The decades at the end of the nineteenth and the beginning of the twentieth centuries saw the movement of millions of men, women, and children out of central and eastern Europe. Where many earlier immigrants seem to have been little more than vagabonds moving from town to town on the continent and finally unloaded upon Britain by their respective communities, these features were to be changed in the 1870s and thereafter by a number of political, religious, economic, and demographic pressures. Most of these later migrants were seeking their fortunes in North America, sometimes in order to make sufficient money to return as rich men to their native countries, more often to make a completely new life in a new world.
For the overwhelming majority their initial steps were closely similar. Those travelling from central and northern Eastern Europe, from Poland and the northern part of the Russian Pale of Settlement, from the Habsburg Empire, and from Rumania, would for the most part make their way to German, Dutch, or British ports in order to take passage across the Atlantic. Some of them had left Russia legally, but according to one estimate 90% of the emigrants from Russia left the country furtively by stealing across the frontier. There grew up a vast network of travel agents, medical facilities, customs officials to deal with all these migrants; there developed a network to exploit them, and another to protect them from exploitation. There was the Hilfsverein in Berlin and the Israelitische Allianz in Vienna, the Montefiore Vereiniging in Rotterdam and the Hebrew Immigrants' Aid Society of New York all playing a very significant part in the migration process. In particular the Hilfsverein in Berlin had been established as the centre of the German network, publishing pamphlets about immigration laws and the journey, English phrase books, repatriating rejected emigrants, tracking down lost baggage, providing kosher food, shelter, clothing and medical care. On the other hand in Paris, where there was an institution, the Asile, set up specifically to give accommodation to Eastern European Jews, there is no evidence of any involvement in transmigration as distinct from the reception of immigrants.
Many of these migrants tended to come through Great Britain; originally the vast bulk of the transatlantic shipping trade had been in the hands of British companies, and even when such companies as the Hamburg-America Line began to demand a share it was still cheaper under various circumstances for migrants to come into Great Britain and travel onwards by British ships. The creation of a 'ring' and a standardised fare structure for the Atlantic crossing did not prevent individual companies from under-cutting, and at one stage the fare from Liverpool fell to £2.10s. [£2.50 pence]. Food was extra but little attention would be paid to that by many. Some of the passengers would have been worried about the standards of kosher food offered on the voyage, and so would have prepared food for themselves in advance of their journey - with baked biscuits, oranges, hard-boiled eggs, herrings, and tea they felt that they could cope. Many travelled 'steerage', and conditions on board the ships were far from good. Many attempts were made, above all by the American authorities, to regulate and improve conditions in steerage. The fact that several such sets of regulations were issued suggests that earlier regulations had not really worked. There was not always adequate provision for safety, while the sexes were often mixed indiscriminately. Certainly there was at least one ship which is alleged to have carried notices directed at the steerage passengers: 'All couples making love too warmly would be married compulsorily at New York if the authorities deemed it fit or should be fined or imprisoned.'
There were two main streams of passage amongst those who came to Britain; many arrived at various ports along the east coast, mainly at the Hull/Grimsby complex but also in Harwich, Newcastle, or Leith, while many others came to London. Those who came to ports outside London seem normally to have intended to travel to America, either immediately as transmigrants or else after having made enough for their onward passage. At all events the small communities who received them were faced with enormous problems. The least they could do would be to meet the boats, feed the new arrivals, and see them on their way, at least to the nearest Jewish community along the railway line from the east coast to the west. Even this represented an enormous financial burden, and the Jewish Chronicle is filled with requests from the various local societies for urgent assistance. One such point of entry was Grimsby, where it was stated that 'The Docks and wharves practically belong to the Great Central Railway Company which also owns several vessels. This company entirely undertakes the forwarding of immigrants and transmigrants. It has built a very large shelter in which aliens are lodged until they are forwarded by train. This ... consists of one huge room capable of accommodating 200-300 persons and a few small rooms for second class passengers. As the men are not separated from the women it can be regarded as a huge waiting room, each individual being provided with a mattress on which he or she can sleep.'
In all these Jewish communities there were Boards of Guardians designed to assist the travelling poor, but some went much further. The Jewish Yearbooks for these years indicate for example the existence in such communities as Dublin, Leeds, and Liverpool of societies intended for the temporary shelter of 'poor strangers' of the Jewish faith. Such associations however were usually on a small scale, often operating on an ad hoc basis, while they often lasted only a short while.
There was one exception to this, London. Here there had been established in 1885 The Poor Jews' Temporary Shelter, designed to meet the needs of Jews who were coming to or passing through London. O J Simon wrote in the Jewish Chronicle in 1885:
"At 19 Church Lane Whitechapel there is a refuge for Jewish people who are out of employment. ... It is obvious that this place of shelter cannot encourage people to be idle, because its abject misery is worse than any workhouse, and it provides less food. There is absolutely no sleeping accommodation except a wooden floor. ... The place, such as it is, is maintained by two co-religionists in very humble circumstances."
One of the best descriptions of the conditions that the Shelter was set up to replace was given by the Shelter's Vice-president and founding father, Hermann Landau, to a Select Committee of the House of Commons in 1888:
"A large number of poor Jews have become homeless ... ; they congregated together in a place in Church-Lane for a considerable number of years past ... I visited their place and I found it in a most insanitary condition. In fact it was a permission given by a Jewish Baker, to whom the premises belonged, for those outcasts to lie down in a kind of loft; and as for undressing, that never could take place there."
He described the fate which would often meet those who had no help on their initial arrival in London.
"There are two classes that arrive [at the Shelter]; one class that comes directly from the ship to the Shelter, and who generally bring money with them, and the other class are those who have been got hold of by so-called porters and lodging house keepers, or, as they represent themselves, shipping agents, and they are generally robbed of every penny they possess and are then set adrift."
On the other hand, care had to be taken that the Shelter should not be regarded as a magnet for Jews to come to the country on a permanent basis, and its Reports continually stress that the institution was not to be regarded as a haven for would-be immigrants to England:
"The bulk of the inmates were passengers in the true sense of the word, for they were en route for various lands across the seas, mainly for South Africa. And the Committee are pleased to add that the class of emigrants is improving in every sense; they possess not only sufficient means to pay their passages but in the large majority of cases have funds enough left to start a new life with."
A considerable amount of information is available for the Shelter. There are a number of Annual Reports which though incomplete give series of figures and comments illustrating the working of the institution. There are a number of Minute books covering the activities of the Executive Committee (though only until 1900) and of the General Committee. But, most importantly, after May 1896 there is a series of registers drawn up by the officers of the Poor Jews' Temporary Shelter. They quite clearly mark a significant new departure for the Shelter, evinced by the fact that the first entry in the first book is numbered ONE. Thereafter the numbers are usually continuous, changing only with the beginning of the Shelter's financial year, i.e. 1st November. Even when there are volumes missing for various periods thereafter none of these ever open with number 1. Clearly there must have been some registers kept before 1896; for instance according to evidence given in 1888 by Hermann Landau, its President, to a House of Commons Select Committee on Foreign Immigration there were lists noting for each inmate his name, his age, the place from which he came, whether he was married or single, and if married the number of children he had left behind him. There were notes about whether he had a trade, a record of the number of days he stayed in the Shelter, and finally where he went on to after he left the Shelter. None of these early volumes survives, and there are important question marks as to why this new series was compiled.
The initial thought behind the original use of this material as a database for undergraduate use was that it would provide an exercise for the students and a demonstration of the uses that could be made of statistical material. It certainly lent itself to straightforward tabulation since it was laid out in such formation. In the event the work has become the foundation for an important piece of historical research, most particularly into the study of patterns of migration at the end of the nineteenth century. It has also opened a number of further questions, in particular concerning the part played by shipping companies in manipulating the potential destinations of possible passengers. Between these years tens of thousands of migrants, largely Jews from Eastern Europe, passed through the Shelter on their way to a new life. The personal details of many of these persons were recorded in the series of thirteen massive volumes already mentioned, spanning the years 1896 to 1914. From these it is possible to construct a fascinating picture of one of the significant population shifts in modern times. However, given the enormous number of individual records and the range of details given for each person, it would have been impossible to analyse the information without the use of a computer database.
The database has been set up using Ingres, on the University's central UNIX system (formerly on a VAX system). Ingres was originally adopted because in 1987, when the project started, it was the only database software available to us. Since then other software has become available, running on PCs and with more modern and attractive interfaces: it is tiresome in Ingres not even to have the use of a mouse. Nevertheless, we continue to prefer the robustness of an industrial-strength system such as Ingres to the 16-bit PC alternatives, particularly given the large-scale multi-user nature of our work. Moreover, the UNIX platform affords better control and monitoring of users' work than would NetWare, and allows cheaper and wider access. At the local level, a UNIX link provides access from student halls of residence, for example, while globally Telnet access is possible from around the world: remote access facilities are currently being set up for a group of fellow researchers in South Africa.
Originally, access was exclusively via VT220 dumb terminals, and these are still used to some extent. More commonly nowadays client terminals are PCs running terminal emulation software. A Telnet VT220 emulator was used for a year with tolerable success, but was found to have significant limitations: its emulation did not extend to 132-column mode, for example, nor to underlining. After trials with Emu-Tek, which we found to be powerfully featured but not very user-friendly, we have finally settled on Kermit as our preferred current solution. Kermit has now been giving us reliable and effective service for nearly two years, providing a good balance between features and ease.
Neither Ingres nor UNIX in their default form could remotely be described as easy for the inexpert, so the system has been substantially customised to facilitate access and use. Input screens - and similarly query and report screens - are driven by application programs specially written in Ingres's 4GL application language. UNIX Korn shell scripts (previously VAX DCL scripts) provide a front-end menu for access to the databases and handle subsidiary tasks such as the setting of keyboard mappings and the routing of print jobs to the print queues. The power and flexibility of the Korn shell scripting language is another reason for preferring UNIX to a NetWare environment.
The database itself is of simple basic design, as set out in the accompanying data structure diagram. The data from the Shelter registers is entered into just two tables, Persons and Moneys:
Persons holds the great bulk of the data. including each person's name, date of entry to the Shelter, sex (alas, rarely clear), age, place of birth, marital status, place last from, occupation, length of stay, date of departure, place gone to, and the ship on which the person left.
Moneys is a small subsidiary table, recording the money the person is declared to have had, sometimes in several currencies, hence the need for a separate database table.
These data input tables are supplemented by a number of look-up tables, used to regularise and expand on some of the information entered:
Occupations allows for the translation of the very variable individual occupation descriptions into standard occupation types
Places similarly provides for the standardisation of place-name spellings and also for the expansion of place information, e.g. by identification of the country
M_codes standardises currency codes, which again can be very variously written in the registers, and allows for (approximate) conversion of non-sterling amounts into sterling equivalents using contemporary exchange rates.
To the above basic structure a number of other more specialised input tables have been successively added as supplementary sources have come to light, such as South African naturalisation lists and shipping records.
Over 40,000 records have now been entered from the Shelter registers, more than half of these by the combined efforts over the last seven years of some 100 undergraduate students. This use of undergraduates as data inputters is an unusual feature of the database. It has been motivated partly by the Department's belief that students should be given some realistic first-hand experience of the use of computers in genuine historical research, as part of their education in computer methods. At a less elevated level, the motivation has been that as a captive labour force students come free. The work is prescribed as a compulsory part of the students' degree programme, and a mark for the care, intelligence, and accuracy of their input contributes a small amount to their degree assessment. Despite the carrot offered by inclusion of the work in degree assessment, the effort and diligence of students has naturally varied quite widely, as has their expertise. A good deal of attention has therefore had to be given to quality control procedures to safeguard the consistency and accuracy of the input.
Students begin by copying out by hand their allocated set of register pages onto transcription sheets. In the early years each student could tackle some 300 entries, but in the tighter constraints of a modular timetable the allocation has been reduced to a little over 100. The transcription sheets are next checked against the originals by the project's research director, who writes in any necessary corrections, after which students enter the corrected transcription sheets onto the computer. As they do so, validation checks built into the input screen guard against obvious errors, such as invalid volume numbers, incorrect date patterns, or faulty capitalisation. Given the variable nature of historical data, however, the amount that can be automatically checked in this way is very limited. Every record added is tagged with a code to denote the inputter, this serving not only to identify who was responsible for the record but also as a security device to ensure that only the person who input a record can subsequently alter it. A printout of the completed input is then checked, first by the research director, who is concerned primarily with its historical accuracy, and then by the database director, who checks additionally for consistency with the database rules. The marked printout is then returned to the student for the corrections indicated to be entered, after which a further printout is run off of the corrected data. The corrected printout is finally checked against the original marked printout by postgraduate students, who highlight any required corrections that the inputter has neglected to make. It is then ultimately the responsibility of the database director to correct any errors still remaining.
A detailed input manual instructs students on the rules to be followed for each field, and endeavours to provide guidance on exceptions and difficult cases. One important datum, for example, is the sequential reference number that the Shelter clerk wrote alongside each entry, since the database uses this to ensure that records can be printed out in the same sequence as in the source. Occasionally, however, the clerk absent-mindedly got the number wrong or neglected to add a number at all. In such cases inputters are given rules to follow for the construction of a decimal reference number, which preserves both the correct sequence of records and the number actually written. A more general problem is that presented by the frequent inconsistency of the Shelter clerks, particularly in their spelling of place names and occupations. Bessarabia, for example, might be written (or misread by the inputter) as Besarabea, Besarabia, Besarabya, Beserabia, Besserabia, etc., while a boot maker might be described as boot m, boot m., boot ma, boot mk, boot mk., boot mkr, boot' m, boot'm, bootmaker, bootmkr, bootmkr., and so on. Although at first sight it seemed desirable to silently standardise spellings and expand obvious abbreviations in the course of input, it quickly became clear that with dozens of inputters consistency would be impossible to achieve, while the attempt would also risk corrupting the data by unwitting misinterpretation. A good illustration is provided by the ‘zing workers' whom we were inclined at first to transcribe as zinc workers until we realised that zinn is the German for tin, making tin workers a more likely interpretation. Our practice has therefore been to instruct inputters to enter precisely what the clerk wrote and resist any temptation to correct or expand. The look-up tables mentioned above are then employed to translate variant spellings into standard forms and probable interpretations when the database is interrogated. The look-up table for place names currently holds over 3,500 variant spellings (or sometimes simply misreadings or mis-typings). Another general issue concerns the transcription of poorly legible text. Inputters were here instructed to enclose in curly brackets any text that could not clearly be deciphered. We have found, however, that the reading of names, particularly, so frequently involves a greater or lesser degree of uncertainty that to have followed this convention rigorously would have been tedious in the extreme. In practice, therefore, the convention has been only lightly used.
It will be apparent from the above account that the input procedures followed have been cumbersome and time-consuming, while it is equally evident from the not infrequent errors that continue to surface that the checking process has been less than foolproof. It cannot be pretended that the use of undergraduate labour is an ideal way to build up a research database. On the other hand, without the more or less willing co-operation of our dozens of student assistants the database would probably never have existed at all. With hindsight there are a number of ways in which the work might have been more effectively handled. In the first place, our experience has highlighted the importance of a close co-ordination between the historical and the technical aspects of the work. It has proved not to be satisfactory, for example, to treat the initial manual transcription of the data as a purely historical exercise. The transcription sheets should preferably reflect the structure of the database rather than that of the register pages, and inputters should transcribe the materials according to the input rules, rather than regarding the input as a separate exercise to be thought about later. The checking of the manual transcriptions should similarly check for consistency with the rules of the database and not just for the accuracy of the reading. Failure to observe these basic points has resulted in much erratic input finding its way into the database, and has cost us much time and effort in the attempt then to eradicate it.
Above all, the supervision of the input could probably have been made very much less onerous, and the results more reliable, by use of the kind of double-entry input methods that are common practice in the commercial world. By having two students independently input each set of data, a computer comparison could then have highlighted errors by pointing up differences, so relieving us of much of the labour of manually checking every detail while at the same time more reliably catching errors that inevitably to some extent slip through even the most careful manual check. A double-entry methodology would also have enabled us easily to estimate the level of residual error remaining in the database after students had completed their work, and identify the outstanding corrections still required. A single-entry methodology leaves one uncertain in this regard, and one is faced with the daunting prospect of another manual scan if further checking is deemed necessary. The value of double-entry methods with suitable historical data has been proven in another of the Department's databases, and would probably have suited the Shelter records well: while much historical material is too complex or amorphous in structure for the technique to be readily feasible, the form of the Shelter materials is relatively simple and well-defined. However, entering everything twice would obviously have slowed the growth of the database, and in the familiar conflict between quantity and quality that increasingly afflicts the academic world - exacerbated by research assessment exercises and other outside pressures - the demands of speed prevailed.
The use of the resulting database falls into two quite distinct categories: data retrieval and data analysis. Data retrieval is simple, following the usual query-by-example methods, and is primarily useful for satisfying the curiosity of inquirers about family members. It is of relatively limited value for broader historical investigation, which calls for the considerably more sophisticated techniques of data analysis. Initially, analysis was carried out by SQL queries in Ingres, but frustrations with the limitations of SQL, and particularly with the awkward and unhelpful Ingres SQL interface, have led us over the years to place increasing reliance on the excellent SAS data analysis system, which also runs under UNIX. To this end, the Ingres database is converted into the form of SAS data sets. SAS offers far more extensive facilities than Ingres for the manipulation, statistical analysis, and graphical display of the data, in the form of line graphs, scatter plots, bar charts, pie charts, block charts, and all manner of complex cross-tabulations. SAS Display Manager also provides a more convenient working environment, and its error messages are a valuable aid in debugging programs, in contrast to the cryptic, uninformative, and sometimes misleading messages thrown out by Ingres. The incorporation in recent versions of SAS of the standard SQL language, alongside its own proprietary language, means that there is now little reason to use Ingres SQL other than for database management, and it is SAS that students now prefer almost exclusively when undertaking the researches for their Finals projects. The resulting division of our database work between two major systems has obvious inconveniences, requiring the data to be held both in Ingres and in SAS formats. However, at present neither system alone seems entirely adequate to the task. Ingres offers superior data input and management facilities, while SAS proves preferable for data analysis and presentation.
Among major areas of analysis that still await further development are those involving data linkage. The database presents two main kinds of linkage problem. In the first place, it is clearly interesting to study the extent to which migration was in family groups, and whether this changed over time or varied with origin or destination. To facilitate analyses of this kind, the database assigns automatic family identification numbers: all the migrants entering the Shelter on the same day, sharing the same surname, and whose names follow each other directly in the register, are assumed to form a family group and are each assigned the same fam_id number. This then allows investigation of the proportion of migrants who travelled in groups, the frequency of different group sizes, the typical composition of groups, etc. Preliminary results indicate that overall about 22% of migrants were in fact travelling in such groups; for the important set of migrants that came from Kovno the proportion was rather higher at 30%. A second and more difficult kind of linkage problem is to identify cases where the same individual occurs in the database more than once, as a result of having figured in more than one source. This applies most notably to the series of supplementary registers, whose coverage extensively overlaps the main series registers, but with some variation in the information given. Establishing links in such cases is complicated by differences in the spelling of surnames, and by the fact that the supplementary registers record forenames whereas the main registers for the most part give only initials. The extensive string manipulation functions available in Ingres and SAS, together with the soundex search facilities in SAS, offer promising tools for tackling such questions, but reliable progress in this direction must await completion and further correction of the database. Similar issues arise in the establishment of links between Shelter records and other sources, such as South African naturalisation lists.
As already mentioned, in their final year students undertake a small research project on the database as a whole, designed to test their abilities to use the computer as a tool of academic research. Many of these projects have involved straightforward sets of comparisons within certain elements of the database, but there have been some projects showing a more imaginative approach to the sorts of analysis that the material makes possible, such as the use made of the various ports in Europe through which migrants travelled, the extent of 'family reconstitution', or a comparison of the nature of the migrations to South Africa and South America. Partly in order to facilitate such projects, an initial decision was made to approach the series of records on a very wide front; material for input was drawn from each of the volumes, thus covering a wide variety of dates between 1896 and 1914. This adoption of a sampling technique enabled the investigator to get a feel of the material as a whole, even if individual students in the course of their projects came up with untypical findings. There was, for example, the student who found himself facing a large number of returnees, and who wanted to know why Kovno had suddenly become so attractive. He had not appreciated that these individuals had been deported from South Africa at the height of the Boer War. Two other students, dealing with a period just a few weeks later, were faced with a wave of Jews from Rumania who had, apparently, walked much of the way from there to Holland.
There was the time when a group of students were working on a section of the registers in 1909/1910 when 1165 Continental agricultural workers from Armenia, Austria, Bulgaria, Germany, Hungary, Italy, Russia, Servia, and Syria arrived from Antwerp and Rotterdam all booked for Canada. The Shelter had agreed to take charge of them, there being no other organisation in London that could handle such a large party. The Jews and some of the non-Jews were accommodated at the Shelter whilst the bulk of the other passengers were boarded and lodged out in registered Lodging Houses in the neighbourhood. They all duly sailed on 5 April, but three days afterwards the Shelter authorities were told that the ship had caught fire off Dover; all the passengers were returned to London with literally nothing at all except what they had been wearing at the time. The Shelter once again had to cope.
As time has gone on it has been possible to concentrate on filling in the gaps created by the widespread approach, and at the same time to introduce a number of other records which throw light upon the database and upon the institution which created the records. Since most of those who passed through Britain during these years were intending to travel to America it was thought at first that many of those passing through the Shelter would have been travelling there. But as further work and analysis was undertaken a very different picture emerged. It was not merely that a substantial number were en route for South Africa but that an overwhelming majority of these were from a particular part of Eastern Europe, the Russian 'governmental district' of Kovno. In addition, examination of the material in the registers as transferred to the computer revealed a close connection between this migration and the ships operated by the Castle and Union Shipping Companies (later amalgamated into the Union Castle Shipping Company). Not only were no other ships travelling to South Africa recorded but the registers contained a very large number of references to accounts being drawn chargeable to Donald Currie and Co., the agents managing the Castle Shipping Company. Reference to the surviving Minute Books of various committees of the Shelter reveal a growing connection between the Shelter and the shipping companies; as early as 1893 the Executive of the Shelter were worried about an 'underuse' of its facilities and there was a call that various shipping companies be approached with a view to encouraging them to act in co-operation with the Shelter. It was shortly after that that the Annual Reports of the Shelter (and the returns made by the Board of Trade annually to the House of Commons) began to record an initial trickle of migrants to South Africa - there were apparently no such migrants before 1893 - and that Donald Currie began to send cheques to the Shelter (all duly recorded in the minutes). The numbers of such migrants rose from 400 in 1892/93 to over 2000 in 1895/96; the minutes record a temporary increase in salary for the Superintendent of the Shelter 'for as long as the present migration to Africa continues', and it can be no coincidence that it was at this point that the registers were initiated. Clearly at this stage a special relationship had been established between the Shelter and Donald Currie, and the evidence of material drawn from the registers, from the minute books, and from a solitary 'Copy Letter book' for 1906 is that that link continued for many years thereafter. The shipping companies sailing to South Africa had invested heavily in a series of modern liners designed specifically for a 'migrant' trade and the evidence of the passenger manifests deposited with the Board of Trade show the extent to which the passengers recorded in the Shelter's registers filled Currie's berths.
What also appears is the extent to which the Shelter itself became increasingly dependent on such transient trade. At the suggestion of Sir Donald Currie the Shelter wrote to the newly founded Royal Mail Shipping Company offering its services for the benefit of migrants to South America; it would meet passengers at the London docks, take care of them during their stay in London, and then see them safely to their onward passage, making sure that under the provisions of the 1905 Aliens' Act 'none would escape'. A similar service was offered the Cunard Company for its trans-Atlantic transmigrants. By 1910 the Shelter had acquired such an accumulation of expertise that it was approached by a number of companies for it to act as their agent in handling large number of individuals travelling from the Balkans to Canada and Australia.
The individual projects have enabled a wide variety of analyses. With the growing number of records now on the computer it has been possible to fine tune such issues as the changing proportions of young children travelling in family groups and the light which this could throw upon issues of 'family reconstitution'. Analysis of the various ports in Europe from which transmigrants travelled suggests a growing use of ports in Russia as against western Europe, a possible demonstration of the desire of the Russian government to promote the use of its own shipping companies. Comparisons of various 'waves' of migration have demonstrated changes in occupation, demographical structure, and destinations in what originally appeared a rather formless mass of data.
One of the advantages in using a relational database is the possibility of relating the basic set of data to other, cognate materials. There is for example a volume which appears at some time to have strayed from the Shelter's own archives. It is a volume covering the period from mid-1909 until early 1910 and listing daily arrivals at the various London docks. Many of the names also appear on the Shelter's registers, but there are in addition a significant number of entries for persons who obviously were able to make their own way to addresses either in London or in the Provinces. There are also two sets of materials which emanated from archive collections in Cape Town. One is a list of some twelve hundred applications for naturalisation in Cape Province between 1904 and 1906. The applicants had to show evidence of residence in the Cape for at least two years; although a number had been there for longer than that enough of them had arrived after May 1896 to make it possible to make connections between the migrants who are recorded as having arrived at the Shelter and those who were now able to demonstrate their place in the new society. There is also a list of all the ships which arrived and landed passengers at Cape Town. No names are recorded, but in determining the total numbers arriving it has become possible to determine the proportion of Jewish immigrants to the total number of arrivals, and thus the significance of this particular ethnic migration.
Many questions remain to be answered, for some of which the computer remains necessary while for others there will be a need for 'traditional' research. The question of why there was such a close connection between South Africa, Kovno, and the Shelter is still unanswered. There were other immigrant organisations in Europe which need to be investigated, while there was a vast Jewish Colonisation Association in Russia whose archives have only been examined superficially. It would be of great help to discover the papers of the Union-Castle Shipping Company for these years. It would be interesting also to be able to assess the image of South Africa which might have been presented in the local press in Lithuania. Nonetheless, none of these issues could even have been raised had there not been a need to create a database suitable for undergraduate investigation.
The registers, minute books, and records of the Poor Jews' Temporary Shelter are the property of the Shelter, which retains the copyright of these manuscripts. The authors gratefully acknowledge the full co-operation given by the Shelter and its permission to make use of this material. Considerable financial assistance has been given through The Kaplan Centre, University of Cape Town. We also put on record our thanks to all the students without whose work, given willingly or otherwise, this research could not have got as far as it has.
