The Nagara temple form

Put a Dravidian vimana and a Nagara shikhara side by side and they look like two solutions to the same problem, which they are. Both mark the sanctum from above. Both are taller than anything else in the compound. But they are built on opposite principles, and the difference is not stylistic — it is structural, and it changes what you are supposed to look at.

Two ways to make a tower

The Dravidian vimana is a stack. Each storey is a real storey: a wall, a cornice, a smaller wall above it, repeated until you reach the cap. You can read the construction from outside because the construction is the form.

The Nagara shikhara is a curve. Its profile rises almost vertically and then bends inward toward the top, and the surface is covered with vertical bands of carving that follow the curve. Nothing in that profile tells you how it is built. And the reason is that the shikhara is not built the way it looks: it is a hollow shell sitting over the sanctum, and the load is carried by the thick inner walls of the shrine below. The tower is, structurally speaking, a roof.

This has a consequence that is easy to miss and worth stating plainly. In the south, adding a storey to the vimana adds structure. In the north, adding height to the shikhara adds weight to a shell. That is why the northern tower is more impressive at a distance and more fragile in an earthquake, and why so many of them are reconstructions.

Lingaraja: the type specimen

The temple of Lingaraja at Bhubaneswar, built in the eleventh century, is the clearest surviving example of the Odishan version, and it is worth using as the reference because it is intact. The deul rises about 55 metres, and the curve is strong: it goes up almost straight for two thirds of its height and then turns in sharply to the amalaka, the ribbed disc that caps it, and above that the kalasha, the finial.

The vertical bands on the surface are not ornament applied to a wall. They are the wall, divided. Each band is a projection with a carved face, and between them the recesses are in shadow, so the whole tower reads as a bundle of vertical elements rather than a single mass. This is the Nagara trick: the tower looks lighter than it is, and the carving is what makes it look light.

Around the base, the wall is divided into horizontal registers — mouldings, friezes, niches with figures — and the density increases as you go up. The same rule as everywhere else in this pillar: the carving collects where the structure is carrying least.

Khajuraho: the tower as a mountain

The temples at Khajuraho, built between about 950 and 1050 under the Chandela kings, take the same form and push it into a cluster. Kandariya Mahadeva, the largest, is about 31 metres high, and its main shikhara is surrounded by 84 smaller spires that step down toward the corners of the platform.

The effect is deliberate: a mountain with subsidiary peaks. The iconography is explicit — the temple is Meru, the axis of the world — but the architectural method is more interesting than the symbolism. Because the subsidiary spires are attached to the main tower and step down in height, they brace it. The cluster is not just a picture of a mountain; it is a buttressing system that happens to look like one.

Khajuraho is also where the erotic sculpture argument gets had, usually badly. The figures occupy the junction bands between the wall planes — the places where the eye needs a transition — and they are a minority of the total carving. Reading the temple as primarily erotic is like reading a newspaper for the advertisements.

Konark: the experiment nobody meant to run

The Sun Temple at Konark, built in the thirteenth century under Narasimhadeva I, was conceived as a chariot: 24 wheels carved with reliefs, about three metres across, drawn by seven horses, with the sanctum as the driver's seat. The 24 wheels are the hours of the day and the seven horses the days of the week, and the whole thing is a calendar you can walk around.

The main tower is gone. It collapsed, probably before the nineteenth century, and what stands today is the assembly hall in front of it, about 30 metres high. This is a loss and also an education. Because the main tower is missing, you can see the section of the surviving hall clearly — the thick inner walls, the hollow space between them and the outer carved skin, the way the roof structure sits on the inner shell rather than the ornamented face.

In other words, Konark shows you what the Nagara tower always was and never displayed: a shell with a carved coat. Every intact shikhara in India hides this. Konark lost the argument and thereby proves it.

North and south, in one sentence

The Dravidian temple puts its construction on the outside and its politics on the inside — the enclosures that hold the land, the staff and the revenue. The Nagara temple puts its politics on the outside, in a tower that announces the patron's reach, and its construction on the inside, where nobody looks.

Neither is more sophisticated. They are answers to different questions: how do you house a god in a place where land is scarce and kings compete on gates, versus how do you house a god in a place where the tower itself is the claim. Once you see that, the two forms stop being regional styles and become two arguments about what a sacred building is for.

How to read one

Stand well back first — 50 metres at least — and look at the profile as a silhouette. If the curve turns in at the top, you are in Nagara territory; if it steps in, you are in Dravidian. Then walk to the base and find the transition between the plain mouldings and the carved registers: that height is where the wall stops carrying and starts performing.

At Konark, do not go straight to the wheels. Walk around the surviving hall and find the joint where the collapsed tower once met it. That scar is the most informative thing at the site, and almost nobody photographs it.

Entry 05 of Architecture. The southern counterpart is entry 04.