Conservation of an Islamic Jug

79.1.169 Before Treatment. The old repairs failed during routine handling. Note the sloppy and discolored glue along the joins and the darker blue fills.

Some of the old repairs on this jug recently came undone when it was picked up. The jug was restored before it came into the collection, so we don’t know what materials were used or how old the repairs are.  It was done in a time when people were less concerned with covering up original material.  Many of the old fills overlapped the glass by as much as 5mm.  Glue from the joins also overlapped the glass.  Overlapping the glass makes it a lot easier to blend in fills and hide breaks, but most conservators these days would consider it unethical and unnecessary.

The jug after all the old repairs were removed.

The old repairs were completely dismantled in acetone and further cleaned under a microscope with a scalpel and a soft brush dipped in acetone to remove all of the old adhesive.  There were a total of 44 fragments and 7 losses in the body and rim in addition to tiny losses along break edges.

Joining fragments under the microscope helps to make sure all joins are properly aligned.

After dismantling the jug, it had to be put back together. This was done with Paraloid B-72, an acrylic glue which is very stable and adheres well to glass.  The microscope was used to make sure every fragment was aligned properly.  Using a microscope is especially important with an object that has as many fragments as this jug did because every tiny misalignment adds up, and can cause the last fragments to not fit properly.  The fragments have to be joined in a certain order to make sure no fragments get “locked out.”

Fills for the missing areas were made using one of our most commonly used techniques.  Plaster intermediary fills were made, molded in silicone, and used to cast colored epoxy fills.  The epoxy fills were then glued in the object with the rest of the fragments.

Losses back with dental wax in preparation for making the plaster fills.

In this case the plaster fills were made with a backing of dental wax which had an adhesive on one side.  The wax was warmed slightly and then pressed against an existing part of the jug with a similar curvature.  After it had cooled the adhesive was exposed and the wax was stuck to the glass.

Making the intermediary plaster fill.

To protect the glass from the adhesive and the plaster, a barrier layer of 30% B-72 in acetone was applied to the glass around the losses first.  When the plaster fills were completed the barrier layer was easily removed with acetone.

To remove the plaster fills, the jug had to be partially dissembled.  The plaster fills were coated with acrylic resin to make them shiny like glass.  They were then molded using a silicone rubber.

The epoxy was colored to match the glass. Blue is one of the most challenging colors to match, especially if it is transparent.  In part this is because of an optical phenomenon called metamerism.

Metamerism (me-TAM-erism) occurs when you see two materials that look like the same color under one type of light, but no longer match under a different type of light.  This happens because of how the different materials absorb and reflect different wavelengths of light.  We perceive the color of an object based on how it interacts (reflects and absorbs) with light.  Simply put, an object that absorbs most red wavelengths and reflects most blue wavelengths will look blue.  But, as you may remember from your kindergarten days, different combinations of colors can end up looking the same.  So two blues may actually reflect slightly different wavelengths and still look the same under some light sources.

The epoxy is colored with dyes and pigments that are pre-mixed with the resin part of the epoxy before the hardener is added.

The light source itself is also a factor, because it determines what wavelengths of light are present.  For example, most indoor tungsten lighting has more reddish wavelengths while sunlight has a full spectrum of colors.  If you have two objects that match in color under daylight but not under tungsten light it is probably because they interact differently with the red wavelengths.  Metamerism can occur with most colors, but blues tend be especially problematic.  Synthetical matches for colbalt blue may look blue under incandescent light, but turn purple under a tungsten light source.

The translucency of the glass and the three-dimensionality of the surface also effect the interaction with light and further complicate the matching process.

The epoxy is colored with dyes which are pre-mixed with the resin part of the epoxy.

Using a mold, the epoxy fills can easily be re-cast if the color doesn't match or if large airbubbles got trapped.

The hardener part of the epoxy is added when the color is matched.

Unfortunately, some of the dyes change color slightly with the addition of the hardener, so there is often a bit of trial and error required to end up with a good fill.

Luckily it is easy to re-cast the epoxy fills with a good mold.

The final castings were re-assembled with the rest of the jug fragments.  The total process took just under 25 hours.

79.1.169 After Treatment

The final result looks pretty good in most light, but looking at the object from different angles and under different light sources make the fills more visible.

Posted by

Astrid van Giffen is the Museum's associate conservator. In 2007, she completed the conservation training program of the Netherlands Institute for Cultural Heritage (ICN) in Amsterdam, with a specialization in glass and ceramics. Her training included internships at the Walters Art Museum in Baltimore, Md,, and The Corning Museum of Glass in Corning. Since completing the ICN program, she’s worked as a private conservator in Oregon and was the Samuel H. Kress Fellow in Objects Conservation at the Straus Center for Conservation and Technical Studies of the Harvard Art Museum (2008-2009). She also holds a BA (2001) in Classical Studies from Willamette University.

4 comments » Write a comment

Leave a Reply