√YE(L) “daughter”
The root √YEL was one of several competing Elvish roots for “daughter”. In The Etymologies of the 1930s ᴹ√YEL “daughter” was first used as the basis for ᴹQ. yelde/N. iell “daughter”, but it was deleted (Ety/YEL). N. iell was given a new derivation from ᴹ√SEL-D “child”, by analogy with N. ionn “son” (Ety/SEL-D), while a new Quenya word for “daughter” was introduced: ᴹQ. yende from a feminine variant ᴹ√yēn of ᴹ√YO(N) (Ety/YŌ). Note that ᴹ√SEL-D itself was initially glossed “daughter”, but was changed to “child” and given derivatives for all genders in Quenya: ᴹQ. selda [n.], ᴹQ. selde [f.], and ᴹQ. seldo [m.].
The picture in later writings is also rather muddled. 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). In several places Tolkien gave Q. Tindómerel “Daughter of Twilight” as the Quenya equivalent of S. Tinúviel, with the final element being derived from primitive ✶-sel(dĕ) > -rel (Ety/SEL-D; PE19/33, 73; VT47/37).
In this period, however, the more common suffix for “daughter” was Q. -iel as in Q. Elerondiel (S. Elrenniel) “*Daughter of Elrond” as applied to Arwen (PE17/56) and Q. Uinéniel “Daughter of Uinen” (UT/182). Furthermore, in a list of masculine and feminine suffixes written around 1959, Tolkien gave (primitive?) yē, yel and (Quenya?) yelde for “daughter”, though in that note the feminine patronymic suffixes were revised from {-yel, iel, -yelde >>} -well-, -uell-, -wend-, -wel, and yen was given as another variant (PE17/190). In other notes from the late 1950s associated with “Changes affecting Silmarillion nomenclature”, Tolkien had feminine patronymic suffixes -en, -ien, but said that Quenya used -ielde, -iel (PE17/170).
Neo-Eldarin: All of the above indicates considerable vacillation between √SEL, √YEL, and √YEN for “daughter” words and suffixes in the 1930s to 1960s: of the three Tolkien seem to favor sel- for “daughter” words but -iel for “daughter” suffixes. For purposes of Neo-Eldarin, I think it best to use √YEL = “daughter”, mainly so that √SEL can remain the basis for “child” words of all genders, at least in Quenya.
ᴹ√YEN “year”
A root in The Etymologies of the 1930s glossed “year” with derivatives like ᴹQ. yén/N. în “year” (Ety/YEN). Tolkien’s ongoing used of words like Q. yén and S. ínias “annals” indicate its ongoing validity (LotR/377; MR/200), but in Quenya at least the meaning shifted to that of an “Elvish long year”, equal to 144 solar years (LotR/1107; MR/471; NM/84).
√YER “desire (for marriage and bodily union), [ᴹ√] feel sexual desire”
The root ᴹ√YER “feel sexual desire” was a later addition to The Etymologies of the 1930s with derivatives like unglossed ᴹQ. yére and N. îr, probably nouns meaning “sexual desire” (EtyAC/YER). The root √YER was mentioned again in notes from 1957 with the gloss “desire (for marriage and bodily union)” (NM/20), serving as the basis for Q. yermë with the sense “sexual desire” specifically within the context of marriage for the purposes of procreation (NM/16). Tolkien said that the Elves had little in the way of sexual interests outside of this context, and that the notion of sexual desire in the absense of love between spouses was an alien idea to the Elves, in keeping with Tolkien’s own Catholic sensibilities. For “improper” sexual desires, the Elves would likely use other terms like ᴹQ. maile/N. moel “lust” [ᴺS. mael] from the root ᴹ√MIL(IK).
ᴱ√YERE(NE) “*iron”
This root appeared in two different forms in in the Qenya Lexicon of the 1910s: as ᴱ√DYEÐE = ‘EÐE equivalent of ᴱQ. anga “iron” (QL/105), and as ERE(N) “iron or steel” = ᴱ√YEREN (QL/36). In the contemporaneous Gnomish Lexicon Tolkien had related forms G. ger “ore, metal” and G. geth “ore of metals (esp. raw iron)” (GL/38); the second of these was deleted, which probably means Tolkien selected ᴱ√YEREN over ᴱ√DYEÐE. In any cases, none of these forms appear in later writings, where Tolkien largely used √ANGA for “iron”.
*√YESET “beginning”
The root ᴹ√ESE(T) “precede” and ᴹ√SET “precede forward” appeared in The Etymologies of the 1930s with derivatives like ᴹQ. esse “beginning”, ᴹQ. esta “first”, and N. seth {“first” >>} “forth”, but the ᴹ√SET forms were rejected and the ᴹ√ESE(T) forms marked with a “?”, being in conflict with ᴹ√ES “indicate, name” (Ety/ESE; EtyAC/SET). The appearance of Q. yesta “beginning” and Q. Yestarë “first day of the year” in later writings (LotR/1108-1109; PE17/120) indicate the root was probably revised to *√YESET.
ᴱ√YḶTḶ “*join”
The unglossed root ᴱ√YḶTḶ appeared in the Qenya Lexicon of the 1910s as a variation of ᴱ√ẎATA “join”, possibly via L infix, having a single derivative ᴱQ. yalta “yoke” (QL/106). In the contemporaneous Gnomish Lexicon, the words G. galt “yoke” and G. ilt- “to yoke, join” are almost certainly related (GL/37, 50). In the 1930s, the words ᴹQ. yanta/N. iant were used for “yoke” as derivatives of ᴹ√YAT (Ety/YAT), but later still Q. yanta/S. iant where used for “bridge” (LotR/1123; SD/129). Thus, I think it is worth positing a Neo-Root ᴺ√YALTA to serve as the basis for “yoke” words.
ᴱ√YṆTṆ “*(en)large, increase”
The unglossed root ᴱ√DẎṆTṆ appeared in the Qenya Lexicon with derivatives like ᴱQ. yanta “large” and ᴱQ. yanta- “enlarge, increase, add to”; Tolkien also wrote and then erased a note saying “or ẎṆTṆ” (QL/106). In the contemporaneous Gnomish Lexicon Tolkien gave primitive yṇt as the basis for words like G. gant “larger, greater, more; (lit.) increased” and G. gantha “more” (GL/37). The superlative forms ᴱN. diant “most” and ᴱN. -iant from the Early Noldorin Grammar from the 1920s were probably related. I think it is worth positing a Neo-Root √GYANTA “(en)large, increase” to salvage some of these early words.