Poll Books: a History

Edmund Green gives the definitive guide to their varieties, formats and contexts [40-minute read]

Voting is a means of aggregating individual preferences into collective decisions, and through which the authority to enforce those decisions is legitimated.[1] The point of voting is to have one’s vote counted. This makes the study of historical voting unusual, in that the historian’s data are used for the purpose for which they were originally collected. From the point of view of the voter, the act of open voting may be seen as saying something publicly about the voter’s core values.[2] Meanwhile, from the point of view of the polity elections were held to be an integrating mechanism which served to bind together the strands of society:

Popular elections procure to the common people courtesy from their superiors. That contemptuous and overbearing insolence, with which the lower order of the community are wont to be treated by the higher, is greatly mitigated where the people have something to give. The assiduity with which their favour is sought on these occasions, serves to generate settled habits of condescension and respect … and deserves to be accounted among the most generous institutions of social life.[3]

That an overtly divisive phenomenon like an election might be seen to bind society together is an example of the way in which otherwise intelligent people (Paley had been Senior Wrangler in his day) use functionalist arguments. Instead, elections might be seen as liminal periods characterised by equality among the actors and bookended by rituals of disaggregation and reaggregation.[4] Still more extremely, religious issues might be seen as schismogenetic in politics, promoting and even maintaining conflict.[5]

Open Voting

High in the hills of rural Switzerland lies the fastness of Glarus, capital of the eponymous canton. In this small town the ritual of the Landesgemeinde continues to be performed from time to time. Traditionally restricted to the adult men of the community, who showed their right to participate by the production of their side-arm, the Landesgemeinde gave open assent to propositions affecting the community. It is the last manifestation in Europe of a practice that was once commonplace, of open voting in elections.[6] Of course, many votes are open. The slight raising of a pencil by which members of a committee indicate their assent to a policy is the most frequently encountered of these, a vestigial reminder of the ‘show of hands’; but we expect that our votes for political office should be secret. It was not always so. Open voting was once commonplace throughout much of Europe. It persisted in Denmark until the beginning of the twentieth century, and in Hungary for elections to the upper house of the legislature until 1939.[7]

The word ‘poll’ indicates the counting of heads, not the counting of votes. In most British parliamentary constituencies until 1885 electors had two votes at their disposal and returned two MPs. But this casting of votes was done in a single act of voting, known as polling. The consequence of this was that voters in parliamentary elections were allowed just one act of polling, whether for just one or for multiple candidates.[8]

Open voting in British parliamentary elections had its origin in the show of hands for a candidate at election time. This was ‘the view’ referred to in the legislation of 1696, which may in turn have supplanted a shouting of voices for each candidate. In many constituencies at many parliamentary elections the show of hands long sufficed to indicate to the returning officer, the sheriff of a county or the mayor of a borough, which candidates should be returned. These elections are generally deemed to have been uncontested, for example when the candidate in a hopeless position would ‘decline the poll’, often citing a desire to preserve the peace of the community.[9] But in larger constituencies, and at closely fought elections, it might be difficult for the returning officer to determine which candidate had a majority of votes. The candidate deemed on the show of hands to have been unsuccessful might demand a poll. From the show of hands it was only a small step to recording who had voted for whom.

The origins and development of poll books

The practice of open voting led readily to the making of poll books, written records of who had voted for whom. The earliest record of a list of those who polled in a parliamentary election comes from the Stockbridge contest of 1614. By the 1660s poll books were becoming commonplace.[10] These early poll lists and poll books were manuscript copies, and the practice of printing poll books comes from the very end of the seventeenth century.

The first printed poll to be published in England was issued in the aftermath of the Essex by-election of February 1694. Ghoulish fascination with a by-election caused by the suicide of the sitting member may have been the motivation behind the publication of this poll, but publication was clearly fulfilling a want.[11] Other printed polls soon followed. Legislation in 1696 aimed at regulating county elections required

In case the said election be not determined upon the view … but that a poll shall be required for the determination thereof, then the said sheriff … shall forthwith proceed to take the said poll … and … shall appoint such number of clerks as shall to him seem meet and convenient for taking thereof; which clerks shall all take the said poll … and to set down the names of each freeholder and the place of his freehold, and for whom he shall poll. And be it further enacted that, every sheriff, under-sheriff, mayor, bailiff, and other officer, to whom the execution of any writ or precept shall belong for the electing members to serve in parliament, shall forthwith deliver to such person or persons, as shall desire the same, a copy of the poll taken at such election, paying only a reasonable charge for writing the same.[12]

There is much to unpackage here. Although the legislation was intended to regulate county contests, it was clearly expected that boroughs as well as counties should take polls at elections. And provision was made for copying the poll to a party unconnected with the administration of the election. Further legislation in 1711 regulating county elections required that the resulting poll books be ‘preserved among the records of the sessions of the peace’,[13] although the requirement that boroughs should preserve their poll books was not introduced until 1843.[14] Legislation in that year required that all parliamentary poll books be sent to Clerk in Chancery, to be preserved among the papers of the Crown Office.

Two shocking innovations came at the very end of the period of parliamentary poll books, and the skies resolutely failed to fall in because of either of them. First, women ratepayers were allowed to vote in local government elections from 1869.[15] Next the innovation of secret voting was tried out in the London School Board elections of 1870.[16] The Ballot Act in 1872 introduced the new regime in parliamentary elections.[17] The first parliamentary election held under the provisions of the new legislation was the Pontefract by-election in August 1872. But the last printed polls were later still: the Bedfordshire by-election of June 1872 led to a printed poll book and two printed poll lists.[18] The very last poll book from a parliamentary election held under open voting was (since the requirement of the ballot was not at first imposed upon university seats) was the London University contest of 1880.[19]

The preservation of poll books

So the story of the preservation of poll books is one of formal processes for central preservation gradually supplanting haphazard local preservation. The policy of centralised preservation had unforeseen consequences when in 1907 the entire run of British poll books from 1843 to 1872 was destroyed. This part of the story has generally been narrated as a disaster. To be sure, the loss was great. But manuscript poll books survive for a century and a half before 1843, the abundance of printed poll books is greater than may have been thought, and historians are no longer so wedded to the study of the origins of a class society. The real problem with the loss of the post-reform poll books is the kind of constituency that is not well represented in printed polls: places like Manchester, Birmingham, and Glasgow.[20]

Poll books verses poll lists

Poll books and poll lists are comprehensive lists of the voting behaviour of identifiable individuals at an election contest. The term ‘poll book’ is generally used of a list bearing a title page, while ‘poll list’ is used of a list contained in another publication. But the substance of both documents is essentially similar. They are comprehensive lists, although they may not include all the voters in a constituency. Some poll books listed the voters for one candidate,[21] for those who polled for a slate of candidates,[22] or voters from one part of the constituency.[23] And they record the voting behaviour of identifiable voters. Beyond this, they are characterised by diversity rather than uniformity. This diversity included both the information contained in the poll book and its presentation.

Contents and formats

All poll books and poll lists record the voting behaviour of identifiable voters.[24] Voters were almost always identified by their names, although occasionally a poll book might be issued alongside an electoral register, using the unique number on the electoral register as the linkage key.[25] But beyond this they may contain a great range of additional information. Lists of unpolled electors were occasionally found before 1832,[26] but the process of electoral registration clearly made it easier to enumerate those who were qualified to vote but who did not poll.[27] Borough poll books frequently included the address and occupation of each voter. Less frequently encountered are poll books recording the wealth[28] or religious affiliation of a voter,[29] or his voting behaviour at an earlier election.[30] Meanwhile, county poll books recorded the place in which the freehold lay the possession of which entitled him to vote, together with his place of residence. Other county poll books recorded the name of the occupier of the freehold.[31]

Many printed poll books follow the presentation format of the manuscript poll book. In this, each record occupies a line, typically beginning with the forename and surname of each voter and ending with two or more columns (one for each candidate) in which the votes were recorded by a printed horizontal bar. Of these, some appear to follow the order in which electors polled, while others are sorted into rough alphabetical order by the surname of the voter. Some are subdivided by place of residence, by parish or even by street, and within that may be given alphabetically.

Other poll books were presented according to the votes given. Poll lists in post-reform newspapers frequently did this, perhaps to save space by displaying all the information relating to a voter in a single newspaper column. Single member seats and the single votes given in by-elections to candidates were perhaps more likely to be displayed in this way.[32] But in a typical election in which four candidates contested two seats the multiplicity of split votes made this form of presentation unduly complicated.

Although poll lists are to be found in newspapers from the first quarter of the eighteenth century they were unusual before the Reform Act. The earliest poll list was from the City of London contest of 1722,[33] but outside London we must wait until the latter part of the eighteenth century for provincial poll lists to appear in newspapers. The Salisbury by-election of February 1765 and the Portsmouth by-election of March 1774 led to the publication of a poll lists in London newspapers,[34] but despite the existence of thriving newspapers in some parliamentary constituencies[35] – the Newcastle Courant (established c. 1710), the Stamford Mercury (1714), the Ipswich Journal (1720), the Norwich Post (1721), the Canterbury Journal (1726) and so on – none seems to have produced a printed poll until the Shrewsbury  election of 1774.[36] Production of poll lists remained sporadic until after 1832.[37] But the publication of provincial poll lists increased dramatically after 1832 as a politicised electorate created a demand and as an equally politicised provincial press created a supply.[38]

Poll lists tend to be smaller than poll books. Corporation boroughs tended to have the smallest electorates in England before the Reform Act: one of the smallest lists is that for the corporation borough of Banbury at the general election of 1831, in which the votes of eight members of the corporation were recorded.[39] But this was an historical curiosity, reproduced nearly forty years after the event. The poll list of the Salisbury by-election of February 1765 recorded the votes of 50 members of the corporation.[40] At the Winchester election on 29 January 1715 the published poll recorded the votes of very few freemen.[41] Other small poll lists include that from the Salisbury by-election of February 1765, in which 50 polled, and the Portsmouth by-election of March 1774, at which 63 freemen polled. In the Devizes by-election of March 1838 just 211 electors polled.[42] A late example of a small poll list comes from the Thetford contest of 1865, in which 215 electors polled. With such small numbers it was hardly worthwhile for a printer to run off even a penny pamphlet, and the list was published in the local newspaper.[43] Elsewhere a local printer might take a risk on issuing a small poll book: the Knaresborough poll book of 1847 enumerated the voting behaviour of 220 voters and recorded the lack of participation of eight other electors.[44]

Large poll lists are also unusual, although they occur occasionally. Examples include the poll for the northern division of Northumberland at the general election of 1847,[45] and the contests for the northern division of Essex in 1865, when the electorate numbered 4,904, and the southern division of Essex in 1865, with a registered electorate of 7,338. But these were dwarfed by the poll list issued after the Liverpool by-election of July 1853, when the registered electorate stood at 16,182, over 17,500 votes were cast, and the number polling appears to have exceeded 10,000.[46]

Before the Reform Act, the largest poll book issued was that for the great Yorkshire election of 1807, in which over 23,000 electors polled.[47] But elections for England’s greatest county were not always this size: the poll list for the attenuated Yorkshire by-election of December 1830 enumerated just 465 voters.[48]

Some constituencies produced poll books as substantial volumes. These included (but were not confined to) the counties. Others produced poll books as pamphlets: York was a regional centre of printing and supported more than one newspaper; but except for the poll at the election of 1837 polls for the city were only produced as pamphlets.[49] Pamphlet editions were almost always from the boroughs. Meanwhile some poll lists appeared in newspapers: these were usually boroughs although poll lists for counties are occasionally found. The production of poll lists in newspapers increased in the years after 1832.

The prices of printed poll books varied widely. County poll books generally contained more information than those issued for boroughs, including the location and nature of the freehold which qualified the elector to poll and the place of residence of the voter, and were produced on a more lavish scale than the borough counterparts. Their retail prices were correspondingly higher, and their survival rates appear to be better. Borough poll books might be large and lavish affairs,[50] or they might be very simple sheets, issued folded but unbound.[51] Advertisements for printed poll books in local and regional newspapers may not be much help in assessing the prices asked for published poll books, as these may commonly exclude the very simplest and cheapest editions. It may be that the median borough poll book retailed as a 6d pamphlet, and the 3d. asked for the Gloucester poll book of 1852 was more typical than the 3s. asked for the poll book for the same constituency after the by-election of October 1816. Eighteenth-century poll books were on the whole more expensive than those in the nineteenth century: mass politics was accompanied by a mass market for political artefacts.

Anonymous, The very wise aldermen of Gotham, scratching for a mayor, 1774, etching on paper, 17.2 x 10.6cm, British Museum, London, 1868,0808.13252. A satire relating to the contest for London mayor, involving Frederick Bull and John Wilkes. Note the polling books on the table © The Trustees of the British Museum
Other kinds of polling records

Similar documents are sometimes confused with poll books. These include electoral registers and burgess rolls, both of which are lists of those eligible to vote.[52] A hybrid document, the marked electoral register, in which a printed list of electors is marked by hand with the voting behaviour of those who polled, may be included in bibliographies of manuscript poll books.[53]

Another document which may incorrectly be referred to as a poll list is a printed ‘state of the poll’. States of the poll resembled cricket score sheets formerly sold for a penny throughout the duration of a Test match. They showed the numbers of votes attained by each candidate at a certain time during the election: commonly at close of play each evening, but sometimes hourly. The only similarity between a ‘state of the poll’ and a poll book lies in the use of the word ‘poll’, which may present dangers when using a word-search in library catalogues.

A third document which might be confused with a poll list is a list of supporters for a candidate. These lists take two forms, first, the list of a candidate’s committee, and secondly a requisition calling upon a gentleman to put his name forward as a candidate. In the years following the Reform Act candidates in county constituencies and in larger boroughs often had extensive committees comprised of the leading figures of the constituency.[54] And signed requisitions could also occupy many column-inches of post-Reform newspapers those calling upon Sir Francis Burdett to stand for the northern division of Wiltshire in 1837 are particularly extensive.[55] On Burdett’s death in 1844 an equally extensive requisition was published urging T. Sotheron to stand in his place.[56]

If a requisition is not a poll list, then it follows that a requisition containing details of the voting behaviour of its signatories cannot be a poll list. The poll list for the election at Weymouth and Melcombe Regis has not yet been found, although it may have been issued with the Weymouth Telegram in late November 1868.[57] Before the nomination, a requisition had been signed by some electors calling on the London barrister John Powell QC to stand. Encouraged by the requisition, Powell allowed his name to be put forward. But the result of the election was a disappointment to him, and he was placed at the bottom of the poll with 452 votes. In his disappointment, Powell published a part of the poll book: the names of those who had signed the requisition, with details of how they had voted.[58] This requisition is not a poll book because it fails the test of comprehensiveness: it does not enumerate all the voters in the constituency, but only a sample of them.

Why were poll books published?

John Vincent drew attention to the diversity of poll books, ranging from (at the simplest) a list of those who polled for a particular candidate to the addition of much other information.[59] ‘In general’, he declared, everything depended on the free will of some small printer looking around for a small profit’.[60] If the desire for profit was the prime motive for the production of printed poll books, what were the secondary motives? We may enumerate seven of them.

Firstly, a very few poll books were official publications. An early example is the poll for Christchurch in 1727,[61] but they were usually issued as part of a House of Commons enquiry into bribery and corruption in the years after 1832. These include the polls for Stafford in 1832, for Roxburghshire in 1837, for Cork County in 1841, for St Albans in 1847, and for Maldon in 1847 and 1852. Other printed poll books may be semi-official: the poll for the southern division of Derbyshire in 1832 bears the imprint of Hansard, while the poll for the London University contest of 1880 bears the imprint of Eyre and Spottiswode ‘for HMSO’. In other poll books a semi-official nature may be indicated in acknowledgements for help given by the authorities in the poll books for the southern division of Derbyshire in 1865 and for Dover in 1868.

Second, poll books were made and printed to provide evidence in the event of a scrutiny (held prior to 1832 by the returning officer before a return was made) or a petition to parliament against an undue return. Legislation in 1696 regulating the taking of parliamentary elections for counties had enacted that the poll clerk should ‘set down the names of each freeholder and the place of his freehold and for whom he shall poll’.[62] In the parliamentary general election for the City of London in 1784 John Sawbridge was returned in fourth place with 2,823 votes; Richard Atkinson was not far behind with a tally of 2,816 votes on the poll. Such a close result indicated scope for a scrutiny to challenge the validity of votes cast. Two printed polls were issued. Neither was a complete poll of the electorate: one was of those who polled for Sawbridge and those who polled for Atkinson,[63] while the other was of Atkinson’s supporters ‘which the liverymen … are desired speedily and carefully to examine’.[64] So this kind of poll book was commonly produced by the opponents of those whose support it recorded.

Third, printed poll books were issued to facilitate exclusive dealing. ‘Who supported the Tories?’ enquired the poll list for The Hartlepools after the first election for that newly enfranchised borough in 1868.[65] The defeated candidate in the Banbury poll of 1835 swore to ‘publish a black book in which the name of every voter shall be entered, his residence and his trade, that every farmer may know with whom he ought or ought not to deal’.[66] And the Warrington contest of 1841 led to the menacing publication of The Black List: a list of Conservative tradesmen in Warrington. The Devizes by-election of February 1844 led to the publication of poll lists firstly in the Conservative Devizes & Wiltshire Gazette and a week later in the reformist Wiltshire Independent. The latter offered a preface to its list worthy of Manchester’s Free Trade Hall:

The Devizes Gazette, in order to carry out the true Tory principles of exclusive dealing, having published the names of those who voted on this occasion, we have thought it right to re-publish those names, and to remark that Mr Burges being a supporter of Monopoly, the tendency of which is to make work scarce, wages low, and food dear, those who voted for him may without any unfairness be classed as enemies of the middle and lower classes, – in short, of all except the great rent and tithe receivers, while those who had the manliness to support Mr Temple, the enemy of all monopolies, may with perfect truth be called the friends of the people.[67]

Fourth, printed poll books were issued as a form of vanity publication: not always for the vanity of the candidates, but for the vanity of the voters. Issued with a wrapper in the candidates’ colours, these were presented by candidates to their voters after an election. This appears to have occurred in Lewes and Colchester, and perhaps in other constituencies. We remember the vanity of Sir Walter Elliott of Kellynch Hall, in whose house the only book on display was of his own pedigree. It it possible that the borough Hampdens of the post-reform period were equally motivated?

Fifth, printed poll books were issued out of civic pride. ‘The poll book of the first contested election’ announced the title page to one edition of the Gateshead poll book of 1837. There is a hint of this civic pride in the poll list issued after the Bolton election of 1847, ‘A list of the persons who voted at the 5th election of members to serve in parliament for the borough of Bolton’,[68] and echoes of it in the reproduction forty and fifty years after the event of the Bolton poll lists from the general elections of 1832 and 1841.[69] Newly-enfranchised boroughs continued to feel a rush of civic pride: the poll of the election of Stockton in 1868 was issued with the title Records of the first parliamentary election.

Sixth, printed poll books were issued as part of the process of ‘nursing’ a constituency and could be used as pro-forma canvassing books for a future election.[70] This last reason presents problems. Printed poll books were undoubtedly used for canvassing at subsequent elections, but after the Septennial Act canvassing at an unknown future date must have been heavily discounted. An exception may be the general election of 1818, when few can have expected the parliament then elected to last for its full term of seven years. The poll book of the Westminster election of 1818 was printed, the first for that large borough since 1780.[71] And if few could have expected that the next election for the constituency would be in 1825, none could have expected that it would come as soon as the following year, precipitated by the suicide of the sitting MP. Printed copies of the Westminster election of 1818 were used for canvassing at the by-election of 1819,[72] but this use can hardly have been a reason for the poll book being printed.

Finally, poll lists may have been issued to fill up space in a periodical. The poll of the 296 voters for Luttrell in the Middlesex by-election of April 1769 (at which Luttrell was declared elected although John Wilkes received 1,143 votes) was issued in printed form.[73] The Barnstaple by-election of March 1824 led to the publication of its associated poll list. This was tucked away in the first volume of a short-lived and undistinguished periodical, the Barnstaple Miscellany. There seems to be no reason for the publication other than to record something apparently interesting about Barnstaple in a failing provincial magazine.[74] The poll of the Cirencester election of 1868 was reissued in the Cotswold Almanack for 1869; meanwhile the poll of the Barnsley polling district of the southern division of the West Riding of Yorkshire was advertised to appear in Lingard’s Illustrated Almanack and Diary for 1869.[75] After the introduction of open voting the re-issue of poll lists remained an occasional feature in local newspapers: ‘How our borough conducted parliamentary elections in olden times’. Some of these include polls which are otherwise unknown: the poll of the Wigan by-election of December 1763, and the poll of the Shrewsbury election of 1852 are both known only from their later publication.[76]

How many poll books were printed?

It is difficult to estimate the number of copies of a printed poll book were run off. The economics of publishing, of relatively high fixed costs for typesetting and relatively low variable costs for printing gives printers an incentive to increase their print runs. But the downside of this was that they might be left with unsellable stock once the excitement of an election died down. Some more expensive county poll books were offered on a subscription basis, by which a printer recovered at least some of his costs prior to publication. Meanwhile there is some evidence of printers selling off old stocks of poll books years after the original publication. An advertisement in the Cheltenham Journal at the time of the general election of 1852 offered old stock of poll books for that constituency: for the by-election of 4 September 1848 (otherwise untraced), for the by-election of 29 June 1848, for the general election of 1847, for the general election of 1841, and for the general election of 1837 (otherwise untraced).[77]

This in turn raises the question of how many editions of poll books have disappeared entirely. There is an adage among bibliographers that ‘Bigger books linger longer; little books last least’.[78] Few of those poll books and poll lists which have been found since the publication of the Sims Handlist have been for county constituencies, for which poll books were issued on a rather more lavish scale than the twopenny pamphlet editions which were typical of parliamentary boroughs. Most of the newly discovered poll books are the twopenny or sixpenny pamphlets from boroughs, or poll lists from newspapers. Many are from the period after 1832. When the History of Parliament Trust issued its Draft Register eighty years ago, the poll for the Liverpool election of 1831 was listed there as being missing. No trace of it has since been found; no newspaper report of it has been discovered. It now seems safe to conclude that this poll never existed (except, of course, as the original poll books made by the poll clerks).[79]

There are many reasons why few new poll books are from counties, and few are from the pre-reform period. In the first place, county poll books were more lavish productions than those from boroughs. Second, the bibliographical record of the ESTC period is outstanding, and there can surely be few printed books from this period which are totally unrecorded.[80] Third, the original conception of the History of Parliament was to go up to 1832,[81] and it is possible that a higher proportion of post-1832 poll books were overlooked during the compilation of the History of Parliament’s original Draft Register of poll books in 1952.

It might be possible to use newspaper advertisements to estimate the number of ‘lost’ poll books. But newspaper advertisements are not certain evidence of publication. Some advertisements solicited subscriptions towards the cost of publishing a poll; if insufficient subscriptions were forthcoming, then the publication might not have gone ahead. The Stroud contest of 1868 was followed by the announcement of a special edition of the Stroud News for 1 December 1868. If this special edition was published, it was not microfilmed. Meanwhile the Stroud News published a ‘corrected’ edition in pamphlet form. Likewise the poll of the Lancaster contest in 1852 was apparently issued with the Lancaster Gazette on 17 July 1852. If a broadsheet was issued, it has not survived, although the Gazette of that date issued a list of corrections and a pamphlet edition was reissued.

What kinds of constituencies tended to produce printed poll books?

Here we must take as the denominator not the return but the incidence of contested elections. Poll books can only come from contested elections. No contest, no poll; no poll, no poll book. A disproportionate number of newly discovered poll books and poll lists come from the last forty years of open voting, from the period after the Reform Act. This disproportionality, already evident, has become much clearer following a systematic search of provincial newspapers for the period 1832-1865.

What kind of constituencies produced poll lists and poll books? Only one poll list is known for a county prior to 1832,[82] and even after the Reform Act they are unusual.[83] From 1700 until 1865 counties produced poll books, not poll lists. But the boroughs produced both books and lists. What kind of borough produced printed polls? Before 1832 they were overwhelmingly from medium-sized open boroughs. Few were from the very smallest boroughs; few were from corporation or burgage boroughs,[84] overwhelmingly printed borough polls before 1832 come from the middle-sized fairly open boroughs with freeman or scot-and-lot franchises. After 1832 a uniform £10 householder franchise eliminated the franchise as a variable here, but it introduced the variable of ‘new’ constituencies. Overwhelmingly after 1832 printed polls came from mid-sized constituencies, both old and new. Geographically the distribution of printed polls follows broadly the distribution of contested elections, with the exception of the lacuna of Cornwall. I think that there was a preponderance towards ‘old’ constituencies, but I don’t yet have the figures for this.

An unexplained phenomenon in the distribution of printed poll books remains. Almost all are from England. Northwards to Carlisle and Berwick upon Tweed there are poll books. Westwards to Bristol, to Gloucester, and to Chester there are poll books. But the production of printed polls virtually ceases at the borders of Scotland and Wales. Why poll books should be almost exclusively an English phenomenon within the United Kingdom is unclear. There may be an element of circularity here: searches of Welsh and Scottish newspapers were more cursory than those of the English press. The handful of newly discovered polls are duly reported.  But it remains true that printed polls are overwhelmingly English.

Early printed polls, from the first 40 years or so (1695-1735) were disproportionately from the counties; late printed polls (1832-1872) were disproportionately from the boroughs.

Printed poll books were published disproportionately for medium-sized boroughs.[85] Before 1832 the smallest boroughs were often corporation or burgage boroughs, closed boroughs in which free elections rarely took place. The generalisation is, of course, imperfect: Gatton (like Westminster) was a Scot and Lot borough, Lymington was a freeman borough. ‘I take the earliest opportunity of informing you’, wrote Edward Gibbon to his stepmother, ‘that in the course of next week I shall be elected for the borough of Lymington in Hampshire’.[86] Few larger boroughs, and few counties, were completely closed; in most at least the semblance of an electoral process took place. The ECPPEC project’s Election Directory reveals exactly how many elections took place in the long eighteenth century, where they were, and how many were actually contested, while the Poll Book Directory lists all surviving poll books and records where they may now be found.

Rival versions

One curious feature of the distribution of editions of printed poll books and lists is the frequency with which there was more than one edition issued for some election contests. The Nottingham contest of 1754 produced not one but three editions of the poll, issued by two different publishers.[87] Sleepy Lymington, the constituency for which Gibbon complacently announced that he would be elected, had no contested elections between 1710 and 1832. But between 1832 until 1872 every election except that of 1835 was contested, together with by-elections in April 1850 and May 1860. Of all these contested elections, only the poll of 1859 led to the publication of a poll book. It is curious that the election led to the publication of not one, but two poll books issued by rival publishers.[88] This was not unusual, and further four kinds of multiple publication may be identified.

The first is the simple ‘second edition’: the poll for New Windsor in 1802 was issued as a pamphlet, and a second edition duly appeared. Sometimes an acknowledged ‘second edition’ survives which implies a putative first edition, such as the second edition of the poll for the Carlisle contest of 1857.[89] The poll for Dorset in 1857 was apparently re-issued in 1859.[90] There are two reasons for these second editions. The first is that the original print run was too small. But it is likely that many ‘second editions’ were in fact corrected second editions. The issue can only be resolved by a detailed examination of first and second editions. The correspondence columns of post-reform borough newspapers would sometimes carry letters from those complaining that their political behaviour had been misrepresented and requesting (even demanding) that a correction be printed.[91]

Occasionally a poll list, having been set up in type for a newspaper edition, was reissued apparently unchanged: the poll list for the Bedford by-election of June 1859 was issued in the Bedford Times on 2 July 1859, and apparently reissued in the same newspaper three days later. The expensive part of publishing a poll list was the extensive typesetting and reissuing it would help fill up a second newspaper and perhaps sell a few more copies. In any event the poll for the next contested election for Bedford appeared in the Bedfordshire Times of 15 July 1865 and was apparently reissued in the same newspaper three days later.[92] On other occasions a poll list would be reprinted with minor corrections: the list of those who polled for Charles Goodfellow at the City of London by-election of December 1724 was reissued with minor corrections in pamphlet form a few days later. The published poll for the City of London election in 1768 was apparently reissued two years later, when the death of William Beckford triggered a by-election. Meanwhile the poll list for the Hereford by-election of 5 October 1841 was rushed out in the Hereford Journal on 6 October; a corrected version was issued in the same newspaper a week later.

The second is the publication of both a poll list in a newspaper and a pamphlet edition: the poll for Shrewsbury in 1774 had its first incarnation in the Shrewsbury Chronicle and was apparently reissued as a broadsheet. Meanwhile the by-election for Clitheroe in May 1853 led to the publication of a poll list in the Blackburn Standard, which appears to have been reissued in pamphlet form. It is assumed that a newspaper edition of a poll list generally preceded its publication in pamphlet form. The poll list for the Bolton contest of 1847 appeared in the Bolton Chronicle of 31 July 1847.  Meanwhile on 7 August that newspaper advertised ‘Copies of the list of votes may be had at the Chronicle office, and of all booksellers and newsmen in a pamphlet, price 1/2d. with covers and stitched, 1d.’ The Bedford poll book of 1857 was apparently reprinted from the list issued in Bedfordshire Mercury on 28 March 1857; meanwhile Edwards’ edition of the Peterborough poll of 1865 was reprinted from the Peterborough Times, and Clarke’s edition of the same poll was reprinted from the Peterborough Advertiser.

The third is the publication of two editions in rival newspapers. Just as parliamentary boroughs from ancient times commonly had two inns, so in post-reform years many supported two newspapers.[93] Before 1832 Banbury had been a corporation borough; after the Reform Act the Banbury Advertiser and the Banbury Guardian regularly issued their own competing versions of the poll list. Similarly the Worcestershire Chronicle and the Worcester Journal each issued their version of the poll list for their city at the election of 1847. Meanwhile the poll for the northern division of Essex in 1847 was issued by the Chelmsford Chronicle, the Essex Standard, and the Essex Herald.[94] A curious hybrid is found in the poll for the Durham city by-election of June 1853, in which the poll was published in one newspaper and the accompanying list of unpolled electors in another.[95] The growth of provincial newspaper publishing went hand-in-hand with the politicisation of the reformed borough electorate: political partisanship encouraged the growth of a provincial newspaper industry, and the provincial newspaper industry encouraged the rise of partisanship.

And fourthly the publication by rival printers of two or more pamphlet editions. The York by-election of December 1759 led to the production of rival poll books.[96] Some of these rivalries were long-standing. The publisher N. Warren is first known to have issued a poll book for Winchester in 1837; the imprint of Warren & Son was still to be found on the published poll of 1868. Warren’s rival was W. Tanner, under whose imprint poll books were issued for Winchester in the 1840s and 1850s. In Maidstone poll books were issued by rival firms of printers: by A. Austen; by various members of the Cutbush family, the proprietors of the South Eastern Gazette; and by J.V. Hall, the proprietor of the Maidstone Journal.[97]

In other cases the publishers were not rivals but collaborators. Perhaps in an attempt to save typesetting costs, the same poll book might be issued with different imprints. The poll book for the Gloucester contest of 1818 was issued with the imprint of J. Roberts, while the same poll book was issued with a different title-page, bearing the imprint of Walker & sons. The poll for Maidstone in 1832 was issued with the imprint of A. Austen, and the same setting of type was issued with the imprint of J. Brown. Meanwhile the poll for Winchester in 1868 was issued not only by Warren and Son, but also by J. Pamplin. Their contents were identical. In the early nineteenth century poll books for Norwich were published by a group of publishers including Crouse, Stevenson, Matchett, and Bacon. Their names were used almost interchangeably on their imprints.

To be clear, most printed poll books survive in only a single known edition. The phenomenon of multiple editions of the same poll was the exception, not the rule. But it was an exception more frequently encountered than might be expected, and the frequency of multiple editions of a given poll appears to have increased in the post-Reform years. This may have had much to do with the development of the provincial press in these years, and occasionally there were three issues of a poll book.[98] Indeed, it is possible that the Bedfordshire election of 1859 led to the production of four printed polls: two lists in newspapers, and possibly two pamphlets.[99]

Poll book discoveries (past and future)

The new discoveries of poll books transform our knowledge of some constituencies: Sims enumerated just two printed polls for Ludlow, from 1865 and 1868. To these may now be added polls from 1832, 1835, 1837, the by-election of June 1839 (in two editions), the by-election of May 1840, and the election of 1852.[100] The political history of the Isle of Wight is similarly transformed, with new polls for the single-member county from 1852, 1857, 1859, 1865, and the by-election of June 1870. Likewise Sims recorded just two printed polls for Stamford, those of the by-election of February 1809 and the contest of 1847. To these may now be added printed polls for 1734, 1830, 1831, 1832, and another two editions for 1847. To the Totnes tally of three (October 1812, 1837, and 1859) may now be added a further five: July 1839, April 1840, 1857, January 1863, and 1865). To the six polls for Maldon recorded by Sims may be added all four of those that he listed as untraced: 1761, 1841, August 1854, and 1857.

The end of the poll book era came swiftly. Arguments for the ballot had circulated since the Reform Act of 1832; the ballot had been one of the six points sought by the Chartists; and George Grote proved to be an indefatigable champion of the cause. Some newspapers, such as the Preston Chronicle, maintained a high-minded objection to printing polls, but poll books and poll lists continued to be published throughout the 1850s. The Reform Act of 1867[101] sounded the death knell for printed polls, not by argument but by cost: the franchise clauses of the Second Reform Act increased the size of the electorate by about 85 per cent, (by 42 per cent in English counties and by 232 per cent in the boroughs)[102] and the publication of poll books rapidly ceased to be profitable. The Ballot Act of 1872 administered the coup de grace.[103]

In the seventy years or so since the History of Parliament Trust issued its Draft Register many hitherto undiscovered printed poll books have been found. The first phase of searching for poll books and poll list culminated in John Sims’ Handlist. This publication was remarkably good, for while it contained many errors of omission what it did contain was very accurate.[104] The ways in which new polls have been found reflect changing technologies. Towards the end of the twentieth century the English Short Title Catalogue brought some new poll books to light, including that for the Surrey election of 1713, a variant edition for the City of London by-election of December 1724, the Surrey poll of 1741, a variant edition for the Shrewsbury election of 1774, and a poll for Montgomeryshire in 1774. But not everything has yet been catalogued in ESTC: lacunae include the poll for New Radnor in 1727, Nickson’s edition of the poll for York in December 1758, and a poll from the Colchester by-election of December 1788.[105]

Microfilming of newspapers brought others to light, such as the poll list for the Portsmouth by-election of March 1774 and the newspaper edition of the Shrewsbury poll at the general election of the same year. The internet transformed the search for poll books, as of so many other things. From a desk in London it became possible to examine the catalogue of the National Library of Scotland, bringing to light the poll for Berwick upon Tweed in 1857. Many of the first tranche of newly discovered polls shared the common characteristic of being poll books rather than poll lists: they had a title page, and many bore the printer’s name together with the place and date of publication. They were catalogued in a standardised form and were easy to compare. In 2005 I rashly delivered myself of the opinion that few of the polls yet to be discovered would be books or pamphlets which have no extant poll books.[106] In fact most of the newly discovered polls are poll lists, most commonly in newspapers.

But the search for new poll lists was further transformed by the digitisation of nineteenth-century provincial newspapers. Even with the search fields restricted by time and place, there remained thousands of newspaper pages to be searched. This could only be achieved using key word searches, and eyeballing dozens of thumbnail representations of newspaper pages. There remain many more to be found. When writing about polls for Devizes below, I did a more thorough search of the newspapers online. This brought to light a reissue of the poll from the by-election of March 1838 and a new edition of the poll from the by-election of February 1844.

Where should the assiduous searcher seek new printed polls? It seems likelier than ever that the tree of printed poll books has been stripped almost bare. The first place to start might be among the echoes of printed polls left by advertisements for poll books. The Windsor poll book of 1837 was almost certainly published,[107] and there is good evidence of the existence of the Tynemouth poll book of the same year.[108] But while the Dewsbury poll book of 1868 was probably published,[109] the proposed publication of the Gravesend poll of 1868, announced for January 1869, may never have happened.[110] These unconfirmed echoes of published poll books are noted in the accompanying bibliography in square brackets. Clearly an announcement that a poll has been published carries more weight than an announcement that a poll will be published, so the publication of the Windsor poll of 1837 referred to above seems more certain than the announcement of a forthcoming poll for the same constituency in 1852.[111]

Beyond this, new polls should be sought where they are most likely to occur. The rivalries of the printers Warren and Tanner in Winchester has been noted. No Tanner edition was noted by Sims for Winchester in 1852 or 1857, but a search has turned up a copy of the latter at Winchester Library, while an advertisement exists for a Tanner edition for Winchester in 1852.[112] The rivalries of the Banbury Guardian and Banbury Advertiser have been noted. Each newspaper printed a poll list after the elections of 1857, 1859, and 1865. But the poll list after the Banbury by-election of February 1859 appears to have been published only in the Guardian: a careful search of the Advertiser has revealed no poll. The Guardian was published by William Potts, who published poll books for Banbury in 1832, 1835, 1841 and 1847 (John Potts, perhaps his son, published the Banbury poll in 1868). Polls for Devizes are known to have been published in the Wiltshire Independent (1857, 1859, February 1863, and 1868) and in the Devizes and Wiltshire Gazette (February 1844 and 1868). The files of these two newspapers are an obvious place to search for other polls. The online files of Gazette contain additional polls for Devizes from March 1838, 1857, 1859, and February 1863, while the Independent published the poll of the by-election in February 1844. In short, from 1844 until 1868 at every contested election except that of 1865 (contested in name only, when two Conservatives walked the course and the hapless Liberal candidate received no votes) each newspaper published its own poll. Most of these are editions of polls which are already known to exist in another form, but that for March 1838 is otherwise unknown. No poll book was ever issued for Devizes, a small borough in which only 211 electors polled in March 1838.  It was thus much the same size as Knaresborough, which did produce a poll book. But Devizes supported two newspapers.

Finally, the search for new poll books should be pursued in the columns of the provincial newspaper press. This is not always as straightforward as may appear at first sight. The publication of a pamphlet edition alongside a newspaper column has been noted. The relationship between the printers of some newspapers and the publishers of the pamphlets is also clear, if only because the newspapers ‘puffed’ their pamphlet editions. The poll of the Lancaster election of 1852 was said to have been reprinted from the Supplement issued with the Lancaster Gazette of 17 July 1852. Indeed, that issue of the Lancaster Gazette included a list of corrections to the poll. But no copy of the poll list survives. It seems possible that the poll book was presented gratis to subscribers to the Gazette, as it certainly was in the case of the poll book issued after the Preston contest of 1868.[113] A ‘List of the poll’ for the Huddersfield contest of 1852 was apparently issued with the Huddersfield Examiner on 31 July 1852. No such issue is in the files of that newspaper, either online or on film, and it is possible that it was the broadsheet edition for that election recorded by Sims.


A personal envoi. Edward Gibbon recorded that ‘It was at Rome, on the 15th of October, 1764, as I sat musing amidst the ruins of the Capitol, while the barefooted friars were singing vespers in the Temple of Jupiter, that the idea of writing the decline and fall of the city first started to my mind’.[114] About 210 years later the United Kingdom was gripped by the second general election within a year, and the present writer sat in the Local History section of the York City Library awaiting the delivery of the books that he had requested. In bookcases around the walls of the room were shelved some of the more frequently used books on the history of that great city. Among them were the poll books of the constituency and taking a couple from the shelf, quite possibly those of 1857 and 1859, he noticed the possibility of tracing an individual across more than one election. The poll books were alphabetically sorted and close in date, and an examination of them revealed the possibility of tracing at least some of the voters from one election to the next.

[1] D.E. Stokes, ‘Voting’, in D.L. Sills (ed.), International encyclopaedia of the social sciences (18 vols, 1968-79), xvi, 387-95.

[2] R. Rose and I. McAllister, ‘Expressive versus instrumental voting’ in D. Kavanagh (ed), Electoral politics (Oxford, 1992), 114-140.

[3] William Paley, The principles of moral and political philosophy (2 vols, Dublin, 1785), ii, 197.

[4] F. O’Gorman, ‘Campaign rituals and ceremonies: the social meaning of elections in England, 1780-1860’, Past and Present, 135 (1992), 79-115. The theory is ultimately derived from A. van Gennep, Les rites de passage (Paris, 1908).

[5] G. Bateson, Naven: a survey of the problems suggested by a composite picture of the culture of a New Guinea tribe drawn from three points of view (Cambridge, 1936; 2nd edn, Stanford, Ca, 1958), 175. Politicians deploy schismogenetic issues because they motivate key groups of electors to vote for them.

[6] The Landesgemeinde is no longer used in elections for representative office, and women have been allowed to participate in it since 1991.

[7] J. Elklit, ‘Open voting’, in R. Rose (ed.), International encyclopaedia of elections (Washington DC, 2000), 191-3.

[8] Newspaper advertisements sometimes asked electors to give a candidate their ‘vote and poll’. This indicated two parts of a process, the raised hand on the ‘show of hands’, followed by a vote for the candidate at the subsequent poll.

[9] J. Rosebank, Partisan politics: looking for consensus in eighteenth-century towns (Exeter, 2021).

[10] M. Kishlansky, Parliamentary selection: social and political choice in early modern England (Cambridge, 1986), 186.

[11] John Honywood made at least three unsuccessful attempts at suicide, before successfully hanging himself. Insofar as suicide can be funny, this certainly is. The autobiography of Sir John Bramston (Camden Society, 1845), 377-8, cited in G. Holmes, ‘The electorate and the national will in the first age of party’ in idem., Politics, religion and society in England, 1679-1742 (1986), 3.

[12] 7 & 8 William III c. 25.

[13] 10 Anne c. 23.

[14] 5 and 6 Victoria, c. 33.

[15] Municipal Corporations Elections Act, 32 & 33 Victoria c. 55 (1869).

[16] London School Board Elections

[17] An Act to amend the law relating to procedure at parliamentary and municipal elections, 35 & 36 Victoria c. 33.

[18] ‘Bedfordshire election’, Bedfordshire Mercury, 7 September and 14 September 1872; ‘Beds. County election … the poll’, Bedfordshire Times and Independent, 24 August, 31 August, 7 September, and 14 September 1872; and Beds. County election … the poll (Bedford: C.F. Timaeus). The pamphlet appears to be a reprint of the newspaper columns in Bedfordshire Times and Independent.

[19] University of London. Parliamentary election, March 1880: list of voters (London: Eyre and Spottiswoode).

[20] M. Taylor, ‘Interests, parties and the state: the urban electorate in England, c. 1820-1872’, in J. Lawrence and M. Taylor (eds), Party, state and society: electoral behaviour in Britain since 1820 (Aldershot, 1997), 50-78. For the century before the Reform Act the best introduction remains F. O’Gorman, Voters, patrons, and parties: the unreformed electoral system of Hanoverian England, 1734-1832 (Oxford, 1989), while there is much of value in H.T. Dickinson, The politics of the people in eighteenth-century Britain (Basingstoke, 1995).

[21] The City of London by-election of December 1773 led to the printing of a poll for the candidate who was returned, A list of the persons that have polled for the right hon. Frederick Bull (Charles Rivington, 1773).

[22] In addition to a complete representation of the poll, the contest for Yorkshire in 1734 led to Persons names that voted for Sir Miles Stapleton, and Edward Wortley, Esq., or one of them singly, within the Weapontake of Claroe.

[23] County poll books towards the end of the period of open voting were potentially unwieldy, and local newspapers sometimes published the poll for a polling district. Examples include polls from the Barnsley district of the southern division of the West Riding of Yorkshire in 1865 and 1868, for the Harrogate district of the West Riding of Yorkshire in 1859, for the Lowestoft district of the eastern division of Suffolk in 1859, and for the Castle Hedingham district of the northern division of Essex in 1865. Meanwhile the southern division of Lancashire saw poll lists published the Leigh district in 1859, for the Bolton district and the Bury and Bacup district in 1861, and for the Wigan district in 1865. Other county polls were published in parts, such as that for Berkshire in 1832, the southern division of Shropshire in 1865 and 1868, the East Riding of Yorkshire in 1837, and the West Riding in 1859.

[24] Voters are usually identifiable by name, although the combination of forename and surname alone was insufficient to identify uniquely all voters in many constituencies. In Westminster’s St James parish in the period 1784-90 there were four voters called ‘William Blake’, one of whom was the poet.

[25] The polls from the northern division and the southern division of Essex in 1832 are of this type.

[26] The lists of the liveries of the fifty-six companies of the City of London (1701), technically a turnout list not a poll book since it does not record the political behaviour of the participants. But it records which liverymen polled, and which did not. Meanwhile poll books from the freeman borough of Derby at the by-elections of December 1748 and January 1775 recorded unpolled electors.

[27] The first poll book for Whitehaven recorded the names of unpolled electors, The poll book of the election (Whitehaven, R. Gibson, 1832). Many other post-reform poll books followed suit. The existence of a known denominator of qualified electors has led the unwary into crude calculations of turnout in nineteenth-century elections.

[28] The poll books for Bath at the by-election of June 1855, and for Newry in 1868 record the rateable value of the property occupied by each voter.

[29] The poll book for Abingdon, Berkshire, in 1734, 1754, and 1768 recorded the religious affiliations of voters.

[30] There seems to have been a flurry of interest in linking electors through time in 1859. Although the poll book for Cambridgeshire issued after the general election in 1802 had recorded the votes of those who had polled two months earlier, the innovation did not catch on. The poll book for the Newcastle-upon-Tyne by-election of June 1859 includes information on the behaviour of those who polled at the general election two months before.  The poll for Banbury at the general election in 1859 records the votes of electors at the by-election three months before: Banbury Advertiser, 5 May 1859. Meanwhile the poll book for Banbury’s elections in 1859 was published in 1865. And the polls for Cambridge for 1852, 1854, 1857, and 1859 were subjected to a longitudinal analysis ‘The poll-book, 1859’, Cambridge Chronicle, 21 May 1859.

[31] The poll books for the Middlesex election of 1802 are unusually comprehensive in this respect, recording the place of residence of the voter together with the location and nature of the freehold and the name of its occupier.

[32] The poll list for the single-member seat of Banbury in Banbury Advertiser, 20 July 1865, records the surname, forename initial, and street of residence of each voter for Bernhard Samuelson, then the surname, forename initial, and street of residence of each voter for Charles Douglas, then the surname, forename initial, and street of residence of each voter for Charles Bell. In this format each column of the Banbury Advertiser economically contained the polls of two voters alphabetically sorted. In total the polls of 531 electors are recorded in less than two whole newspaper columns.

[33] ‘A list of the persons names who have polled’, supplement to Freeholder’s Journal, 27 April 1722.

[34] ‘A true and exact copy of the poll’, Public Advertiser, 11 February 1765, and ‘A list of the burgesses’, London Evening Post, 9 April 1774.

[35] R.M. Wiles, Freshest advices: early provincial newspapers in England (Columbus, Ohio, 1965), 373 and accompanying chart usefully summarises the state of the English provincial newspaper press up to 1760. Of the provincial towns shown in H. Barker, Newspapers, politics, and public opinion in eighteenth-century England (Oxford, 1998), 112 as having newspapers in the 1780s, only Leeds, Manchester, Whitby, Whitehaven, Wokingham, and Sherborne were not parliamentary constituencies.

[36] ‘A correct alphabetical list of the burgesses and freemen’, Shrewsbury Chronicle, 12 November 1774.

[37] Few things bring the specialist local historian more pleasure than chastising the presumption of the generalist; and few things bring the generalist more pleasure than the keen eye of the specialist. I would welcome information on any printed polls that have been overlooked.

[38] H. Barker, Newspapers, politics and English society, 1695-1855 (Harlow, 2000).

[39] ‘A copy of the poll of the electors of the borough of Banbury’, Banbury Guardian, 15 July 1869.

[40] ‘A true and exact copy of the poll’, Public Advertiser, 11 February 1765.

[41] A list of the corporation of Winchester [1715]. As this has a title is might properly be considered a poll book, although it was issued as a single sheet. This contested election was not noticed in Hist. Parl. 1715-54.

[42] ‘Devizes borough election’, Devizes and Wiltshire Gazette, 29 March 1838.

[43] ‘Thetford borough election’, Bury Free Press, 22 July 1865.

[44] Borough of Knaresborough. The poll on the election of two burgesses (Knaresborough: William Parr, 1847).

[45] ‘North Northumberland election’, Newcastle Courant, 20 August 1847. According to Stooks Smith, 2,490 electors polled.

[46] Supplement to Liverpool Mercury, 15 July 1853. The previous election having been declared void, electors had two votes at their disposal. A Conservative headed the poll with 6,034 votes; his running-mate received 5,543. The highest-placed Liberal received 4,673 voted, and the hapless John Bramley-Moore was at the bottom of the poll with 1,274.

[47] J. Gray, An account of the manner of proceeding at the contested election for Yorkshire, in 1807 (York, 1818), xxiii. See also E.G. Wilson, The great Yorkshire election of 1807: mass politics in England before the age of reform (Lancaster, 2015).

[48] ‘The poll’, Yorkshire Gazette, 11 December 1830

[49] The poll ‘book’ for York in 1837 held at the IHR appears to be columns cut from the York Chronicle. If this is not the case, then it is evidence of the practice of newspaper publishers recycling their columns into pamphlet form.

[50] The published polls of the Westminster elections of 1749 and 1774 were necessarily large, recording the votes of over 9,000 and 7,500 electors respectively. They were issued at 2s. 6d.

[51] A list of persons who voted in the election of members for the borough of Blackburn (Blackburn: J. Burrell, 1835).

[52] Two bibliographical guides to electoral registers are J. Gibson, Electoral registers, 1832-1848 and burgess rolls (Bury, 2008), and R.H.A. Cheffins, Parliamentary constituencies and their registers since 1832 (1998). The latter includes a list of poll books held in the British Library. In United States usage the term ‘poll book’ is used of a list of those eligible to vote.

[53] An example is the marked electoral register from the eastern division of Worcestershire in 1868. It is held in TNA, PRO 30/1/3.

[54] ‘Burnley election. Committee for promoting the return of Lt. Gen. Sir James Yorke Scarlett’, Burnley Advertiser, 14 November 1868 occupies seven columns of newsprint and lists many hundreds of names. Scarlett received 2,238 votes, suggesting that much of his support came from members of his committee.

[55] ‘To Sir Francis Burdett, Bart.’ Devizes and Wiltshire Gazette, 27 July 1837, and in many other newspapers.

[56] ‘To T.H.S. Sotheron, Esq.’, Devizes and Wiltshire Gazette, 8 February 1844.

[57] Its publication was reported in Tadcaster Post, 3 December 1868.

[58] ‘Weymouth election’, Dorset County Express, 1 December 1868.

[59] Simple lists of all the voters for one candidate are remarkably uncommon. An example is [H. Campbell], The names of those patriots, in the parish of St John, that supported Mr Tierney (1796).

[60] J.R. Vincent, Pollbook: how Victorians voted (Cambridge, 1967), 1. After about 1780 poll books were usually printed in the constituency, but a few were printed in London. The poll for New Shoreham and Bramber in 1784 and that for Queenborough in 1826 bore a London imprint, as did the polls for Maldon in 1841 and 1857. Meanwhile among county contests the polls for the southern division of Derbyshire in 1832 and the northern division of Lincolnshire in 1852 bore London imprints.

[61] In The case of Mr Richard Holoway (1727), ESTC T 20087.

[62] 7 & 8 William III c. 25.

[63] A list of the liverymen of London, who voted for Mr Alderman Sawbridge and John Atkinson, Esq., at the late election of Members of Parliament for the City of London (London: W. Lane [1784]). ESTC T 198099.

[64] A list of the persons who have polled for Richard Atkinson, Esq., which the liverymen of the City of London are desired speedily and carefully to examine [London, 1784], ESTC T 200237.

[65] ‘First election for The Hartlepools. Who supported the Tories?’, South Durham and Cleveland Mercury, 5 December 1868.

[66] In A copy of the poll (Banbury: William Potts [1835]).

[67] ‘Devizes election’, Wiltshire Independent, 22 February 1844.

[68] Bolton Chronicle, 31 July 1847.

[69] W. Brimelow, Political and parliamentary history of Bolton (Bolton, 1882), 156-92 and 417-44.

[70] John Wilkes assiduously nursed his Middlesex constituency and annotated his copy of the Middlesex polls for 1768-9. The polls at, and since, the last general election for the county of Middlesex (London, J. Swan, 1772). BL pressmark C.45.g.3.

[71] The Westminster poll book of 1780, presented cadastrally on a street-by-street basis within each parish, was ideally suited to the purposes of canvassing. However, in all the rich documentation of the Westminster election of 1784, there remains no trace of any such use.

[72] P. Jupp (ed.), British and Irish elections, 1784-1831 (Newton Abbot, 1973), 127-9.

[73] ‘A list of the freeholders who voted for Colonel Luttrell at the election for Middlesex’, Oxford Magazine, III (1769), 91-5.

[74] ‘List of the freemen, who voted for a representative in parliament for the borough of Barnstaple’, Barnstaple Miscellany, I (1823-4), 345-50. This sort of poll list is among the hardest to find.

[75] Barnsley Chronicle, 21 November 1868.

[76] ‘Ancient poll book’, Wigan Observer, 27 April 1860; ‘Wigan’s ancient poll book’, Wigan Observer, 30 April 1910; and ‘An alphabetical list of the poll for the borough of Ludlow at the general election in 1852’, Ludlow Advertiser, 25 March – 15 April 1905. It seems likely that the last of these was derived from a printed poll book which is no longer extant.

[77] Cheltenham Journal, 5 July 1852.

[78] Cited in M.F. Suarez, ‘Towards a bibliometric analysis of the surviving record, 1701-1800’ in M.F. Suarez and M. Turner (eds), The Cambridge history of the book in Britain, v, 1695-1830 (Cambridge, 2009), 57.

[79] John Phillips drew attention to the missing poll for Liverpool in ‘Poll books and English electoral behaviour’, in J.M. Sims (ed.), A handlist of British parliamentary poll books (Leicester, 1984), vi.

[80] J. Allen, Bibliotheca Herefordiensis (Hereford, 1821) notes the otherwise unrecorded poll books for the borough of Leominster in 1741, 1784, and 1790. The systematic study of county bibliographies may bring a few more of these bibliographic echoes to light.

[81] The original remit of the committee appointed in 1929 was to report on ‘the materials available for a record of the personnel and politics of past members of the House of Commons from 1264 to 1832’, L.B. Namier and J. Brooke (eds), The House of Commons, 1754-90 (3 vols, 1985), i. v.

[82] For the attenuated contest for Yorkshire in December 1830.

[83] The eastern division of Cumberland in 1837; the northern division of Northumberland in 1847; Bedfordshire in 1857 and again in 1859; and the northern division of Essex at the by-election of May 1835 and at the contests of 1847 and 1865; and the southern division of Essex at the by-election of June 1836 and at the contested elections of 1837, 1847, 1852, 1857, 1859, and 1865.

[84] The printed polls from Banbury in 1831 and from Petersfield in 1727 are later printings.

[85] M. Taylor, ‘Interests, parties, and the state: the urban electorate in England, 1820-1872’ in J. Lawrence and M. Taylor (eds), Party, state and society: electoral behaviour in England since 1820 (Aldershot, 1997).

[86] J.E. Norton (ed.), The letters of Edward Gibbon (3 vols, 1956), vol. 2, 270. The by-election of 25 June 1781 was uncontested.

[87] An alphabetical list of the burgesses & freeholders (Nottingham: Samuel Cresswell, 1754), ESTC T 161928; A copy of the poll of the burgesses and freeholders (Nottingham: Thomas Collyer, 1754), ESTC T185193; and An exact copy of the poll (Nottingham: Samuel Cresswell, 1754), ESTC T 183386.

[88] A correct alphabetical list of voters (Lymington: Edward King, 1859) and An alphabetical list of the voters who voted (Lymington: Frederic L. Watson, 1859). Each survives in only a single known copy.

[89] The poll book for the borough (2nd edition, Carlisle: Charles Thurnam, 1857).

[90] Advertisement in Southern Times, 7 May 1859. However, the record in R. Cheffins, Parliamentary constituencies and their registers since 1832 (1998), 235 of a poll list for Dorset in 1859 being held by the British Library Newspaper Library is spurious: there was no election contest for Dorset in that year.

[91] Correspondence in Huddersfield Chronicle, 7 August 1852, and in Huddersfield and Holmfirth Examiner, 7 August 1852, refers to a ‘List of the poll’ announced in Huddersfield and Holmfirth Examiner on 24 July 1852 for appearance the following week. It is possible that this is the broadsheet edition listed in Sims, and that this broadsheet was issued with the newspaper rather than being a formal supplement to it.

[92] The phrase ‘apparently reissued’ is used because it is not possible to establish the identity of the issues without the most careful comparison.

[93] The ancient parliamentary borough of Hindon boasts two inns, at opposite ends of the main street running through the village. And we remember that in pre-Reform Middlemarch Mr Brooke engaged Will Ladislaw to edit The Pioneer as a rival newspaper to the established Trumpet.

[94] Careful comparison of these may yet reveal that they were printed from the same setting of type. This work has not yet been done.

[95] ‘The poll at the city of Durham election’, Durham Chronicle, 1 July 1853, and ‘Freemen on the register who did not poll’, Durham County Advertiser, 1 July 1853.

[96] The poll for a member (York: John Jackson) and The poll for a member (York: N. Nickson, 1759).

[97] With the exception of the by-elections of March 1838 and June 1838 these polls appear only to have been produced as pamphlets. It is curious that a town so richly endowed with printed poll books, and having two rival newspapers, should have eschewed the publication of poll lists.

[98] Examples include Stamford in 1847 (two pamphlets and a newspaper list), and the northern division of Essex in 1847 (one pamphlet and two newspaper lists).

[99] Bedfordshire Times, 14 May 1859; Bedfordshire Mercury, 16 May 1859; General election, 1859 … The poll (Bedford: Frederick Thompson, 1859), and The poll, taken at the Bedfordshire election, 1859 (Bedford: Rowland Hill, 1859). This last is known only from an advertisement in Bedfordshire Mercury, 23 May 1859.

[100] This last known only from its being reprinted in Ludlow Advertiser, 25 March 1905, 1 April 1905, and 15 April 1905.

[101] 30 & 31 Victoria, c. 102.

[102] C. Cook and J. Stevenson, A history of British elections since 1689 (Abingdon, 2014), 252.

[103] 35 & 36 Victoria, c. 33.

[104] The only systematic error in the Sims Handlist was the exclusion of those polls to be found in British Parliamentary Papers, although they were included in the bibliography to H.J. Hanham (ed.), Electoral facts from 1832 to 1853 impartially stated (Brighton, 1972). These include the polls for Hindon in 1774, Stafford in 1832, Roxburghshire in 1837, and Maldon in 1847 and 1852. It is possible that the ‘missing’ poll book for Poole in 1837 is in fact the municipal poll for that year in B.P.P.

[105] A true copy of the poll, Bodl. G.A. Radnor b.1; The poll for a member of parliament to represent the City of York, Bodl. Vet. A5e. 7433; and The poll taken at Colchester, Bodl. Vet. A5d. 1778.

[106] E.M. Green, ‘New discoveries of poll books’, Parliamentary History, 24 (2005), 334.

[107] Advertised in Windsor and Eton Express, 5 August 1837, and other issues of the same newspaper.

[108] Poll book of the electors who voted at the election for the borough of Tynemouth, 1837, noted in Newcastle upon Tyne public libraries committee, Catalogue of material concerning Newcastle and Northumberland (Newcastle upon Tyne, 1932), 370.

[109] ‘The poll book for the borough of Dewsbury has just been published, price 6d’ announced the Barnsley Chronicle on 27 February 1869.

[110] Gravesend Reporter, 5 December 1868.

[111] Windsor and Eton Express, 10 July 1852.

[112] Hampshire Chronicle, 4 April 1857.

[113] Preston election, 1868. The poll list (Preston: Herald Steam Printing Office). Meanwhile the Preston Chronicle maintained a conscientious objection to printing poll lists.

[114] Edward Gibbon, Autobiography (World’s Classics edition, 1907), 160.