We don’t tend to do a lot of object reconstruction here in the NMSC Archeology Lab, but every now and then something comes along that needs just a little bit of extra love and we get to put Humpty Dumpty back together again. It might be a teacup that’s broken in half and is interesting enough for research or exhibit, or it might be significantly easier to date and catalog the artifacts if we can see them as relatively complete vessels. In the case of the bottles from the cellar of the Longfellow House, for example, being able to see the complete vessel forms really helped date and interpret the site.
Before we can start to work with the different pieces, everything must be cleaned and labeled with catalog numbers. Often pieces from the same vessel can be found in different parts of the site, so it’s important that each piece has a distinguishing number.
Most archeologists and conservators love puzzles, and generally the first step to solving a puzzle is to classify the different types of pieces. Is it a corner? An edge? Is the color the same as other pieces? This is the exact same method we use to evaluate fragments of vessels. Pieces that are clearly part of a neck or a rim go to one side, base fragments to another station, and body sherds to a station in between.
In the beginning it can be hard to see how we will ever make sense of it all amidst the thousands of sherds. I have to say that after staring at bits of glass or ceramic for a few hours you get surprisingly good at noticing subtle differences in glass color and thickness, shades of white, unusual features like inclusions and bubble formation patterns, subtle surface striations formed during the bottle-blowing process, or even how a particular glaze spalls or crackles. Even the way different fragments of glass exhibit signs of glass disease provide clues to possible vessel matches.
As matches are found we use small pieces of low-adhesive blue painter’s tape to temporarily keep the “mend” together. Once we exhaust all the possible matches we evaluate whether there’s enough of a vessel left to reconstruct, and whether doing so will make it easier or harder to store.
Some vessels, like the teapot above, may seem pretty close to complete, but are missing key structural pieces. Trying to reconstruct these partial vessels may lead to stress in new areas of the vessel, and new fragmentation. In cases like this, it’s better to photograph the temporarily reconstructed vessel as we’ve done above. We then store the sherds as we would any others, after removing all the blue tape and any trace of adhesive, and then making notes in the catalog records so that future researchers will know that the pieces are associated.
We may also decide that a piece isn’t meaningful enough for a full reconstruction. The Saratoga National Historical Park collection had at least 15 whiteware chamberpots that met our reconstruction criteria of 90+% complete, but many of them were so similar to one another that we would not have gained any new insight from permanent reconstruction. We must also consider safe and efficient storage of the vessel within the collection. A reconstructed vessel needs to be handled a lot more delicately than a whole vessel or sherds, and how it will be stored is very important. If we don’t think it can be stored safely, it won’t be reconstructed. A reconstructed vessel can also take up quite a bit of storage space, which is always at a premium. A complete chamberpot will almost always take up more space than a bag of sherds.
The Canton porcelain serving dish above had already been partially reconstructed, so locating additional pieces was pretty exciting. Mending was easy, since the intact partial reconstruction did not obstruct any of the rediscovered sherds.
This Canton platter was less complete, but we were able to find some of its other pieces amongst a sea of blue and white Canton porcelain sherds. Unlike those 5000 piece puzzles, where you start by creating the outside “frame,” when reconstructing a vessel it’s important to rebuild it from the inside out. In evaluating the sherds it’s possible to see in what order the vessel was broken, and they will need to be mended in reverse order to make certain they fit back together. For this platter in particular, you can tell that it first broke in half, and then each half broke in half, and each quarter broke down further, and so on. To reconstruct it, all those smaller pieces had to be put back into quarter sections, and then the quarters merged to halves, and then the two halves fused back into a mostly complete vessel.
As an aside, porcelain is probably one of the dreamiest materials to mend. The breaks are usually very clean, and it’s easy to line up two sides of a fracture. Sometimes the mend fits back together so tightly that it’s hard to see where there was a break in the first place!
Why go through all this trouble? In some cases it’s really hard to identify what a vessel might have been from fragments alone. In the case of the item above, looking at rim sherds alone may have led us to think it was a bowl, and looking at just the basal or body sherds suggested a flower pot. But once everything came together it was clear we were looking at something pretty unusual. The current best guess is that it’s an insert for a commode; definitely the first of its kind I’ve seen in the lab!
While it’s not practical to mend every single possible vessel, there’s certainly a lot of good reasons to selectively reconstruct particular artifacts. Identification, cataloging, and storage are all considerations from that first “A-ha!” moment when two otherwise dissimilar sherds magically fit together. When we do choose to reconstruct, we further our understanding of a collection, as well as provide the amazing satisfaction of helping something broken for hundreds of years come together again.