FINAL REPORT

Non-Destructive ImagingOf Worn-off Hallmarks and Engravings

From Metal Objects of ArtUsing Scanning Acoustic Microscopy

GRANT NUMBER MT-2210-0-NC-21

November 15, 2004

By

Paul L. Bensonand

Robert S. Gilmore

1

TABLE OF CONTENTS

EXECUTIVE SUMMARY ..........................................................................................................2

PART 1 INTRODUCTION ........................................................................................................31.1 Historical Development of Acoustic Imaging ...........................................................5

PART 2 METHODS AND MATERIALS.....................................................................................72.1 Theoretical Background.............................................................................................7

2.1.1 Stress Annealing ............................................................................................72.1.2 Anisotropy......................................................................................................8

2.2 Coupling Fluid ...........................................................................................................92.3 Acoustic Transducers...............................................................................................102.4 The Scanning Acoustic Microscope ........................................................................112.5 Initial Experimentation ............................................................................................11

2.5.1 Sterling Silver Coupons ...............................................................................112.5.2 Gold Coupons ..............................................................................................13

2.6 Objects Scanned.......................................................................................................14

PART 3 RESULTS .................................................................................................................163.1 Sterling Silver Spoon Handle...................................................................................163.2 Sterling Silver Fish Knife Blade ..............................................................................173.3 Dessert Knife ...........................................................................................................203.4 Apostle Spoon..........................................................................................................203.5 Sterling Silver Fork..................................................................................................213.6 American Silver Dime .............................................................................................223.7 English Teaspoon.....................................................................................................23

PART 4 DISCUSSION ............................................................................................................23

PART 5 CONCLUSIONS ........................................................................................................25

PART 6 ACKNOWLEDGMENTS ............................................................................................26

PART 7 REFERENCES ..........................................................................................................26

PLATES……………….........................................................................................................28

2

EXECUTIVE SUMMARY

The goal of this project was to determine if worn-off or illegible hallmarks on silver andgold works of art could be imaged using scanning acoustic imaging techniques. Theproject was quite successful, up to a point, and represents the first time acoustic imageshave been made of remnant deformation in silver and gold objects of art.

This imaging technique utilizes the residual plastic deformation left in the metal after thehallmarks or engravings have been struck to form a sonogram image representing thedifferences in the acoustic response of the struck area and the surrounding metal.Theoretically, this should be possible based on the calculated degree of anisotropy andthe stress annealing temperature of both silver and gold. In practice, these residualstresses in gold of the historical alloys of 18K-22K (75%-92% gold) could not be imagedwhile silver, of various experimental alloys, did yield good images.

Numerous combinations of coupling fluids, lens geometries, and sound frequencies weretried in both surface-wave and back-wall imaging modes. In the end, FC-40 (an inertfluorocarbon liquid) used in the surface-wave imaging mode with an input frequency of20 MHz through an F/1 lens produced the best results. These then are the recommendedmaterials and parameters for imaging worn-off marks on silver and gold when using ascanning acoustic microscope.

Hallmarks and engravings from a total of twenty-six silver, two gold, and one copperalloy (bronze) objects were imaged during the two-year course of the project. Multipleimages were made of the marks on these objects so that well over one hundred imageswere made during the project. The silver objects ranged in age from the 17th to the 20th

centuries, the gold objects were modern coupons, and the one bronze object was a coindating from the 1790’s. The silver content of the objects ranged from 80-95% silverwhile the gold objects were of 18K and 22K. The silver objects examined came from thecollections of the Nelson-Atkins Museum of Art, private collections, and from purchasesfrom dealers.

Imaging of worn-off or illegible hallmarks and engravings was successful one hundredpercent of the time when residual flow was still present in the metal. Searches of theavailable literature on hallmarking techniques and consultations with silversmiths and theSuperintendent Assayer of the Worshipful Company of Goldsmiths in London revealedthat heat was probably used during most of the final cleaning and shaping processes ofsilver objects after the hallmarks were applied. It now appears that in most cases the heatwas sufficient to anneal the metal, thus eliminating the residual plastic flow. In thesesituations there was nothing to image except metal grain structure. Unfortunately, thereare no visual clues on the surface of the metal to indicate whether the requisite plasticflow is still present in the subsurface.

3

1. INTRODUCTION

The use of hallmarks on silver has a longhistory dating back to the fourth century ADand represents the oldest known form ofconsumer protection. A series or system offive marks has been found on Byzantinesilver dating from this period though theirinterpretation is still not completely resolved(Dodd, 1961).

Hallmarking of European silver probablyoriginated in France in the 13th century andspread from there to other countries. Thealloy that is today universally recognized as‘sterling silver’ (92.5% silver) originated inan English statute of 1300 and was based onthe alloy of the English silver coinage in useat that time. The English gold alloystandard was based on an existing alloystandard known as the “Touch of Paris” or19.2 carats/ 80% gold (Hare, 1978). Incontrast to these long established standards,the American standards were not formalizeduntil 1906.

As hallmarks were a form of consumerprotection there were strict penalties fortheir misuse. For example, a goldsmithcould have his substandard wares seized, hecould be fined, jailed, maimed, banished, oreven put to death (Jackson, 1921).

The complete history of Europeanhallmarking of silver and gold is far fromcomplete as some historical records havebeen lost through time. For example, theLondon guildhall records prior to1681 werelost in a fire at the Assay Office, and inHolland records were destroyed when theguild system was abolished in 1807. As thehistory and standards of hallmarking silverand gold objects from various countries iscomplex, it should be consulted on anindividual basis (Figure 1).

Figure 1. Typical set of English hallmarks indicatingthat the object is made of Sterling quality silver,made by Hester Bateman in London in the year 1787,and a duty has been paid on the finished piece.

Hallmarks on silver and gold objects can fixthese pieces in history by providing directevidence of the maker, the place and date ofmanufacture, and the quality of the metalalloy at a particular time. To some extentthen the historic, monetary, and intrinsicvalue of the objects are directly linked to theability to discern the hallmarks. Silver’ssusceptibility to tarnishing means that itmust to be polished regularly to maintain itsdesired bright metallic surface finish. Thepolishing process removes a thin layer ofsilver metal so that over time the hallmarkswill be gradually reduced to the point wherethey are either illegible or completely wornaway, resulting in the loss of valuablehistoric information. The ability to read theoriginal marks would greatly aid in placingthese objects back into their rightful place inhistory.

Even though the hallmark may becompletely worn away, there may still besufficient residual plastic deformation withinthe metal from the original act of striking thesurface to create an image of the vanishedhallmark. This residual deformation can becharacterized in the form of an acousticresponse when the surface is insonified witha focused acoustic beam; the amplitude ofthe response is then used to create an imageon a computer screen. The highly polished

4

silver surface provides a nearly idealmedium for the utilization of acousticimaging techniques. As a general rule, thesurface roughness of the material to beimaged should be less than one-third thewavelength of the input acoustic frequencyto provide sufficient returned energy toproduce an image; when the surfaceroughness exceeds this parameter there istoo much surface scattering of the acousticsignal to produce a meaningful image.

Other methods have been successfullyemployed to images worn-off informationfrom metal. Recovery of filed-off serialnumbers from firearms is a well-established

procedure in law enforcement forensiclaboratories. Unfortunately, all of thesetechniques are destructive to the metal tosome extent (Table 1). The most commontechnique involves polishing the area to beimaged then etching the surface with an acidto bring out the latent serial numbers;needless to say, the use of this techniquewould not be tolerated on works of art. Thenewly developed acoustic imagingprocedure is non-contact and does not harmthe metal in any way. It is the only knownnon-destructive technique that has thepotential to recover lost information fromsilver and gold works of art.

Table 1. (compiled from Treptow, 1978)

5

1.1 Historical Development of Acoustic Imaging

The ultrasonic imaging technologies forvisualizing the surfaces and interiors ofopaque solids are well established (Gilmore,1999). Between 1929 and 1931, Sokolovand Mulhauser independently proposed theuse of ultrasonic waves to form images ofthe interior of materials for materialscharacterization and nondestructiveevaluation (NDE). During the 1930’s allefforts to develop ultrasonic imagesinvolved the development of acousticamplitude sensitive screens that displayedvisible contrast in proportion to the acousticamplitude incident on the screen. Theseimage converter screens (such as thePohlmann Cell and the Sokolov Tube) hadsuch poor sensitivity and resolution thatlittle use was made of them other than ascuriosities. Pulse-echo and pulse-transmission C-Scan images, using bothfocused and unfocused ultrasonic beams,were introduced in the early 1950’s. Theprimary use was for industrial NDE. Theseinitial C-Scan images were displayed onphotographic or voltage sensitive paper andwere acquired by scanning a singletransducer back and forth over the subjectmaterial. The image was built up line byline. By the early 1970’s ultrasonic C-Scaninspections of both the surfaces and interiorvolumes of industrial materials were ingeneral use and C-Scan images had beenproduced as high as 50 MHz in frequency.In the early 1970s work at StanfordUniversity under the direction of C.F. Quate(Lemons and Quate, 1979) combined zincoxide on sapphire transducers, C-Scan dataacquisition, and microwave electronics tocreate very small ultrasonic images at GHzfrequencies. These images rivaled opticalmicroscopy in resolution, detail, and field ofview; therefore, the devices that made themwere called Scanning Acoustic Microscopes.The GHz frequencies, low depths ofpenetration, and very small fields of view

limited the industrial usefulness of scanningacoustic microscopy except formicroelectronic assemblies. However, thenear optical resolution of the acousticmicroscope images provided a newemphasis and enthusiasm for ultrasonicimaging in general. This renewed effort,combined with the collateral advances in thecomputational power, storage, and displaycapabilities of small computers, resulted inthree decades of rapid progress in ultrasonicimaging devices, methods and applications.By the start of the 21st century ultrasonicimaging methods were well established tocharacterize material microstructures, bonds,defects (flaws, voids, cracking, porosity,layer delaminations), coating delaminations,elastic modulus and density variations, heataffected zones in welds and other fusionprocesses, stress distributions in isotropicmaterials, and in vitro carious lesions.Materials examined include ceramics,composites, glass, metals and alloys,polymers, plastics, semiconductors,electronic components, geological materials,coffee and soybeans, bone, teeth, softbiological tissue, and organic compounds.However, a recent literature search hasfound only three references to the use ofacoustic microscopy for evaluating metal orceramic art objects (Stravoudis, 1989;Benson, 1991; Ouahman, 1995).

Several texts are available that clearlydescribe ultrasonic imaging and acousticmicroscopy (Lemons and Quate, 1979;Briggs, 1982; Gilmore, 1999); therefore thecharacteristics and operation of the systemswill only be summarized here. A typicaltransducer used for acoustic imagingconsists of a piezoelectric layer cut to aspecified frequency and bonded to a plano-concave lens to focus the ultrasonic beam.For high frequency operation the lens isusually fabricated from single crystalsapphire or fused quartz. Alternatively,eliminating the lens and spherically curving

6

the piezoelectric layer itself can also focus theultrasonic beam. In the case of pulse-echo C-Scan data acquisition, the transducer acts asboth the transmitter and receiver of theacoustic energy. A short electrical pulse isapplied to the piezoelectric layer to create theacoustic pulse and return acoustic echoesinteract with the layer to create electricalsignals. The object to be scanned is placed atthe focal point of the ultrasonic beam. Whatmakes the acoustic microscope unique is theability to place the focal point of the acousticenergy either on the surface of the object orsubsurface in the object’s interior. Again, aswith all C-Scan type data acquisition, theimage is acquired by raster scanning theultrasonic beam and acquiring echoamplitudes at an increment along the scanlines equal to the line-to-line spacing (Figure 2).

Figure 2. Schematic of an ultrasonic imagingsystem; higher frequencies and higher imagemagnification would make the same schematic anacoustic microscope.

The contrast changes in acoustic images areproduced by variations of elasticity, density,and acoustic attenuation within the materialbeing imaged. When the hallmark is appliedby striking the surface of the metal, a localizedchange in the silver’s physical propertiesoccurs. Some of these properties that arealtered include yield strength, tensile strength,hardness, and ductility. Any or all of thesechanges effects the acoustic attenuation of theinsonified signal directly under and in nearproximity to the struck area (Figure 3).

Figure 3. Cross-sectional view of deformation instamped hallmark. Slip lines, twinning bands, andsmaller grain size result where metal absorbsstamping compression (shown by arrows). FromTreptow, 1978.

In the specific case of imaging wornhallmarks, it has been demonstrated thatimages of the residual deformation in themetal from the stamping process can beobtained by two methods: (1) Backwall orback surface imaging where an acousticbeam is focused through the full thickness ofthe silver and on the back surface containingthe hallmark deformation (Figure 4a); (2)Surface wave imaging of the surfacecontaining the hallmark deformation (Figure4b). In other words, surface waves are usedto produce images of the entry surface, i.e.,the struck surface, whereas backwall imagesare obtained from the surface opposite to thestruck surface.

Figure 4. Schematic showing (a) back surfacereflection imaging (or backwall imaging) and (b)mode converted surface wave imaging.

(a) (b)

7

Only that part of the reflected signal thatcorresponds with the surface of interest isused to create the image; this process isreferred to as ‘gating’ (Figure 5). Correctly

choosing the proper reflection to gate andprocess requires experimentation andexperience on the part of the operator toobtain a properly resolved image.

Figure 5. A typical acoustic pulse showing a ‘gate’ around that portion of the pulse to be processed toform an image. (source of image unknown)

2. METHODS AND MATERIALS

2.1 Theoretical Background

2.1.1 Stress Annealing

A first step in determining if the amount ofresidual deformation in silver or any othermaterial is sufficient for acoustic imaging isto determine the stability of this deformationover time. The lowest temperature thatmight affect this stability is the residualstress annealing temperature. Thistemperature is generally considered to beapproximately 4/10ths (0.4) of the absolutemelting temperature as expressed in degreesKelvin (oK) (Callister, 2003). The highesttemperature below the melting pointaffecting the retention of the deformation isless exact, but is in the range oftemperatures at which recrystallizationoccurs. Here the grain boundaries in the

silver migrate and the microstructureentirely recrystallizes. Any residual plasticflow remaining from striking a hallmarkwould begin to relax at the stress annealingtemperature and could totally disappearduring recrystallization. Since the meltingpoint (Mp) of sterling silver is 893oC(1166oK) the stress annealing temperaturewould be approximately 0.4 x 1166oK =466oK, or approximately 93º above theboiling point of water (100ºC or 373oK).Room temperature is typically approximatedat 300oK; since the lowest criticaltemperature for sterling silver (466oK) iswell above this temperature, it seemsreasonable to expect the residualdeformation produced by a hallmark stampto be relatively stable over a few hundredyears of time, even if repeatedly washed inhot water. Stress annealing temperatures forother metals are given in Table 2.

8

Table 2. Approximate Stress Annealing Temperatures for Various Metals

METALMELTING POINT (MP) IN

DEG. CELSIUS

STRESS ANNEALINGTEMPERATURE IN DEG.

CELSIUS

RATIO OFMAX AMBIENT

TEMPERATURE TO MP INDEG. KELVIN*

Gold** 24K 1063 261 0.2422K 1003 237 0.2518K 905 198 0.2714K 845 173 0.28

Silver pure 962 221 0.25Sterling 863 193 0.27

Copper 1082 269 0.23Lead 327 -33 0.50Bronze 10% tin 1005 238 0.25

20% tin 890 192 0.27Brass 10% zinc 1040 252 0.24

20% zinc 995 234 0.25Iron 1538 451 0.18Steel 1515 442 0.18* at ratios less than 0.40 the plastic flow surrounding the hallmarks should be stable at temperatures up to the

stress annealing temperature** from Smith, 1978; all other melting point figures are from Lide, 2002; other figures were calculated

2.1.2 Anisotropy

A second consideration in imaging residualdeformation is to determine the acousticproperties of the subject material and anypossible anisotropy of the material. Unlessthe deformation process produces micro-fractures there is no reason to anticipate thata truly isotropic material would be renderedanisotropic by plastic deformation.Anisotropic materials, however, shouldundergo considerable change duringdeformation, since a local deformationwould significantly rearrange thatmicrostructure. It seemed appropriate toestimate the anisotropy in silver todetermine if ultrasonic backscatter from thesilver microstructure itself might be used totrack the deformation underlying hallmarks.The three elastic constants for single crystalsilver (cubic system) are C11 = 1.239 Mbar,C12 = 0.939 Mbar, and C44 = 0.461 Mbar(Simmons and Wang, 1971). Isotropicmaterials have only two independent elastic

constants instead of the three required todescribe the cubic system. A typical test forisotropy (again within the cubic system) isthe Zener Anisotropy Ratio [C11 – C12] /2.0 to C44 (Chung and Buessem, 1968).Clearly [1.239 – 0.939] / 2.0 = 0.150 and isnot equal to 0.461 so silver possessesconsiderable anisotropy. Thereforeultrasonic backscatter from the silver grainsshould be able to track the modifications inthe microstructure caused by the plastic flowin the silver around the hallmarks.Anisotropy for other metals is shown inTable 3. Having established this possibility,one should immediately state thatbackscatter imaging of the silvermicrostructure has not proven effective todate for displaying residual deformation inthe silver. The probable explanation for thishas to do with the small size of the silvergrains since, even at 50 MHz, the grains aretoo small to provide any backscatteramplitude.

9

Table 3. Estimation Of The Degree Of Anisotropy Of Various Metals

METAL C11 C12 C44 RATIO OF [C11-C12]/2.0 TO C44*

Gold 1.923 1.631 0.420 0.15Silver 1.239 0.939 0.461 0.15Copper 1.684 1.214 0.755 0.24Lead 0.495 0.423 0.149 0.05Brass-4% zinc 1.633 1.177 0.744 0.23

9% zinc 1.571 1.137 0.723 0.2217% zinc 1.499 1.098 0.715 0.20

Iron 2.314 1.346 1.164 0.49* If this ratio is less than C44 then the metal exhibits anisotropy. All data are from Simmons and Wang,1971.

2.2 Coupling Fluid

For frequencies much above 1 MHz,acoustic waves are rapidly attenuated inair so it is necessary to utilize a couplingfluid between the transducer and object tobe imaged. The acoustic properties of thecoupling fluid are a significant factor indetermining the resolution that can beachieved by the acoustic imaging system.The most widely used fluid is water butother fluids have acoustic properties(namely a higher or lower velocity) thatmake them superior to water particularlywhen surface wave imaging is used (Table4).

Distilled water at ambient roomtemperature was initially used to produceback-wall acoustic images of thehallmarks. While this producedreasonable images it was felt that theimage resolution could be improved byprocessing surface wave information from

the metal. A search was made for a liquidwith a relatively slow acoustic velocitytransmission and that met certain healthand safety considerations. A line ofperfluoro liquids called Fluorinert®,manufactured by 3M Specialty Materialsmet the health and safety criteria and arenon-corrosive when in contact withprecious metals. They contain nohydrogen or chlorine making them un-reactive to most metals and they areenvironmentally safe, non-toxic, have alow volatility, and have an acoustictransmission velocity of about one halfthat of water.

After initial experimentation with variousFluorinert® liquids it was determined thatFC-40 was best suited to the imagingrequirements. This coupling liquid hadthe least attenuation of the acoustic signalwhile its acoustic properties best matchedthe transducer lenses available at the time(Table 5).

10

Table 4. Relative Acoustic Velocities Of Some Coupling Fluids

Table 5. Acoustic Properties Of Select Fluorinert® Liquids

COUPLING LIQUIDACOUSTIC VELOCITY

(MM/µSEC)DENSITY(G/ML)

VISCOSITY(CENTISTOKES)

FC-40 0.656 1.87 2.2FC-70 0.687 1.93 14.0FC-72 0.512 1.68 0.4FC-77 0.505 1.78 0.8FC-84 0.542 1.73 0.55Distilled Water 1.495 1.00 1.0

The propagation of longitudinal soundwaves in water is approximately 1.48mm/sec versus 0.656 mm/sec in theFC-40. This slower acoustic velocitytranslates into a smaller angle of incidencefor the generation of surface waves insilver allowing for the use of standardtransducers for the imaging instead ofrequiring custom-made transducers ifwater was the coupling medium (Table 6).

For this work FC-40 was used to makesurface wave images in the sterling silverobjects discussed here. The imagesacquired in this work were all made byimmersing the silver objects in FC-40 forsurface wave imaging and in distilledwater for back-wall imaging.

2.3 Acoustic Transducers

After some initial testing it was determined thatthe best (highest resolution) images for back-wall imaging was obtained with a Panametrics

Type V390 50 MHz F/2 transducer with 0.5”focal length, 0.25” aperture opening, and a 1.5”diameter quartz buffer-rod. For surface waveimaging a Panametrics polymer film 20 MHzF/1 transducer with 0.4” focal length and 0.4”aperture opening was used. These transducerswere already available at the facilities where allimaging took place.

With the hopes that the resolution of thesurface wave images could be furtherimproved, a custom designed transducer waspurchased. It was a Panametrics PZ25 33MHz PVDF film transducer with 0.25” focallength and 0.25” aperture opening. It was feltthat the shorter fluid path of this transducerwould allow for higher frequency returns(improved resolution) but the frequency of thereturned signals still did not exceed much morethan 7.5MHz. This was only a slightimprovement over the 20 MHz F/1 transduceroriginally used and did not produce visuallysuperior images.

COUPLING FLUIDTEMPERATURE(DEGREES C)

VELOCITY(MM/µSEC)

ABSORPTION(S2/M)

Distilled Water 25 1.495 22.0Distilled Water 60 1.550 10.2Acetone 30 1.158 54.0Ethanol 30 1.127 48.5Methanol 30 1.088 30.2FC-40 25 0.656 Not available

11

Table 6. Acoustic Velocities And Angle Of Incidence To Generate Surface Waves InSterling Silver And Gold*

CHARACTERISTICSTERLING

SILVER18K GOLD 22K WATER FC-40

Longitudinalvelocity(mm/sec)

3.89 3.55 3.39 1.48 0.656

Shear velocity(mm/sec)

1.73 1.46 1.31 0 0

Rayleighvelocity(mm/sec)

1.63 1.33 1.15 Frequencydependent

Frequencydependent

Angle ofincidence togeneratesurface waves

Water

65.3o

FC-40

23.7o

Water

>90o

FC-40

29o

Water

>90o

FC-40

35o

*All figures were experimentally derived

2.4 The Scanning Acoustic Microscope

The microscope utilized for allexperimentation is owned by General ElectricResearch and Development and located inSchenectady, New York. The scanner is anAnorad ruling engine design. The electronicswere developed by General Electric and have abandwidth of 100 kHz to 100 MHz with 70 dBof available gain. The dynamic range of thissystem is 40 dB and its output data is digitizedat 8-bits. The display software is calledWinview and was also developed by GeneralElectric.

2.5 Initial Experimentation2.5.1 Sterling Silver Coupons

Initial experimentation was conducted on ablank sterling silver (92.5 % silver) couponmeasuring approximately 25 mm x 25 mm x 3mm. An experienced silversmith then placedthree different hallmarks on one surface. Asilversmith was employed to produce thehallmarks thinking that he would strike thesilver with approximately the same force usedby silversmiths for the past several hundred

years so that the marks would be neither toodeep nor too shallow. The struck surface ofthe coupon was then polished with variousgrades of diamond paste on a StruersPrepamatic rotary lapping machine until thehallmark was no longer visible; approximately0.3 mm of silver was removed. This was doneto approximate the slow removal of the silversurface in much the same manner as years ofpolishing. The polishing was done at thepremises of Honeywell Federal Manufacturingand Technologies in Kansas City, Missouri.

Once the marks were completely removedfrom the surface, the coupon was placed in acontainer with some keys and the containervibrated to produce scratches on the silver tosimulate the surface on a genuine aged silverobject. The three ultrasonic images shown inFigure 6 illustrate the detail ultrasonic imagingcan produce on both intact hallmarks and thedeformation remaining after removal bypolishing. Figure 6a shows a 50 MHz F/2back wall image of a coupon that still retainsalmost all of the hallmarks placed on it asstruck. Figure 6b shows a 50 MHz F/2 back-wall image of the residual deformation in thecoupon where almost all of the original

12

hallmarks have been polished away. Figure 6cshows a 20 MHz F/1 surface wave image ofthe same deformation as figure 6b exceptviewed from the surface containing the residualdeformation. Both back-wall images wereacquired using water to couple the ultrasonicbeam into the part. The surface wave imageused FC-40 in order to mode convert alongitudinal wave in the fluid into a surfacewave on the silver coupon’s surface.

In order to prove that this imaging techniquewas not unique to the apparatus at GeneralElectric, another research facility that employsacoustic microscopy, Honeywell FederalManufacturing and Development in KansasCity, Missouri, was asked to replicate theexperimental procedure used on the testcoupon. The images compared quite favorablyto the images produced by General Electric soit is felt that the imaging procedure isapplicable for use on other acousticmicroscopes. See Plate No. 31.

Figure 6. Three ultrasonic images of a sterling silver test coupon. (a) A 50 MHz F/2 back-wall image of the original hallmark. (b) A 50 MHz F/2 back-wall image of the residualdeformation remaining in the coupon after the hallmark has been polished away. (c) A 20MHz F/1 surface wave image of the same deformation in 6b except imaged from the surfacecontaining the deformation

13

Figure 7. Acoustic images: (A) an 18K gold coupon and (B) a 22K gold coupon

2.5.2 Gold Coupons

Coupons of 18K and 22K gold, measuringapproximately 25 mm x 25 mm x 3 mm, werepurchased and stamped with hallmarks by anexperienced silversmith. These alloys werechosen because the earliest historical Europeanstandard for gold work was 19.2carats (80%). Inlater years additional standards were establishedbut it was felt that 18K and 22K would representthe alloys that would be most often encounteredin works of art.

Once the hallmarks had been placed on thecoupons they were polished off as previouslydescribed. Surface wave imaging of the goldcoupons was done only with FC-40 as thecoupling medium; no back-wall imaging wasdone. Converted surface waves were captured

but surprisingly, either no images or barelyperceptible images of the polished-off hallmarkswere recovered (Figure 7). A possibleexplanation for these results can be accounted forby one of the physical properties of gold- itsmalleability. High purity gold is extremelymalleable and does not produce plastic flowwhen struck during the hallmarking process.Instead, the metal simply pushes aside and wells-up around the hallmark leaving no deformation toimage. It is proposed that a lower purity gold,such as 14K (58.3%), that has been alloyed withan increasing quantity of copper will exhibit someplastic deformation making it possible to recoverhallmarks using the acoustic imaging technique.No further experimentation with gold was donebecause of the lack of results on these twocoupons.

14

2.6 Objects Scanned

A total of twenty-nine silver objects were scanned(Table 7). Objects were selected to cover a broadrange of ages and to represent silverwork fromseveral countries. It was assumed that not allobjects were “sterling” silver as alloy standardsvaried from country to country and changedthrough the course of time. For example, theEnglish silver standard was changed from sterling(92.5%) to Britannia (95%) for the period 1697-1719. However, the Britannia standard wasnever abolished leaving the silversmith free toproduce wares that were marked with the sterlingstandard but could in fact be of the higher silveralloy. It should be remembered that the alloystandards provided for minimum silver contentbut not a maximum silver content (Jackson,1921).

Other countries established their own standards,which, again, varied over time. Imaging ofAmerican silver objects was initially avoided, astheir silver content could not be verified withoutadditional testing. Standards for American silverwere not formally codified until 1906, whichmeant that silver could be called ‘sterling’

without meeting the generally accepted 92.5%silver content. In fact, there are examples ofAmerican ‘sterling silver’ that contain no silver atall. In the end, these variations in the silvercontent proved to have no effect on the ability toimage the residual deformation in the silver.

Objects chosen for imaging were limited to afairly small size for two considerations. First wasthe quantity of the coupling fluid available. Afterthe coupling fluid, FC-40, was chosen only asmall quantity, approximately two liters, waspurchased due to the very high cost of thematerial and for fear that it could quicklyevaporate if there were no mechanism to preventit from doing so. Secondly, it was much moredifficult than initially thought to find silverobjects with worn-off hallmarks that the ownerswere willing to lend for the experiments.Insurance considerations prevented certain piecesfrom leaving the Nelson-Atkins Museum andprivate owners were hesitant to loan their silver tothe project without additional insurance. In theend, only five pieces from the Nelson-AtkinsMuseum were tested, with the rest coming fromloans from silver dealers, private owners, orpurchases.

15

Table 7. Objects Scanned For Hallmark Recovery Project

OBJECT DATE OWNER ORIGIN PLATE NO.Sterling silver test coupon Modern Purchased 1, 2 & 31Sterling silver test coupon(annealed)

Modern Purchased Figure 12B

Teaspoon (Bateman) 1792 NAMA no. 72-45/4B English 3Spoon (Christina) Early 20th

cC. Nelson American 4

Teaspoon (salt) 1780 Purchased English 5Teaspoon (Paul) Unknown P. Benson Unknown 6Teaspoon (EW) 1790’s (?) Purchased English (?) 7Teaspoon (MH) Unknown Purchased Unknown 8Teaspoon (JRD) Unknown Purchased Unknown 9Teaspoon (1 Mono) Unknown Purchased Unknown 10Paten cover 1569 NAMA no. 52-16B English 11Dessert knife 1800-1810 NAMA no. 76-34/12C French 12Fish knife 1875-1925 NAMA no. F83-76/10 French 13Spoon (Churchman) Unknown M. Churchman English 14Fork (Belgium) 18th c. W. Wilkinson Belgium 15Spoon (Apostle) 1603 J. Shredds English 16Spoon (Apostle) 1603? J. Shredds English 17Fork (Barbados) Colonial W. Wilkinson Barbados 18Spoon (Rattail) 1716 J. Shredds English 19U.S. Silver Dime 1840’s G. MacCurdy American 20 & 21French coin 1790’s K. Garland French 22Spoon (BH) 18th c J. Weidman Scottish 23Spoon (George III) 1790 W. Wilkinson English 2422K gold coupon Modern Purchased 2518K gold coupon Modern Purchased 26Nichols Silver Coupon A Modern Purchased 27Nichols Silver Coupon B Modern Purchased 27Trivet 1787 NAMA no. 71-45/1 English 28Fork (Monogram) 1808 W. Wilkinson English 29Spoon (1809) Unknown W. Wilkinson French

Caribbean30

16

3.0 RESULTS

To eliminate duplicate discussions of theresults of scanning all the individual objects,only a few of the results that illustrate aparticular point will be presented here.Acoustic images of all of the objects can befound in the attached plates.

3.1 Sterling Silver Spoon Handle

Figure 8 shows a set of surface wave imagesof a sterling silver spoon wrought by Peterand Ann Bateman dating from 1792. Thisteaspoon, one from a set of eight, waschosen because its four hallmarks variedfrom perfectly readable to completelypolished away. Also, the four hallmarks are

legible on the other spoons from this set,making it easier to target the desired imagequality. In Figure 8a the makers’ initials areclear but much of the remaining hallmarkshave been removed. Figure 8b showsisolation, magnification, and partial recoveryone of the hallmarks believed to be that of alion. The image of the “lion” shown inFigure 8c is the best result of a series oftrials where the focus of the transducer waschanged slightly for each trial. Theimportance of even very small changes inthe system focus has been repeatedlydemonstrated in the course of this work. Tothe untrained eye the figure of the lion is notclear in the acoustic image but to an expert itis readily discernable (Wilkes, 2001).

Figure 8. (a) Ultrasonic images of the handle of a sterling silver teaspoon wrought by Peter and AnnBateman dating from 1792. (b) The initials of the makers are clear but much of the remaininghallmarks have been removed. (c) Shows the isolation, magnification, and partial recovery of a figurethought to be a lion.

17

Subsequent imaging processing comparisonswith a visible hallmark from anotherteaspoon from this same set of spoons hasconfirmed that the recovered image of thelion mark is identical to the visible hallmark.A series of six acoustic images of the worn-off lion hallmark were adjusted for size andoverlaid on top of a digital image of thevisible lion hallmark. The acoustic imageswere made by adjusting the focal spot of thesound waves either a little higher or a littlelower in the metal. With only a twohundred-nanosecond difference in the traveltime of the sound waves from the firstacoustic image to the sixth one there was asurprising difference in the quality of theimages (200 nanosecond travel timeconverts to an actual distance difference of0.0026 inch). The composite acousticimages had near perfect registration on thevisible hallmark which demonstrates that theillegible hallmark was struck with the samedie as the visible hallmark on the teaspoonfrom this set of teaspoons (Figure 9).

Figure 9. A recovered acoustic image of a worn-offlion hallmark overlaying a digital image of anidentical hallmark from the same set of teaspoons.

Figure 10 shows the best results of a seriesof trials on recovering the date letter ‘r’.Here, a series of three acoustic images of theletter ‘r’ were adjusted for size andorientation then overlain on top of a digitalimage of the corresponding hallmark from

another teaspoon from the same set. Again,there is near perfect registration of therecovered acoustic image on the visiblehallmark. This demonstrates that therecovered hallmark image is actually theletter ‘r’ and that it was struck with the samedie as the visible hallmark on the otherteaspoon.

Figure 10. Recovered acoustic images of the dateletter ‘r’ overlaying a digital image of a visible ‘r’hallmark from the same set of teaspoons.

3.2 Sterling Silver Fish Knife Blade

Figure 11 is intended to show the lack ofsubsurface deformation where one wouldnaturally assume that it should be present.Shown is a set of ultrasonic back-wallimages of the sterling silver blade of aFrench fish knife dated approximately 1875to 1925. One hallmark has been isolatedand magnified (b) for comparison to theback-wall image of the deformation in thetest coupon (c). Clearly no deformationappears to extend from the fish knifehallmark, suggesting that it has either beenimproperly struck by the silversmith, it hasbeen worn away through the subsurfacedeformation zone, or the residual

deformation has been ‘relieved’ during anannealing process. The annealing processcould have occurred during manufacture orthat the heat from the process of solderingthe handle to the blade was sufficient tocause a localized annealing of the hallmarkssince they are placed quite close to theattached handle.

A literature search found that three of thefour hallmarks applied on French silvermanufactured prior to 1789 were actuallyapplied to the roughed–out silver sheetbefore the object was completed. The

finished object would therefore have beensubjected to multiple annealing steps duringits manufacture thereby relieving the metalof any remnant deformation from thehallmarking procedure (Bimbenet-Privat andde Fontaines, 1995). This is in contrast tothe English system of applying thehallmarks only after the object had beencompleted or nearly completed, thus theresidual plastic deformation in the metalwould be expected to be retained. Anexception to this procedure will be discussedlater.

Figure 11. Ultrasonic back-wall image of the sterling silver blade of a French fish knife datedapproximately 1875 to 1925. One hallmark is isolated (b) and magnified for comparison to the back-

18

wall image of the coupon (c). No deformation appears to extend from the fish knife hallmark.

19

To confirm the historical accounts of theFrench hallmarking procedures a sterlingsilver coupon was stamped with severalhallmarks as described earlier. Again, themarks were polished off and images of thesehallmarks were produced with the acousticimaging technique. The coupon was thenannealed in an oven at 700 ºC for twelveminutes and then subjected to the imagingprocedure. After only one annealing theremnant deformation has been ‘relaxed’ andthe hallmarks can no longer be imaged(Figure 12). Incidentally, the dark area inFigure 12 B shows an enrichment in coppercompared to the lighter colored more silverrich area around it. (See also the SEM-EDXanalysis on Plate 32)

Regrettably, this means that the acousticimaging technique will not work on pre-

1789 French silver objects (after this datethe French hallmarking procedures werechanged).

By chance, Figure 11 also shows additionalinformation recovered by acoustic imagingconcerning the quality of the solder join ofthe handle to the blade. The light coloredspots inside the attachment area representgaps/flaws in the solder join. These areashave a different acoustic response than thesurrounding well-soldered metal so they arereadily visible.

Another interesting chance image wasobtained from a 16th century paten cover(not illustrated) during the course of imagingthe hallmarks. In this case, flaws or bubblesin the coating that were not visuallyapparent but were quite obvious inthe acoustic image.

Figure 12. Effects of annealing on sterling silver. (A) A recovered acoustic image of polished-off hallmarks on a sterling silver coupon before annealing. (B) The same area after annealingthe coupon.

20

3.3 Dessert Knife

Acoustic images of a steel bladed silverFrench dessert knife dating from thebeginning of the nineteenth century areillustrated in Figure 13. The partially worn-off maker’s name is ‘engraved’ on the steelblade and appears to be an ideal candidatefor the recovery of an engraving using theacoustic imaging method. The directreflection image in (a) clearly shows themaker’s name with the central sectioncompletely worn away; Figure 13b is asurface wave image of the same area. No

Figure 13. Silver dessert knife with a steel blade.(a) Direct reflection acoustic image. (b) Surfacewave acoustic image of the same area.

remnant reformation is seen in (b) becausethe maker’s name has not been applied tothe blade by stamping but probably byetching or through the mechanical removalof metal. In either case, there is nosubsurface deformation so acousticalmethods cannot recover the missinginscription.

This object serves as a reminder that inscriptionscan be created by several means and only thosecreated by stamping the metal produce asubsurface deformation that can be recovered byacoustical imaging; the process of removingmetal does not create the requisite deformations.

3.4 Apostle Spoon

Figure 14 is an acoustic image of the hallmarkson a silver Apostle spoon supposedly dating from1503. The date letter “F” was visible and clearlyreadable but the other mark could not bediscerned. A series of acoustic images weremade of the hallmarked area by focusing thesound at progressively deeper intervals into themetal. Unfortunately, none of the images wereable to resolve the illegible mark but someunexpected information was gained. In Figure14it can be seen that (1) there is a noticeable bulgein the handle where the hallmark has beenapplied and (2) there is a well defined join linejust to the right of the letter “F” where it appearsthat the handle was either broken and repaired orthat two individual handles have been joined toform one complete handle. In either case, theplacement of the hallmarks is rather odd. Itshould also be noted that this join and bulge arenot apparent in a visual inspection of the spoon.

An experienced dealer was able to immediatelydetermine that the spoon was a forgery based onstylistic grounds while the acoustic imageprovides hard evidence of the deceit.

21

Figure 14. Acoustic image of the hallmarks on a silver Apostle Spoon supposedly dating from 1503.

3.5 Monogram 1808 Silver Fork

Figure 15a is an acoustic image of apurposely-removed hallmark from anEnglish fork by Thomas Barker dating from1808. At some period during the spoon’slifetime an owner decided to add themonogram ”M” to the back of the handle.In order to accommodate this addition, thelion hallmark was removed. Traditionally,there have been four ways of removinghallmarks and engravings. If the markswere shallow they could simply be polishedaway. Deeper marks could be hammeredout with a subsequent thinning of the metal.They could also be filled with silver solderand finished to seamlessly blend with thesurrounding metal. Finally, they could befilled by a process know as ‘stoning’. In thiscase the surface of the silver is literallyrubbed with a stone, pushing thesurrounding metal into the indentations ofthe hallmarks or engravings. Marksremoved by hammering and stoning cannotbe recovered by acoustic methods but markserased by polishing and filling should berecoverable. In the case of the lionhallmark, its recovery probably meant thatthe hallmark was simply polished away.

Note that in the acoustic image presentedhere, the hallmark is difficult to decipher.At best it can only be recognized that therewas a hallmark present at some time in thepast (Figure 15b). The interpretation of this

Figure 15. Silver fork with a hallmark deliberatelyremoved. (A) Photograph of the visible hallmarks onthe fork. (B) An acoustic image of the hallmarksshowing that a fifth mark was present at one time.The now-missing hallmark has been interpreted as alion.

being the lion hallmark was based on amuch clearer computer image at the time theimage was acquired.

Image processing using Adobe PhotoshopCS was used to try to enhance the image ofthis recovered hallmark. Various filterswere applied to the original acoustic imageto bring out any residual information thatcould be present in the acoustic image and it

22

is now possible to positively identify thehallmark as that of a lion, as was expected(Figure 16).

Figure 16. A. Original acoustic image of removedhallmark from the spoon in Figure 15a. B. Photo-enhanced image of the same hallmark indicating aworn image of a lion. C. Photograph of a lionhallmark for comparison with B.

The use of photo-enhancement softwaregreatly improved (in some cases) the qualityof the acoustic images to the extent thatpositive identifications of images of worn-off hallmarks could be made by non-experts.Subsequently, all of the acoustic imagesacquired during the course of thisinvestigation were re-examined and theseenhanced images are included in the Plates.

3.6 American Silver Dime

One important application of the acousticimaging technique involves the potentialability to image worn-off inscriptions oncoins. Coins are commonly encountered onarchaeological excavations and can providecritical assistance in dating and determiningthe history of the site.

Two coins, one copper alloy and the otherone silver (Figure 17), were imaged to tryand recover the date stamped on them. Inneither case was this successful but differentpreviously illegible markings on the coinswere seen in the acoustic images. In thiscase, the ability to image the markingsprobably depended on whether or not therewas a design on the opposite side in thecorresponding position. Since both sides ofthe coins were struck at the same time thereis a mixing of the subsurface deformationsthat the acoustic signal could not resolve.Even though the dates of the coins could notbe directly ascertained, the images of theother markings were sufficient to allow anexpert to make an informed identification ofboth coins.

In the case of the dime illustrated in Figure17A the surface wave image reveals theword ‘one’ which is not visible in the directreflection image Figure 17B.

Figure 17. American silver dime dating from the 1840’s. (A) Surface wave acoustic image made in FC-40 asthe coupling medium. (B) Same image except water was used as the coupling medium for the direct reflectionimage

23

3.7 “EW” English Teaspoon

The teaspoon had a partially readablemaker’s mark with the first initial being a“T”, the second initial too worn too be read,and two other illegible hallmarks present onthe reverse of the handle (Figure 18).

Figure 18. Photograph of a teaspoon with onepartially readable hallmark and two illegiblehallmarks.

The acoustic images revealed that the twoillegible hallmarks were probably the datemark and the mark for sterling silver, a lion,but both images lacked the resolution tomake a definitive interpretation of the marks(Figure 19).

Figure 19. Acoustic image of the ”EW” teaspoonwith illegible hallmarks and a partial hallmark.

Image processing using PhotoShop CSclarified the acoustic image to the extent that

the previously unidentified teaspoon couldnow be identified as being made by ThomasWallis (initials TW) and the date mark as theletter “h” for the year 1783/84. The thirdhallmark was not distinct but resembles themark for sterling quality metal; i.e. the lionpassant. (See Figure 20.)

Figure 20. Image processed acoustic image with thedate letter “h” and the maker’s initials “TW”outlined for clarity. The third hallmark is stillillegible but resembles the lion passant mark forsterling quality metal.

4. DISCUSSION

A total of twenty-nine silver, gold, andbronze objects from widely varying timeperiods have been subjected to the acousticimaging techniques. Objects imagedincluded spoons, forks, knives, coins, apaten cover, a trivet, and coupon blanks.Results from the modern sterling silverblanks have been very encouraging. Thehallmarks were placed on the blanks in theearly summer of 1997 by an experiencedsilversmith. These hallmarks were well andtruly struck, i.e., their original existence iswell documented. After the hallmarks wereremoved by polishing, ultrasonic imagingproduced clearly decipherable images of theremnant deformation on the surface of thesilver. Both surface wave imaging andback-wall imaging were clearly effective atdisplaying residual deformation in the silver.Where only part of the hallmark wasremoved, the imaging methods are able toshow remnant deformation extending outfrom the remnant surface dents in the

24

surface. The blanks are now approachingseven years in age. Repeat images showresults in 2002 that reproduce the resultsshown in the initial 1997 images. However,despite the clear anisotropy in silver,backscatter imaging of the silvermicrostructure has not yet proven effective.Neither the silver microstructure itself nordeformation of that microstructure has beenshown by backscatter imaging at the 20MHz or 50 MHz frequencies used to date.The failure of the backscatter imaging isconfusing since both the back-surfacereflection images and the surface waveimages clearly indicate that the acousticproperties of the silver showed significantchanges at the hallmark locations. Thesmall size of the silver grains may be onefactor in the inability to produce backscatterimages. More work needs to be done tofully understand this.

Work to recover partially obliteratedhallmarks on antique silver objects has beenless encouraging than the work on thecoupons. But in these cases one cannot becertain that the hallmarks were properlystruck in their original condition. The silverblade of the French fish knife (Figure11)demonstrates this case in point. The fishknife is ideally configured for back-wallimaging and yet no remnant deformationcould be shown to extend from the dentedmarks remaining on the blade. Severaldifferent scenarios could account for this.First of all, the hallmark could have beenimproperly struck so that the entire markwas never there in the first place. It is alsopossible that the heat from the solderingattachment of the handle annealed the silverthus removing the residual deformation ofthe hallmark. Use and/or polishing mayhave partially removed the residualsubsurface deformation or repeated washingin boiling water over an extended period ofyears partially annealed the silver. This last

possibility is most unlikely as the theoreticalstress annealing temperature of sterlingsilver (193o C) is well above the boilingpoint of water.

One other scenario based on the hallmarkingprocedure itself may also be possible. Whena hallmark is applied to a thin piece of silvera ‘witness mark’ may appear on the reverseside of the silver from where the mark wasstruck. This witness mark is a raised areawith the same shape as the hallmark. If thismark is visible the silversmith may wish toremove it; this procedure is called ‘settingback the hallmark’. The silversmith maysimply hammer the witness marks flat or canapply localized heat to that area first to makethe hammering process easier and less likelyto cause any damage to the surroundingmetal. In the case of flatware, the hallmarkswere frequently applied to the back of thehandles before they had been wrought totheir final shape. This allowed the assayoffice to place their marks completely on thesilver and still allow the metalsmith thefreedom to produce a slender handle that inthe final shape would not provide sufficientspace for the hallmarks; the final shapingwould have certainly involved heating themetal. This local annealing effect wouldthen diminish the ability to produce animage of the hallmark using acousticmethods.

As mentioned earlier, it was not possible toexperiment with silver objects of greatervalue than flatware due to insurancerestrictions. It is possible that objects thatare of a size that would allow for thehallmarks to be fully struck withoutalteration of the shape of the piece after themarks have been applied would be moreamenable to the imaging technique. Forexample, the shape of a plate or vesselwould have been fully developed before thehallmarks were applied (except in France)

25

such that only a final cleaning would havebeen required to complete the piece and it isnot envisioned that cleaning, even in a hotacid bath to remove fire scale, would besufficient to remove the subsurfacedeformation of the hallmarking process. Itis quite regrettable that this could not havebeen proven by experimentation on antiquesilver objects.

Surface wave images of two antique coinssuggest that downward or compressivedeformation, i.e., a dent, is more readilydefined than the up welling of material.Efforts to image the originally upraisedpatterns of the years in which coins werestruck have not yet been successful. Thissuggests that the deformation under dents ismore readily detected than bulges. Also,since both sides of a coin are struck at thesame time there will be some mixing ofsubsurface deformations making it moredifficult to separate individual elements ofthe design, e.g., the date.

5. CONCLUSIONS

The success rate for acoustic imaging ofworn-off hallmarks on the twenty-nineobjects in this project has beenapproximately ten per cent. While thisinitially appears to be a fairly unsuccessful,the project has succeeded in producingimages one hundred per cent of the timewhere remnant plastic deformation exists.Subsequent imaging processing of theimages utilizing a standard softwareprogram further clarified about fifty percentof the acquired images. When thedeformation no longer exists either throughbeing poorly struck, being annealed out, orcompletely worn/polished through the zoneof deformation, acoustic methods cannot

produce an image. Unfortunately, there areno visual clues on the surface of the metalthat will permit speculation on the successor failure of the acoustic imaging technique.Each object will have to be imagedindividually to determine if there is anyresidual deformation to be found.

Images of removed or worn-off engravingsmay be recoverable depending on theirmethod of removal. The inscription wouldhave to have been applied by chasing, i.e. byhammering the lines into the metal asopposed to removal of the metal with ascribe. Then they would have to be‘removed’ by either filling them with silversolder or by being polished away. If theinscription was ‘removed’ by eitherhammering or by stoning then the acousticmethod cannot be used for their recovery.

Worn hallmarks on objects manufacturedfrom high purity gold cannot be imaged withthe acoustic methods as described here. Thehistorical standards for objects made of goldhave been 18K or greater. It is suggestedthat at this purity gold is too malleable toproduce plastic flow when struck in thehallmarking process. As the quantity ofalloying metal in gold increases(corresponding to a decrease in the purity ofthe gold) the chances for the acousticrecovery of worn hallmarks and inscriptionsshould increase but this has not yet beenproven experimentally.

Acoustic imaging of archaeological artifacts,e.g. coins, has great potential to be avaluable aid in dating and re-constructingthe history of excavated sites. The onlycaveat being that the objects must have avery smooth surface to permit successfulimaging.

26

6. ACKNOWLEDGEMENTS

This work described in this paper waspartially supported by a grant from theNational Park Service and the NationalCenter for Preservation Technology andTraining. Its contents are solely theresponsibility of the authors and do notnecessarily represent the official position orpolicies of the National Park Service orNational Center for PreservationTechnology and Training.

A Coolidge Fellowship Grant from theGeneral Electric Research and DevelopmentCenter to R. Gilmore supported part ofGilmore’s experimental research, and theauthors also thank GE for providing theCenter’s ultrasonic imaging facilities. Again,the contents of this paper are solely theresponsibility of the authors and do notnecessarily represent any position orpolicies of the General Electric Company.

The authors would like to thank Dr. RobertOrgan, the Superintendent Assayer of TheWorshipful Company of Goldsmiths,London, for his contributions onhallmarking technology; Marty Cunninghamat Honeywell Federal Manufacturing &Technologies, Kansas City for polishing thehallmarks off of the silver and gold coupons;Eric Jamieson also at Honeywell FederalManufacturing and Technologies forproducing additional acoustic images; DaleBenson, Conservation Assistant and JohnLamberton, Digital Imaging Specialist at theNelson-Atkins Museum of Art, for theimage processing of the acoustic images,Christina Nelson, former Curator ofEuropean and American Decorative Arts atthe Nelson-Atkins Museum of Art, for theuse of her silver collections and WynyardWilkinson for the generous loan of silverobjects from his private business.

7. REFERENCES

Benson, P. L. 1991. Scanning acousticmicroscopy and potential application inarchaeological conservation. Unpublisheddissertation in partial fulfillment of therequirements for the Degree inConservation. Institute of Archaeology,University of London.

Bimbenet-Privat, M. and G. de Fontaines,1995. La datation de l’orfèvrerieparisienne sous l’ancien régime: poinçonsde jurande et poinçons de la marque 1507-1792. Paris: Paris Musées.

Briggs, A. 1982. Acoustic Microscopy.Oxford: Oxford University Press.

Callister, W.D. Jr. 2003. Material Scienceand Engineering: An Introduction. 6th

edition. New York: John Wiley and Sons.

Dodd, E. C. 1961. Byzantine silver stamps.Dunbarton Oaks Studies, 7: 23-35.

Chung, D.H. and Buessem, W.R. 1968. Theelastic anisotropy of crystals. InAnisotropy in Single-Crystal RefractoryCompounds, Volume II. Ed. F.W.Vahldiek and S.A. Mersol. New York:Plenum Press.

27

Gilmore, R. 1999. Industrial ultrasonicimaging/microscopy. Physical Acoustics,Volume XXIV, ed. E. Papadakis. NewYork: Academic Press. 275-346.

Hare, S. 1978. Touching Gold and Silver:500 Years of Hallmarks. Catalogue of anexhibition at Goldsmith’s Hall, FosterLane, London EC2, November 7-30, 1978.London: Worshipful Company ofGoldsmiths.

Jackson, C.J. 1921. English Goldsmiths andTheir Marks. London: Macmillan and Co.Limited.

Lemons, R. and Quate, C. 1979. AcousticMicroscopy. In Physical Acoustics,Volume XIV. ed. W.P. Thurston and R.N.Thurston. New York: Academic Press. 1-92.

Lide, D.R. 2002. Handbook of Chemistryand Physics. 83rd edition. Boca Raton,Florida: CRC Press LLC.

Ouahman, R. et al. 1995. Application de lamicroscopie acoustique à l’étude desproduits de corrosion du ferarchéologique. In Metal 95: Proceedingsof the International Conference on MetalsConservation, Semur en Auxois, France,September 25-28, 1995. 50-54.

Pickford, I., ed. 1989. Jackson’s silver andgold marks of England ,Scotland &Ireland. 3rd edition. Woodbridge, Suffolk,England: Antique Collector’s Club.

Simmons, G. and Wang, H. 1971. SingleCrystal Elastic Constants and CalculatedAggregate Properties: a Handbook.Cambridge, MA: MIT Press. 85-86.

Smith, E.A. 1978. Working in PreciousMetals. London: N.A.G. Press Ltd.

Stavroudis, C. 1989. Possible applicationsof acoustic emission and scanningacoustic microscopy to the field of artconservation: phase I- preliminary report- final draft. Malibu, California: GettyConservation Institute. 43-61.

Treptow, R.S. 1978. Handbook of Methodsfor the Restoration of Obliterated SerialNumbers. NSAS Contractor report CR-135322. Cleveland, Ohio: NASA.

Wilkes, W. 2001. Personal communication.Wynyard R.T. Wilkes F.S.A. Scot, SilverMerchant, 165-169 Portobello Road,London W11 2DY England.

Wyler, S.B. 1937. The Book of Old Silver,English, American, Foreign: with allAvailable Hallmarks, Including SheffieldPlate Marks. New York: Crown.