[17] See above, p. [9]. Of my other pictorial ‘finds’ from Central Asia only the mural paintings of Mīrān, approximately dating from the third to fourth century a.d., show this use of ‘high lights’; cf. Serindia, pp. 504, 508, Pls. xl-xlv.
[18] These two-lobed tufts of hair recall those shown on the heads of the angels and putti in the wall-paintings of the shrines excavated by me at Mīrān; see Serindia, Figs. 134, 138, 140; Plates xl, xli.
[19] Cf. Serindia, p. 850 sq.; also below, p. [23].
[21] For details on these points and on the question of style, cf. Serindia, p. 847 sq.
[22] Cf. Serindia, p. 848.
[23] Cf. Serindia, p. 850; Chavannes, Mission archéologique en Chine, i. Planches 207–10.
[24] Cf. Serindia, p. 849, note 18.
[25] This is against the fixed iconographic convention of Indian tradition which shows the right hand raised and the right shoulder uncovered by the under-robe. The explanation may be sought for in the fact that in the case of banners both sides of the silk gauze had to be painted. Here and in the Buddha of the banner in the middle of the Plate we have obviously cases of a mistake made by the artist as to which side was to be treated as the one intended for contemplation and properly finished.
[26] Cf. Serindia, p. 858, and the reproduction of the banner, Ch. lv. 0012, Pl. lxxv.