² Cf. below, p. 192 ff., nos. 4 and 6. Both the uninscribed fragments mentioned below have bases.

³ Schaeffer, Missions en Chypre (1936), 14.

⁴ In l. 5 the actual reading is ADIAR and RONTIFICO.

⁵ L. 1: υἱὂ. In l. 12 ἓτους was probably written in full as in LBW 2806 (though abbreviated in no. 5 below); probably included in l. 12 and not in l. 13.

⁶ In l. 1, DDDNNN is clearly earlier than DDNN of no. I (d), for the latter is in part cut over it. L. 2: C of ‘Cl(audio)’ has subsequently been changed into V by the addition of a stroke. On the third N of ‘Constantino’ a large A has been superimposed, while the final O is almost lost under the later VG. L. 3: XIM of ‘[ma]ximo’, AVG, the M and probably the last three letters of ‘triumfatori’ have not been tampered with; alteration of the remaining letters has not affected the certainty of the reading. From the confusion of l. 4 each letter of ‘Constantio’ can be recovered. The upright which precedes the name may have been an L of which the lower stroke has later been deleted. Into the lower part of l. 5 the last line of no. 1 (d) has trespassed, but with no attempt to re-use its letters. L. 7: AVGG. L. 8: , an abbreviation found in nos. 2, 3, 4, and 5, below.

⁷ For the first two lines, cf. n. 6 above. Lines 3 and 4 are well illustrated in pl. xxiii (left). In l. 4 ‘Iuliani’ is preceded by a probable ‘M’ and a certain ‘D’, the latter cut over an ‘O’, the former joining the ‘C’ of the earlier ‘Constantio’ to the upright (an L ?) which precedes it. In l. 5 ‘ESA’ of ‘Caesaris’ is inscribed in a flaw in the surface of the stone which the earlier inscription avoids. I am much indebted to Professor J. G. C. Anderson for his great kindness in examining squeezes of 1 (c) and (d). In essentials he is in agreement with the text offered above.

⁸ Height of letters in (a), 0·03 to 0·052 m. A diacritical point is freely used, especially after abbreviation. In (b), height from 0·025 to 0·05 m., the height increasing as this inscription advances.

⁹ Cf. Kubitschek, Grundriss d. antiken Zeitrechnung, 69. The 30th August as an alternative day for the commencement of the Egyptian year is not preferred by Mommsen.

¹⁰ Cf. also PIR ² i, p. 358, 76. Septimius on his departure for the East in 193, left a certain Domitius Dexter as praefectus urbi in the place of a person of that name. In Epit. de Caes. 20, 6, a Bassus Severi amicus is mentioned. It may well be that Bassus accompanied the Emperor and subsequently became the proconsul of Cyprus.

¹¹ Height, of L, 0·07 m.; of other letters, from 0·043 to 0·06 m. These measurements also apply to no. 1 (d).

¹² Excluding CIL iii, Suppl. 6732, only three milestones from Cyprus have hitherto been published. Two are discussed below. The third is CIL iii, 12111 (Hogarth, Devia Cypria, 112, no. 34: IGR iii, 936). Found in a very mutilated condition between Hg. Epiktitos and Bellapaise in the north of the island, it has now vanished. It is bilingual, and apparently earlier than the Septimian era. Hogarth reads in his last line ΛΕ, and considers the stone in all probability the thirty-fifth milestone on the road from Salamis to Lapethus. It is, however, unlikely that this is in fact the last line, for we should expect some mention of the city or proconsul through whom the stone was erected. Further, the locality indicated is between 31 and 32 English miles in the direct line to Salamis, and from this line the road, whether it took the Kyrenia or the Lefkoniko pass through the Northern Range, must have diverged considerably. Lastly, it would be remarkable if, in what was doubtless the territory of Lapethus, distances were measured from Salamis. I fear that CIL iii, 12111, can give no useful information on the Roman road in north Cyprus.

¹³ Thus no instance is given by Cagnat, Cours d'Épigraphie latine 272 ff. But cf. LBW 2807 (d) from Curium to the Emperor Jovian.

¹⁴ I examined this milestone before it was taken to Nicosia, and was unable to reach the ends of lines 3, 4, 7, 8, 10, and 11, with my squeeze. The omission is only serious for l. 10. The margin of the Greek text is spaced some two letters to the left of the Latin, (ἔτους) in l. 10 is represented by L, a sign of Egyptian origin, regularly used in Cyprus both in the Hellenistic and in the Roman period.

¹⁵ The Latin lettering is from 0·035 to 0·06 in height. Height of Greek letters, from 0·029 to 0·05 m.

¹⁶ Attalus is included among the proconsuls of Cyprus by Hogarth, o.c., 118; but is wrongly placed in the reign of Elagabalus.

¹⁷ In the Acta Sanctorum, De S. Auxibio, 3, it is said that Auxibius landed ‘in pagum qui Limne … dicitur … quarto ab Solorum urbe lapide’. On the silting up of the harbour of Soli in the early Empire, Limne (Limenia) became the port for that city. Limne is generally located at the modern Limnitis, 4½ miles as the crow flies to the west of Soli and at the mouth of a river of the same name, where indeed antiquities have been found in some bulk (JHS xi, 1890, 88–91Google Scholar). But this site is not on the main road, and is probably 6, certainly not 4, Roman miles distant from Soli.

¹⁸ I first examined this inscription at the Police Station at Polis.

¹⁹ Letters very narrow, from 0·042 to 0·052 m. in height.

²⁰ L. 1: outset, DDNN. L. 7: NOBBCAESS. Of the cross above the inscription, only a portion of the lower arm survives.

²¹ Height from 0·035 to 0·041 m. Letters regular and well formed, sigma in particular suggesting the first century (below, p. 198,fig. 11). Thus sigma of the classical shape occurs in some 39 datable inscriptions of the Empire, 32 of these belonging to the first century, 7 to the second; and of these 7, 4 are Hadrianic or earlier, the remainder slightly suspect. Lunar sigma appears thrice in the latter half of the first century, 15 times in the second. Square sigma, 7 times in the second century, once probably in the first.

²² Each arm of the cross which surmounts the inscription branches into 5 points. Height of cross, 0·175 m.; of letters, from 0·04 to 0·056 m. Forms of letters as in no. 3 (b).

²³ L. 9: after ΚΑΙΣΑΡΙ two faint letters NA of the same size and character as the main inscription may be accidental markings on the stone. L. 10: Φλα(υία) Πα- is difficult, an upright immediately preceding the final alpha being unexplained. This stroke may be accidental; otherwise Φλαο[υ]ία [Πα-] should be read.

²⁴ No. 5 is an excellent example of the square alphabet not infrequent in Cyprus in the second century. This alphabet was in common use at Gerasa in the first century, is apparently of Syrian origin, and is found in its purest form at Dura-Europos and at Susa (Bradford Welles, Gerasa, City of the Decapolis, Inscriptions, pp. 358–360). Height of letters, lines 1 to 11, 0·03 to 0·038; 1. 12, 0·032 to 0·042 m. They are freely telescoped, e.g. in δημαρχικῆς; in Ἀδιαβηνικοῦ; in ἀνέστησεν; in Θεοῦ; in μητρόπολις; in Νέρουα; in ἀνθυπάτω; in ᾿Αντωνίνου; in ἐγγόνω.

²⁵ Cf. C. Müller, Cl. Ptolemaei geographia p. 958. The missing name between Paphos and Soli is clearly Arsinoe.

²⁶ The Roman road did not, therefore, serve the town on Cape Drepano, of which the extensive ruins, dating to the Roman period, are described by Hogarth, o. c. pp. 10 ff.

²⁷ A glance at any modern map will show the road from below Kouklia to Akhelia as straight as a die. If this line be continued, it would appear that the ancient road left Paphos on the landward side.

²⁸ In l. 2 of the Latin text, AB of ‘ab<n>[epote]’ is followed by what looks like a modern capital L; of this the lower stroke may be accidental. In l. 9. ‘B[ass]ou[m]’ is very indistinct, so that the spelling of the name is not certain. In both the Latin and the Greek, the name of Geta has been obliterated. In l. 9 of the Greek, it is interesting to find the abbreviation Ṁ = υί(λια), for this does not occur again in Cyprus in the text itself. A certain part of the Greek I was unable to reach with a squeeze; hence the last lacuna in l. 6, the lacunae of lines 7, 10, and 11, and the second half of that in 1. 8. Both for the Latin and the Greek the text is generally bad. I have not thought it necessary to point out each case of blurred or weathered lettering.

²⁹ Height of Latin letters, with V some-what shorter, 0·025 to 0·04 m. Height of Greek letters, for the last line, as high as 0·052; for the rest, from 0·024 to 0·04 m. Punctuation seems to have been general in the Latin after abbreviation. Ligatures are frequent in the Greek, much less so in the Latin. Thus the second hasta of N becomes T in Antonini. The Greek examples are much more varied. In Γερμανικοῦ of 1. 1, my and alpha are telescoped; in υἱωνῶ of 1. 2 ny is simply a slanting stroke joining the two omegas; eta in ΆΑδιαβηνικῶ a horizontal stroke between beta and my. Other ligatures are: eta with my and with rho; kappa with eta; ny with epsilon and with tau; omega with ny.

Whether this inscription and LBW 2806, which stood 2 Roman miles to the east, were cut by the same hand is doubtful. Although several of the same ligatures occur in the Greek of both inscriptions, the forms of A and alpha differ considerably; such solecisms as ‘tribouniciai potestate’ are found in both, but against this should be set omissions and divergences which are equally significant.

³⁰ Of lines 1 and 2, nothing is visible. Towards the middle of l. 5, are deep and uncouth markings, reminiscent of FT, ET; these are not helpful and are probably in the main accidental. L. 7: [NOB]BCAESS, of which the last two letters are difficult, surviving only as two uprights.

³¹ Height from 0·046 to 0·035 m. Letters narrow and straggling.