^{page 73 note 1} Diuus Augustus, 5.

^{page 73 note 2} Ib., 94.

^{page 73 note 3} Pp. 114–9.

^{page 73 note 4} When Mr. Garrod says (p. 117) that ‘ the translation of pre-Julian into Julian dates in- volves problems of immense intricacy, upon which no two scholars agree,’ he goes too far. German scholars agree—and I agree with them —that 90 days were intercalated in 708, and that every recorded date between a. d. V. Kal. Apr., 696 (B. G., i, 6, § 4), and the commencement of the Julian calendar can be reduced, with a possible error of not more than one day, to its corresponding date in the Julian calendar.

^{page 73 note 5} I make this reservation because, as I shall presently show, Dr. Fotheringham himself has since rejected the most important part of that evidence. But in rejecting it he is almost alone.

^{page 74 note 1} See my Ancient Britain, p. 709, n. 2

^{page 74 note 2} P. 118.

^{page 74 note 3} Att., iv, 3, §§4–;5.

^{page 74 note 4} xl, 47, § 1.

^{page 74 note 5} Ancient Britain, p. 709, n. 2.

^{page 74 note 6} Cicero's letter also disposes of the theory of Stoffel (Hist, de Jules César, ii, 387), that between 57 and 52 B.C. there were two intercalary months. For if there were, the number of days that elapsed from November 21, 57, to the last day of December, 53, must have been 1,504, in which case three successive intercalary months would have consisted each of 23 days.

I do not anticipate that anyone will struggle to escape from the conclusion which has been drawn in the text from Cicero's letter by resorting to the desperate and still-born suggestion of De La Nauze (Mém. de litt de I'A cad. des Inscr., etc., xxvi, 1752–4 [1759], p. 259); for if he does, he will find himself between the Devil and the deep sea. De La Nauze argued that the eightday nundinal week belonged only to the Julian calendar, and that previously the week had consisted of nine days. Test this hypothesis by Dion's statement that January 1, 52 B.C., was a nundinal day, combined with his implied statement (xlviii, 33, § 4) that December 31, 41 B.C., was a nundinal day, and you will find that it will not work.

De la Nauze's suggestion was based upon a blunder of Macrobius (i, 16, § 34), the origin of which will be obvious to anyone who compares his statement with that of Varro (Rerum rust., ii, 1)

^{page 74 note 7} Georg., i, 33–5. Cf. Manilius, iv, 548. I owe these references to my friend, Mr. E. J. Webb.

^{page 74 note 8} Neue Jahrb. f. klass. Philologie, cxxix, 1884, pp. 569–72.

^{page 74 note 9} The alternative theory is not only irreconcilable with the testimony of Cicero, of Augustus himself (Mon. Ancyr., § 1, with which cf. Cicero [Att.t xvi, 8, § 1]), and of Dion, but it does not even save the credit of Suetonius; for, as Mr. Webb remarked to me, it compels its advocates to assume that when Suetonius (Diuus Augustus, 100) stated that Augustus died at the age of 76 years all but 35 days—a statement which is confirmed by Dion (lvi, 30, § 5), and which exactly tallies with the view that Augustus was born on September 23 of the Julian calendar— he made a miscalculation of about three months and stultified himself.

An attempt has, however, been made to explain away Suetonius's statement of Augustus's age by suggesting that it was based upon a computation of the civil years. It is true that after the Julian reform Cicero (Att., xiii, 42, § 2) gave the date of his own birth according to the unreformed calendar (cf. Att., vii, 5, § 3); but if Suetonius was thinking of the civil years, he not only took the unreformed date for the startingpoint of Augustus's life and the reformed date for its termination, but also perverted the unreformed date, reckoning a. d. VIIII. Kal. Oct,. as equivalent to September 23, whereas it was equivalent in the unreformed calendar to September 22. A mode of reckoning which can only justify itself by a self-contradictory double entendre is surely inadmissible.

^{page 75 note 1} Pp. 713–4.

^{page 75 note 2} xl, 47, § 1.

^{page 75 note 3} xlviii, 33, § 4.

^{page 75 note 4} P. 117.

^{page 76 note 1} See Ancient Britain, pp. 716–7.

^{page 76 note 2} Livy, as far as I know, only mentions three,— those of 189 B.C. (xxxvii, 59, § 1), 170 (xliii, 11 § 13), and 167 (xlv, 44, § 3).

^{page 76 note 3} Holzapfel argues, in the face of Dion's statement, that Caesar intended to make his first intercalation in 713; for he supposes that the Caesarian cycle did not begin until after the Terminalia of 709, that is to say, the sixth day before the Kalends of March; but he holds that the pontiffs, misunderstanding Caesar's regulations, made the first intercalation in 712. See Ancient Britain, pp. 717–23.

Solinus says (Collect, rerum memorabilium, i, 45–6) that ‘whereas it had been enjoined that they [the pontiffs] should intercalate one day in the fourth year, and this ordinance ought to have been carried out on the completion of the fourth year they intercalated at the beginning of the fourth year, not at the end ’ (nam cum praeceptum esset anno quarto ut intercalarent unum diem, et oporteret confecto quarto anno id obseruari illi incipiente quarto intercalarunt, non desinentc). If by ‘ the completion of the fourth year’ Solinus meant the fourth year of the Julian calendar, that is to say, 712, and if he had good authority for his statement, then Caesar intended that the first intercalation should take place in 713, after the error which needed correction had amounted to one whole day. But observe the looseness and inconsistency with which Solinus expresses himself. Immediately after saying that the inter calation ought to have taken place ‘ in the fourth year,’ he says that it ought to have taken place ‘on the completion of the fourth year.’ To state the facts correctly required extraordinary precision and nicety of expression; and this requirement he failed to satisfy. His meaning may have been that in whatever year the first intercalation took place, the second ought to have taken place four years later, and so on. This interpretation (though my argument is independent of it) is corroborated by the passage in Dion Cassius, to which Dr. Fotheringham refers. But the natural meaning of Solinus's words—with which compare Macrobius, i, 14, §§ 6, 13—seems to be that Caesar intended to make his first intercalation in 713, and that the pontiffs made it in 712. Anyhow he evidently means that an intercalation took place in one of the four years that preceded 713.

^{page 76 note 4} See Ancient Britain, pp. 718–9.

^{page 77 note 1} See p. 76, n. 3.

^{page 77 note 2} Ancient Britain, pp. 714–26.

^{page 77 note 3} Further proof is given in Ancient Britain, pp. 712–3

^{page 77 note 4} Einleitung in die Chronologic, i, 1889, p.75.

^{page 77 note 5} Dr. Fotheringham's words are:' I have abana doned the nundinal reckoning, and I now endorse Lersch's doubt about the continuity of this reckoning.'

^{page 78 note 1} I have not been able to see the second edition (1899), to which Dr. Fotheringham refers, of Lersch's work: it is not in the British Museum, and neither it nor the first edition is mentioned in Bursian's Jahresbericht.

I have heard the remark (traceable to Lersch ?) that Cicero's words, Ante diem X nundinae: contio biduo nulla (see p. 74, supra) suggest that the nundinal reckoning was not continuous. Why, I cannot understand. No doubt Cicero implies that, as a rule, contiones were not held on market-days. But can any instances be given of contiones held on days which in the ordinary course would have been market-days, and if so, should we be justified in concluding that they were not market-days, and therefore that the nundinal reckoning was disturbed ? I know of only one instance,— that of a market-day on which a contio was held (Att., i, 14, § 1). It would seem, then, that the rule was not invariable (see Th. Mommsen, Röm. Staatsrecht, i, 1887, p. 199, n. 3; iii, 1, pp. 372–3 and P. Huvelin, Essai hist, sur le droit des marchés etc., 1897, pp. 90–2). Besides, there is evidence that it had no effect upon the nundinal reckoning. Caesar is said to have enacted that no contio should be summoned on a nundinal day (Macrobius, i, 16, § 29); and all the extant calendars, in which the nundinal succession is absolutely continuous, belong to the period between 31 B.C. and A.D. 46.

^{page 78 note 2} There is a third instance if I am right in believing that the Lepidianus tumultus, of which Macrobius speaks (i, 13, § 17) occurred in 43 B.C. See Ancient Britain, pp. 719–21.

^{page 79 note 1} B. G., i, 6, § 4.

^{page 79 note 2} See Ancient Britain, pp. 707, 709–10.

^{page 79 note 3} See p. 74, n. 6, supra

^{page 79 note 4} Mém. de lift de 'Acad. des Inscr., etc., xxvi, 1752–4 (1759). P 249.

^{page 79 note 5} My calculation is confirmed by F. K. Ginzel's Handbuch der mathematischen und technis Chronologic, ii, 1911, pp. 272, 545

^{page 79 note 6} De la Nauze (op. cit., p. 249) implies that the vernal equinox would not have been favourable for the emigration. Perhaps it is unnecessary to notice such an objection. The Usipetes and Tencteri migrated into Gaul, with the wagons which housed them, in the depth of winter (B.G., iv, 1, § 1; 4, § 7; 14, § 4); the Bellovaci with five other tribes assembled and marched, accompanied by ‘ a great multitude of wagons’ (ib., viii, 14, § 2), in the depth of winter (tempore annidifflcillimo [ib., 6, § 1]); several other instances are to be found in the Commentaries; and Stonewall Jackson moved troops with numerous wagons in January, 1862, along ‘mountain roads,’ which were ‘little more than sheets of ice'’ (Stonewall Jackson, by Col. G. F. R. Henderson, i, 189S, p. 235). Surely, then, the Helvetii might reasonably expect that the 24th of March would not be too early.

^{page 79 note 7} Neue Jahrbucher fur Philologie, etc., cxxix, 1884, P 578

^{page 79 note 8} B.G., i, 5, § 3.

^{page 79 note 9} Ib., 10, § 1.

^{page 80 note 1} B.G., 11; 15, § 4.

^{page 80 note 2} In 1868 and again in 1896 the English harvest ‘was early in July!’ See Whitaker's Almanack1897, p. 597

^{page 80 note 3} B.G., i, 16, §§ 1–4.

^{page 80 note 4} Hist, de Jules César, ii, 57 n. 2, 95.

^{page 80 note 5} See Caesar's Conquest of Gaul, 1911, p. 635

^{page 80 note 6} Actually, in order to make this reduction we should have to assume that the troops marched more than 27 kilometres a day. The route which Napoleon traced for Caesar's march was partly wrong. Ocelum (B.G., i, 10, § 5) was not, as he supposed, Usseau, but hard by Avig liana, in the valley of the Dora Riparia (Caesar's Conquest of Gaul, 1911, pp. 430–2); and Caesar did not march to Lyons by way of Grenoble but past Embrun and Chorges (ib., pp. 613–6),— a longer distance

^{page 81 note 1} B.G., i, 7, § 6; 10, § 3. Napoleon omitted to take account of the time which elapsed between the Ides of April, when Caesar forbade the Helvetii to pass through the Roman Province, and the day when he started for Italy to fetch reinforcements.

^{page 81 note 2} See the preceding note.

^{page 81 note 3} More than 1,000 kilometres in 27 days. Deducting the 7 days in which they had to repel attacks, and therefore could hardly have covered more than 24 kilometres a day, they would have been obliged to march more than 840 kilometres in 20 days, or 42 kilometres a day, even if we allow not a single day for rest!

^{page 81 note 4} B.G., v, 23, §§ 2–5 point to the same conclusion In Ancient Britain (pp. 712–3) I showed that this passage, compared with Cicero, Att., iv, 18, § 5, and Q.fr., iii, 3,§ 1, cannot be explaine unless 90 days were intercalated in 708.