Distribution in numbers
The largest known Kennedy surname population is currently the United States of America. The US Census Bureau state that in the year 2000, Kennedy was ranked 130th most common surname with 171636 individuals. 83% of these were self-identified 'white'. (Previously it had been ranked 137th in a 7.2M sample from the 1990 census). In England and Wales it is less common, only ranking 174th with 35954 individuals in 2002 according to the Office of National Statistics. In Scotland a recent survey of births and deaths by the General Register Office for Scotland for the years 1999-2001 ranked Kennedy 51st most common. In Ireland my own survey of current telephone books show the surname ranks approximately 15th, in line with other estimates.
Howard Mathieson and I have now (3.1.12) mapped all the Kennedy tenants in Griffiths Valuation at the parish level. Or rather I extracted all the data (over 6000 Kennedys!) and Howard did the map. It fills in the detail of my earlier country level map and shows the extraordinary concentration in Co. Tipperary which had 1700 Kennedy tenants in 1851 when the county valuation was done. The top parish however is Shankill in Belfast, Co. Antrim where most of the Kennedys are of Scottish origin. Interestingly Nenagh, which might be considered the ancestral parish for the southern branch, only shows 35 Kennedy tenants and is some way down the rankings although if you count properties with a Kennedy landlord too it places 3rd in the whole country behind Shankill and Kilrossanty in Co. Waterford, where Sir Charles Edward Kennedy baronet was county sheriff and a major landholder. This map is copyright Howard Mathieson and must not be reproduced, likewise all the maps on this page are copyright to Howard, Steve Archer of Archer Software and/or myself.
Howard Mathieson of the Guild of One-name Studies has kindly produced two maps using a subset of the Scottish parish register entries for Kennedy. As the overall dataset is so large (12,487!) Howard used the records for 1800 to 1850. This is still sufficient to show all the main Kennedy strongholds - Ayrshire, Glasgow, the western Highlands and Strathtay. The top 3 parishes for this period are Glasgow, Tyree and Kingussie&Insh.
Most of the maps shown on this page relate to Scotland and England; if you want to see Ireland, I have uploaded a county level distribution map from Griffiths Valuation 1848-1864, which shows the expected strongholds in the counties surrounding Tipperary in the south, and those subject to Scottish settlement in the north-east. I have now (Dec 2011) added the Ireland 1911 census map too.
Historic distribution in maps
Ireland 1848-64 (Griffiths Valuation)
The National Archives of Ireland have now finished publishing the entire 1911 census , which enables another distribution plot. Co. Tipperary still comes out top with 2470 individuals with Co. Louth at the other end of the scale with just 62. The map is not too different from the Griffiths Valuation plot above from 60 years earlier. Ireland and her Kennedy population have not urbanised in the way the Scottish have.

Readers may have seen the distribution maps that have been produced using the easily available 1881 census of Great Britain (but not Ireland). The Surname Atlas by Archer Software will plot county by county distribution of a surname, either by total numbers or normalised against the total population of the county. UCL produce a similar plot with a 1998 comparison chart. This has now been taken over by the National Trust and rather bizarrely appears to plot both charts using modern postcode areas instead of counties. Another charter has shown how the Kennedy surname spread through successive US census returns (note, based on census sampling not total counts), there is also one for France etc.
The next three maps are Copyright © Archer Software and displayed under the terms of the Surname Atlas User License. Please do not republish them.
Kennedy distribution in 1881 by total numbers
Kennedy distribution in 1881 per 100,000 of the population
What do these charts tell us? The first chart shows that total numbers are highest in Lanarkshire and Ayrshire, the second that as a proportion the surname is commonest in Argyllshire and Inverness-shire.
Of course, the Lanarkshire total includes Glasgow, then and now Scotland's largest population centre. In addition Lanarkshire and Ayrshire were the two top counties for mining and weaving which attracted immigrant Irish Kennedys to Scotland from the start of the industrial revolution. (It is insufficient to judge this just by examining place of birth, since many of these are Scottish born children of Irish parents. These figures are still being worked on.)
The second chart showing high concentrations in Inverness and Argyll may surprise some, given how often we read that Kennedy is a, if not the traditional south-west Scotland surname. There is certainly no Irish effect here: there is only one Irish native Kennedy in Inverness-shire in 1881 (a garrisoned soldier from Westmeath) and surprisingly only 10 in Argyll, mostly in Campbeltown. In fact the Kennedys have been in these highland counties for several centuries; depending on which story you prefer, they are either anciently descended from Dunure, indigenous to the area or come from the Western Isles ;-). They are still strongly represented in Inverness county, particularly in Lochaber, less so in Argyll.
Full analysis of the concentration in Lancashire will be published at a later date; suffice to say that this is a mixture of Scottish and more numerous Irish Kennedys. The former gravitated to Manchester and the latter to Liverpool although these distinctions are not absolute. The double hot-spot within the county can be seen by switching to mapping by Poor Law Union in the Surname Atlas.
Kennedy distribution in Lancashire 1881 by Poor Law Union.
Two points stand out requiring further work. One is that we need a parish level mapping display, however Archer Software's database does not have the parish outlines required to do floodfilled maps of Scottish parishes. The less desirable alternative is to use another mapping style and display the parish centre - which in the western Highlands may be a small off-centre location (eg Fortingall village is nowhere near the centre of its parish and had no Kennedys of note). Howard Mathieson's map at the top of the page goes a long way to correcting these mapping deficits.
The second issue we need to deal with is the need to separate out the Irish and their descendants, without losing site of the complexities of the various migrations back and forth that render the tag 'Irish' of questionable meaning. For example a trip through the records of Maybole, a town which the Scottish Kennedys played a pivotal role in, shows Irish born weavers with the name Kennedy - but they may turn out to be Scottish Kennedys coming back home after going to Ulster in the plantation era.
How might we separate out the Irish? By going back to before the growth of the social changes that brought these particular Irish Kennedys to Scotland - in the case of the 1881 census we would like to see what Scotland looked like before the mining and textile industries, and the road and canal building work became available. But the earliest surving census is from 1841 and that is already a bit late, and the 1851 census is clearly too late. These two early censuses rank Lanarkshire as top county for the Kennedy surname but now Inverness ranks 2nd ahead of Ayr, using total numbers. Scotland lacks the kind of 'pseudo-censuses' that other countries have, the only comparable survey over the whole country is the 1691 hearth tax returns and they omit too many people to be of much use. At that time the Kennedy returns were strong in Ayrshire but much less apparent in the highlands, from what I have seen so far.
The best idea I can offer, with caveats, is to examine the rate of births and baptisms recorded in the parish registers. Their advantage is that they go back into the 1500s (the first Kennedy parochial record is from 1561); their biggest disadvantage is their highly variable availability, with some highland parishes only going back to the early 1800s, barely beating out that earliest census. There is another selection effect to watch out for; in a census everyone is returned but people had to pay a fee to get registered in the parish and so the very poorest may be disadvantaged. This factor may partly but not totally explain why at the county level, Lanarkshire drops to fourth ranked behind Ayr, Perth and Inverness by OPR birth totals. The chart at the top of the page shows the results using six graded shading colours. The full table of figures follows. Despite the significant caveats, the chart provides a meaningful indicator of where recorded Kennedy ancestry most likely lies (and hence where the bulk of research for the Kennedy study takes place). The most disadvantaged of the major Kennedy counties in this table is Ross and Cromarty (the only top 10 placer to have more Kennedys living in the first census than have ever been recorded as being born!); this is almost entirely due to the fact that the parish register for Lochs parish in southern Lewis, where most of the Ross and Cromarty Kennedys congregate, only goes back to 1831.
County | Births |
---|---|
Ayr | 2447 |
Perth | 1796 |
Inverness | 1193 |
Lanark | 1106 |
Midlothian | 925 |
Dumfries | 648 |
Argyll | 634 |
Aberdeen | 472 |
Renfrew | 414 |
Angus | 386 |
Fife | 323 |
Wigtown | 321 |
Roxburgh | 312 |
Ross & Cromarty | 193 |
Stirling | 176 |
Kirkcudbright | 164 |
Bute | 161 |
Dunbarton | 131 |
Caithness | 101 |
Kincardine | 99 |
East Lothian | 71 |
Banff | 52 |
Selkirk | 46 |
Moray | 42 |
Shetland | 32 |
Orkney | 31 |
Peebles | 31 |
Sutherland | 29 |
Nairn | 28 |
Berwick | 26 |
West Lothian | 14 |
Clackmannan | 13 |
Kinross | 5 |