I was preparing a family this week that I thought included the English word wall. But as I researched it, I found that Germanic *waigaz: "wall" connects to that family, but English "wall" is from the unrelated Germanic *wallą from Latin vallum: "wall, rampart".
June's theme is linguistic effects of the Celtic Hegemony of the early Iron Age. The rest of the month will focus on words that were borrowed from Gaulish into Latin (things like very iconic Latin word gladius being borrowed from Gaulish, which appears in …
Proto-Indo-European *h₂éwis: "bird" and Proto-Indo-European *h₂ówis: "sheep" appear to be umlaut variants, presumably through the more sparsely attested root *h₂ew-: "to put on clothes, to cover". *h₂éwis: "bird" would be "a clothed (in feathers)" and *h₂ówis: "sheep", "a clother (in wool)".
Germanic really went to town with this root, creating a bunch of morphological forms that aren't attested in any other branches, and in a few cases I can't even relate them to any standard derivational forms I know in either PIE or Proto-Germanic.