After a year of extensive conservation treatment in the studio, the Nzilani team has begun reinstallation of the Charles Connick New Testament Window.

IMG_7922Empty crates ready to return to the studio

The 302 conserved panels will be reinstalled in the cast concrete tracery of south-facing window #41, utilizing traditional and modern stone-setting techniques.

Comparing the before and after condition of the panels (below), you can tell what a vast difference the conservation has made, which will ensure the window will be enjoyed for generations to come.



OLYMPUS DIGITAL CAMERALancet 2 “St. James” panel pre-and post-conservation in reflected and transmitted light

Those of you who have visited Grace Cathedral during the interim may have noticed the warm hue emanating from the transept. In keeping with previous restorations at the Cathedral, we inserted temporary streaky amber glass “plugs,” which provide weatherproofing while still allowing light to emanate through the church. As you may recall from a previous blog post, creating the plugs wasn’t straightforward.  It took numerous measurements to make the templates, which informed the shape and size to cut the temporary glass.


Temporary Glass “Plugs”

After the plugs were installed, the templates became invaluable during the rebuilding of the non-rectilinear shapes found in the rose windows.

The conservation of this monumental, multi-paneled window afforded us the experience to streamline our treatment process; including: developing optimum releading strategies, usage of slip joints, even organization protocols like a detailed spreadsheet of the inside dimensions of the transport crates (seen in the truck above).


The Old Testament

Immediately after installing the New Testament Window, we will remove and start conservation of the Old Testament Window; following the original Connick installation order. Situated directly across the aisle from each other, the New Testament was installed in 1931 & the Old Testament in 1932.

The Old Testament window (#34 pictured above) has an identical assemblage of five lancets with seven roses above, as the New Testament.

Treating the windows “back-to-back” will offer us the unique experience of putting our acquired hands-on, “muscle memory” and organizational knowledge to use in the conservation of the Old Testament Window.  The systems we have already put in place will facilitate its speedy and efficient conservation.

This is a rare (conservation) opportunity, as the work will allow us to compare and contrast the installation and original building techniques in both windows.  We already have some theories about what the studio did with the second phase, we can’t wait to prove or discount them. As we’ve before, the art conservator wears many “hats” – in this case the art historian will inform the glazier and vice versa – it’s very exciting! We feel honored to be following in a master studio’s footsteps.

We look forward to sharing our discoveries and conservation solutions as we embark on this new stage of work at Grace Cathedral.

















The South Wall of St. Vincent de Paul is currently being installed!

The center rose presents a nice case study of the challenges of installing multiple panels that must work together to create a single continuous image.  The center of the rose has a “Christ the King” figure with rays of yellow and red light emanating outward.  These rays begin in the center panel and must continue “underneath” the metal framework and into the adjacent panels.  If the color rays and the lead lines are misaligned, the image would look disjointed and unharmonious.IMG_9484
At Nzilani we take special care to install the panels such that all the rays and their lead lines align perfectly.  However, as with all installation projects, many factors can thwart our pursuit of perfection.

Which panel goes in first?

The metal frame divides the rose into 9 sections.  The center square panel must be installed first because it provides the “map” for how the other sections are aligned.  The center panel is held on all four sides by metal frame, so there is no room for adjusting the panel later on.  Once the center panel is locked in, we move on to fitting the adjacent panels.

IMG_9200The panel directly below the center panel is held on two sides by metal frame and along the bottom by a curved wood frame.  If the panel is too short or if a slight rotation is needed to align the panel with the one directly above, rubber shims can be placed between the bottom of the panel and the wood frame.

The alignment of the bottom panel was nearly perfect: IMG_9297

We see lead lines and glass that flows smoothly “underneath” the metal bar and putty bevels.  The eye is not tripped up by any glaring misalignment.  The two adjacent panels are also aligned correctly.

The true test is how the panels look illuminated:IMG_9472

There is a slight shift in the red and yellow rays, but this should be negligible when viewed from afar.

Flaws in panel design

The panel above the center panel proved to be much trickier.


There are noticeable points of discontinuity in the junction between the these two panels.  In this case, we are limited in our ability to correct the orientation.  We cannot shift the top panel because there is no extra room between the two metal bars on either side.  Even if that were possible, the alignment would not be improved because the top panel was originally built out of alignment with the center panel.  IMG_9470

Viewed from inside the church, the transitions are not perfect, but the overall flow seems fine. This section shows how important careful panel design is to the finished product. However, there is a noticeable design flaw on the upper left corner (see if you can find it!). The upper tips of two of the rays have been switched during the glazing process. Because panels are built in the studio on a table that is not illuminated; even the most careful glazier may switch a piece with one that is similar in shape.

At Nzilani, we check multiple times during the conservation process to ensure the final outcome is of highest quality. To err is human; to not fix a known error is unprofessional. After this photo was taken, we exchanged the pieces in situ… this is how it looks now.IMG_9578

We often quote the German proverb:

“Vier Augen sehen mehr als zwei” – Four eyes see more than two”


It takes a little more time to do this final review, but it’s worth it. Our work should last a generation; this is the time to leave the window in the best shape (both structurally and visually) we possibly can.


Last week, we completed installation of temporary glass in the 7 roses of the New Testament Window (#41) at Grace Cathedral.  This week we focus on the 5 tall, narrow spires called lancets below the roses. Each lancet is dedicated to an individual saint (Peter, James… John & Thomas), with Jesus in the center.before_after lancet

Unlike the tricky U-channels found in the roses, the large rectangular panels of the lancets are mounted with the simpler T-bar system.  The panel is centered between the two vertical channels and sits on top of the metal bar.  Metal tabs on the top and bottom hold the glass in place.tabs

Of course, not every opening has the same dimensions, so trimming of the glass panels was often necessary.

Perhaps the most difficult section of the entire window to install was “head and shoulders” at the very top of each lancet.  The four sections had to be installed in a particular order and carefully joined with slip joints.


Another interesting feature discovered at Grace Cathedral was the system used to capture and remove condensation on the windows.  Long copper trays were found at the base of each lancet.  The water is drained away with pipes that exit to the exterior:IMG_8730

Meanwhile, Back at the Studio…

With hundreds panels of glass leaving the Cathedral and entering the Nzilani studio, it is extremely important to have a clear system of monitoring where each panel is at any moment in time and what stages of treatment have been completed.  At Nzilani, we have developed a graphic/color-coding system that helps quickly identify the location of each panel.  A visual map helps us keep track whether the panels have been photographed, rubbed, soaked, dismantled, rebuilt, etc.

Lancet maps
Lancet maps
Command Center
Command Center

Rose 2_1
After stained glass panels are removed for treatment, clients have a number of options for how the void is filled during the interim period. Plywood, plexiglass, or glass are among the options.  Glass is the most labor-intensive option because it is the least forgiving material, thus presenting a number of challenges. It also is very appealing in that it comes in an endless variety of colors and textures; and can be used again, in similar configurations in other areas of a building (i.e. used for multiple, successive treatment campaigns).  Grace Cathedral has chosen to fill their space temporarily with textured, streaky amber glass.


The New Testament Window (#41) has 6 “roses” similar to the one in the picture above. Each rose consists of 8 shapes which are repeated.  Though the shapes are consistent for all of the roses, there are minor variations in the actual dimensions of each space.  The most labor-intensive option would be to measure every space and cut unique pieces of glass. This is obviously time (and cost) prohibitive.  Instead, we chose to find a master template for each of the 8 shapes that would be sufficient for all 6 roses. Multiples identical pieces of glass were cut from these master templates.

The Goldilocks Effect

Since each space is different, the glass pieces were either perfect, a little too small, or a little too large.  Take for example shape B, what we call the “stingray”:

Shape B

Notice how we cut two pieces of glass for this shape.  This is because the (stained) glass is set in a traditional stone setting. If you looked at a cross-section of the building, the space in which they are set is a “U” channel, not an “L”. If one piece of glass were cut, it would be impossible to fit into the channel.  To solve this problem, we cut the glass in half and add a piece of lead to connect them.  This is called a “slip joint”.

A Fit that is “Just Right”

For the opening above, we found that the shape B glass was a little too small for the space.  To resolve this, we cut an extra strip of glass and lead to span the distance. This per basis modification takes a little extra time, but it ensures a perfect fit for every opening.  The master template works for the majority of the spaces, however, maximizing efficiency and reducing costs.

Despite the challenges presented by glass fills, the process is fun and exciting… misc_01

… and the results are breathtaking.









From the ground, a visitor to San Francisco’s Grace Cathedral will see beautiful, multi-colored stained glass windows. However, once those windows are removed from their stone niches, the crew at Nzilani see the windows in a much different light, literally.  Take for example the following panel, (the bottom of St. Thomas’s feet, of the New Testament):

Transmitted light


Reflected light


Viewers from the ground see the windows with light transmitted through the glass, while we see the windows with light reflected off the surface.  This reflected light allows us to identify the many structural and cosmetic problems that can develop over many years of weathering.  One such problem can be seen in the lower left corner of this panel:


Detail: linseed oil migration


Notice the amber-colored globs on the surface of the lead.  In the past, fabricators used to hand mix the putty used to seal the windows.  The putty is a mixture of linseed oil and calcium carbonate (chalk powder).  If the mixture ratio is incorrect, (or the windows installed during a particularly hot day) the linseed oil can migrate to the surface over time, resulting in these unsightly globs.  This is a good indicator that the putty is failing and that the window needs conservation treatment.


Detail: water damage


In this image, we see a number of issues.  There is a white, hazy film on the surface of the glass.  When lead oxidizes, it becomes white and powdery.  This film could also be whiting (calcium carbonate powder) used to clean the glass surface when it was originally fabricated.  Additionally, there is a noticeable white drip coming from the top of the right foot. This indicates water damage through a hole in the lead.

This panel also gives us clues as to how the panel was originally installed.  At regular intervals along the surface of the lead, we can see bits of copper wire soldered the to the lead joints.  These wires were twisted around steel, metal rods (called “support bars” or in stained glass parlance “rebar”) attached to the stone, providing structural support for the large panel.


Another interesting discovery was illuminated in the studio.  When the panel is seen head on, something strange is revealed about the design of the feet:

Detail: elongated feet


Notice how the artist exaggerated the length of the toes.  From the ground, the viewer sees the panel at a very steep angle.  This angle makes everything above look compressed.  To compensate for this, the artist intentionally painted the toes much longer than normal.  Such careful thought and ingenuity is seen in nearly all the stained glass windows at Grace Cathedral.

Grace Cathedral Extensively Restoring Stained Glass Windows

We are so excited to announce that our work at Grace Cathedral has made it into the news!

For a sneak peak of what Nzilani will be working on until March 2016, watch the video above, or click the link. We hope to do another in progress news segment to keep everyone up to date, but for now, you can follow us here. We will be posting our process with this project and any new exciting finds.



Nzilani continues to be excited with our new project at Grace Cathedral. We are overjoyed with our spread in the San Francisco Chronicle this week, which shows a sneak peak on site and in the studio.

SF Chronicle 1SF Chronicle 2SF Chronicle 3

About half a year ago; we were contacted by Grace Cathedral in San Francisco. They had received a gift from an anonymous donor who had earmarked the funds to go towards the preservation of the building, particularly the windows. We recommended surveying all 66 art glass windows before we embarked on any work.There are two main types of windows, dalle de verre and traditional stained glass. You can see a page from our report above. It was the perfect way to acquaint ourselves with the current condition of every panel within the church – and what we recommend to every client (regardless of the project size).
Screen Shot 2015-06-15 at 5.19.44 PM


The majority of the stained glass was designed and made by the studios of Charles Connick (1930-1950) or William Willet (mid-1960’s). Not surprisingly, the early Connick windows, though expertly executed, are starting to fail due to their age and positioning in the architecture. With the delivery of the 100+page survey, we decided along with the church that we should start to systematically conserve the windows according to age and condition.

This June, we began the conservation of the New Testament Window. It is a once in a generation opportunity, and one we don’t take lightly.”


It is located on the South side of the Church on California Street. Locals – keep an eye out for us as you drive by – we’ll be on site for a few more weeks. The New Testament Window, or “Window 41” according to the Cathedral’s mapping system, was created by Connick Studios and installed in 1931. 303 panels collectively coalesce into Connick’s grand vision of the seven “wayward” churches of Asia; with Christ the brother and his closest disciples Peter, James, John & Thomas below.



Those not familiar with the stages of conserving stained glass may not know that a conservator working on windows within architecture has to wear more than one hat. Some are figurative (artist, detective, archivist, historian, scientist, to name a few) and others quite literal (hard hat anyone?).

The beginning process of removing the sections of a “window” as large as this entails much planning, project management and coordination. To succeed, one must be in the mode of ensuring every panel is accurately labeled, carefully removed and stabilized for its journey down the scaffolding, into its custom-built crate for its voyage across the bridge to our studio in Oakland.



Before the removal even starts, we photograph each panel on the exterior and interior to ensure we have an accurate record of all the components like the subtle painterly nuances in the borders that these Neo-Gothic windows contain. But, keeping the budget of our client in mind, our on-site time is limited to the safe action of removal. We leave detailed examination of what the windows depict for later.


Once the panels are safely at the studio, we can allow ourselves the luxury of examining each of the 5,000 pieces as they are individually cleaned by hand. Look for future “Grace Cathedral”blog entries that do just that! We look at glass colors, paint styles and techniques (elongation of feet and faces to account for parallax – quick finger painting that “reads” as tiny jewels of light when viewed from 85 feet away), lead came profiles and any anomalies, so that we can accurately conserve the panels exactly as Connick originally created them.

Sometimes we are rewarded for our diligence, a tromp l’oeil in the corner of a depiction of a bible scene; or the holy grail, a signature of the artist himself etched into a border piece.


Charles Connick.. BOSTON 1931 © “



Here’s a sneak peak of the angel at the very top of the window, up close and in detail.
GC41_RS2_1C_Before Conservation_Transmitted




Window 48 was finished last week and we’re well on our way to finishing Window 45. Our process is to start from the top of the window and then work left to right. We are carrying on with a structural change we made on the last round of similar windows (33-36), in order to keep things consistent and structurally sound.

Here is the panel numbering system we have created for the three, arched windows 45-47.SVdP45_Finished Overall

When the long panels (ie. panels 2, 4, 11, and 13) of Windows 45-47 came into the studio, they tended to be flimsier than the smaller panels (ie. 1, 3, 5, etc.).


The image on the left is before conservation, the image on the right is after.

The fragility is due to their length. These panels are approximately 45” long and were originally built with no support bars. These panels were falling apart before we started any treatment. The border lead would detach with any movement, and many of the fillet pieces were cracked. The middle sections of the panels were bowed – the only thing keeping them together was the temporary tape secured around all edges (top to bottom and side to side) upon removal in order to secure them during transport.

At the end of completely re-leading the panels, we added two rebars to every long panel. The rebars follow the stylistic lines of the original design, while making them stronger and more stable in the long run.DSCN2852_copy

Stabilizing bars, often referred to in stained glass conservation as “rebar”,  are used in stained glass as a way to reinforce the stability of the panels. The rebars we are using for St. Vincent de Paul are flat steel bars. The original bars varied in size from 1/8″ to 1/2″ wide and 1/8″ thick.

Rebars are necessary to increase the longevity of a panel, but they can deter from its overall image. A subtle alteration in the placement of the bar can greatly improve the overall affect.”

Panel 6 (pictured below) is part of St. Vincent de Paul’s Window 48’s tracery.This panel originally had a straight rebar across the middle of the angel. When a panel (or window) enters the studio for treatment, the first step in conservation is to photograph the it in its original state.


Afterwards, we remove the rebar in order to create an accurate rubbing. The areas where the rebar was attached to the panel are marked with a red “X” within a red circle. We mark the solder joints so we know how to best recreate the panel’s rebar after it is rebuilt.

The images below depicts Panel 6 after it was built, soldered, puttied and polished. The top photo shows exactly what the original rebar would look like after conservation. However, as mentioned above, the original rebar cuts directly through the image of the angel’s body, creating a disruption in the image. The bottom image is our remedy to this problem. By bending the rebar, we are able to follow the lead lines already in place in the panel and not disrupt the image.


This is the final image of the panel AFTER conservation. Not only is the entire panel cleaner and brighter, but the image has been improved by bending the rebar.

SVdP48_6_After Conservation_Transmitted(small)