^{page 303 note 1} Bühler readmute.

^{page 303 note 2} The variœ lectiones of the Kālsī, Mānsehrā, and (partly) Girnār versions are unimportant and need not be repeated here.

^{page 303 note 3} The rendering of dhamma by “ morality ”, etc., is senseless. Dhamma in the Aśoka inscriptions never means anything but “Buddhist doctrine, Buddhism”; with this I propose to deal in another connection.

^{page 303 note 4} It is unintelligible to me why Hultzsch rendered the single punaḥ in this sentence by “ repeatedly ”, a translation that cannot be upheld.

^{page 303 note 5} This “here ” undoubtedly reminds us of Rock Ed. V M, where the other versions have hida (K, M, Dh.) or ia (Sh.) while G has the explanatory Pāṭalipute.

^{page 303 note 6} With paraṃ ca tena A. cf. Rock Ed. V E, paraṃ ca tena (in a temporal sense).

^{page 303 note 7} Bühler, , Epigr. Indica, ii, 466Google Scholar, translated sāmantāḥ by “ vassal-kings ”, which is undoubtedly the common meaning of the word. Previously Wilson, , JRAS. (O.S.) xii, 169Google Scholar, rendered it: “ and those princes who are near to (or allied with) that monarch ”; Kern, , IA. v, 272Google Scholar: “ his neighbour kings ” (with a foot-note: “ in the first place Baktria ” ); and Senart, , Inscriptions de Piyadasi, i, 74Google Scholar: “ des rois qui l’avoisinent.” Thus Professor Bhandarkar, D. R., JBBrRAS. xxi, 398Google Scholar, in taking exception to the translation of Bühler, was not without predecessors; pointing to the various reading sāmīpaṃ of the Girnār version he strongly advocates the translation “ neighbours ”. This view was endorsed by Smith, V., IA. xxxiv, 245Google Scholar, who had previously (Asoka, 1st ed., p. 115) adopted the translation of Bühler. According to my humble opinion there can be no doubt that Bühler was right; it is only natural that Aśoka should think those other princes to have been the vassals of Antiochus, who was, besides himself, the most powerful monarch of the period, and he certainly drew conclusions from the state of his own dominions where there were undoubtedly numbers of half-subdued Sāmanta's. As for sāmīpaṃ (or opa) cf. the remark of Hultzsch, , CII.² i, p. 3, n. 3Google Scholar (according to Michelson, , AJPh. xxx, 183 ff.Google Scholar, it is ═ Skt. sāmīpyam).

^{page 304 note 1} The identification ═ Subhagasena was suggested already by Schlegel, A. W. von, Indische Bibliothek, i, 248; ii, 301Google Scholar. There exists no known Indian prince of that name; cf., however, Subhaga, prince of Gāndhāra (with whom cf. CHI. i, 512Google Scholar) in the Mahābhārata, vii, 6944 (Bombay)Google Scholar.

^{page 304 note 2} To suggest that, we should want the phantasy of Wilford who in Asiatick Researches, v, 285 sq.Google Scholar, concluded that rendered an Indian Śivakasena, which would again be ═ Aśokasena (cf. also Prinsep, loc. cit., p. 162). Already Wilson scoffed at this rather adventurous idea.

^{page 304 note 3} JASB. vii, 225 sqq.Google Scholar (reprinted Essays, ii, 20 sq.).

^{page 305 note 1} This sentence contains two rather apparent mistakes: Magas was not the son-in-law but the stepson (and perhaps also the adoptive son) of Ptolemy Soter; his mother, Berenike, was also the mother of Ptolemy Philadelphus.

^{page 305 note 2} Wilson's arguments were criticized by General Cunningham in The Bhilsa Topes, p. 110 sq., which was an easy enough task. Cunningham was right in eliminating Antiochus III; but he states, with a slight exaggeration, that Prinsep had definitely fixed upon Antiochus II (unless we have here possibly a misprint— II for I).

^{page 305 note 3} Cf. Inscriptions de Piyadasi, ii, 256 sqq.Google Scholar; IA. xx, 242.Google Scholar

^{page 305 note 4} Cf. Asoka, 3rd ed., p. 162.

^{page 306 note 1} Cf. e.g. Bevan, , The House of Seleucus, i, 298, etc.Google Scholar

^{page 306 note 2} It would, of course, be theoretically possible to think also of Ptolemy III Euergetes (24–221 B.C.). That would, however, seriously dislocate the chronology of the three first Mauryas. Ptolemy III, it is quite true, was not, as a ruler, a contemporary of either Magas or Alexander of Epirus; but that would probably be of little importance in this connection.

^{page 306 note 3} Cf. ZDMG. xl, 137.Google Scholar

^{page 306 note 4} Cf. CII. i ², p. xxx, note 2.Google Scholar

^{page 306 note 5} Cf. the literature quoted in CII. i ², p. xxx.Google Scholar

^{page 306 note 6} Cf. Plutarch, , Pyrrhus, p. 4.Google Scholar

^{page 306 note 7} Cf. Beloch, , Griechische Geschichte, iii, 2, p. 105.Google Scholar

^{page 307 note 1} Appianus, , Syr. 65.Google Scholar

^{page 307 note 2} Cf. Bevan, , The House of Seleucus, i, 168 sq.Google Scholar; the date given here is in accordance with the Cambridge Ancient Hist, vii, 709.Google Scholar

^{page 307 note 3} Ap. Athenæum, x, 438c; cf. also Aelianus, , Var. Hist., ii, 41.Google Scholar

^{page 307 note 4} Cf. Bevan, loc. cit., i, 172.

^{page 308 note 1} He seems to be known also by at least two other names, viz. Agathocles or Andragoras, cf. CHI. i, 438. It is not quite sure that they all refer to the same man, though, of course, nothing definite can be suggested here.

^{page 308 note 2} Justin, xli, 4.

^{page 308 note 3} Cf. CHI. i, 439 sq.Google Scholar

^{page 308 note 4} As for Diodotus the following circumstances, even if quite hypothetical, may well be taken into consideration. It seems to me fairly probable that Diodotus was really the satrap of Bactria who about 274/73 B.C. furnished Antiochus I with some twenty elephants during his war with Ptolemy (CHI. i, 437). If that were the case it seems quite likely that Diodotus had been appointed satrap of his important province already during the viceroyalty of Antiochus I in the East, which came to an end in 281/80 B.C. Diodotus, whose reign seems to have been rather short (cf. above, p. 308), must then have been a fairly old man in 250 B.C.—at least about or well above sixty. The reasons for his rebellion are, of course, unknown; but they may have ultimately been connected in some way or other with the execution of the young Seleucus, the elder son of Antiochus I, who was probably viceroy of Iran, and must have been put to death in the year 263 B.C. (cf. Bevan, loc. cit., i, 150, n. 3, 169; Cambridge Ancient Hist, vii, 709 sq.)Google Scholar. What I mean is that Seleucus may have been popular and perhaps even have tried to reign on his own, while Antiochus II was perhaps less well liked throughout the East.

^{page 309 note 1} Even if such were the case there is no reason for the remark sometimes put forward about Diodotus (and even Arsaces) not being mentioned by AŚoka. For AŚoka, even if he had happened to hear about some upraising in Bactria, would scarcely have considered its leader worthy of mention as one of the kings connected with Antiochus.

^{page 309 note 2} Cf. Arrianus, , Anabasis, vii, 22, 5.Google Scholar

^{page 309 note 3} Seleucus, according to the latest available evidence (cf. Cambridge Ancient Hist, vii, 98, n. 1)Google Scholar, was murdered some time between 30th November, 281, and March, 280 B.C.

^{page 310 note 1} Cf. Arrianus, , Anabasis, vii, 4.Google Scholar Antiochus I thus most probably was born in 323 B.C. and cannot, at the time of his death, have been sixty-four years old (Bevan, loc. cit., i, 168, quoting Eusebius, i, 259).

^{page 310 note 2} Cf. Bouché-Leclercq, , Histoire des Séleucides, i, 7.Google Scholar

^{page 310 note 3} In this connection let me quote the following passages: “ Antiochus … had some things to his favour. In the first place, his hold upon the eastern provinces was firm. His mother, it must be remembered, was of Irânian race, and those peoples might naturally cleave to a king who, by half his blood, was one of themselves. Through his mother, many perhaps of the grandees of Irân were his kindred ” (Bevan, loe. cit., i, 74). “ Antiochos avait sur son père l’avantage d’ètre à demi iranien par sa mère Apama et, peut-être pour cette raison, moins impopulaire dans l’Iran ” (Bouché-Leclercq, loc. cit., i, 40).

^{page 310 note 4} Cf. CHI. i, 434Google Scholar, with pl. ii, 1. The Cambridge Ancient Hist, vii, 93Google Scholar, correctly remarks that the appointment of Antiochus as viceroy of the East was not without precedence in Achsemenian times.

^{page 311 note 1} Cf. Gutschmid, von, Geschichte Irans, p. 26 sq.Google Scholar; the greatest of authorities, the late Ed. Meyer, , Hermes, xxxiii, 643Google Scholar, speaks of Antiochus as “ der grosse aber in der Ueberlieferung fast verschollene Stādtegründer ”. Cf. also Bevan, loc. cit., i, 163.

^{page 311 note 2} That this name should be transliterated into Amitrakhāda, not ghāta, I have tried to prove, following older interpretations, in JRAS. 1928, p. 132 sqq.Google Scholar On Bindusāra—or whatever was his name (CHI. i, 495)—cf. the elever but utterly hypothetical article by the late Professor Gawroński, in Rocznik Orientalistyczny, ii, 21 sqq.Google Scholar, which, according to my opinion, affords no tangible results.

^{page 311 note 3} Cf. Fragm. Hist. Grœcorum, iv, 421.Google Scholar The story of the Indian king wanting to buy a philosopher, which seems strikingly un-Indian, is apparently meant for a witty sneer at the far-off barbarians, but does not interest us here.

^{page 311 note 4} The slight discrepancy between CHI. i, 495Google Scholar, where Seleucus and i, 433, where Antiochus I is said to have sent this Daīmachus to India is probably of no consequence at all. For he may in reality have been sent by Antiochus acting as the viceroy of his father in the East (cf. ).

^{page 311 note 5} Cf. Bevan, loc. cit., i, 69 sq.

^{page 312 note 1} Most of these princes were closely related to each other. Berenike (I), the daughter of Lagus and Antigone, daughter of Kassander (cf., however, Beloch, , Griech. Geschichte, iii, 2, 128Google Scholar), first married a certain Philippus, the father of Magas and of Antigone, wife of Pyrrhus of Epirus. Berenike then married her half-brother Ptolemy I and became the mother of Ptolemy II. Magas thus was the cousin of this ruler; he himself married Apama, the daughter of Antiochus I. Pyrrhus and Antigone again were the parents of Alexander of Epirus.

^{page 312 note 2} Aśoka, as governor of some of the western provinces of the empire during the lifetime of his father, may already then have entered upon relations with Antiochus, at that time possibly still the viceroy of the East.

^{page 313 note 1} Cf. Les Inscriptions de Piyadasi, ii, 243 sqq.Google Scholar

^{page 313 note 2} In the following I am not concerned with any inscriptions except the fourteen Rock Edicts and the two separate ones of Dhauli and Jaugaḍa. Of the new Mysore version, the discovery of which was announced in the IHQ. v, I have, unfortunately, not been able to gather even the scantiest information.

^{page 313 note 3} vijinitu Bühler; but cf. tiṭhiti, aloceti (CII. i ², p. xcviiGoogle Scholar).

^{page 314 note 1} Kālsī correctly vijayaṃ.

^{page 314 note 2} We are not here deeply concerned with either the date or the mode of Aśoka's conversion, which have been much discussed. That the conversion occurred immediately after the Kālinga campaign there cannot be the slightest doubt. And as even those virtues which Aśoka does elsewhere (cf. Rock Edicts IV, IX, etc.) praise as the most meritorious ones are said in xiii, J, to have been practised even among the people of Kālinga, it would be a perfectly justifiable conclusion that Buddhism was at that time widespread in that country, and that the conversion of Aśoka did really originate from there.

^{page 315 note 1} The reason why it was not published in Kālinga is, of course, quite conspicuous and has been pointed out long ago. It would, however, be still more obvious if the edict waa really published immediately after the conquest and not several years afterwards.

^{page 315 note 2} It seems peculiar that the epithet Priyadarśin should occur nowhere in the two separate edicts. For this some local reasons unknown to us may account. It is also somewhat remarkable that in the second separate edict Dh. has everywhere Devānāṃpriyaḥ where J. uses the word rājā (cf. the parallel conditions prevailing in Rock Edict VIII, A; cf. CII. i² p. xxx).

^{page 315 note 3} We are strongly reminded of the existence even to this day of uncivilized hilltribes within the frontier districts of Orissa, etc.

^{page 316 note 1} On this place cf. Deo, B. S., Quart. J. Andhra Hist. Res. Soc., iii, 41 sqq.Google Scholar

^{page 316 note 2} It seems somewhat remarkable that several names containing that of the constellation Tiṣya belong to the Maurya time. There is Aśoka's wicked queen Tiṣya-rakṣitā, and his brother Tiṣya (on this name cf. Pāṇini, iv, 3, 34). There is further the contemporary king Tissa of Ceylon (Dīpavaṃsa), and the great divine Tissa Moggaliputta (cf. Geiger, , Mahāvaṃsa, p. xlvii sq., etc.).Google Scholar Still further there is Puṣyagupta, a viceroy of Candragupta (Epigr. Indica, viii, 46 sq.)Google Scholar; and there may be even more names of which I am not aware. The fifth Pillar Edict further tells us that on Tiṣyā castration and branding of animals must not be performed. Unfortunately, I cannot suggest any probable connection of the Maurya family with this constellation though there may well be one.

^{page 317 note 1} Cf. IA. xlii, 159 sq.Google Scholar

^{page 317 note 2} With this use of the word sambodhi cf. Jātaka, iv, 236, 2Google Scholar: mahāyitvāna sambodhiṃ (with mahāyitvāna cf. mahīyite in the Rummindēī and Nigālī Sagar inscriptions).Cf. also Mookerji, , Asoka, p. 105 sq.Google Scholar

^{page 317 note 3} Cf. e.g. Lüders, , Sitz. ber. Preuss. Akad. d. Wiss., 1914, p. 846.Google Scholar

^{page 318 note 1} Somewhat similar measures were at times taken by Akbar, cf. Smith, , Akbar the Great Mogul, p. 167.Google Scholar

^{page 318 note 2} To peacock's flesh no doubt magical qualities were ascribed; it was believed to convey immortality, not to decay, etc. Cf. Jātaka, ii, 36 sq.Google Scholar; Johansson, , Solpfågeln i Indien, p. 78 sq.Google Scholar; Charpentier, , Festschrift E. Kuhn, p. 283, n. 4Google Scholar; Mookerji, , Asoka, p. 62.Google Scholar

^{page 319 note 1} It is certainly remarkable that this rescript contains at least two words which strongly remind us of Jain terminology, viz. vaci-gutī (vaca-guti) in D and kalāṇāgamā in J (this, by the way, must mean “ possessed of good scriptures ”, not “ pure in doctrine ” as rendered by Hultzsch). Of the officials mentioned here the dharmamāhamātra is in all probability the special supervisor of the Buddhist saṃgha (cf. Delhi-Tōprā VII, Z); the ithījhakha certainly has got nothing to do with the gaṇkādhyakṣa of Kauṭilya (thus CII. i², p. 22, n. 4)—he may possibly be some sort of overseer of the nuns; the vaca-bhūmika is the supervisor of the holy cows (and probably of the pinjrapols, cf. Rock Ed. II), a purely Brahmin official.

^{page 319 note 2} Cf. CHI. i, 430, 472, 698.Google Scholar

^{page 319 note 3} It must, however, be observed that these words do not necessarily involve that Candragupta was still alive during all the visits, though the text says παρά . The successor of Candragupta, as we know, was not even known to the Greeks by his real name.

^{page 320 note 1} There remains the possibility that the four years during which Aśoka is said to have reigned before his anointment do in reality mean nothing but a co-regentship with Bindusāra (cf. also CHI. i, 503, n. 1.Google Scholar). the case the latter's regnal years would come in somewhere between 299/97 and 274/72 B.C. But all this is pure guess-work.

^{page 320 note 2} On the date of Candragupta cf. also the able paper of Dr. Stein, O., Indologka Pragensia, i, 354 sqq.Google Scholar

^{page 320 note 3} It is to be sincerely hoped in the interest of Indian ancient history, which is mainly constructive, that the emendation Nandrum for Alexandrum is really the correct one. Otherwise the passage from Justin would tell us an absolutely different tale.

^{page 320 note 4} Here the Mudrārākṣasa, which may be of considerable historical value, is especially illuminative.

^{page 320 note 5} It seems to have been always taken for granted that Agrammes or Xandrames (on whom cf. Thomas, E., JRAS. 1865, p. 447 sqq.Google Scholar), the despicable sovereign of the East who had murdered his predecessor, was in reality a Nanda. But we look out in vain for definite proofs of such a suggestion. Xandrames, as Professor Thomas has rightly remarked (CHI. i, 469 sq.Google Scholar), most probably renders a Sanskrit form Candramās, and this is certainly not far from Candragupta. That Candragupta should have visited Alexander while in the Punjāb (Plutarch, Alexander, lxii) sounds suspiciously like a myth.