The Crosby Garrett Helmet

A copper-alloy Roman cavalry parade helmet, its face-mask visor cast as the calm face of a young man beneath a griffin-crested Phrygian cap, found broken into more than sixty fragments in a Cumbrian field in May 2010. Because it contained no gold or silver, it fell outside the UK's Treasure Act, and the finder and landowner were free to sell it. Christie's restored it, in about 240 hours, and sold it on 7 October 2010 for a hammer price of GBP 2,000,000, more than 2.28 million with premium, to an anonymous telephone bidder, defeating a public campaign by Tullie House Museum in Carlisle that had raised roughly 1.7 million pounds.

It is one of only a handful of Roman face-mask cavalry helmets ever found in Britain, and its case became a byword for a real gap in UK heritage law: an object judged internationally important by the British Museum could still be extracted, restored for sale rather than for scholarship, and sold into private hands, all entirely lawfully. It has been shown four times since, on loan from its anonymous owner, and remains outside public collections.

What it is

The Crosby Garrett helmet is a two-piece Roman cavalry parade helmet of the type used in hippika gymnasia, ceremonial mounted exercises rather than battlefield equipment. It stands 407 millimetres high in its reconstructed state. The lower part is a face-mask visor, hinged at the browline so it could be raised, cast and chased in the likeness of a clean-shaven young man with elaborately curled hair and a calm, slightly parted-lip expression. The upper part is a domed headpiece shaped like a Phrygian cap, its peak curving forward, on top of which stands a separately cast griffin, one forefoot resting on an amphora, wings raised.

Nothing else found in Britain looks quite like it. The griffin crest in particular has no exact parallel among the surviving British cavalry helmets, though a related face mask that surfaced on the London market in 2019 has been proposed, by the conservator who examined both, as a product of the same workshop or even the same hand.

How it was made

X-ray fluorescence analysis carried out at Christie's identified two different alloys in the object. The helmet bowl and the face-mask visor are a copper alloy of roughly 82 percent copper, 10 percent zinc and 8 percent tin. The griffin crest is a different, leaded bronze of roughly 68 percent copper, 4 percent zinc, 18 percent tin and 10 percent lead, a mix better suited to fine casting than to sheet raising.

The bowl and mask were raised and chased from sheet metal over a former, then joined by a hinge so the mask could be lifted clear of the wearer's face. The visor was originally tin-plated, giving it a silvery sheen that would have contrasted with the warmer, coppery-yellow colour of the unplated bowl and crest, a deliberate visual effect built to catch sunlight and draw the eye of spectators at a distance, which is exactly the purpose ancient writers give this class of helmet.

The find

In May 2010 a metal detectorist, a young man in his twenties from Peterlee in County Durham, was searching pastureland at Crosby Garrett in Cumbria, on a farm owned by Eric Robinson, land he and his father had been detecting for several years without finding more than Roman coins and small fittings. Sally Worrell, then the Portable Antiquities Scheme's finds adviser for later prehistoric and Roman finds, later gave the fullest account of what happened next: the helmet came out of the ground broken into 33 larger fragments and a further 34 smaller ones, the visor lying more or less complete and face down, and it reached Christie's in June 2010 still packed with soil, the face-mask resting face-down on cotton wool in a plastic box.

Some secondary press accounts instead describe the find as made by two people, or by a father-and-son pair working together; Worrell's account, based on direct contact with the finder, describes one named-to-PAS individual who had detected the land with his father over preceding years. No finder has ever been named publicly, and this record notes the discrepancy rather than resolving it, since none of the accounts consulted settles it with primary documentation.

Why it was not Treasure

The UK's Treasure Act 1996 protects specific categories of find: hoards of coins, and objects of gold or silver, along with a narrow set of other categories added since. A single object made of copper alloy, however important, however clearly of the Roman period, however obviously a unique national find, is not Treasure under the Act if it is not gold, not silver, and not part of a qualifying hoard. The Crosby Garrett helmet is bronze. It was found alone, not among other objects. On both counts it failed to qualify, and the finder and landowner were free, entirely lawfully, to keep it, sell it, or dispose of it however they wished.

Archaeologist Paul Barford made the point bluntly at the time: a find that any archaeologist would call nationally important had fallen through what he called a loophole in the law, precisely because the law measures importance by the metal an object is made of rather than by what it is. The case was later cited in official reviews of the Treasure Act as an example of the gap between an object's evident significance and its legal protection, part of the record that eventually fed into proposals to widen the Act's scope beyond precious metal.

The disputed findspot

Because the helmet was not Treasure, there was no legal requirement to report it to the Portable Antiquities Scheme at all, and no legal requirement to reveal precisely where it was found. Finds Liaison Officers for Cumbria and Lancashire did not obtain a twelve-figure grid reference for the site until 30 August 2010, three months after the find and after the object had already gone to Christie's and been substantially restored. When archaeologist David Gill visited the question in a Looting Matters post that October, he noted that the site had not been independently verified: a hole had been dug and some ground disturbance was visible, but neither could be dated, and no other Roman material from the same field had been recorded on the local Historic Environment Record.

Gill returned to the question in a 2014 article for the Journal of Art Crime, titled, deliberately, Context Matters: The So-called Crosby Garrett Helmet, after further work, including a trial excavation by Minerva Heritage Ltd at the alleged burial spot, failed to remove the doubt: any cut in the ground that might have been made when the helmet was buried had itself been destroyed by the digging that recovered it, and a PAS finds officer for the region, Stuart Noon, questioned whether burial at the reported depth of around 50 centimetres was consistent with the pattern of damage on the object. That reported depth is itself inconsistent across sources: some secondary accounts of the finder's own description put the burial at around 25 centimetres, resting on two flat slabs of rock inside a pit, with a further large stone and more soil packed in above. This record could not reconcile the two depth figures against a single primary account, and lists the inconsistency itself as evidence of how thinly documented the recovery was.

Gill's conclusion was not that the Crosby Garrett findspot was fabricated, but that it had never been, and by 2014 could no longer be, independently confirmed. Commentary on the case since has been split between archaeologists who treat the reported location as probably genuine but poorly recorded, and those who regard the entire question as symptomatic of what a PAS database entry can and cannot actually prove.

What later investigation of the site found

Once the findspot had been identified to the Portable Antiquities Scheme, Tullie House Museum sponsored further investigation of the field, after the helmet itself had already left the ground. That work reportedly identified low earthworks forming part of an enclosure roughly 1,600 feet long, combining ditch-and-bank fortification methods associated with both native British and Roman building traditions, and consistent with a Romano-British farming settlement whose origins may reach back to the Bronze Age. Nothing of this had been visible or recorded before the helmet's discovery prompted archaeologists to look.

The same investigation is reported to have recovered further copper-alloy and iron objects, a bead, and two Roman coins dated to between 330 and 337 AD, found within the same artificial stone feature associated with the helmet's burial. Those coins date roughly a century or more after the helmet's own likely manufacture in the later 2nd or earlier 3rd century, and this record has not found a source that resolves what that gap means: whether the feature was reused or added to well after the helmet was first buried, whether the helmet itself was deposited later than its manufacture date would suggest, or whether the coins belong to a separate phase of activity at the same spot that happens to share the feature. This record states the puzzle rather than resolving it, because no source consulted resolves it either.

From field to saleroom: the restoration

Christie's commissioned an independent conservator, Darren Bradbury, to restore the helmet ahead of sale. The work, which took around 240 hours, involved repairing cracks and holes with cyanoacrylate adhesive and resin fills, retouched to match the surrounding metal's colour and texture, and rejoining the fragments into the substantially complete object now known. The British Museum, whose senior curator of Romano-British collections, Ralph Jackson, examined the piece during this period, asked that full scientific conservation and recording precede the cosmetic restoration. That request was not acted upon, and the restoration went ahead on Christie's timetable rather than a conservation one.

Sally Worrell, who with Ruth Fillery-Travis carried out X-ray fluorescence analysis on the object at Christie's, noted afterward what had likely been lost: formal conservation might have identified impressions of textile or other organic material preserved in the metal's corrosion products, evidence of how the helmet had been wrapped, packed or dressed before burial, and any such evidence is now gone. The restoration achieved what it was commissioned to achieve, an object fit to be sold at its best. Whether it achieved what a national collection would have wanted, a fully documented object with its burial evidence intact, is a different question, and the record on that point is now closed.

The sale at Christie's

Christie's announced the helmet's discovery publicly in mid-September 2010 and made it the centrepiece lot of its 7 October antiquities sale in London, with a pre-sale estimate of GBP 200,000 to 300,000. The estimate was passed within seconds of the bidding opening. Six bidders took part; by the time Tullie House Museum's consortium had to withdraw, at 1.7 million pounds, two bidders remained in the room and on the telephones.

The hammer fell at GBP 2,000,000. With Christie's buyer's premium the price came to approximately 2,281,250 pounds, and with VAT chargeable on the premium at the rate then in force, the total paid by the buyer came to approximately 2,330,468.75 pounds, a figure this record derives by reconciling several press reports rather than from a single confirmed Christie's statement, and flags accordingly. Press coverage at the time mostly quoted a rounded 2.3 million pounds, or its dollar equivalent of roughly 3.6 million, without separating hammer from premium; this record separates them because the difference between a 2 million pound hammer and a headline 2.3 million pound sale is exactly the kind of distinction that gets lost in retelling. The buyer was an anonymous UK resident, described by Christie's as an experienced collector of fine art and antiquities, bidding by telephone.

The fight to keep it in Cumbria

Tullie House Museum and Art Gallery in Carlisle led a public campaign to buy the helmet for Cumbria, with the support of the British Museum. The National Heritage Memorial Fund pledged GBP 1 million toward the purchase. The Art Fund added 200,000 pounds, the Headley Trust and the Monument Trust a further 300,000 pounds between them, and the J Paul Getty Jr Charitable Trust 75,000 pounds. An anonymous overseas benefactor offered 50,000 pounds on condition the public matched it, which the public did, and local fundraising, including a double-your-money pledge from a Carlisle businessman, brought in a further tranche of small donations. Tullie House had assembled roughly 1.7 million pounds by the time of the sale.

It was not enough. The hammer price alone, before premium, was 2 million pounds, and once premium and VAT are added the gap between what Cumbria had raised and what the anonymous buyer paid comes to something over 600,000 pounds. Mike Mitchelson, then leader of Carlisle City Council, called the response overwhelming and the outcome clearly disappointing. Roger Bland, then director of the Portable Antiquities Scheme at the British Museum, was blunter about what the result meant for the object's public future: to have it bought by a U.K. private buyer was the worst possible outcome.

What the finder and the landowner received

There was no statutory reward to divide, because the Treasure Act did not apply and no ex gratia treasure reward was ever in question. Press reporting at the time described the sale proceeds as split, with the landowner Eric Robinson receiving roughly half and the finder, or finders, receiving the remainder, consistent with the voluntary finder-landowner arrangements long recommended by the National Council for Metal Detecting's code of practice even where the Treasure Act does not bite. This record has not been able to independently confirm the exact split, the contractual terms, or the amounts each party actually received net of Christie's charges, from a primary source, and flags this as unverified in detail even though the broad outline is widely reported.

After the hammer: the loans

The anonymous buyer has, on several documented occasions, lent the helmet back for public display, each time through Christie's or the borrowing institution rather than through any direct public statement by the owner. It was shown for the first time since the sale at the Royal Academy of Arts in London, as part of the Bronze exhibition, from 15 September to 9 December 2012. It returned north to Tullie House in Carlisle from 1 November 2013 to 26 January 2014, then travelled to the British Museum from 28 January to 27 April 2014, where it was displayed alongside the Ribchester helmet, allowing the two most complete British face-mask cavalry helmets to be seen side by side for the first time. It was shown once more at Tullie House from 8 April 2017, as part of Hadrian's Cavalry, a programme spanning ten venues along Hadrian's Wall between April and September 2017 that brought together cavalry equipment from collections across Britain and the near Continent.

No public loan, sale or exhibition has been reported since 2017 in the sources consulted for this record. Whether the helmet remains with the same anonymous owner who bought it in 2010, or has changed hands privately since, is not established.

The face, the cap and the griffin

What the visor's face and the crest's griffin were meant to evoke is argued rather than documented. The helmet has no inscription and no accompanying text of its own to explain its imagery, so any reading rests on comparison with other Roman material and with what little survives of ancient writing about these helmets and the displays they were made for. Specialists have proposed several identities for the youthful, curly-haired face: a generic idealised soldier or hero, the Trojan prince Paris, the god Attis, or a deliberately ambiguous eastern Mediterranean type meant to signal the exotic, non-Roman opponents that cavalry sports mock combats were often staged to represent. None of these readings is settled, and this record treats each as a proposal rather than a fact, because no ancient source names the figure and no comparable inscribed example survives to confirm any one identification over the others.

The griffin is on firmer ground, if not by much. In Roman visual culture the griffin, a composite of lion and eagle, was a widely used emblem of vigilance, strength and guardianship, and appears elsewhere in military and funerary contexts without needing to be tied to one specific cult. Its presence on a cavalry parade helmet is best read as a general statement of the wearer's ferocity and status in the display, rather than evidence of a specific religious allegiance, though this too is an inference from the way the motif is used elsewhere, not a documented statement of intent by whoever commissioned this particular helmet.

Hippika gymnasia: what the helmet was for

Helmets of this class were made for hippika gymnasia, formal mounted exercises performed by auxiliary cavalry units in front of their commanders and, on occasion, the emperor. The fullest ancient description comes from the Greek writer and Roman cavalry commander Arrian, whose second-century military treatise describes riders competing in javelin-throwing and manoeuvre displays wearing gilded or polished iron and bronze helmets specifically chosen to catch the eye of onlookers, with the most skilled and highest-ranking riders given the most eye-catching equipment. The tin-plated visor and gilt-toned bowl of the Crosby Garrett helmet fit that description precisely: a face built to gleam white against a body built to gleam gold, seen at a gallop from a spectators' stand.

This is a genuinely documented function, not a speculative one; what remains uncertain is only the specific identity of this helmet's original wearer, unit, or garrison, none of which survives in any record.

The narrow company of British face-mask helmets

Only a small number of Roman cavalry helmets with face-mask visors have ever been found in Britain, and the Crosby Garrett helmet takes its place in a short, specific list. The Ribchester Helmet, found in Lancashire in 1796 and now in the British Museum, is the oldest known English find of the type and, like Crosby Garrett, was recovered damaged and required reconstruction; it is a different object, found 214 years earlier, in a different county, and has its own separate history and archive. The Newstead Helmet, excavated in 1905 at the Roman fort of Trimontium near Melrose in the Scottish Borders and held by National Museums Scotland in Edinburgh, not Wales, is made of iron rather than copper alloy and is the third member of this small British group; it was shown at Vindolanda, not alongside Crosby Garrett, during the 2017 Hadrian's Cavalry programme.

The Guisborough Helmet, found in North Yorkshire in 1864 and held by the British Museum, is not a face-mask helmet, but it is the closest documented parallel for one specific detail of the Crosby Garrett find: it too had been deliberately folded and dismantled before burial, a pattern that a 2017 technical study argued was consistent with formal, deliberate deposition rather than casual loss or burial for safekeeping ahead of an intended recovery that never happened. A separate, much more fragmentary parade-helmet face-mask, excavated in 1924-25 from a late first-century pit at the Roman fort of Brecon Gaer in Powys and held by National Museum Wales, is sometimes loosely invoked as a Welsh comparandum for the same reason; this record could not verify any distinct object properly identified as a Trawsfynydd Helmet at National Museum Wales, and treats the Brecon Gaer fragment, not a Trawsfynydd find, as the genuine Welsh parallel worth citing.

A further, separate face mask, judged by the conservator who examined it to be from the same workshop as, or closely related to, the Crosby Garrett helmet, was offered for sale by TimeLine Auctions in London in September 2019. It depicts a female or Amazon-type face rather than a young man, is around 95 percent complete, and has an estimate rather than a confirmed hammer price recorded in the trade press consulted; it is a different object, in different, private ownership, with no confirmed findspot of its own, and this record notes the proposed workshop link as the restorer's opinion rather than an established fact.

The Treasure Act reform this case helped drive

The Crosby Garrett helmet became a standing example, cited repeatedly in the years of policy debate that followed, of the central weakness in the Treasure Act 1996: a test based on what an object is made of rather than what it means. Official reviews of the Act, and subsequent parliamentary and museum-sector commentary, returned to this case specifically as a demonstration that a copper-alloy object judged by the British Museum's own curators to be of national importance could still leave the country's public collections entirely lawfully, simply because it was not gold, not silver, and not part of a qualifying hoard.

That pressure eventually produced a change in the law. The Treasure (Designation) (Amendment) Order 2023 introduced a new, significance-based category into the Treasure Act's definition, alongside the existing precious-metal and hoard tests, covering finds of outstanding historical, archaeological or cultural significance regardless of the material they are made from. The Order took effect in July 2023, thirteen years after the Crosby Garrett sale. Commentary on the reform, including academic analysis in the Modern Law Review, has repeatedly used the Crosby Garrett helmet as the clearest illustrative case for why the change was needed, on the reasoning that a find with the same facts, made after July 2023, would now very likely qualify as Treasure and be offered first to a public museum at a price fixed by independent valuation rather than sold at open auction to the highest bidder. The helmet itself, sold under the old law thirteen years earlier, gained nothing from the change; its case is part of why the change exists.

What is still unresolved

Four separate gaps run through this object's documented history, and none of them can be closed from the sources available to this record. The findspot itself, though probably genuine, has never been independently verified to the standard archaeologists would expect of a national find, and the delay before a precise grid reference was recorded means the question can no longer be settled by further fieldwork at the reported site alone. The number and identity of the finder or finders is reported inconsistently across otherwise reliable sources, and no name has ever been made public. The scientific evidence that a properly sequenced conservation, ahead of cosmetic restoration, might have recovered from the corrosion products and fragment positions was not collected, and cannot now be recovered by any later examination of the reassembled object. And the anonymous buyer's identity, and whether the helmet remains in the same hands as in October 2010, are not publicly known and have not changed in any source this record could locate since 2017.