A root for “brother” that Tolkien introduced in notes on finger-names from the late 1960s as a companion to √NETH “sister” (VT47/14, 26, 34). It conflicts with, and possibly replaces, earlier uses for √KHAN such as √KHAN “back” in notes from around 1959 serving as the basis for the prefix Q. han- in hanquenta “answer” (PE17/166). The root ᴹ√KHAN also appeared in The Etymologies of the 1930s glossed “understand, comprehend”, with various derivatives in both Quenya and Noldorin of similar meaning (Ety/KHAN).
It is unlikely that all these uses of √KHAN coexisted, but I think at a minimum both √KHAN “brother” and ᴹ√KHAN “understand, comprehend” should be retained for the purposes of Neo-Eldarin, as the latter has no good replacements in Tolkien’s later writing. As for hanquenta “answer”, it might be reinterpreted as “a saying providing understanding”, and so be derived from ᴹ√KHAN “understand”.
A root in Quenya Notes (QN) from 1957 given as the basis for the “blood” words Q. serke and S. sereg as well as the flower name S. seregon “blood of stone” (PE17/184), a flower name that also appeared (untranslated) in The Silmarillion (S/203). It may replace the root ᴹ√YAR “blood” from The Etymologies of the 1930s with derivatives like ᴹQ. yár and N. iâr of the same meaning, the later an element in the surname of Túrin: N. Iarwath “Blood-stain” (Ety/YAR). In later Silmarillion drafts, it became Iarwaeth (WJ/83) and then S. Agarwaen “Blood-stained”, the form it took in the published version of The Silmarillion (S/210). The etymology of S. agar- “blood” is unclear.
Neo-Eldarin: For purposes of Neo-Eldarin, I’d stick to √SEREK = “blood”.