July 1998 Issue 28
Archaeopteryx - a wing and a prayer
Risk analysis or risk assessment can be described as 'the study of decisions subject to uncertain consequences' 1 and is applicable to any system, product or equipment where risks lurk, including the care of museum collections. In this context, risk assessment has been developed particularly in natural history collections, probably because these collections tend to be very large - nature is a large and complex entity - and in comparison to collections of cultural objects they are vastly under-resourced.
While the V&A may boast a conservator to object ratio in the region of 1:70,000, the Natural History Museum could still claim a ratio of only about 1:700,000 if it deemed all its curatorial staff to be conservators. And natural history objects do not deteriorate less than paintings, sculpture or furniture. Indeed, given their organic nature and the routinely invasive study methods employed, natural history collections are arguably at a constantly high risk of destruction.
Decisions about risks are greatly determined by perceptions of what there is to lose. In his 'Wager' 2 , the 17th century mathematician and theologian, Pascal, considered what decision we should make on belief in the existence of God. His assessment of the risks involved ran, more or less, thus: God is such a big thing that he is beyond reason and therefore, when it comes to belief or disbelief in God, existence and non-existence are like the faces of a coin when tossed - equally likely to turn up. Pascal reasoned that it is better to bet on belief than disbelief since, if you get it right, you are on your way to heavenly happiness and if you get it wrong, you have lost nothing. However, if you opted for disbelief and you are correct you gain nothing but 'noxious pleasures, glory and good living' 3 . However, if you are wrong you have risked - and will meet - infinite damnation.
In the theological world of risk assessment there was no scope for hedging your bets, but how do we fare in the real world? Take Archaeopteryx lithographica (Figure 1) for example. This, the earliest bird known - the 'ancient wing from the lithographic stone', lived about 150 million years ago and is now known from only seven fossil examples. The best specimens display almost undisturbed and perfectly preserved skeletal remains and delicate feather impressions in stone long prized for printing, but still mud when this animal was alive. Since only birds have feathers, Archaeopteryx is classified as a bird, but it has some skeletal features which are indistinguishable from those of reptiles. Archaeopteryx has an importance beyond most relics of past life:
'Possibly no other zoological specimens, fossil or Recent, are considered so important as are those of Archaeopteryx lithographica... Certainly few other specimens have generated such widespread interest or provoked as much speculation and controversy... Archaeopteryx may well be the most impressive fossil evidence of the fact of organic evolution.' 4
The London Natural History Museum's specimen of Archaeopteryx was found in 1861and purchased a year later as part of a larger collection after much negotiation by Trustees. Archaeopteryx was insured during transit for approximately £200. At the time Ruskin lamented the trustees' inability to recognise a bargain when they saw one:
'This collection of which the mere market worth, among private buyers, would probably have been some thousand or twelve hundred pounds, was offered to the English nation for seven hundred.' 5
Controversy has been a feature of Archaeopteryx since its acquisition. From 'Which way up is it?' to 'Did it take to the air as a ground-running flapper or a tree-climbing jumper, and was it capable of flight at all?' to 'How does it relate to other animal groups, and how far removed is this bird, with reptilian features, from being a reptile?'. In our understanding of the history of life these are major issues but, when scientific conflict is not interesting enough, there is always some mileage to be gained from alleging that Archaeopteryx is a hoax because it flies in the face of biblical creation 7 .
Until recently, the Natural History Museum's understandably reluctant estimate of a financial value of the London Archaeopteryx was £2 million - understandable because there is nothing with which to compare it in the fossil world other than the other six examples, none of which have been market-tested. However the astonishing, and some would say inflated, price of $8.3 million paid for a dinosaur last year suggests that Archaeopteryx might be revalued upwards towards the £10 million mark 8 , the top of a wide range of estimates made ten years ago 9 .
Archaeopteryx, therefore has an accumulated value in scientific, cultural and financial terms. So what about the risks to this national and international treasure?
A semi-quantitative methodology for assessments of risks to collections has been in use at the Canadian Museum of Nature for some years, based on the recognition of ten agents of deterioration considered to be comprehensive, and their characterisation as rare and catastrophic, sporadic and severe, and constant and mild (Table 1) 10 , 11 . A numerical magnitude of risk for defined collections is calculated as a product of the loss in value, fraction susceptible, extent and probability. The value of this assessment lies in the comparison of the numbers, allowing decisions to be made on priorities and resource commitment.
In this case, although Archaeopteryx is a rather small and specific sample when viewed as a 'collection', the calculation of risk magnitudes provides some insight into the significance of the various risks. The values of risk magnitude shown in Table 1 were calculated as an illustration of the approach and cannot be considered as absolute values. Their purpose is to offer some means of comparison.
While the results in Table 1 seem to give a clear message, and this may be correct, there are problems with this method which need to be acknowledged. Firstly, there is a lack of data, both on probability and frequency of occurrences. There is therefore a dependency on estimates, expert or otherwise, based on personal experience, knowledge and - ultimately - judgement. In four cases confidence, and knowledge of deterioration processes, was considered to be so weak that numerical values were not assigned, although the risks were considered to be relevant. And, for ten of the risks, deterioration was considered not to result.
The results outlined in the paragraph above are full of value judgements, as is the methodology used to produce them. This is not surprising. Personal knowledge and perceptions draw attention to one area while avoiding another. Equally, the energy required to make a judgement is as relevant a factor as the judgement itself. The more difficult the information is to acquire, such as the probability of an aeroplane crash on the Palaeontology Building of the Natural History Museum, the more likely it is that a lower level of rigour will be acceptable in judgements in that area. In the professional world of safety analyses, the quality of the results obtained are recognised to be dependent on the analysts, methods and models employed and comparative studies have shown the extent to which the same situation is open to a variety of results. 12
As a measure of confidence in, and consensus on the results in Table 1, ten colleagues were asked to rank in order the risks in Table 1 on the basis of their knowledge and experience, and to exclude those which they considered to be irrelevant. These colleagues all have direct knowledge of Archaeopteryx and, at some time, have direct - though varying - responsibilities for its storage and use. Figure 2 illustrates the lack of consensus within the group and a lack of agreement with the ranked results of Table 1. (The four risks for which values were not calculated in Table 1 were secondarily ranked tenth to thirteenth).
While nine of the group agreed that Physical Forces Type 2 was the most significant risk (the other ranked it in second place), only five agreed on the second most significant, five on the third and only two on the fourth. The level of consensus was equally low for the risks considered to be irrelevant. Only seven of the twenty three risks were not considered irrelevant by someone, and the greatest number who agreed on what was irrelevant was seven for Custodial Neglect - Type 3.
There are obvious reasons for this poor level of agreement. The ten colleagues who participated in this exercise have a variety of specialist knowledge but their specialisms do not overlap much. So, most people identified the risks of sporadic and severe physical forces because these are characteristic of the use of Archaeopteryx through its removal, transport and access for study, and this is the risky area in which most people are involved. More specialist knowledge is required to understand the risks from incorrect relative humidity or contaminants and, since fewer people have this knowledge, more people are likely to score this as a lower risk.
The high level of consensus in selecting Physical Forces - Type 2 as the most significant risk for Archaeopteryx might be a measure of the likely accuracy of this decision, but the poor consensus both within and between the group's rankings and those obtained by the semi-quantitative methodology, indicate the attention which needs to be given to choosing a source and measure of comparative analysis. Comparisons may introduce more uncertainty than they resolve if they are not of equal or greater accuracy.
The absence of reliable data is not an obstruction to risk assessment although it is an obstruction to more objective assessment. In these situations we should opt for a rational approach and determine the level of consensus which is achievable - or even desirable - and, where there is agreement on the relevant risks, assume the worst until such time as a better consensus can be achieved. Similarly, where there is a lack of consensus on the magnitude or rankings of risks, consensus may be more readily achieved not by drawing on varied areas of expertise, but by focussing on the more personally valued issue of consequence 13 .
It may be difficult to appreciate the height from which Archaeopteryx needs to be dropped in order to damage 50 or 60 or 70% of the specimen, and from there to determine what fraction of its value has been lost, and to measure the significance of the risk of dropping, but the consequences of such an event occurring are likely to be more readily appreciated. In the case of Archaeopteryx, the consequences of doing otherwise will bring a damnation beyond that wagered by Pascal.
Thanks to the staff of the Palaeontology Conservation Unit, Dr. Angela Milner, Head of Vertebrates and Anthropology Division, and her curatorial colleagues.
6. Waller, R., Conservation risk assessment: a stragegy for managing resources for preventive conservation, in Preventive Conservation Practice, Theory and Research. Preprints of the contributions to the Ottawa Congress, IIC. 1994.
10. Waller, R., Conservation risk assessment: a strategy for managing resources for preventive conservation, In Preventive Conservation Practice, Theory and Research. Preprints of the contributions to the Ottawa Congress, IIC, 1994, pp. 12 - 16.