Comparative Tables 1856-1996
Current Leading 500 Names

History of research into leading surnames

The first publication to analyse the relative frequency of leading surnames was the Sixteenth Annual Report of the Registrar General of England and Wales, published in 1856. The following extracts raise questions that even 150 years later have not been answered:

"While it is obvious that the original adoption of a particular surname was the result in most cases of arbitrary circumstances - since, John Smith, instead of being called after his occupation, might equally have chanced to become John Johnson from his father's Christian name, or John Wood from the situation of his abode, or John Brown from his complexion - it is curious to remark the predominance of certain names, which seem to have been adopted preferentially by large numbers of the people, or conferred upon them by others, and now prevail in every county of England. Do these common names hold the same rank in point of numbers which they had at first, or have some of them spread and multiplied more than others? For instance, is the present predominance of the Smiths amongst English surnames due to the original numerical strength of that great family, or to some special circumstances acting upon the ordinary laws of increase, owing to which the descendants of the hammer-men have multiplied at a greater rate than the bearers of any other name? Has the progeny of the tawny Brown increased faster than that of the fair complexioned Whites, relatively to the original numbers of each race, so as to account for the excess of the former over the latter; or were the Browns in a majority in the first instance? Various are the surmises and speculations to which such questions may give rise."

I will give his results further on, but here are his findings for three leading surnames:

Smith Taylor Brown The figures for Taylor and Brown are incorrect in my source (0.67 and 0.51), and have been corrected here
1.36% 0.68% 0.57% My thanks to ME for pointing this out.

In 1933, the then Registrar-General- S. P. Vivian conducted a similar survey on the same names:

Smith Taylor Brown
1840/41 1.441% 0.712% 0.600%
1885/56 1.442% 0.702% 0.618%
1930/31 1.699% 0.651% 0.704%

The 1933 historical findings show a certain consistency and the 1853 survey does not fit in easily with these. Between 1840 and 1885 did the relative frequency of the name Smith decline, and then rapidly recover, or is the 1853 methodology suspect?

Vivian, the 1933 Registrar General, concludes that:

"If a surname increases faster than the general population, it must be at the cost of another surname, and vice-versa. There must be some reason for this which, given sufficient evidence, it might be possible to identify. There is, on the other hand, the law of probabilities... There is the possibility of differential fertility or mortality rates; although anything in the nature of selection in regard to physical characteristics must be ruled out in the case of large samples of the population distinguished from one another by little, if anything, but their surnames. The 1930-31 figures connote an actual Smith population of well over half a million and a population of Browns and Taylors a little above or below a quarter of a million each."

Another possibility is that any fluctuations are not real, but merely a reflection of the imperfections of the various sampling systems used

Since then, there have been various other counts on high-frequency surnames as reported in Anthony Camp's The frequency of common surnames. (Genealogists' Magazine, 1998)

  • HM Treasury (Organisations and Methods Division) 1944
  • Civil Service Central Staff Records - pilot study 1964
  • National Insurance count 1966

However, with the above surveys, it is not clear what areas they covered. Was it just England and Wales, or Great Britain, or the United Kingdom? So direct comparisons between the figures ought to be treated with caution.

The next major development was a count of all the marriages registered in England and Wales from January to March 1975 by the human biologist Gabriel Lasker. His results were published in a 1983 paper and will be shown further on in this article.

The 1990s have seen the appearance of the UK electoral registers on CD-ROM. For those who could afford, or have access to it, this will result in an accurate count of those bearing the name With the following provisos:

  • The count is only of those over the age of 17. An adjustment has to be made to account for those aged 1 to 17. Government statistics usually give figures for those aged 1-18. From these I deduced that the proportion of the population aged 1-17 in 1996 was 21.95%.
  • Any count is not compatible with the 1881 census of GRO counts, as it is on a much wider area, and leading surnames have a marked regional distribution.

Finally, 1999 saw the appearance of the 1881 Census on CD ROM, which allows a much faster method of retrieving levels of incidence. A quick count on 'Smith' and 'Brown' (restricted just to England and Wales) resulted in 1.4% for Smith and 0.84% for Brown. Compare this with the survey of the 1884/5 indices above.

Backtracking for a moment, Peter Christian has alerted me to his file from the March 1849 GRO marriage indexes, with soundex codes and frequency. This is a csv file (comma-separated variable) which lists 14674 surnames with the number of occurrence of each. (The file size is 227 kB). It is useful for measuring isonymy in 1849 (people marrying those with the same surname), but not for measuring the national frequency. For example, the Joneses have a marriage frequency that is twice what is normally expected, well above that of the Smiths.

As David Hey indicates, surveys should really be made just on death registrations, as birth registrations were seriously under-recorded (before the 1870s) and marriage registrations are affected by the frequency of isonymic marriages.