ᴹ√SUB “sink (esp. in water)”
The root ᴹ√SUB “sink, esp. in water” appeared in a rejected page of verbal roots from the Quenya Verbal System of the 1940s where it had a derived past tense form ᴹQ. sumbe “*sank” as well as a distinct verb form ᴹQ. sumba- “to submerge” (PE22/127). It might be a later iteration of the unglossed root ᴱ√SUQU in the Qenya Lexicon of the 1910s, with derivatives like ᴱQ. sunqa “going down, descending; abysmal, deep, profound” and ᴱQ. suq- “fall, fail, go down, die down” (QL/87).
Neo-Eldarin: I think it is worth retaining ᴺ√SUB “sink” as a Neo-Root.
√SUK, √SOK “drink, drain, gulp, quaff”
The first appearance of this root was unglossed ᴱ√SOKO in the Qenya Lexicon of the 1910s with derivatives ᴱQ. soko- “drink” and ᴱQ. sokto- “give to drink, drench” (QL/85). In the contemporaneous Gnomish Lexicon the base form was given as sog- which had derivatives like G. sog- “drink” and G. suith “a drink, a draught” (GL/68).
In The Etymologies of the 1930s the root was given as ᴹ√SUK with derivatives like ᴹQ. suhto/N. sûth “draught” and ᴹQ. suk-/N. sog- “drink”, along with a variant root ᴹ√SUG with derivatives ᴹQ. súlo/N. sûl “goblet” (Ety/SUK). The root appeared as vocalic variants √SUK and √SOK “drain, drink” in both the first and second versions of Tengwesta Qenderinwa from the 1930s (TQ1: PE18/45) and circa 1950 (TQ2: PE18/94). Finally the root √SOK “gulp, quaff, drink” appeared in notes associated with the Quendi and Eldar essay from 1959-60 where it had a single derivative: Q. soika “thirsty” (VT39/11).
In Tolkien’s later writings, the roots √SUK and √SOK had competition from √YUL “drink”; see that entry for details. As both √SOK and √YUL coexisted in the Quendi and Eldar essay, I am of the opinion that √SUK/SOK may not have been abandoned.
ᴱ√SUKU “*resin, gum”
ᴱ√SUKU⁽¹⁾ appeared as unglossed root in the Qenya Lexicon of the 1910s with derivatives like ᴱQ. súke “resinous tree, pine or fir” and ᴱQ. sulka “sticky, viscous” (QL/86-87). It had derivatives in the contemporaneous Gnomish Lexicon such as G. thuctha- “adhere to, cleave” and G. thugli “resin” (GL/73), indicating the true root form was *ᴱ√ÞUKU. I think it is worth positing a Neo-Root ᴺ√THUK “resin, gum” to salvage some of these early words.
ᴹ√SUS “hiss”
A root in The Etymologies of the 1930s glossed “hiss” with the derivative ᴹQ. surya “spirant consonant” (Ety/SUS). The root ᴹ√SIS was written nearby and was likely intended to be a variant as suggested by Carl Hostetter and Patrick Wynne (EtyAC/SIS). These two variants might be later iterations of primitive ᴱ✶sṣt- from the Early Noldorin Dictionary of the 1920s with derivatives like ᴱQ. histe- “hiss” (PE13/163).
ᴱ√SUẈU “*feminine patronymic”
An unglossed root in the Qenya Lexicon of the 1910s with derivatives like ᴱQ. sui “daughter”, ᴱQ. súyon “nephew, daughter’s son”, and ᴱQ. fion < ᴱ✶þẉ-iı̯on-d, hinting at a variant root form ᴱ√ÞUẈU (QL/87). In the contemporaneous Gnomish Lexicon there were forms like G. fwi- and G. -thwi as feminine patronymics as wells as G. fwion “nephew” and G. fwîr “niece” as children of one’s sister (GL/36). G. suil “daughter” was also related (GL/68), and deleted form thuil “daughter” (GL/73) seems to indicate ᴱ√ÞUẈU >> ᴱ√SUẈU.
√SYAD “compress; [ᴹ√] shear through, cleave (through)”
The most notable use of this root was in the derivation of the name Q. Sangahyando “Throng-cleaver”, the name of a sword in Tolkien’s earlier writings back in the 1910s (QL/81) but ultimately becoming the name of a warrior of Umbar in The Lord of the Rings appendices (LotR/1048). The first iteration of this root was as ᴱ√HYAÐA “plough through” in the Qenya Lexicon of the 1910s, with derivatives like ᴱQ. hyanda “a blade, share” and ᴱQ. hyar “a plough” (QL/41). In the contemporaneous Gnomish Lexicon it had derivatives like G. hanna- “mow, cleave” (GL/48). In The Etymologies of the 1930s, it appeared as ᴹ√SYAD “shear through, cleave (through)” with derivatives like ᴹQ. hyar- “cleave”, ᴹQ. hyatse “cleft, gash” and N. hast “axe-stroke” (Ety/SYAD); it replaced a deleted variant ᴹ√KHYAD “cleave” (EtyAC/KHYAD).
However, in the first version of Tengwesta Qenderinwa (TQ1), also from the 1930s, Tolkien gave ᴹ√SYAD the gloss “compress” (PE18/51), and in the contemporaneous Outline of Phonetic Development (OP1) {ᴹ√KHYAD >>} ᴹ√SYAD had the derivative ᴹQ. hyarna “compact, compressed” (PE19/45). Similar forms appeared in the second version of Tengwesta Qenderinwa (TQ2) and Outline of Phonology (OP2) from the early 1950s (PE18/101; PE19/92). However, in between these two sets of documents, in the Quenya Verbal System of the 1940s, Tolkien gave ᴹ√SYAD the gloss “cleave through” as the basis for the verb ᴹQ. hyar- “cleave” (PE22/102).
The other element Sanga- of the name Sangahyando was derived from the roots √THAG or ᴹ√STAG, which Tolkien only ever gave glosses like “(com)press”. Thus I think it best to assume the glosses in TQ1, OP1, TQ2 and OP2 were aberrations, and that √SYAD generally meant “cleave”.