to regard for a moment the evidence of the bigoted Danish monk where it differs from that of Snorri, an honest and intelligent witness who knew his subject and had no parti pris. Another practice from which the subject has suffered is that of ascribing to Celtic influence everything interesting in Norse sources which touches on the supernatural.
The materials for the present study are collected chiefly from those Icelandic sagas which deal with the time before the establishment of Christianity in Iceland in the year 1000, and from the Lives of the Kings of Norway, as far as the death of Olaf Tryggvason in the same year, in the Heimskringla; with occasional reference to the Prose Edda where this serves to throw light on the other sources. In the Icelandic family sagas the references are scattered. The saga-writers are sparing of detail, and mention nothing that does not actually bear on some feature of the story: there is also a general vagueness in the matter of dates and seasons. The references are arranged as far as possible in order of date of occurrence, which can usually be approximately fixed by reference to genealogies; the date of occurrence being naturally of importance, where the question is of the growth or decay of a custom.
The present literary form in which the sagas are cast dates in the case of the greatest sagas to the thirteenth century. The longer sagas are compilations from a number of smaller ones; in the case of Njála, Laxdæla and Egla, the compilation is carefully done, Laxdaela especially arranging its material with considerable attention to artistic effect; the other two are more to be depended on, therefore, in matters of detail. In Eyrbyggja there is little attempt to present an artistic whole. The shorter sagas which supplied the material for these must, even in their written form, have been composed a century earlier; and to the twelfth century at latest