84 



JOURNAL, R.A.S. (CEYLON). [VOL. XIX. 



Greek patterns because we are more familiar with them than 

 with those of the other nations of antiquity. As pointed out 

 above, it is not the actual decorative motifs so much as certain 

 modes of associating them that are traceable to Greece. 



Before returning to Ceylon we may notice one other Indian 

 example, viz., the decoration of the lat (1605 a.d.) at Allaha- 

 bad (see fig. 13, from Ferguson's " Indian and Eastern Archi- 

 tecture"), as it shows, in addition to the lotus and palmette 

 forms, the real and bead pattern which is also common in 

 Ceylon (see Plate I., fig. 3 D). 



Returning now to Ceylon, we shall find that "palmette" 

 patterns are less rare than at Bharhut.* Their actual existence 

 however is significant. The following is a list of a few 

 unmistakable examples of ' ' palmette " patterns comparable 

 with the forms (figs. 10, 11, 12) occurring at Bharhut : the 

 embroidered armlet (Plate I., fig. 1) already described ; the 

 almost identical pattern on the border of an embroidered betel 

 bag in my possession (Plate I., fig. 3) ; a characteristic form 

 found on Kandyan eaves- tiles of recent make (Plate II. and 

 fig. 14) — this is an example of what has been termed 

 gegenstdndige lotus-bluth und palmette, and is suggestive of 

 Greek influence ; some very similar tiles in the Museum (one 



* It must be taken into consideration that if examples of the minor 

 arts survived from Bharhut, it would be much easier to compare the 

 Kandyan patterns with those found there ; our knowledge of modern 

 Sinhalese patterns would be very limited if we had to confine ourselves 

 to the study of work in stone. 



13 



FROM Idt AT ALLAHABAD. 



