^{page 509 note 1} See Mr. Hoey's remarks, loc. cit. below.

^{page 510 note 1} As there is no certain instance of a long ī or ū haying been intended in this record, perhaps it would be more accurate to say “without any use of the mark for the long ā except in atiyāyikaya, line 4, and as a component of the o of Koṭhagalani, line 2, and nō, line 4.”

There may perhaps be an imperfect presentation of ā in (from that point of view) sāsane, line 1, and medītna, line 3. But it is very dubious.

^{page 511 note 1} The Gaṇa includes Vārāṇasī. I find its derivative, vārāṇasēya, ‘produced or born in, or belonging to, Benares,’ in the form balanaseya, in the Jōgīmārā inscription, ARASI, 1903–4. 128.

It may be added, in connexion with a puzzle which still remains unsolved, that the Gana further includes, not only Kausāmbī, but also the name Vana-Kauśāmbi, “Kauśāmbī in the forest.” To this latter detail my attention was drawn, some little time ago, by Dr. Barnett.

^{page 512 note 1} There is no need to assume that this character was at first omitted by accident, and was then inserted on revision. The smaller letters in this record were evidently so formed as a matter of convenience in arranging the lines of the text, as is often done (e.g., in a book-plate with a long record on it) in the present day. The second letters in lines 3, 4, were evidently placed as they are, and were formed small, to allow room for the nail-hole or rivet-hole at that corner; and the last two letters in that line were spaced out for the same purpose.

^{page 513 note 1} There are four instances in this record, in which l stands for an original r. In the name Ḍasilimata, it may or may not do so. The record does not present an r.

^{page 514 note 1} This instance is cited by Gray in his Indo-Iranian Philology, § 177; but it is the only one given by him.

^{page 515 note 1} For the change of bh to v, coupled with a throwing back of the aspiration, Professor Bühler quoted the Marāthī gādhav = gardatha, ‘a donkey.’ Mr. Gray, in his Indo-Iranian Phonology, § 317, stamps the change as extremely rare: he has not given any Indian instance of it there; but he has mentioned gāḌhav in § 140 and other places. I cannot trace any other instance.

Professor Bühler cited also the Marāthī tighaī, ‘a building having three rooms or divisions along its length,’ which Professor Pischel gave him, with the suggestion that it may be derived from tighava and may stand for an inferential form *trighraba. Molesworth has given tighaī, with that meaning, and dughaī, ‘a building having two apartments along its length,’ but has not offered a derivation of them, or shewn any separate word ghaī.

^{page 516 note 1} It may be remarked that triyava, ‘having, weighing, or measuring three barley-corns,’ will not help us.

^{page 516 note 2} Even that condition might perhaps have been not unreasonably complied with, from Professor Bühler's point of view, by accepting mathu as standing for matthu, the established corruption of mastu, ‘whey.’ I mention this by way further illustrating the kind of puzzle presented by spelling such as that of our record.

^{page 517 note 2} The word may possibly have some connexion with the root mā, ‘to measure,’ from which, it seems, we have in Pāli mētabba as well as mātabba = mātavya, ‘should be measured.’ But a me could be obtained in other ways: it might come from a mavi, as we have thēra = sthavira; and Professor Pischel has given us, in his Grammatik der Prākrit-Sprachen, § 166, māḍambha from *maīḍambha = *mṛrigīḍambha, and mēhara, alongside of maīhara, from *matidhara.

^{page 518 note 1} It is at any rate certain that we have here some derivative from vṛi, ‘to surround, cover, conceal,’ etc. The gerund vāratn would give vālam, and so. vala, in the spelling of our record; hut it would hardly construe. The infinitive, vartum, varitum, would construe; but it would give vattum or vaṭṭuin, which, cannot he found here.

^{page 519 note 1} We have the same assimilation and shortening in also the ordinary Prākṛits; see Pischel's Grammatik der Prākrit-Sprachen, § 284, which gives us ayya = āryar kayya = kārya, and suyya = sūrya.

^{page 520 note 1} The character is, in fact, exactly similar to a reversed li, which might of course mean either li or ri. But I do not find that an acceptance of it in that way would help.

^{page 521 note 1} Other similar forms, as well as the same words in other passages, are citable from the Kharōshthī edicts.

^{page 521 note 2} The published treatments of these two words (IA, 6. 159; 20. 163, para. 15; 22. 305, para. 34) are based on assumptions that in lākhāpetavaya the lā is a mistake for li or le, and that in both words the terminations tavaya, tavāya, stand, by mistake or otherwise, for taviya as = taviye or taviyam. And Professor Bühler cited the first word, so treated, in support of his explanation of gahitavaya by grahītavyam.

The actual readings of these two words in the Rūpnāth edict are quite clear in the facsimiles; see the plates in IA, 6. 156; 22. 299. And what the passage says is as follows:— “And this matter has been caused to be engraved on rocks (both) in other localities and here (at home), (and, where) there is a stone pillar, on (that) stone pillar,— with a view to causing it to be observed, and, by this same intimation, with a view” etc. I refrain from offering at present a translation of the remainder of the sentence, which includes vivsetavvāya; the meaning of that word (and of certain other peculiar derivatives from apparently the same root) will have to be more fully considered, with the help of my explanation of lākhāpetavaya, lakkhāpetavvāya, in conjunction with the treatments by Professor Kern and Dr. Vogel (EI, 8. 169, 170 f.) and by Professor Venis (JASB, 1907. 2, 4 ff.) of the analogous passage in the Sārnāth edict.

^{page 521 note 1} Just as in the case of Travellers' Bungalows; the usual rule is that a newcomer may evict an occupant who has held any room or set of rooms for twenty-four hours.

^{page 526 note 1} According to Monier-Williams' Sanskrit Dictionary, Trivēnī is a name of Prayāga, Allahabad, as being the confluence of the Ganges, the Jamnā, and the subterranean Sarasvatī. There is also a Tribēṇī on the Hūglī, a sacred place, about twenty-five miles above Calcutta, so called, according to the Imperial Gazetteer of India, 13. 353, because there is a confluence there, also, of the Ganges, the Jamnā, and a local Sarasvatī; the Jamnā here being (see the Imperial Gazetteer, 7. 134) the lower section of the Brahmaputra, from its entrance into the plains to the confluence in question.

At Tribēṇī Ghāt, two or three small nullahs seem to flow into the Gaṇdak. But I cannot obtain any indication of there being, there a real confluence, sufficient to suggest that the name was always Trivēnī.

The matter would become simple if, in our record, in the third syllable in line 3, we might read ve instead of va; thus obtaining tiyaveni, for tiyavēnī, tīvēni. While, however, the photo-etching gives a slight suggestion in that direction, it is not endorsed by the electrotype.

I would say explicitly that I do not suggest any etymological connexion between the names Tryavani and Tribēnī Ghāt. I hold that the latter supplanted the former, as the result of some suggestion or invention, leading away from the original appellation of the locality, but combining a reminiscence of it.

^{page 527 note 1} Just as the notices in a Traveller's Bungalow are stuck up in each room or set of rooms.

^{page 529 note 1} See, for instance, Cunningham's Coins of Ancient India, plate 1, figs. 26, 27, 29, and plate 2, figs. 6 to 9, 11 to 17, 19, 20; and Rapson's Indian Coins, plate 1, figs. 11, 13, and plate 3, fig. 6.

^{page 529 note 2} A mark which the photo-etching shows, between this symbol.and the railed, enclosure or box, is only an exaggeration of a fault in the making of the plate.

^{page 529 note 2} Can any of them be carried back, in their present meanings, to any date earlier than mediæval times?