The following is quoted from the website, http://forum.stirpes.net/archaeogenetics/24911-ancient-origins-halogroups-found-shetland-islands.html

Like so many helpful sites it has disappeared. Unfortunately, I did not note the author.

R1a1 (Hg3, Eu19) is, along with Q, the only haplogroup that can unequivocably be linked to a Norse ancestry, and more specifically to the west coast of Norway. It is virtually unknown in the Celtic regions such as Ireland (only 10 in a recent sample of 1200), and barely makes an appearance in Friesland, but occurs at a relatively high frequency in the Scandanavian countries: over 25%, in Norway, 20% in Sweden, and 15% in Denmark. Curiously those who have a haplotype within this haplogroup often have fairly close matches in Mongolia, India, Siberia, and Eastern Europe (where it attains its highest frequency of over 50% in Poland and the Ukraine). It is believed that the haplogroup emerged among the Kurgan peoples of the Eurasian Stepes (the Ukraine), where their ancestors had dalied during the Last Glacial Maximum (although India has also been proposed). From there they spread north, south and east - but little further than what is today Poland. It is anticipated that about a third of individuals from Shetland with "son" names will be assigned to the R1a haplogroup, an as yet unknown number of them being of European origin, and others with Asian or Siberian - like haplotypes found in 40% of Norwegian R1a1, and matching the general "Somerled" (Lord of the Isles) type seen in his descendants of the surname McDonald.

If perchance someone with a Scottish surname is R1a, then clearly their ancestors on the Scottish Mainland were descendants of a Norse Viking who settled in that part of the world. It is anticipated that very few with Scottish surnames will have this haplogroup.