Starting from the vocalic phenomena of the 8th Century Japanese we can set up a table of reconstructed vowels of the Pre-Japanese as that shown on the last page of this paper. The vowels of the domain A are the back, those of the B being the front ones. There had been a strong tendency toward the so-called vocalic harmony : a primary base-word (or a primary root-word) was vocally constructed only from among the vowels either of the domain A or only of the domain B. (Historically attested evidences of the 8th Century onward showed already the relax of this rule). E.g. ao 'blue', kamo 'duck', yoru 'night', tukï 'moon'; but: koto 'cause, thing, word (s)', mono 'thing', toki 'time, tide', kokoro 'soul, mind', etc. (Later forms of "relaxed" formations : kasö 'father', morö 'round', töfo 'far, distant, remote', usi 'ox, cow', etc.). Interchange (Interchanging alternation) There had been a seemingly curious phenomenon or tendence to differentiate semantic functions of the primary forms by way of interchanging their constituent vowels across between the two vocalic domains : [Table omitted] Intra-change But the internal exchange of the vowels within each domain respectively did only provoke a slight, rather stylistic, but not categorical, differenciation of the semantic functions of the forms.[Table omitted] Similar phenomena may well be found also in Korean, e.g. mat 'taste (of the palate)': mɔt 'taste, flavour (aesthetic)', sal-da 'be alive': sɔl-da 'be raw, unripe', pulda 'blow (of the wind)': pAl-(A)m 'wind', etc. But the scarcity of evidences in ancient Korean prevents us from finding out an adequate principle of their semantic differentiation. Professor Alo Raun, Indiana University, has kindly informed me that in Finno-Ugric there may tentatively be distinguished three kinds of vowel alternations : (1) onomatopoetic-descriptive, e.g. in Estonian kilin vs. kolin (the former denoting a higher pitched noise), in Hungarian, kever 'mix in a gentle or reasonable way' vs. kaver 'mix carelessly, mess up', (2) deitic : in Finnish, tässä 'here' vs. tuossa 'there', in Hungarian, itt 'here' vs. ott 'there', ez 'this' vs. az 'that', (these procedures are no more productive), (3) 'paradigmatic': in Finnish, palaa 'burn (intr)' vs. polttaa 'id (tr.)'. Being a highly leading scholar in the field of Finno-Ugric linguistics, he is very cautious about the matter because of the paucity of evidences and remarks rightly that 'these phenomena must be treated separately for each language family because there is no exact analogy.' The same might be the view of Prof. Bjorn Collinder in his recent 'Comparative Grammar of the Uralic languages' (Stockholm 1960). Was the semantic differentiation by way of vocalic alternations a specific phenomenon of Ancient Japanese only? or was it here a accidental one only? …… I would much like to be instructed in.