This project successfully achieved its three goals: (1) to identify more exactly the nature of the materials used to create the imagery on copper; (2) to identify the artistic processes used to create the imagery; and (3) to develop a systematic, integrated set of digital photographic techniques for effectively recovering, enhancing, and displaying the art works on Hopewellian copper artifacts, as one approach to preserving the images and guiding their conservation. A team of eleven researchers, with specialties in archaeology, remote sensory systems, digital image photography and enhancement, applied metallurgy, mineralogy, petrology, paleoethnobotany, and prehistoric textile analysis carried out the project work.
Ultrahigh resolution (3360 x 2253 pixel) color digital photographs of 219 sides of copper breastplates, celts, and headplates, and near-infrared and midrange-infrared digital images of 263 sides of these kinds of copper artifacts, were captured with three different sensor systems. The sensor systems included a Leaf Lumina color digital camera, a Cohu 4810 near-infrared camera with sensing ranges of 0.7 to 1.0 microns, and a Hamamatsu C1000-03 midrange infrared camera with sensing ranges of 1.0 to 1.8 microns.
The mineralogy of all inorganic materials on 44 sides of copper artifacts were inventoried and found to include cuprite, malachite, azurite, chrysocolla, turquoise, hematite, and hydroxyapatite (bone), some of these materials in multiple morphological variants. Soil, mica, mother-of-pearl, calcite or aragonite, a powered bone and calcite pigment, a powdered talc/serpentine pigment, and ash were also identified. Many surface characteristics of the powdered bone and calcite pigment showed that it had been painted on at least one artifact. Most examples of the copper-based minerals were found to have the surface characteristics, diversity, and contextual association with organics typical of copper patinas, like those produced by contemporary copper artists and jewelry makers. A major finding of this project was that the artistic compositions found on Hopewellian breastplates, celts, and headplates were made commonly by chemically inducing patinas on these artifacts.
The organic surface materials on both sides of 77 copper artifacts were inventoried and found to include textiles composed of Group I plant fibers (herbaceous plants like Indian hemp and milkweed), leather and hide, feathers, fur, bark, carbonized wood, uncarbonized seeds, bast, monocot stems, cut-up plant stems, a plant-fiber plaster of a kind, and unknown organics. Organics were observed much more commonly on the copper artifacts than previously thought. Their presence can be attributed in part to their use in making art works on the copper through their arrangements as collages and their role in creating copper patinas, specifically as a means for holding corrosive acids or acid-salt mixtures on the copper surfaces in prescribed areas of the artistic compositions. The occurrence of some kinds of organic materials on the artifacts, in a limited number of cases, can possibly be attributed to their having been placed on the copper artifacts during decommissioning rituals, probably to the artifacts having been wrapped in textiles prior to burial, and possibly to the artifacts having been placed next to clothing comprised of hide, fur, and/or feather.
Textiles of four different weaves, including oblique interlacing, spaced 2-strand twining, alternate pair twining, and spaced alternate pair twining, as well as cordage were found on about half (68 of 132) of the surveyed sides of copper artifacts. Their occurrence on the copper artifacts can be explained by the same five processes cited above for organic materials in general.
Quantitative modeling of corrosion processes using Pourbaix thermodynamic quantitative models confirmed that the great variety of copper corrosion minerals on the artifacts cannot be attributed to natural processes of corrosion of copper in the ground, but can be explained by patination and/or painting with corrosion pigments. The mineralogical studies showed that painting was very infrequently done, leaving patination as the primary explanation.
Experimental replication of the Hopewellian art works on copper using contemporary copper patination methods, along with salts and acids easily available to Native Americans of the Eastern Woodlands, confirmed that all of the observed kinds of corrosion products and their patterning as images on the Hopewellian objects can be explained by artistic patination. Given what is known about the distribution and kinds of organic materials on the artifacts, most of the patinas were probably produced intentionally and directly, while some may have resulted unintentionally from the decoration of the artifacts with cutouts made of textiles, hide, and other moisture-holding organics, and with intentionally shaped arrangements of fine plant materials.
Microscopic examination of the broken edges of 20 of 49 broken copper breastplates showed that nearly all the breaks were well-covered with cuprite and malachite, and that the breaks are ancient and were made prior to burial. This finding supports the idea that, upon being decommissioned, some breastplates, like may other kinds of artifacts in other media, were broken into pieces resembling animal and human forms of the kinds found in the artistic compositions created on them and in other formal Hopewellian art.
Two methods for capturing digital images were innovated during the project. Curved objects were photographed multiple times from different, overlapping, controlled vantages and joined together in Adobe Photoshop to create continuous, minimally distorted, flat layouts of them. Photographic lighting oriented at prescribed angles allowed an optimal documentation of the surface relief of objects as well as their colors.
Methods of digital image enhancement and the five captured spectra were evaluated for their capabilities in discriminating among different kinds of surface materials and in discerning art works made of those materials. Improvements in the photographic definition of the artworks were made by two kinds of contrast enhancements, both of which were found productive and applied to all of the color images captured: a total histogram stretch, in which the histogram of all three bands of the RGB image were stretched at once, and an individual color band histogram stretch, in which the histogram of each band was stretched separately. Histogram equalization was seldom found helpful in resolving the art works. The contrast and clarity of all near- infrared and midrange-infrared digital images were improved with a sharpening filter in Adobe Photoshop followed by a histogram stretch–the experimentally determined, optimal order of operations. Other image enhancement methods that were explored and found fruitful were seven different kinds of calculations among stretched bands of different color (R x B, R x B-inverse, R-inverse x B, R-inverse x B-inverse, B x G, B- inverse x G-inverse, and B – G), formed in Adobe Photoshop, and the creation of hybrid infrared and color images, formed in the GIS program, IDRISI. The latter required modifying the resolution, resizing, and registering of the color, near-infrared, and midrange-infrared bands so as to coordinate with each other.
Qualitative evaluations of the above image enhancement procedures showed which specific band calculations are more commonly effective and which less in revealing the art works on the copper artifacts, although all seven band calculations were helpful in some material context, considering the entire corpus of artifacts. In addition, the Red and NIR bands and combinations of these with other colors generally seemed to provide the art works their greatest visibility. Different hybrid combinations of color, near-infrared, and midrange-infrared bands and color band calculations were found necessary to maximally reveal the different art works made of different materials, and sometimes the different features comprised of different materials within an art work, as might be expected. Specific material and image conditions were also found to determine the effectiveness of using supervised versus unsupervised clustering methods of image enhancement, cluster analytic versus palette boundary redefinition methods of image enhancement, and palette blending versus a palette with crisp boundaries among colors to enhance images. Principle components methods of redefining spectral dimensions of an image were not found useful in enhancing the visibility of the art works compared to using color or infrared bands or band calculations selected for their directly observable, individual effectiveness in enhancing features of an art work. The use of cool, neutral colors rather than primary or secondary colors to display images was found helpful in displaying art works.
A canonical discriminant function analysis of the five color, near-infrared, and midrange-infrared spectral bands of 52 kinds of materials found on Hopewellian copper artifacts indicated the green and midrange infrared bands to encompass most of the variability in image brightness that discriminated among the kinds of materials and to be essential to digitally enhancing and defining art works. The red and near-infrared bands were also found important in these regards, as found in the qualitative evaluations of enhancement procedures.
Together, the qualitative and quantitative studies provide clear guidelines for enhancing digital color and infrared photographs of Hopewell copper artifacts so that art works on them can be maximally discerned.