists question the presence of the root of blowing (although a "blow-twister" is a possible colloquial compound) in favour of alpha intensive.
At the same time Methodius [OCD(4) s.v., p.942; mu 432], as reported by Etymologicum Genuinum 15 and Etymologicum Magnum 37.4f., took the adjective, when used of motion or reflected light, to designate the "self-generating movement" (a)nupo/staton ki/nhsin) of a snake, and derived it from a verb ai)/ein (apparently hypothetical, probably from ai)ei/, a)ei/, 'always'), together with a)i+/ssein 'to dart, flash (of light)'. Thus ai)o/lh nu/c might be 'night in perpetual motion' (cf. alphaiota 247). In 1937 Émile Benveniste, perhaps independently, advanced a similar etymology and sense from the IE root for 'life-force', aiw-, seen in ai)w/n, ai)ei/, Latin aevum, Vedic ayu- 'lively, mobile'. Those who followed him saw in the use of the headword for inanimate objects such as bronze helmets (kappa 2122), shields and mitres (alphaiota 250) a sense of vital force, to translate as 'shimmering, flashing' (see Mette in LfgrE), exciting to translators but unacceptable to archaeologists (see Page, note 97) and to etymologists, as having no analogy or explanation for the ending ?olos, "which may seem embarrassing for the theory" (Chantraine).
In 1950 Ernst Fraenkel advanced the most persuasive etymology. The word, originally *vaivolos, is from the IE root uel-, seen in a)/ella, ei)le/w (as Porphyry saw), Latin volvo and English wheel, meaning 'roll, undulate, turn'. The reduplication in ai is intensifying and parallel to that in dai/dalos, daida/llein, paipa/lh, paipa/llein = sei/ein. Fraenkel explained the loss of initial digamma in Homeric verse as dissimilation. This history suits a snake (Iliad 12.208; Paean Delph. i in Apoll. 1.19 Powell, Coll. Alex.; Callimachus fr. 575.1 Pfeiffer; Nicander, Theriaca 155, alphaiota 246), the undulating body of Echidna (Hesiod, Theogony 300, cf. scholia, polue/likton), and a mass of flesh-eating worms (Iliad 22.509), and might apply visually to the shapes of a wasp (Iliad 12.167) and a Mycenaean figure-of-eight shield (alphaiota 250), both pinched at the "waist". This etymology did not, however, explain, any more than Porphyry, the uses glossed as poiki/los ('multicoloured, intricate') or skoteino/s ('dark, shadowy': alphaiota 245), or the use of the related verb (alphaiota 246) of ripening grapes and livid flesh.
These latter meanings are now explained as from a Mycenaean word, a3-wo-ro on a Linear B tablet from Cnossos, almost certainly for mingled colours (cf. alphaiota 252). The tablet lists two bulls or horses by name or colour; cf. the use of a colour word as the name of Achilles' horse Xanthus. One is "Black" from the adjective kelaino/s. Ours is then, almost certainly, either "Roan" (of patchwork colouring; cf. the word’s use for a tortoise shell at Homeric Hymn to Mercury 33) or "Piebald" (black and white). Now this name cannot come from the same root as *vaivolos, i.e. uel-, for Mycenaean does not lose initial digamma, as Geiss points out. It belongs to a distinct word.
Concerning Homer, the odd thing is that one can take any of these four or five interpretations of the basic meaning and, with a certain ingenuity, apply it to most or all of the instances in the two epics. An Homeric shield (Iliad 7.222, 16.107) may be imagined and explained as 'easy to turn or move, multicoloured, intricately worked, with its bronze or gold outer coating vibrantly shimmering or flashing in the sun, wave-like in its shape'. When Pandarus falls from his chariot dead (Iliad 5.294-96) and his gleaming armour clangs on top of him, is it also multicoloured, discoloured by blood, rolling, well-crafted or (unlike its owner) still flashing with life? Which of these does Homer mean? The answer is all, or at least some. The British poet William Empson has explained the ways in which poets exploit the various meanings and intertextual echoes of given words, giving them a resonant richness of reference (see bibliography; cf. Dyer). Who cares what colour Hector’s helmet may really have been, if it danced to and fro or its crest trailed undulating behind like a hoopoe's, if it was of bronze and flashed in the sun or contained the force of life (kappa 2122)? The poet wants the mind of each hearer to travel out into his own imaginary world in a sequence of images not necessarily related to some historical or religious truth. Of a yellow-jacket wasp (Iliad 12.167) we see simultaneously the motion in the air, the colours and the wasp waist. To most it does not matter what colour the feet of Achilles' horse Xanthos [Myth, Place] (Iliad 19.404) were; the description po/das ai)o/los interacts with a)/ella and suggests that the horse's feet went like a hurricane in speed, noise and rolling motion. Yet Homer probably expected a horseman to take the name as the colour chestnut, sorrel or bay and 'multicoloured' in the adjectival phrase to conjure up a handsome sorrel with white "socks".
The same technique of deliberate ambiguity is found in Athenian theatre and perhaps elsewhere: e.g. Aeschylus, Seven against Thebes 494 (of smoke, the sister accompanying fire and twisting in an intricate design of mingled colours, brilliant and sombre); Sophocles, Ajax 1025 (where the point of Ajax's two-edged sword is no longer flashing or whirling with life but multicoloured, stained with his drying blood), Trachiniae 94 (alphaiota 247, of the colours of night, but also night as ‘ever-turning’); Aristophanes, Frogs 248 (of the flashy choreography and colours of the frogs' dance, alphaiota 244), Thesmophoriazusae 1055 (a pastiche of Euripides on the dark journey to Hades, alphaiota 245).
Outside the Homeric tradition, the headword means primarily 'changeable' (see LSJ at web address 1; cf. eu)ki/nhtos in the definition here), of music of varied modes, of gods and heroes capable of metamorphosis, chiefly in the Orphic Hymns (alphaiota 249), or of unreliable people, whose words and opinions cannot be trusted; it may also 'wily, quick-witted' in a positive sense. The sense of Mycenaean 'multicoloured' is also enduring, e.g. of fine clothes at Callimachus, Aetia 7.11.
[1] The adjective poiki/los, used so often of the headword, has several distinct meanings (web address 2), all deriving ultimately from the craftsmanship of intricate weaving or objects, often inlaid; cf. pi 3082, tau 435, alphaiota 10, delta 106, delta 107, eta 386, etc. We should beware of thinking that it always means 'multicoloured', even if that appears to be the intention in the definitions here (but note Eustathius on Iliad 7.222, daidalqe/n, 'crafted'). The sense 'changing colour' is not documented, although a known sense of ai)o/los.
[2] See alpha 545, alpha 546, alpha 547; and the discussion in Porphyry.
[3] The verb a)/ein is very rarely used in the sense of a)h=nai (for a)vh=nai) 'to blow', required here; it means 'to sleep' (or other homonymous meanings: see LSJ).
[4] For the various homonyms of this verb see LSJ ei)/lw (web address 3). See "C" ei(le/w, used here, and "D". Our text of Porphyry sometimes uses the smooth breathing.