Every account of kath-kuni begins with the same sentence, and it is always the wrong sentence: the houses survived the 1905 earthquake. They did, but stating it that way makes the technique sound like a talisman. The useful question is narrower. The earthquake did not test whether these walls were strong. It tested whether they were allowed to move, and where.
A wall made of layers that are not bonded
A kath-kuni wall is not a stone wall with some wood in it. It is a stack of alternating courses — dressed stone, then deodar beams, then stone again, then beams — with no mortar anywhere and no iron fastener anywhere. The beams run in pairs along the wall, one inside and one outside, with rubble and small stones packed into the cavity between them, which brings the finished wall to somewhere between forty-five and sixty centimetres.
At the corners the beams cross and lock, and this is where the name comes from: kath, wood, and kuni, corner. Two joints do the work. The kadil is a wooden dowel; the maanwi is a dovetail. Both are cut so that the members can rotate slightly against each other without separating. Nothing in the assembly is rigid, and nothing in it is glued.
That last point is the whole design. A mortar joint is a promise that two stones will behave as one piece. A kath-kuni joint is an admission that they will not, and a decision about where they will be permitted to disagree.
What the survey officers actually saw
After April 1905, officers of the Geological Survey walking the Kullu valley wrote down something that reads oddly in an engineering report. The houses looked, they said, top-heavy; the construction seemed wrong at first sight. Then they noticed what they called the natural resisting power of the timber-bonded walls, and revised the assessment.
Top-heavy is the right first impression and the wrong conclusion. The walls are thick, the roof is heavy slate, and the upper floors cantilever out over the street to catch the sun — a silhouette that promises to topple. But the weight is doing two jobs. The slate roof holds the wall down and stops the courses spreading; the mass of stone sits low, in a ground floor given over to cattle. The building is not light. It is low, and it is loose.
The foundation is part of the same argument. Rather than digging to something solid, builders cut a trench and fill it with loose stone. A foundation like that cannot resist a shock, and is not asked to. It settles, re-arranges itself, and passes a smaller, slower movement up into a wall that is already designed to absorb it.
How high it goes
The technique is usually illustrated with houses, which understates it. The same valleys have tower temples and multi-storey granaries — the shrine towers at Chehni and the Bhimakali complex at Sarahan, the tall kothi structures — built to a dozen levels or more on the same system: stone courses, timber bands, no mortar anywhere.
Height is where the argument gets tested, because the overturning moment at the base grows faster than the wall's capacity to resist it. A bonded pier under that load develops one long vertical crack and separates. A stack of unbonded layers cannot do that; it can only spread outward, and the timber bands plus the weight of the slate roof are exactly what prevents the spread. The same three elements that let a house sway let a tower stand.
That is the real evidence that this is a system and not a habit. A habit stops working when you change the scale. A system tells you what to add when you do.
The parts that are not about earthquakes
It would be tidy if every element of the form had a seismic explanation. Most of them do not, and the technique is more interesting for it. The double wall with rubble in the cavity is a thermal device: stone takes the sun in the afternoon, timber slows the transfer, and the interior holds its temperature through a night that drops below freezing. The small openings with heavy timber frames are structurally useful in a quake, but they are also what you build when the wind carries snow and the fuel budget is a forest you have to walk to.
The cattle on the ground floor are the clearest case of a form with several reasons. Livestock under the living room is normal across the western Himalaya, and the animal heat rising through the floor is part of how these houses are warmed. It also puts the heaviest occupancy at the bottom, which is what a wall of this kind wants. Nobody designed it for the earthquake. It happens to be good for the earthquake.
This matters for how the technique is described now. Presented as earthquake-proof heritage, kath-kuni becomes a curiosity to be photographed and occasionally imitated in a resort. Presented as what it is — a construction system that solved cold, slope, snow load, cattle and fault movement with the same set of moves — it becomes a set of decisions that can still be argued with.
Why it is being replaced anyway
Concrete is cheaper per square metre, needs no skilled mason, and can be ordered. Deodar cannot. The trees are protected, the seasoned timber is expensive, and the knowledge lives in the hands of a mason — a tavu — who learned it over years and is now often the only one in a valley who did.
There is also a subtler reason, and it is the one worth carrying away. Kath-kuni is a maintenance architecture. The timber courses need replacing on a cycle of decades; the slate needs re-laying; the whole assembly is designed to be taken apart and put back. A concrete house asks for nothing for thirty years and then asks for everything. To a household deciding how to spend a small sum now, that deferral is not irrational. It is the same trade every tradition of repairable building has lost in the last seventy years.
The houses that survived 1905 were not chosen. They were simply the ones whose walls had somewhere to move. That is a thin margin to build a philosophy on, and it is also exactly the margin that a crack in a bonded wall does not have.
Sources and checks
- Middlemiss, C. S. The Kangra Earthquake of 4th April 1905. Memoirs of the Geological Survey of India, Vol. XXXVIII, 1910
- Geological Survey of India Seismic zoning of India — Zone IV and Zone V coverage of the western Himalaya
- Bhushan, Rahul / NORTH Documentation and rebuilding work in kath-kuni, Naggar, Kullu district
- Himachal Pradesh Tourism Development Corporation Naggar Castle — conservation record and monument notes, Kullu
- Archaeological Survey of India Bhimakali temple complex, Sarahan, and the tower temples of the Sutlej valley
- Indian Standards IS 1893 — criteria for earthquake-resistant design, and the timber-reinforced masonry literature that cites this technique
Last checked 1 November 2026 Yellowsea Huang Corrections