A root in The Etymologies of the 1930s glossed “smoulder” with derivatives like ᴹQ. yúla/N. iûl “ember(s)” and ᴹQ. yulma/N. iolf “brand” (Ety/YUL). In all likelihood it was supplanted by the later root √YUL “drink”.
Neo-Eldarin: Despite it’s probable replacement by √YUL “drink”, I’d still recommend using some derivatives of ᴹ√YUL “smoulder” for purposes of Neo-Eldarin. Noldorin forms like iolf and iûl are not viable as derivatives in Sindarin, since initial yu- became the (vowel) y, for example: ýneg < ✶yūneke (VT47/41). But the Noldorin forms could be adapted into (Neo) Sindarin if they were derived from a Neo-Root ᴺ√YOL. Quenya forms could likewise be transferred to this root, or retained unmodified if we assume an occasional √YOL/YUL variation, similar to vowel variations like √SOK/SUK (PE18/45, 94). Assuming the existence of such a ᴺ√YOL/YUL Neo-Root would allow many of these 1930s forms to be salvaged without directly conflicting with √YUL “drink”.
A root in The Etymologies of the 1930s, initially glossed “daughter” but later “child” with derivatives ᴹQ. selde, ᴹQ. seldo, ᴹQ. selda = female, male and neuter “child” (Ety/SEL-D). In Notes on Names (NN) from 1957 Tolkien gave sel-de “daughter” (PE17/170), while S. sel(l) = “daughter” appeared in both the King’s Letter from the late 1940s (SD/129) as well as the Túrin Wrapper from the 1950s (VT50/5). The diminutive form for “daughter” appeared as Q. selyë in notes from the late 1960s (VT47/10).
Neo-Eldarin: For purposes of Neo-Eldarin, I prefer √YEL for “daughter” as a variant of ᴹ√SEL(D) under the influence of √YON “son”, mostly so I can still use the 1930s “child” words for other genders, at least in the Quenya branch. I would still use Q. seldë and S. sell for “daughter”, however, with a bit of semantic drift, with “girl” words becoming Q. nettë and S. neth.