For nearly three years our team worked inside the Golden Temple in Amritsar on the first scientific conservation of its wall paintings. This is an account of what we found, how we treated it, and what working in a living shrine taught us.
At a glance
- Until 2013, deteriorating paintings at Sri Harmandar Sahib ji were repainted, not conserved.
- The paintings visible today are overpaint on top of the original mohrakashi frescoes.
- We grouped the damage into three surface conditions, and each one set the treatment for its area.
- Most cleaning was done with a 1–2% solution of triethanolamine; the old varnish needed an acetone-based gel.
- In the busiest areas we worked in a forty-minute window at night.
- The greatest risks to the paintings come from devotion itself: wet cleaning during seva, wet garlands at Gurpurabs, and festival crowds.
Being trusted with the art of an active religious site brings gratitude and anxiety together. Gratitude comes from the faith many of us hold for the place. Anxiety comes from the responsibility: millions of devotees are emotionally attached to every surface of it, and we dealt with those feelings throughout the project. There is also always a religious body involved. Our intentions and those of the Shiromani Gurdwara Parbandhak Committee (SGPC) led to the same goal, but at first each side had to learn the other’s approach to it.
How the paintings were made
Sri Harmandar Sahib ji is the high point of Sikh architecture, and nearly all of its walls and ceilings carry painted work. The style is known as the Sikh school of art. Its most prominent pattern category, dehin, is an imaginative study of idealised forms; it borrows extensively from Hindu and Muslim traditions but depicts them in its own way and in a new setting, as Dr Madanjit Kaur has written.
The paintings were first executed in mohrakashi, a true fresco technique. Red-baked bricks were laid in sand-lime mortar. Plaster of slaked lime and sand was driven into the joints and beaten flat with a long wooden strip called a garmala until it was ready for the intonaco, the fine plaster over the coarser arriccio. When the surface became slightly dry and sticky, it was burnished with agate and the drawing was cast.
What visitors see today is not those frescoes. It is later paint laid over them.
A long history of repainting
On our first visit the earlier retouching was obvious. Backgrounds had been repainted to cover their dirty white look, and flowers, birds and other motifs had been overpainted wherever they were flaking. On the parikrama walls most of this was in water-based colour; the backgrounds of the ceilings and arches carried a thick coat of a synthetic medium.
The earlier frescoes were by Bhai Gyan Singh, created meticulously in true fresco. When they began to deteriorate, the people in charge did not know how frescoes could be restored, so artists of the calibre of Bhai Atma Singh were commissioned to overpaint the walls. Over the following decades, various people retouched his work to hide flaking paint. All of it was done in good faith.
Up until 2013 the concept of treating old deteriorating paintings was not known at Sri Harmandar Saheb Ji.
Three kinds of damage
We grouped the paintings by the condition of their surface, because each condition called for a different treatment.
Behind glass, unvarnished
Most of the wall panels. They were cleaner, but flaking in many places, with fine craquelure caused mainly by moisture trapped behind the glass. Many had been retouched where paint was lost. The north-east walls had the most damage and the most intervention; the south-east walls the least.
Varnished, unglazed
The arches of the window niches. The varnish had yellowed with age and turned brittle, cracking and lifting the paint. It had to come off without breaking the hard flakes it held.
Unprotected, under grime
Mainly the painted arches, open to weather and to touch. Beneath thick grime they had been overpainted in several media of different solubility, and very little original paint remained. The only way to save the design was to trace its patterns, follow the surviving particles of original pigment, and redo the work.
Microscopy confirmed the earlier retouching and showed the under-paintings. Chemical tests identified ten pigments, most of them earth colours.
How we treated them
The site set its own limits on materials. Everything had to be non-toxic, non-flammable and free of strong odour, and we avoided volatile solvents and heat wherever we could.
Flaking areas were secured first, with facings of Japanese tissue or a spray of cyclododecane. Because the paint was mixed media, no single cleaner worked everywhere. An ethanol and water mixture lifted lighter grime. Most panels needed triethanolamine, at 1–2% on the walls and stronger on the ceilings. Acetone alone evaporated before it could soften the old varnish on the window arches, so we made an acetone-based gel that stayed in contact long enough to work. Ion-exchange resin removed thick inorganic deposits.
Lifting paint was fixed with Paraloid B-72, applied by syringe or brush; calcium hydroxide nanoparticles worked in some places but not in others. Losses and voids were filled and grouted with a conservation-grade lime mortar that carries no efflorescing salts, does not harm the colours and can be removed if needed.
We followed the principle of minimum intervention. On most panels we in-painted only the lost areas and never painted over original paint. In the window arches, pillars and staircases, where loss was extensive, we agreed with the officials to retouch the backgrounds so that a patchy surface would not disturb the ambience of the shrine.
For protection, the window arches received Paraloid B-72 and the staircases Fluoline HY. The staircases are narrow and always crowded, so people will touch the walls, and Fluoline HY resists graffiti, water and oil. New mouldings and glass went back over most panels, but that work was outside our scope.
Working inside a living shrine
The SGPC’s sevadaars manage the flow of pilgrims, and they kept distractions away from us, which were enormous at the start. Our harder task was explaining the difference between renovation, restoration and conservation, and between conservation and the cleaning and repairs done as housekeeping. Harder still was making the case for analysis before treatment, such as holographic speckle interferometry, infrared thermography, FTIR and XRF.
A few months in, we agreed to work in traditional clothes. We kept our heads covered, kept our work from disturbing the prayers and those sitting in meditation, and followed maryada on cleanliness, conduct and vegetarianism. Where the sangat is present all day, we were given forty minutes at night; one corner could not be completed in that window. The SGPC’s own workshops for woodwork and electricals helped us whenever we needed them.
Devotion itself is the paintings’ greatest risk. We saw devotees doing seva by wiping paintings with wet cloths. At every Gurpurab, wet garlands and bouquets are fixed to painted walls and ceilings. Festival crowds raise a structural question too: the building, ground floor included, stands on columns rather than solid ground, and we recommended a study of the load its vertical and horizontal elements can bear.
The project in the record
The contract was won by Drona, a Gurgaon-based company, which brought our studio in to carry out the work.
- Until 2013: deteriorating paintings at the temple are repainted rather than treated.
- October 2014: nine months into the work, Namita Jaspal presents it at the Dilbir Foundation’s seminar on legacy conservation at the Bhai Vir Singh Museum (The Tribune).
- May 2017: we publish our first account of the project.
- June 2020: we publish the full case study of the wall paintings.
In the press
- The Tribune reported a team of seven conservators on site, the gold on the parikrama walls and roofs restored, and the arches still to do (“Golden Temple repair work nears completion”).
- The Tribune reported the first floor almost 90% complete and night work beginning on the gold inside the sanctum (“Restoration begins at Golden Temple’s sanctum sanctorum”).
- SikhNet profiled our team of seven women from different parts of India and different faiths (“Women Take Charge in Preserving Sikh Art at Golden Temple”).
Questions we are still asking
- How much load can the structure bear when festival crowds fill it?
- What glazing, sealing and moisture control should protect the panels that went back behind glass?
- Very little has been published about the artwork of Sri Harmandar Sahib ji. How can it be documented?
- How rightly is maryada understood, and do young devotees need better education in it?
Religious sites keep history alive in their rituals as well as their structures. Preserving their art links the culture’s past to its future, and it needs devotees, the religious body, business associations and conservators working together. This project taught us that at a site of this importance we have to study beyond the current theory of conservation.
We are grateful to every member of the teams who worked with us at the temple, to the staff of the SGPC, and to Ms Harsimrat Kaur, Minister of State, for initiating and supporting the project.