Language and Folklore. (Supplement). 549 
concerning family variants. The eventual stabilisation of the variants | 
in the tradition is a problem not yet solved, and I doubt whether 
“my material is sufficiently large for its solution. After having worked 
out the list of parallels which is found at the end of this section I 
am able to make the following summary, which is more complete 
and correct than the earlier one. 
Variants of Melodies. 
Jtalized numbers = phonographic records. Upright ones = DD records. 
Text | Mel. | Text | Mel. Text | Mel. Text Mel. | Text | Mel. || Text | Mel. 
| [A is {| CE 5 | asf] 2 | sof 
Se PH hall в 
Не arte 
(EN 
Texts 1—11 include petting songs and nursery rhymes; 16—33 epic-lyrical poems; 
115—196 drum-songs (uaajeertut and juridical songs). 
A closer investigation of the single groups within nos. 1—33 seems 
to me to substantiate my own immediate impression from Ammassalik, 
that the old poems, both those chanted in recitative and those with 
decided melodies, have somewhat fixed, lasting musical forms. 
It must be remembered that my so-called variants largely depend 
on the heterogeneous manner of their transcription, some being DD 
texts, others Phon. texts. It is clear that the dictated note merely 
forms a hasty sketch, as compared with the elaborate phonogram ; 
while the latter may be repeated as many times as desired in exactly 
the same form, the former is heard but once, and noted down in 
"haste. In the question of provincial or family variants, this cause of 
variation is irrelevant. We cannot thus use the DD records to sub- 
stantiate the true melody variants, for only the texts and melodies 
are of consequence which were made on а phonographic record in 
which various reciters and singers have sung the same number to 
the phonograph independently of each other, so that we are enabled 
_to make an exact transcription and then compare on an equivalent 
| basis. No. 11 shows great similarity in the two variants which have 
the same informant (at different times of the year) but only a slight 
similarity between them and a third variant of another singer. In 
no. 33 there is indeed only a slight resemblance between variants of 
the same informant, but in this case— just as in nos. 115, 120 and 
140, where, however, the resemblance is greater — one of the variants 
is a Phon. the other a DD text. 
