Mamallapuram – Mahishasuramardini Mandapa

Mamallapuram – The Workshop of the Pallavas

Mahishasuramardini Mandapa

This cave temple was known as Yamapuri Mandapa by the local as told by Kavali Lakshmiah when he visited the town in 1803. However soon this name went into oblivion and now this excavation is popularly known as Mahishasuramardini Mandapa, due to the presence of an exquisite bas-relief panel depicting the goddess in motion slaying demon Mahishasura.

This east facing cave temple measures 32 feet long, 15 feet deep and 12.5 feet high1. The cave temple is excavated on a rock crop above which is constructed a structural temple. The cornice of the cave shows roughly cut chaitya arches (dormer windows) and mini shrines above. The front façade is supported on four pillars and two pilasters. One of the pillar was broken and replaced with cut-stone pillar as we see today. Similarly its adjacent pillar’s top portion, abacus, was also broken and removed. Srinivasan2 suggests that these damages might be done during the Vaishnava resurgence when they tried to make a wide gap to bring the inner shrine facing directly to a viewer.

Photograph taken by a photographer of the Archaeological Survey of India around 1900-01 (image courtesy – British Library)
Mahishasuramardini Mandapa (image courtesy –

The pillars resemble to those found in the inner row of the Koneri Mandapa. The capital of these pillars is a bulbous cushion piece similar to those in the Koneri Mandapa however there is an additional abacus member found here. This suggests, from architectural perspective, this excavation is a further advancement on to top of the pillar styles of the Koneri Mandapa.

Plate 20 from James Fergusson’s ‘Illustrations of the Rock Cut Temples of India’, by Thomas Colman Dibdin 1839 (image courtesy – British Library)
Lion and Vyala Pillars (image courtesy –

Once you enter the cave hall, you are welcome with the large relief panels on the lateral walls, which have made this cave very famous among the visitors and scholars alike. These large scale reliefs are considered among the masterpieces of the Pallava art by various scholars. Longhurst3 writes, ‘The visitor to Mamallapuram will be struck by the artistic merit, originality of treatment and power of execution displayed in most of the sculptures, particularly with regard to these tableaux of Vishnu and Durga‘.

Photograph taken by Nicholas and Co in 1880 (image courtesy – British Library)

This panel, showing Durga battling with Mahishasura, is famous for its exquisite carving, grace, beauty and virility. Durga is shown riding a prancing lion and holding her bow stretched and ready to attack. She is shown with eight hands, holding a bow, sword, bell, discus, knife, pasha (noose) and conch. She is followed by army of nine soldiers, eight dwarf ganas and one female. The female figure is shown carrying a sword, ready to attack, and is identified as yogini Jaya by Srinivasan4. Her eight ganas are also armed with sword and bows, except two who are shown holding a plate of offerings and a parasol.

Durga Panle (image courtesy –

Mahishasura is also shown with his retreating army. His army, defeated and down in moral, is shown very vividly, one soldier is shown falling cut in half, few soldiers are hiding behind the bulky Mahishasura and few are already tasted the dust of the battlefield. Mahishasura, shown with a large body, is trying to hold his ground and holding his head club. His attitude suggests that he has already suffered much at the hands of Durga and her army and now only the final blow is in waiting.

Vishnu Anantashayi (image courtesy –

The another relief panel shows Vishnu, lying over Adi-shesha, in his yoganidra-murti5. He is shown with two arms. Near his feet are shown Madhu and Kaitabha. Srinivasan5 mentions that in this relief, we do not see flames coming out of the hood of Adi-shesha however Nagaswamy6 disagrees stating that the blazing flames emitted from Shesha can be seen on the back wall. The attitude of the two demons also suggests that they are perturbed by the flames and thus bending back to avoid the flames.

Vishnu (image courtesy –

Below Vishnu is shown Bhu-devi, stead near his feet, with her hands joined in adoration. Opposite to her are shown two male figures, who could be identified with Sudarshana (discus) and Nandaka (sword), two weapons of Vishnu in their animated form. The figure of Nandaka, identified by Srinivasan, as Nagaswamy7 identifies it with Garuda and Lockwood8 with lotus. Gopinath Rao9 identified the two figures below Vishnu as Markandeya and Bhrigu, however this seems incorrect in this context.

Vishnu Panel (image courtesy –

Two flying figures, above Vishnu, also represent his other two weapons, the male is Panchajanya (conch) and the female is Kaumodaki (club). Beck10 suggests that the flying female figure may be identified with Yoga-nidra, a form Shakti assumed and entered into Vishnu.

Lion Pillars

On the back wall of the cave, three shrines have been excavated. The central shrine is provided due importance by projecting it out into the hall. This shrine is also provided with its own madapa, supported on two seated-lion pillars and two seated-vyala pilasters. The lions of the front pillars are real lion like images while the vyalas on the back pilasters have protruding horns from their heads. This cave is a good example showcasing the architectural evolution in the pillar styles where the phase of slender cylindrical pillars evolved into the seated-lion base pillars carrying similar cylindrical shaft above its head. The pillars with seated lion base have been assigned to the period of the Pallava king Narasimhavarman I by many scholars.

As we have seen in a previous chapter, that study of the form of dvarpalas helps understanding the nature of the temple. This cave temple is a very interesting study in this aspect. It was Gift Siromoney and Michael Lockwood11 who first brought this anomaly into public view. Let us follow them and understand the issue.

Dvarpalas of the leftmost shrine

The first cave, on viewer’s left, has its set of dvarpalas carved on the either sides. The dvarpala on the left has trishula prongs protruding behind his headdress and his one hand is resting on a club. The dvarpala on the right has a protruding axe-blade on his headdress and he points one finger towards the shrine. These two dvarpalas suggest that this shrine would have been dedicated to Shiva.

Dvarpalas of the rightmost shrine

Dvarpalas of the rightmost shrine do not have trishula or axe-blade attributes. They are both shown in similar attitude and wear long dhoti coming till their ankles. They both are wearing the sacred thread, one on his left shoulder and one on his right shoulders. As they do not show Shiva or Vishnu association therefore it may be concluded that this shrine was dedicated to a non-Shaiva deity, probably Brahma.

Central Shrine (image courtesy –

Now left shrine dedicated to Shiva and the right shrine dedicated to non-Shaiva deity, with relief panels of Vishnu and Durga, it is evident that the central shrine should be dedicated to Vishnu. However, we find a large Somaskanda panel inside the central shrine! How the dedication of the shrine got changed from Vishnu to Shiva and why? Let us have a look at the dvarpalas of this shrine.

The dvarpalas of the central shrine show very much the Shiva attributes, one have protruding trishula prongs and other one has an axe-blade, both are holding clubs and points a finger to the shrine. However looking at the niche in which these dvarpalas are carved, Lockwood tells that it appears that these niches were designed to house dvarpalas without clubs, as the case with Vaishnava dvarpalas. The clubs of these dvarpalas are carved where a pilaster should be, thus suggesting that it was an afterthought.

Somaskanda Panel

On the back wall of the central shrine, a Somaskanda panel is carved out. This is a unique panel as the features we find here are not found or rarely seen on other similar panels. A figures of Shiva, Parvati and baby Skanda is shown seated on a lion throne, under which is shown Nandi seated. Shiva is shown seated in a sukhasana posture. Pendent legs of Shiva and Parvati are resting on the back of Nandi. Baby Skanda is seated in the lap of Parvati. Behind Nandi is shown a devotee, whom Srinivasan12 identifies with Chandikeshvara, however this is not agreed by Lockwood and other scholars. Behind Shiva are shown Brahma and Vishnu, both standing. Above them is shown Surya.

Lockwood and Siromoney also point to the different of style in the Somaskanda panel and the relief panels of Vishnu and Durga. He compared this Somaskanda panel with that of at Dharmaraja Ratha and concludes that the Somaskanda panel find in this cave does not match with that of the ratha however it matches with several ones in the Kailasanatha Temple, Kancheepuram.

Thus, Lockwood and Siromoney concludes that the Somaskanda panel is a later work in comparison to the reliefs of Vishnu and Durga. And this Somaskanda panel would be done around the period in which Kailasanatha temple was built.

They suggest a possibility that originally this central shrine was designed to be dedicated to Vishnu with corresponding dvarpalas. Somehow the work was not finished in time, and when Shiavism was in ascendancy, the nature of the original design got changed resulting central shrine being dedicated to Shiva. This ides of central shrine dedicated to Vishnu is also supported by Dehejia13.

Srinivasan14 suggests that the three cells were meant to house three out of five forms of Shiva and the central cell could only be completed during the reign of Parameshvaravarman I. However, from the concluding remarks of Lockwood and Siromoney, it appears that the cave was originally designed to be dedicated to Hindu Trinity, with Vishnu shrine in the center, and it later got changed with Shiva taking place of the central shrine. This may have happened during the reign of Parameshvaravarman I, who in his own inscriptions, is mentioned as staunch devotee of Shiva.

From the above discussion, we have seen that this cave temple have undergone vandalism twice in its lifespan. The first vandalism, most probably during the period of the Pallava king Parameshvaravarman I, was when its Vishnu shrine was converted into a Shiva shrine. It would have happened during the last quarter of the seventh century CE.

The next act of vandalism was during the Vijayanagara period of fourteenth-fifteenth century CE, when its Shiva character was changed to Vaishnava. Vaishnava symbols of conch and discus were carved on its pilasters. The two middle front pillars are removed or damaged and a cavity was carved into the porch of the central shrine to accommodate an image.

Back to Index


1 Longhurst, A H. Pallava Architecture part II
2 Srinivasan, K R. Cave-Temples of the Pallavas. p 150
3 Longhurst, A H. Pallava Architecture part II
4 Srinivasan, K R. Cave-Temples of the Pallavas. p 154
5 Srinivasan, K R. Cave-Temples of the Pallavas. p 155
6 Nagaswamy, R. Mahabalipuram. p 33
7 Nagaswamy, R. Mahabalipuram. p 34
8 Lockwood, Michael. Mamallapuram. p 88
9 Gopinath Rao, T A. Elements of Hindu Iconography part 1. pp109-110
10 Beck, Elisabeth. Pallava Rock Architecture & Sculpture. p 131
12 Srinivasan, K R. Cave-Temples of the Pallavas. p 153
13 Dehejia & David. Addition, Erasure and Adaptation: Interventions in the Rock-cut Monuments of Mamallapuram. Archives of Asian Art vol. 60 (2010), p 6
14 Srinivasan, K R. Cave-Temples of the Pallavas. p 156