«Deteriorating Paper in Sweden A Deterioration Survey of the Royal Library, Gothenburg University Library, Uppsala University Library and the National ...»
A Deterioration Survey of the Royal Library, Gothenburg
University Library, Uppsala University Library and the
Originally published in Swedish: Ett vittrande kufturarv.
Uppsala University Library
ISSN 0284-5636 Stockholm, September 1988
Deteriorating Paper in Sweden
A Deterioration Survey of the Royal Library, Gothenburg
University Library, Uppsala University Library and the National Archives Originally published in Swedish: Ett vittrande kulturarv.
Jonas Palm Per Culihed Uppsala University Library Fo U-projektet for papperskonse rye H ng Report No.3- ISSN 0284-5636 Stockholm, September 1988
CONTENTSPreface 3 Summary 4 The Survey Methodology 5 The Surveys 8 The Royal Library 8 The National Archives 12 Conclusions of the Stanford Survey 14 The Comparison Between Uppsala and Gothenburg University Libraries 18 The P-values at UUB and GUB 18 The P-values and their Relation to the Chemical Properties of Paper 19 Groundwood - An Unstable Paper Fiber that is Sensitive to Different Environmental Conditions 22 References 23 Appendix 24
The first of a series of reports from the project is a survey of literature on internal and external factors of aging and deterioration of paper, including a review of the most important deacidification methods. The report has a short summary in English.
This report deals with the actual state of some major Swedish collections. The surveys are based on the Stanford method which was evaluated in a study by the Uppsala University Library last year. The results of the Uppsala study are integrated in this report.
The surveys have been performed by personnel at the different institutions and elaborated by a project group at the Uppsala University Library.
As the results of the Swedish deterioration surveys will probably be of wider interest, the report is also being published in a special English version with the previous Uppsala study as an appendix.
Stockholm, September 1988 Ingmar Fröjd Project Coordinator rd
SUMMARYThe results from a previously published survey done at the Uppsala University Library (Uppsala universitetsbibliotek - UUB), see Appendix, has been the basis for the present surveys done at the Royal Library (Kungliga biblioteket - KB), the National Archives (Riksarkivet - RA) and the Gothenburg Universty Library (Goteborgs universitetsbibliotek - GUB).
The surveys have been made within a national research and development project on paper conservation and carried out in two ways.
KB and RA made a Stanford survey with random sampling of the books to be tested and GUB made a survey on the same books, present in their library, that had been tested in the UUB survey in Uppsala with the purpose of detecting possible effects of different preservation environments.
The results of this report are, in short, the following:
-The number of books with paper in bad condition (P 3) is roughly the same at RA and KB, about 20%. The UUB survey showed a similar result.
-The highest number of books with bad paper - in proportion to the amount of tested books - is from a period between 1860 to 1890. The same goes for UUB.
Regarding the amount of deteriorated paper and the timing, the results are unambiguous and probably representative on a national level.
-RA has a great amount of paper with P 2-grading (worn), 30%. One explanation could be the nature of archive materials - it is "worn" already when it arrives at the archives. Another explanation could be that RA has a relatively large number of samples from the first half of the 19th century, which means they are almost entirely made out of rag. Especially the survey results from GUB showed that paper made of this fiber has a tendency to fall into the P 2 category owing to unfavorable visual appearance.
'The papers tested in the comparison between UUB and GUB that have the same P-values show low content of groundwood. It is the reverse with paper that have different P-values in UUB and GUB. Here there is a greater amount of groundwood present, which confirms other results on the instability of groundwood and its sensitivity to preservation environments.
'The distribution of P-values in the group of papers with more or less content of groundwood shows that the preservation environment has been more favorable in UUB than in GUB.
Extensive deterioration of paper in books and documents has been detected in the Swedish collections investigated. The main picture of the situation differs quite a lot, however, from a recently published survey made in some libraries in the Nordic countries. The differences can be explained on methodological grounds and the way the survey was performed. The results in this report show more resemblance with results from known international surveys though the state of the deterioration is not as alarming in Sweden.
A conclusion is that the results stress the importance of deepening our knowledge about the influence of climate and the preservation environment on the deterioration of paper.
THE SURVEY METHODOLOGY
During the last two decades several international reports have been published on the subject of deterioration of paper in archives and libraries. It is a disaster which has already started in some places and in other places it is expected in the not-so-distant future. The reason for this is the use of paper of inferior quality made since the second half of the 19th century. As the demand for paper increased the manufacturing processes were rationalized and wood fibers were introduced in paper making.
The international reports have published results with high percentages of books with deteriorated paper. The Library of Congress, Washington, says that about 30% of the collections are in such a state that they no longer can be used by the public; Harvard University Library reports that 40% of their collections are in the same state; and Stanford University Library 26%.
Similar figures come from other libraries.
During autumn and winter 1986/87 a survey was carried out at the Uppsala University Library.' The survey was based on the method used at the Stanford University Library, the one we found with the best documented methodology.2 The survey was based upon the hypothesis that if you can randomly choose a representative number of books in a large collection and judge them objectively you could tell how large a share of the entire collection was in a bad state. 384 books would give a statistical precision of 95% ± 5% and that number was chosen both by the Stanford University Library and Uppsala University Library.
The Uppsala survey started with the exclusion of certain types of publications. These were periodical prints, not cataloged prints and books printed before 1800. No exclusion was made on the basis of what country the books had been printed in.
To establish the rack, shelves and eventually the books to be tested, random numbers were used.
The Stanford survey judged the condition of paper, binding and cover.
Each book was to be judged on these three parts which each had three grading alternatives, 1 = good condition, 2 = worn condition and 3 = bad condition. These gradings were later to be weighted against each other and the results would give a picture of the state of each book. The condition of the paper was considered to weigh more than that of the binding and cover, which is quite natural as the usability of a book is highly dependent on it.
An important part of the survey was the folding tests made in the corner of a page. This folding was done six times. If the fold broke when lightly pulled, the paper was graded 3, regardless of the results of the other parameters. The parameters and gradings are listed below.
1 Palm and Cullhed, Papperskvalitet.
The results of the paper gradings from this testing procedure were later compared with results from chemical analyses done in the UUB survey of the paper in the tested books. The purpose was to see if the results from the Stanford method survey coincided with current criteria for paper qualities.
Thus paper samples were taken from the chosen books and analyzed according to fiber content, type of sizing, lignin content, type of fillers and pH. (We later excluded the results of the identification of the presence of fillers, as we were uncertain about their importance for paper durability.) It is possible to judge the quality of a paper quite objectively with the help of these parameters.
Measuring of the pH in a type of paper has often been over emphasized in evaluating the quality of single paper sheets. This measurement must be put into a context with other characteristics of the object. A paper with pH 5 can be more brittle than one with pH 4. But the average pH of a number of papers shows that the more brittle papers have lower pH than the ones in better mechanical condition.
Thus pH measuring is a blunt instrument when it comes to determine paper condition. One example of how far too general conclusions have been drawn from a survey based on pH measurement alone, is Ivar Hoel's survey in some libraries in the Nordic countries. 3 The pH as a tool for determining paper condition has been greatly exaggerated. Ivar Hoel's report will be discussed further on in this report.
The extended UUB survey with the chemical analyses showed that the Stanford method's paper grading, we call it P-grading - P 1, P 2 and P 3, coincided well with the different qualities paper has based on its chemical composition.
We also found that the Stanford method was a good tool for doing surveys of the state of libraries' and archives' collections.
Against this background the authors were asked to perform a Stanford survey at the Royal Library (KB), the National Archives (RA) and the Gothenburg University Library (GUB). The survey was to be a part of the national research and development project on paper conservation (FoU projektet for papperskonservering) funded by the Swedish government.
At KB and RA the Stanford surveys were to be made with random sampling and performed in the same way as the survey done at Stanford and UUB.
The survey at GUB was to be a direct comparison with the Uppsala survey, as GUB was to examine the same editions of the books tested at UUB. The purpose was to detect any differences between UUB and GUB which might tell us something about different preservation environments.
By putting together the results from the different surveys at UUB, KB, RA and GUB it would be possible to establish a basis on which further decisions on mass conservation could be founded.
Personnel from each institution carried out the gradings, and the results were elaborated at UUB.
The results are divided in two parts, one covering the results from the surveys at KB and RA and the other the results of the comparison between UUB and GUB.
4Percentages are published without decimals throughout this report.
The paper grading - P-value - was used to describe the paper condition. Figs.
3 and 4 show the P-values at KB and UUB.
The P 3-categories at KB and UUB are similar. The main difference is found in the P 2-category; KB has 22% and UUB 14%. This may be due to
-more tears and worn edges through extensive use or
•a higher degree of air pollution which gives the paper a yellow tone around the edges.
Fig. 3. KB P-values.
If the P-values are spread over time and put into relation to the total number of books per decade tested, the following graphs (Figs. 5 & 6) will show the same tendencies in KB and UUB.
The KB-graph has an unexpected dip in the total number of books tested from the 1950's to the 1960's. As mentioned before, this is due to the fact that foreign monographs from this period are stored in a depot outside Stockholm and thus not included in the survey.
A similar dip appears in the UUB-graph during the 1980's, which probably is due to an introduction of numerus currens-signing of books in
1979. These books are stored together in one part of the building, and since the random sampling was based on shelves all over the building, this part of the collection is not properly represented.
As a certain amount of time is required for a paper to deteriorate, we do not consider these irregularities to be of any importance. Almost every sample from these periods is graded P 1, irrespective of its chemical properties. This is obvious when the P 1-curve is compared with the totalcurve.
F owe 1850 1900 1950 The relationship between the total number of books tested and the ones with P 3 shows that there is a great increase of P 3 during the 1860's to the 1890's. After that the P 3-curve falls while the total number of books increases steadily. There is a noticeable bulge in P 3 in the KB survey, during the 1950's. None of the other surveys we made showed similar results, neither could we see anything like this in the Stanford and Yale surveys. We think that KB's bulge can be the result of individual differences among the testing personnel which may show clearly in a uniform sampling collection from the same period for example.
The National Archives (Riksarkivet - RA)
A modified version of the Stanford method was used at the National Archives as the character of archival collections differs from library collections. The sampling of shelves from which to pick objects for testing was the same as in UUB and KB. The archive collection consists of bound volumes and boxes from which samples were taken in the same way. The modification was in the sampling of the individual sheet of paper to be tested. The content was weighed and with the help of a random number a point between zero and maximum weight was chosen. With the use of a scale it was possible to obtain to the sampled sheet.