552 W. THALBITZER. 
1915. ! In the matter of tonality I find here some resemblance to the 
Greenland songs. None of them, however, are identical with any other 
Eskimo melody known in literature. ? 
5. 
When I place my texts deciphered from the phonographic records 
beside those taken down from dictation, they stand out from each 
other in sharp criticism. The same number written first from dictation 
and later recorded on a phonograph, shows, contrary to expectations, 
great deviations, — the two texts do not cover each other. The present 
volume contains scores of illustrations of this same fact. To be sure 
the discrepancies between the variants are due in many instances to 
the usual and natural reason that they were reported by various in- 
formants, so that the divergence in the corresponding texts may be 
readily explained by different lines of tradition, which I, as an em- 
pirical collector, had merely to note down as they were, — eventually 
too, in other cases, to the individual changes made by the different 
informants, to personal “mistakes” in the repetition or to local pro- 
nunciation, — these latter classes of deviations are already more serious, 
because due to so-called or actual sources of error. However there 
are others which are even worse. It happened — and I was astonished 
the first times it occured — that even in instances in which the native 
informant was the very same person, who, on a given occasion had 
sung into a phonograph, and later dictated the text of the song to 
me — or even one who had, on the same occasion, done first the 
one and then the other, — (and as a rule as soon as a song or tale 
was recorded on the phonograph I took care to have so slow a repeti- 
tion given that I could write it down on paper) — that even then, 
various discrepancies appeared between the dictated text and that de- 
ciphered from the phonographic wax-rolls. 
Deciphering took place first on my return home to Copenhagen 
several years after the phonographic record was made. It was thus 
impossible to obtain from the informant an explanation of the causes 
of the variations (in his texts of the same song or tale). The phono- 
graph can not lie, no doubt could exist as to the correctness of the 
record. On the other hand I trust my own ear; my dictated text is 
1 E. Hague: Eskimo songs. Journ. Amer. Folklore, vol. 28 (pp. 96—98). 
2 I have touched on the question of identity between distant melody-variants 
within the Greenland area on p. 56 (with foot-note). Up to date no variant of 
a true Eskimo melody has been found more than a few miles from the place 
in which its pendant is found. Which of them is the original is impossible to 
prove, as the composer (the same as the author) is always unknown, and pre- 
sumably dead many years previous. 
