# Risk analysis

Physical damage to films, while of concern, tends not to get worse in storage. The long term factors that effect the usability of a film relate to biological factors and the base and image stability (if we include any magnetic film or magnetic striped film as image).

Long term storage is expensive. Maintaining suitable environmental conditions is at best a compromise between the desired life expectancy of the object and running costs. Included in running costs are both energy and net-letable-area (floorspace). Energy costs particularly are readily identifiable expenses with constant pressure to be reduced. However, maintaining a suitable storage environment is critical to prevent biological attack and chemical deterioration. Benchmarking the state of a collection is an important tool in being able to assess the success of any set of storage conditions or for providing evidence for the necessity in improved storage.

Surveying the collection to set the benchmarks is a major task but will provide a sound footing on which to measure the performance of decisions regarding the collection. While it is theoretically possible to measure or benchmark several factors in a single survey it is generally safer to concentrate on one or two factors.

Even if there have been no benchmarks established a survey can be used to determine the size of any major projects on the collection, e.g. planning resources for the duplication of severely deteriorating films.

## Sampling

To accurately determine the state of a collection, every object should be examined and reported. However, for most collections, this would prove to be an impossible task requiring resources far beyond those available. A statistical approach can be taken that samples the collection and from the smaller sample provides a clear picture of the state of the collection. The degree of confidence in results of this approach can be very high.

To determine the sample size required for the degree of confidence required the following formula can be used:

n=(Nz2pq)/(E2(N-1)+z2pq) where:

n = the size of the survey sample
p = the estimated number of “decomposing” films
E = error
z = standard deviations for the confidence level (for 95% z = 1.96)
q = p-1
N = the total number of films in population

Confidence level is the probability that the sample results are correct. 95 per cent is the usual degree of confidence in such tests. Confidence interval, also known as acceptable error, is the range that can be expected in the results.

For example, if a survey was carried out with a 95 per cent confidence level and a 3 per cent confidence interval and it was found there was a 10 per cent incidence of vinegar syndrome, this would mean that between 8.5 per cent and 11.5 per cent of the collection would be affected.

## Base stability

One of the greatest concerns for film collection managers around the world is the decomposition of the film base. Rapid decomposition is known to occur in all cellulose esters. The decomposition reaction by-products provide a ready clue to how far the reaction has proceeded. By measuring the acid levels around the film a fairly good picture of the degree of base deterioration that has occurred can be gained.

For cellulose nitrate films the test used to involve artificial aging and measuring the time taken to cause a colour change in an indicator test strip (Alizarin Red). This test is rarely used today as there are concerns over the accuracy of any artificial aging tests and the test was destructive requiring a small sample to be removed from the film at the head, middle and tail of the film. Currently there is no generally accepted non-destructive test for cellulose nitrate stability.

Cellulose triacetate, the overwhelming bulk of film collections, can be readily tested by several products. The Image Permanence Institute in Rochester NY, USA and Dancan in Denmark both produce test strips that reflect the acid content of the film under test. The test strips are strips of paper that have been saturated with a buffered indicator solution that changes colour around pH 4.5. Research has indicated that this pH corresponds to a free acid level of 0.5 a level considered to be the 'onset’ of deterioration.

## Image stability

The human vision system is good at comparative measurements in density and colour but is very poor at determining differences without a reference. Any benchmarking of image stability would require some standardised method of measuring changes in density and colour balance.

## Moulds and other biological factors

Moulds are readily controlled by controlling the storage environment, especially in respect to relative humidity. Below 60 per cent and moulds will cease to be viable. However, it is not always possible to maintain these conditions and the occasion may arise when the possibility of mould infestation in a collection needs to be surveyed. Moulds and the extent of the damage caused by their feeding and growth are readily seen by visual examination.