Lord Skanda-Murugan

An Illustration of Iconographic Contact between Kārttikeya and Manjusrī in China

Kārttikeya-Manjusrī (5th century AD)
Kārttikeya-Manjusrī (5th century AD). From Yun-Kang Caves, Shensi Province, China

by B.N. Mukherjee
from Buddhist Iconography, pp. 138-141.

(Superscript numbers refer to end notes.)

An interesting male figure can be noticed by the side of the entrance to one (no.8) of the caves at Yun-Kang (or Yungang) near Datong in the province of Shensi (or Shanxi) in China. The main caves are dated to the period of the Northern Wei Dynasty (c. 385-534 or 550 AD). Artistic activities began here in about 460 AD.1 The male figure in question sits to the front probably on a high stool, resting his feet on a peacock sitting to the right and turning its face towards him. It holds a ball in its mouth. The male figure is five-headed. But it may be considered as six-headed if we concede the possibility of conceiving a head at the back of the main one, which the viewers cannot see in a panel of sculpture in relief. The male figure has six arms. The two uppermost hands display the symbols of Sun and Moon. The middle left hand grasps a bow and the lowest left hand carries a cock held closely to the chest. The lowest right hand, which shows a mudra, rests on the right thigh. The middle right hand is broken; but it seems to have carried an arrow, since the corresponding left hand holds a bow (Fig. 26).2

All these attributes and the mount tend to identify the male figure as an icon of Skanda or Kārttikeya (or Kumara).3 The deity is indicated here as having six heads (five visible and one invisible). Skanda is referred to as having six faces in different texts.4

The appearance of a brahmanical deity in a Buddhist shrine in China in the 5th century AD need not cause surprise. In another panel in the very same cave we can notice a three-headed male seated on a recumbent bull. The latter figure may be considered to represent Mahesvara.5

It may, however, be pointed out that peacock was associated also with bodhisattva Manjusri,6 who, according to Chinese Buddhist legend, was ordained by Gautama Buddha to turn the Wheel of Law for the salvation of the Chinese.7 Wen-Shu-Shi-Ii (i.e. Manjusri) is said to have chosen the Wu-tai-shan (i.e. Pancasirsha or Mountain with Five Peaks) in the province of Shensi as the place of his manifestation.8 His activities were believed in certain sources to have begun in the first century AD In any case, Manjusrī could have been well-known in Shensi before the date of the Yun-Kang caves, in which the male figure under discussion was sculptured.9 Manjusri, the God of Transcendent Wisdom,10 can indeed be considered to have a characteristic shared also by Kārttikeya as Brahmanyadeva or God devoted to sacred knowledge.11 That the deity of Buddhist pantheon established a syncretistic relationship with the Brahmanical god is also indicated by such appellation for Manjusrī as Manjukumara.12 Manjusrī is referred to as Kumarabhuta ('He who has become Kumara or a Kumara'), which expression may allude, among others, to Kumara Kārttikeya.13 Again, Manjusrī is also described, like Kārttikeya or Kumara, as a Kumara and as 'having the appearance of a Kumara,14 Manjukumara, according to a dhyana in the Sadhanamala, carries (like the deity at Yun-Kang) a bow and an arrow.15 In some sculptures, Manjusrī has five heads,16 which is at least the visible number of heads in the representation of the deity at Yun-Kang. These data may indicate the feasibility of accepting the latter as a syncretistic icon involving Kārttikeya and Manjusri.

The earliest iconic descriptions of Manjusri, indicating such syncretism, appears in the Arya-Manjusri-mulakalpa,17 datable to a post-Gupta period and so later than the date of the Yun-Kang deity. In that case, we cannot deny the possibility of one of the earliest contacts between Kārttikeya and Manjusrī having taken place in the latter's traditional domain (viz., the province of Shensi) in or before the latter half of the 5th cent. AD (to which period the Yun-Kang icon can be dated).

It is difficult to determine the proportion in which Manjusrī and Kārttikeya exerted influence on the Yung-Kang image. It has been found in a cave associated with Buddhism. We also know of the relationship of Manjusrī with Siva (or Nilakantha) also represented in the same cave. Several Siva lingas in Nepal are engraved with icons of Manjusri.18 However, it must be borne in mind that Siva is basically a brahmanical deity. Moreover, it is not known whether a regular and close relationship between Manjusrī and Siva had been developed by the date of the two Yung-Kang icons, i.e. by about the third quarter of the 5th cent. AD Again, this type of relationship was most probably prompted by the coalescence of the concepts of Kārttikeya and Manjusri. It was only natural that Manjusri, under the influence of Kārttikeya, developed a relationship with the father of the latter (according to a belief). It has also to be noted that the Yung-kang image with a peacock is carrying a cock, a well-known cognizance of Kārttikeya.

Whether the image concerned is to be recognized as Kārttikeya-Manjusrī or Manjusri-Kārttikeya (or Manjukumara), it betrays clearly the influence of the concept of Kārttikeya. The related concept was certainly older than that of Manjusri.19 The evidence of the Yung-Kang icon indicates the familiarity of that area with the concept of Kārttikeya by about the 5th cent. AD

If Kārttikeya was known in Shensi of the Chinese mainland in the 5th century AD, he should have been a recognizable deity from a still earlier period in Chinese Central Asia, through which area his cult or iconic concepts could have reached from the Indian borderlands to Shensi. That the direction of the travel of this concept was from the Indo-Iranian borderlands to China via Central Asia is further indicated by the appearance of a popular and earlier Iranian iconic feature like the holding of the Sun and Moon symbols20 in hands in the representations of Kārttikeya in Central Asia and China.

We have shown elsewhere21 that with the spread of Buddhism in Central Asia and China, Kārttikeya (or Mahasena or Kumara) was gradually adopted as a minor guardian deity (sometimes having demonic appearance) in the Buddhist pantheon. But the deity was popular enough to make some impact on the iconography of an important Buddhist divinity like Manjusrī in Central Asia as well as in India.


  1. China Reconstructs, September 1982, Vol. XXXI, No.9, pp. 60 and 63; Hart Hurling, I. and A., Chinese Art, New York, 1953, p. 234.
  2. China Reconstructs, September 1982, Vol. XXXI, no.9. pp. 60 and 63. See also the Expedition, Summer, 1983, Vol. XXV, no.4, p. 45, where the bird is wrongly described as an eagle.
  3. Agni Purana, ch. 50, vv. 27-29; Matsya Purana, ch. 260, vv. 46-50.
  4. For an example, see the Agni Purana, ch. 50, vv. 27-29.
  5. Expedition, Summer 1983, Vol. XXV, No.4, p. 45.
  6. S. K. Saraswati, Tantrayana Art -- An Album, Calcutta, 1967, p. XIX and fig. 10.
  7. A. Getty, The Gods of Northern Buddhism, 2nd edition, reprint, Tokyo, 1962, p.110.
  8. Ibid.
  9. Ibid.
  10. Getty, op. cit., p. 110.
  11. One of the epithets of Kārttikeya is Sanatkumāra, meaning 'Son of Brahma'. In the Chhandogya Upanisad (VII, 26) Sanatkumara is referred to as the instructor of Narada in Brahmavidya. The deity is shown in a panel at Ellora as teaching Siva the significance of pranava (1. N. Banerjea, Development of Hindu Iconography, 2nd edition, 1956, pp. 255 and 363, fn. 1).
  12. B. Bhattacharyya, sādhanāmāla, Vol. I, Baroda, 1925, No. 70, p. 151; Saraswati, op. cit., p. XVIII.
  13. Arya-Manjusri-mulakalpa, ed. T. Ganapati Shastri, pt. II, Trivandrum, 1922, pp. 253.304, 315, 332, 441,460.
  14. Bhattacharyya, op. cit., Vol. I, No. 70, p. 142. "Manju, according to certain authorities, may possibly be a Tokharian word corresponding to Sanskrit Kumara" (Getty, op. cit. , p. 110, fu. 1; E. Elliot, Hinduism and Buddhism, Vol. III, p. 221).
  15. Ibid., no.76, p. 151; Saraswati, op. cit., p. XXII.
  16. Getty, op. cit., p. 113.
  17. T. Ganapati Sastri, op. cit., pt. II, pp. 317, 318.
  18. D. C. Bhattacharyya, Iconography of Composite Images, New Delhi, 1979, p. 42, fig. 34.
  19. Skanda-yaga (Atharvaveda-parisista, 20); Journal of the American Oriental Society, 1890, Vol. XV, p. 5; Patanjali's Mahabhasya on the Astadhyayi, v. 3, 99; Skanda-kumara and Bizaga (Visakha) on Kusana coins; reference to the same deity as Kārttikeya and Kumara in a Nagarjunakonda inscription, Banerjee, op. cit., p. 143.
  20. For an example, see O.M. Dalton, The Treasure of the Oxus, 3rd edition, London, 1964, p. 57; pl. XXXII, No.203.
  21. Our paper entitled "Kārttikeya (Mahasena) in Central Asian Iconography" will be published shortly.

See also
Wei Tuo P s (Bodhisattva Skanda) | Skanda in Japanese Buddhism | Kārttikeya in ancient Java | Kārttikeya in ancient Cambodia
Research articles from International Conference on Skanda-Murukan