page 259 note * The small ruined brick edifice on the top of the rock can hardly be considered an exception

page 264 note * The changes which time has produced are in some letters very great; and where characters are so simple as those of the Tamil language, even slight alterations in form give rise to perplexity. I may adduce, as an example, the letter ca, the most ancient form of which was a Latin cross . In the course of time a top was added to the left side, and the cross bar was curved thus The next alteration was in the addition of a perpendicular line falling from the left extremity of the top . The top was then extended to the right , and by prolonging the extremities of the curved line, the modern letter has at lenglh been formed, or in a still more complicated manner as in the Grantha thus In its modern form it might easily be confounded with the which, though it now has a tail, was anciently written without one, thus . Other examples might be given, but they suggest themselves on an inspection of the alphabet itself.

I cannot touch on the subject of ancient Tamil characters without remarking, that their extreme simplicity seems one among many circumstances, which indicate that the language is of very high antiquity. The Sanscrit of the South of India is written in characters (the Grant'ha) derived from the Tamil, but they are much more complicated, and therefore probably posterior in point of antiquity. The peculiar structure of the Tamil language, wholly dissimilar from the Sanscrit, its deficiency in aspirated consonants, its possession of letters and sounds not found in Sanscrit, its division into dialects, one of which contains but few words of Sanscrit derivation; and lastly, its locality at the southern extremity of India, would seem likewise to indicate an independent origin, and one of at least equal antiquity with the Sanscrit itself; but this is a subject foreign to that now under consideration, and deserving a more lengthened discussion than the limits of a note will allow.

page 265 note * Mrs. Graham gives an engraving of this small pagoda, which she says is called the Teer of Arjun, and she explains the word Teer (properly Zér), to mean a place of religious retirement. The explanation is erroneous, as the word Ttér signifies, in Tamil, a car or sacred vehicle corresponding with the Sanscrit Rat'ha.

page 265 note † When Mr. Goldingham wrote his account, this pagoda contained a lingam (see his first paragraph), so that it has passed from the Sáivas into the possession of the Vaishnavas sine, that period. Brahman families of both sects reside on the spot.

page 265 note ‡ In this character there are two forms of affix for long a, used indifferently as convenience may suggest. Thus, in the first stanza, we find the syllable () cáran'am () made thus ; while in the eighth stanza the same syllable in the word cáma is made thus

The letters p (), and s (), seem similar in form; but perhaps some slight difference may have escaped my observation. In modern Grantha, an inflection in the middle of the character makes the difference between s and p, thus . So also in the ancient character, the inflection may have been greater in the s than in the p, thus and ; but I have not upon this supposition felt authorized to depart from the copy which I made on the spot.

page 268 note * There is one instance, as will be seen on a reference to Plate 17, in which the characters are of the same kind as those in the inscription of the Gánés'ā Pagoda, a proof that both were in use at the same period. It is probable therefore that one was the round, and the other the square form, analogous to the two varieties of Páli and Ariyam.