^{page 640 note 1} In the records this name is spelt varyingly with ū or with u.

^{page 640 note 2} The Aihole inscription here refers to the three adjoining countries, Laṭa, Rajputana, and Malwa, as submitting to Pulikeśin II.

^{page 641 note 1} For the sake of simplicity I give here, and throughout, only single equivalent years of the Christian era, which for the purpose in hand is quite sufficient.

^{page 642 note 1} It has become usual to call such figures ‘ numeral letters, ’ because of their curious resemblance to letter-forms. I believe the practice of calling them so originated with the late Pandit Bhagvānlāl Indrajī (Ind. Ant., vi, 42 ff.). But the resemblance does not really become noticeable before the eighth and following centuries (especially in Jain, Nepalese, and Buddhist manuscripts), and the farther back one traces the symbols the more the resemblance disappears. Whatever the origin of the numeral figures may be, I do not believe that they have any connection with the letters (simple or compound) of the Brāhmī alphabet, in the sense in which this connection has been ordinarily understood. I hold, therefore, that it is misleading to distinguish between numeral figures and numeral letters. There really exists only one set of symbols—numeral figures ; and their growing resemblance to letters is due merely to the growing whimsicality of scribes who exaggerated a fancied resemblance.

^{page 643 note 1} The same anomalous form saṁvatsro is supposed to occur in a Khajurāho inscription (Ind. Ant., xxvi, 30, 31). But the word really reads correctly saṁvatsare. The akshara re is somewhat indistinct, but the down-stroke shows, in the middle, a slight indentation (cf. ri at the end of line 3 in Sir A. Cunningham's Survey Reports, vol. xxi, pi. xvi A; the mark is very obvious in vol. x, pl. ix, i), and it is therefore not the down-stroke of the vowel o, but the indented body-stroke of the consonant r. In any case, even if it were o, the akshara would have to be read tso, not tsro ; there is no underwritten r in it.

^{page 648 note 1} The beginning of the Parihar-Gūrjara occupation might be traced back even to the earliest conquest by Vatsarāja.

^{page 650 note 1} Professor R. G. Bhandarkar translates (E.I. iv, 287) “the preceptor charging the Gauḍas with the vow of humility,” which conveys no very intelligible meaning. The notice seems to indicate either that Krishna, before his accession, lived at Dharmapāla's court and superintended the education of Devapāla, or that the latter, before his own accession, lived at Krishna II's court, where he received his education.

^{page 651 note 1} The reference cannot be to Bhoja I (as suggested in E.I. i, 253), for under that monarch the Gūrjara power was at its zenith, and it is out of the question that he could have been a protégé of Kokkalla I.

^{page 653 note 1} The attribution of this inscription to Harsha is clearly wrong. Unfortunately it is badly mutilated, but its general purport is unmistakeable. The reference to Harsha is finished in line 7, and the sovereign referred to in line 10 must be his successor, Yaśovarman.

^{page 657 note 1} The Udepur praśasti of about 1080 A.D. (E.I. i, 223) gives a much longer hut, on the face of it, mythical ancestry.

^{page 657 note 2} The original text runs as follows : galitā Gūrjara-hṛdayāt Kālaṁjara- Citrakūṭ-āśā. Mr. Bhandarkar translates this, “the hope of conquering Kālanjara and Citrakūṭa dropped away from the heart of the Gūrjara prince ” (see p. 5 of his paper on the Gūrjaras). This apparently reflects an earlier translation, even more strongly expressed, in Professor R. G. Bhandarkar's Early History of the Dekkan (2nd ed., p. 75): “The Gūrjara prince who was preparing to take the fortresses of Kālañjara and Chitrakūṭa in the north had to give up the enterprise.” The compound is translated much more soberly and correctly by the same Professor in E.I. iv, 289: “the hope about Kālañjara and Chitrakūṭa.” (The italics throughout are mine.) The idea of “conquering” or “preparing to take” is not suggested by anything in the compound, and it is wrongly imported into it. For Kalanjar and Chitor did not require conquering; they had already belonged to the Gūrjara empire for a very long time. At this time, it is true, one of them, Kalanjar, had passed into the power of the Chandel Yaśovarman, but that prince had placed Mahīpāla on the throne and was on his side. Naturally Mahīpāla trusted for security to the stronghold of the Chandel. In any case, considering the relation in which these two monarchs stood to each other, one cannot suppose that Mahīpāla could have contemplated “ conquering ” from him the stronghold of Kalanjar.

^{page 659 note 1} Vijayapāla receives the full imperial titles Paramabhaṭṭāraka, Mahārājādhirāja, Parameśvara. For himself Mathanadeva only claims Mahārājādhirāja and Parameśvara, while to his father Sāvaṭa he gives only the title Mahārājādhirāja.