The percentage of the bridled variant of the common guillemot [Uria aalge] was checked by yearly sample counts at four cliff-breeding colonies along the east coast of Scotland between 1950 and 1957 with the purpose of discovering how this percentage varied from colony to colony and from subcolony to sub-colony within each colony in time and place. Over the whole of the study area the percentage of bridled birds fits roughly with the general trend over the Atlantic range of this pelagic auk. The frequency of the bridled variant rises from less than 1% in the south of England to about 10% in north Scotland and to 50% in Iceland. The 4 colonies investigated along the coast-line from Stonehaven, Kincardineshire, northwards and round into the Moray Firth as far as Troup Head, Banffshire, ranged between 2% and 8% of bridled birds. The distances between the colonies and between the sub-colonies in each colony were measured and a rough estimation in broad categories made of the number of breeding birds in each subdivision. Of the 4 colonies the southernmost at Fowlsheugh had by far the greatest numbers, and occupation along the 2 miles of cliff was practically continuous. At all the other colonies the subcolonies were discrete. Counts of the proportion of bridled birds were made at all these colonies from 1950 to 1957 with some variation. The purpose was to determine whether change in this proportion was consistent with changes observed at various colonies over the Atlantic seaboard between decennial surveys made since 1938. These counts showed a notable consistency in bridled ratio throughout the period of observation, which extended to subcolonies as well as to colonies. Further evidence for this stability in gene ratio is seen in the consistency of the pattern of distribution of bridled birds over the study area. Differences in bridled ratio between colonies and even between some subcolonies only 100 yd (91 m) or so apart were statistically significant and, even though some of these small-scale differences showed a reversal of the general northward increase of bridled birds, this mosaic pattern was preserved throughout the observations. This difference in space of the percentage of bridled birds followed an expected arrangement in which colonies differed among themselves more than did subcolonies The constancy with which subcolonies differed, even though some of them were quite close together, showed that gene flow between them must have been small. These results, taken in conjunction with those of a population study by color-ringing, show that the turnover of these guillemot populations is slow (adult survival rate is about 87% per annum) and that movement of birds, even between close subcolonies, is slight. Therefore, the quite large changes in bridled ratio recorded at some colonies in the Atlantic between decennial surveys could not have been achieved under the stable conditions revealed in the study area and must have been due to factors other than the selective elimination of one genotype. Apart from the possibility that some of these changes may be due to observer bias, these changeable colonies are probably much less stable in numbers and turnover than the ones investigated in this study and future surveys will need to inquire into conditions there.