zurück zu Kolumnen
Peter Herrmann. October 2008
vorige Seite
Age Classification

The latest stand in the analysis of West African bronzes from the point of view of free art brokers and the upmarket art trade.

In various articles published in the last two years, Dorina Hecht and Peter Herrmann promulgated aspects of a new interpretation of West African bronzes with a focus on Nigeria. Reactions varied widely. On the one hand, a discerning group of fellow dealers and collectors welcomed our work, happy to have been provided with new material for argumentation. On the other hand, a small group of doubters regarded our pieces and accompanying remarks with deep mistrust and underhandedly spread all sorts of rumours. As usual, there was also an undecided group caught somewhere in the middle; in the absence of detailed knowledge, it waits with interest to see where the discourse leads.

Because we recently put together a few small experimental series with bronzes from Cameroon, came across new information regarding Nigerian bronzes and were able to evaluate objects from the Paul Garn collection in Dresden, which was acquired in 1920, I decided to resumptively tackle the entire topic of age classification anew.

My remarks are based on the assumption that thermoluminescene analyses are widely recognized and accepted. At present, I know of no one who would seriously challenge these analyses, to which I will henceforth refer as TL analyses for brevity's sake. And even if there were someone who doubted their accuracy, it could surely be attributed to either misunderstanding or the influence of the systematic denunciations of certain market actors.

The world's most distinguished museums and collections use the TL technique, which also constitutes the basis of much international research. For us, TL is the foundation of all arguments. The present problem thus begins at the point where TL findings contradict findings derived from metallurgical analyses.

At the behest of one of our sellers, three objects from Benin with clear TL-determined ages were tested on the basis of their metallic composition and patina by a German laboratory. The new age determinations for all three objects contradicted the TL determinations. It is not hard to imagine that technicians in both laboratories experienced a certain degree of nervousness related to professional acceptance of their findings, which was then passed on to the respective customers, who, for their part, were polarised.

This topic has always been a focus of mine, and naturally, I was determined to arrive at a conclusive result – not least in consideration of my reputation. Again I had to ask myself if items could be forged by means of modern methods.

In the 1990s, Dr. Hermann Forkl at the Linden Museum in Stuttgart provoked several law suits when he argued that there were radiation methods with which one could attach any desired age to any given object. Unfortunately, we have never been able to locate the ominous device with which one can work so precisely that a serious possibility of making copies exists. Scientific research revealed that while radiation methods could theoretically be used to determine an old age, in practice, there was no easy way of circumventing all other testing methods. The quartz and feldspar curves analysed by TL are too complex to be forged, and tracks of an internal alpha radiation can always be detected. What's more, any radiation technique for forging old age would be limited by the stylistic details of the given object. Not surprisingly I know of no one who still believes in the existence of this counterfeiting method.

And thus the pendulum swung in the direction of the lack of credibility of metallurgical methods. To us, the existing expertise seemed exceedingly unscientific.


At first, we weren't planning on challenging the comparison of the analyses as such. The way that the alloy was identified seemed correct; it was the conclusion derived from this identification that seemed misguided. Because old English sequence analyses specify zinc proportions, the analyst stuck unquestioningly with the antiquated proportions and concluded that the ages we had determined for the objects "could not be". A too-high proportion of zinc, the analyst argued, meant that an object dated at 400 years by TL analysis "could" actually date only to the 18th century, at the earliest. A low concentration of antimony and arsenic "suggested" an even more recent date of creation. After that, the material analysis became more and more incorrect. The analyst failed, for example, to differentiate between aluminium oxide and metallic aluminium. The analyst's "it could not be" reasoning was based on the fact that one of the objects contained 0.5% aluminium, which has been manufactured in metallic form only since the end of the 19th century. Based on certain related presumptions, the analyst reached an "absolute" conclusion and definitively classified the object as "new."

In doing so, he ignored the fact that traditional material analysis techniques do not differentiate between aluminium oxide and metallic aluminium. The alloy contained aluminium oxide because the bauxite used in the smelting of West African ores – and, by extension, the corresponding casting moulds – contained aluminium oxide. When we examined recently hand-crafted objects from Nigeria, however, we found no traces of aluminium; that is because these objects, unlike the older ones, were made using imported metals, which are far more economical in this day and age.

An additional explanation for the presence of aluminium holds that certain local West African casting methods involved temperatures high enough to generate metallic aluminium in the alloy. I have not yet verified this theory, but an explanatory report offered by a chemist makes it seem so plausible that I wanted to mention it here as a side note. (*)

The fact that the presence of aluminium has so long been accepted as an indication that an object is "new" raises a very important question. Why should an ominous forger go to the effort of interpreting alloys, counterfeiting a patina, somehow introducing a fake casting core onto an object and throwing a soda can into the alloy to assure that the analysis reveals traces of aluminium? From a technical point of view, that makes no sense.

The sequence analyses from old sources are postulates, not hard facts. Indeed, in an article in the catalogue Benin – Könige und Rituale (Benin – Kings and Rituals), Dr. Peter Junge points to several existing sequences that contradict age classifications based on the old sources.

In addition to undergoing metallurgical examination, patinas were also judged by the aforementioned metallic analyst according to their elements. Here, too, the results were based on a succession of assumptions that culminated in the assertion that a forger had used potassium permanganate and heating methods to achieve a black-brown colour meant to simulate an older age. One of the assumptions that formed the basis of this conclusion was that the object had been stored underground; from that, the analyst deduced that all accumulations of chlorine, sulphur and tin were misleading. Most West African bronzes, however – including the three examined – were not stored underground. In fact, most bronzes from Benin and Cameroon that we know of were found above ground, exposed to fresh air for centuries. The analyst also seemed to lack knowledge about how the objects were made. Throughout West Africa, the manufacturing of bronzes instantaneously produced a fine, brown, translucent, synthetic patination of organic substances – due mostly to the fact that a fresh gold-yellow or red-yellow colour was not usually desired.

In fact, this method of patination is still used today in the manufacturing of hand-crafted artefacts in Cameroon. And not, we must emphasise, for counterfeiting purposes, but rather because it produces objects that are preferred by customers and thus easier to sell. No one would ever dream of claiming that a piece of European antique-styled furniture were a counterfeit just because it had been patinated in a way to promote sales. In Europe, as in Africa, an experienced and serious dealer can recognise the difference between old and new at a glance.

What none of us knows is what types of cleaning agents were used in West African houses in the 20th century. Judging by our own recent past, we can hardly assume that gentle organic products were used. Since the 19th century, cleaning agents have contained whatever chemical manufactures want them to contain. It would certainly be worth examining the residues that these agents leave behind and the chemical reactions they thereby provoke. At various times in West African history, moreover, furbished metal objects were favoured over unfurbished ones. A measurable patination thus often only set in years after the creation of an object.

By now, then, it should be clear that patina analysis can only shed a very limited light. It presupposes that the object was never handled, that it was cleaned gently and that its storage conditions were consistent, though almost nothing is known about these factors beyond the aesthetic tastes of the time and the resulting surface handling.

For the time being then, analytical examination of metal probes and patinas should be cast aside. That is not to say that there is no reason to pursue that type of analysis in the future, but rather that in its current form, it is not yet sufficient for making exact determinations. So far, the TL analysis is more exact because it analyses absolutely.

Most experts agreed to these results. But there is another argument levied against the usefulness of TL analysis with which we have been very much concerned. The TL technique evaluates fragments of clay core that usually still remain inside the castings. The fact that these are not a part of the alloy itself gives rise to the suspicion that a forger could insert older fragments into the castings of newer objects of his own accord. This is an obvious suspicion and, in fact, a well-founded one: such insertion can be and is performed, but cannot go undiscovered. One absolutely cannot talk about an alleged forgery glut. For our part, we've attempted the insertion one time only. Pulverised core fragments dissolve in water; they adhere the way sand adheres and can be easily flaked off of surfaces. Thus core fragment powder can only be attached to a casting with the help of a bonding agent. A forger, in other words, would need additional materials in order to affix the powder to the core as firmly as it would have been through natural casting – that is, firmly enough to require a great deal of strength to scratch the fragments off the casting. TL analysis, however, detects this additional bonding agent and senses the no longer homogenous connection between core fragments and casting. The experienced probe examiner notices this immediately and conducts a follow-up organic or inorganic test for adhesive agents to clear up the discrepancy.

A forger would also face very practical problems. He would have to own a giant shelf on which to store a selection of TL-tested fragments. He would have to pulverise these shards and use them to create a figure fine-tuned to the stylistic particularities of a given age. Because the literature only provides unspecific dates, the forger would have to be an incredibly well-equipped person with extensive skills. Ignoring, for the moment, the fact that the bonding agent problem is still not solved, it's still hard to imagine that counterfeits have been circulating the market for decades and that a forger has nonetheless never been caught. Moreover, for an experienced sample-taker, an artificial terracotta sample is perfectly distinguishable from a casting core sample.

Bronzes, however, cannot be compared to terracotta objects. When, for example, one puts together excavated shards from the Nok culture like a puzzle in order to form a figure, then these fragments – with definitive, compatible ages – are congruent with stylistic classifications. Burned shards, however, cannot be inserted into bronzes. Pulverised material reacts diffusely.

Another theory, similar to the aforementioned radiation theory, is based on the high-pressure attachment of pulverised core fragments to castings without the use of a bonding agent. How this is supposed to be accomplished eludes us. The friction resulting from high pressure creates a heat that starts at 220 degrees; depending on how hot that heat actually gets, the measurable material would be partially or completely destroyed. In cases of such interference, TL analysis will determine a younger, even recent age. As yet, investigation into whether or not there is an alternative variation of this method has yielded no results.

In many discussions, doubters of our object analyses expressed a concern that commercial interests lie at the bottom of our laboratory's TL expertise. Inconsistencies in our analyses were compared with analyses performed at Oxford University, whose alleged independence was offered as proof of its superiority. Ignoring the fact that so-called "independent research" barely exists any more, the populist argument that if it did exist, it would be somehow superior surprised us. Let us now take a closer look at this supposed independence.

We delivered a commemorative head in the Ife style from the Paul Garn collection in Dresden to the Begbroke Science Park analysis laboratory at Oxford University. A previously conducted TL analysis of the head, which was acquired between 1920 and 1930, pinpointed the object's age at plus or minus 550 years. The emphasis on provenance and date of purchase is important because in 1920 in France, no one could have thought of manually tampering with a casting core. At that time, after all, TL analysis did not yet exist; no one, then, would have thought to recover the casting core remains necessary for TL of analysis. Nor did the notion of forgery in its current form exist. Our seller's other metallic analyses, however, were performed on objects that were imported to Germany in the last decades, so that theoretically, there was a remote possibility that the objects were forged even though we were able to refute all suspicious allegations in great detail. Still, we wanted to take every possibility into consideration in order to assure a bullet-proof examination.

Before going into detail about the Oxford analysis, I'll present the results upfront. Unfortunately, this analysis, too, was an extremely dubious affair.

At first, I assumed that the metallic analysis itself would be accurate, but that the conclusions drawn from it would not be. I expected the head to contain aluminium and to exhibit a too-high proportion of zinc – these, after all, are the primary causes of the entire dispute over metallic analysis and TL.

The introduction to the Oxford analysis asserts that we delivered the object for comparison to two Ife-style heads that were excavated in 1938 and described by Willett. That was the first mistake – we did not, after all, deliver the object for a comparison. We wanted a metallic analysis and nothing else. Nonetheless, the comparison with these two heads is very interesting. The heads, both originally used as examples by Dr. Peter Northover, come from contemporaneous digs in Ife. Our head was probably never buried, but rather exposed to air for centuries. The two Northover heads had a zinc proportion of 10%; our head, a proportion of 33%. The analysis presented this fact as one of two main reasons for its characterisation of the TL analysis as "wrong" and the categorisation of our head as "new." But if one compares the 33% proportion of zinc with all existing examinations of bronzes from Benin and Ife, one will see that though it is certainly high, it is still very clearly within the realm of possibility. A 33% portion is present in numerous style variants across several centuries. Zinc is smelt in Nigeria.

The second argument for the "new" determination was, as expected, a 0.3% presence of aluminium in the alloy. What's important here is, first of all, that the two Ife heads used for comparison were not examined for traces of aluminium; in other words, there was no true basis for a comparison. Shockingly, the analysis furthermore failed to differentiate between oxide and metal. The simplified Oxford analysis simply cannot distinguish between these two states. It is thus clear that not only were the analysis results falsely interpreted, but the analysis itself was flawed.

A surface dezincification due to corrosion – a normal process that results from the varying tension between copper and zinc when exposed to air – is also inaccurately interpreted in the Oxford analysis. The object surface would be burned with a torch for the creation of a fake patina, which allowed the zinc portion to vaporise. An iron-nickel impurity discovered by the analyst in an inaccurately executed enlargement allowed that analyst to again arrive at the conclusion that this simply could not be in an Ife head.

But a closer look at the analysis reveals that those were not the only errors made by the Oxford experts. The analyst speculated wildly on quantities of metal in the alloy and traces of corrosion on the surface without any relevant basis. He used irrelevant methods and applied a totally antiquated analysis. In a separate article, we examine all of these mistakes much more closely, and on the basis of that examination, categorically determine that these Oxford analyses cannot be used.

On a particularly bizarre note, the expert determined the age of the head to be 50 to 60 years. There is no clear basis for this conclusion, which is as accurate as a result determined by dice thrown across a bar counter. If the object were truly 50 to 60 years old, after all, the head would have been bought before it even existed!

Now, it begins to become clear why certain "experts" use the term "forgery glut," which we could never before understand.

Can it be true that the biggest auction houses – which most often do not provide analysis results in their catalogues, but notoriously work together with Oxford – are in possession of a large waste pile of analysed objects? For usually, both a TL analysis – also offered at Oxford (!) – and a metallic analysis are conducted on auction house objects, and more often than not, the results are inconsistent. Are these objects then separated from the stock and returned to the seller with the notice, "Counterfeit?" Can it be true that accordingly, only a small portion of the objects – those whose analyses are consistent with the lead and zinc sequences, usually because they were made using aluminium-free alloys imported from overseas starting in the 15th century – would have the chance to land in an auction?

But let us set conjecture aside and return to the supposedly forceful logic on which the determination of 50 to 60 years is based. Fifty to 60 years ago, TL analysis – and, by extension, would-be forgers who could bring casting cores into play – did not yet exist. TL analysis has only been around since the 1970s, used at first as a supplement to metallic analyses performed at universities. The true problem arose about ten years ago when commercial metallic analysts hit the market, and the various expertises increasingly contradicted one another. The search for consensus was no longer a primary concern; instead, cutthroat competition reigned.

The allegations made by these laboratory operators on the basis of inaccurate analyses had a fatal effect on the market and thus also on the author of this article. Followed to their logical conclusions, these allegations insinuate that even before 1989, the year of my first analysis orders, I had manually inserted casting core remains into objects from Nigeria and Cameroon. For that to be true, I would have to have acted with unbelievable foresight, furnishing objects that at the beginning of the 1990s were not yet stylistically classified with prophetic ages that contradicted all then-current research.

The Oxford conclusions and, by extension, the German metallic analyses presently available to us are simply irrelevant. Nigerian objects imported via Denmark in the 1950s, bronzes imported from Cameroon in the 1940s, bronzes imported from Mali in the 1920s, objects from inheritances, from private collectors, from European and African dealers – according to the metallurgic theory, all of them landed in an ominous laboratory where, via a method that no one can specify, they were manually endowed with casting core fragments. The entire process, moreover, was so flawlessly executed and well concealed that after 30 years of commercial TL analysis, still no one can identify a forger or explain his method. So well-concealed that there is no evidence of what method was used to achieve the fragment insertion, the only theory being that it was rubbed on as a bonding-agent-infused powder. So well-concealed that hundreds – thousands – of objects that belong to collectors and dealers across the world are part of a counterfeiting cartel whose structure has never been determined.

The age classification of "new" or "50 to 60 years old" is as absurd as the contention that we, by implication, were involved in an alleged counterfeiting ring that succeeded in using dealers, TL analysts, collectors and scientists for its own ends so successfully that the parties involved never noticed that they were being used. Where is the logic behind this supposed global network of people who don't know one another, have never seen one another, belong to no common organisation and yet still have in common the fact that they are involved in a method of forgery whose existence no one can prove?

In the 1990s, several court decisions confirmed the occurrence of aluminium in bronzes based on the testimony of prominent expert witnesses with doctorates and professorships. They confirmed the existence of zinc proportions of varying sizes. Without reservation and based on analyses of objects from Benin and Ife, they dismissed as absurd the obsession with lead components and their reactions that Oxford still takes as a basis for analysis today.

Courts, expert witnesses and scientists. Are they, too, all united in the enormous counterfeiting cartel?


One must reject all such monstrous notions so that a mutually acceptable common ground can be achieved for the classification of objects. The laboratories specialising in metallic analyses must change their examination criteria. It is clear that analysis failures exist and even more clear that deductive conclusions are inaccurate. Legally, scientifically and art historically criticised errors must not be kept alive by these laboratories. It is clear, after all, that these errors are the source of the current counterproductive argument amongst collectors, dealers, ethnologists and art historians.

As described in several previous articles, the resistance of these analysis laboratories furthermore reveals a clear discrimination against the craftsmen, artists and art historians of the objects' countries of origin and thereby extends to the political and diplomatic sectors.

Already once before, when bronzes from Nigeria arrived in Europe in 1900, were important insights into African artistic work rejected on the basis that the savages over there were not capable of such achievement. Some ethnologists subtly choose to take the devious route of attributing this art to a divine kingdom in order to construct an ominous explanatory model featuring less exceptions and assured cultural demise, which was then flushed out by the colonialist-absolutist spirit of the 20th century and today lives on in the mindset of structurally conservative western collectors and scientists.

Why do we struggle so much with Africa?

If we want to establish a modern, equitable relationship to various African countries, we must also learn to accept their positions. Bronze casting in Cameroon is at least 300 years old, which means that there are therefore also objects that are that old. There are many more bronzes in Nigeria than described in the literature, and they are considerably older than many western scientists currently accept.

I have outlined the background of these issues concerning bronzes in various articles, most of which coincide with the positions of many Nigerian art historians. Their research work, in turn, is fundamentally dependent on cooperation with Germany because we have the world's largest stock of Benin bronzes. For now.

Unfortunately, cooperation is hindered when German dealers are labelled as unserious by their own countrymen and when research stagnates. Just as the Garn collection is now being offered in the United States, more and more of the objects that form the basis of cooperation are disappearing from Germany because their value is underestimated here. For at least two decades, large German collections and important individual pieces have been migrating to neighbouring Belgium and France as well as to the United States.

Let's hope that the newly created professorship at the Free University of Berlin will blow a fresh new wind and assure that the issues discussed in this article become the focus of further research. Conducted together with African scientists.


© Peter Herrmann. October 2008
Translation: Jenna Krumminga



* Supplement, October 2012

It is now accepted that in West Africa, pure - that is, metallic - aluminum is present in the earth. Regardless of what elementary state the existing aluminum is found in bronzes, it is not an indication of age. (Back to aluminum.)

Wikipedia >> Aluminium >>

zurück zu Kolumnen Texts