The Germanic initial sound combination wr (e.g. in *wrītan 'write') has not been preserved in any standard language. In the Scandinavian languages the development of this sound combination has resulted in five different initial sounds or sound combinations, namely rw, r, w, br and vr. The aim of this study is to describe the occurrence of these sound combinations, and to explain their age and origin. A limited number of words, mainly from Swedish dialects, has been investigated. The Swedish material has been assembled in a collection that has largely also been mapped, and so has a collection of Norwegian dialect material. The result shows that br is frequent in the Swedish dialects and that it also occurs in Trøndelagen and Østlandet in Norway, and on Jutland. R is widespread in Norway and occurs in Sweden in some words in Norrland, Dalarna, Värmland, Dalsland, Uppland and on the island of Gotland. Rw has been documented in Upper Dalarna, and w in the same area, as well as in the Kalix dialect in northernmost Sweden. The author demonstrates that the Norwegian loss of w might have started already in the 6th century in western Norway. The loss on Gotland is independent and can be demonstrated in Old Gutnish. The metathesis rw is found in Swedish and Norwegian 13-14th century sources from Uppland, Västmanland, Östergötland and southeastern Norway. The developments wr &gt; r and wr &gt; rw probably have their roots in a difference between the west Norwegian wr and the east Norwegian and Swedish war which is documented from PrimScand times. W is a secondary development of rw. The change wr &gt; br can be dated by place-name material to the early 15th century. The change wr &gt; vr appears to have taken place at approximately the same time. The author demonstrates that the result, wr &gt; br or wr &gt; yr, is apparently governed by certain phonetic factors. The failure of the Germanic wr to survive depends on the combination being phonetically complex. The risk that w would be lost was therefore considerable. However, differentiating forces worked for its retention. All developments emanating from wr, apparently disparate, can be seen as features of a larger process where a general linguistic tendency towards a weakening of sounds is confronted by conservative forces, with the above results as a consequence. / <p>Doktorsavhandling vid Uppsala universitet 1991</p>