month of April, because five or six months before the time of the battle of Salamis, which was Boëdromion, August or September, B. C. 480. The march of Xerxes, then, could not have begun later than B. C. 481, exeunte, on the one hand, nor earlier than his sixth year ineunte, on the other. If so, his sixth year ineunte must have coincided at the latest with B. C. 481, exeunte: and therefore his first ineunte with B. C. 486, exeunte. But even this is too late a computation; for as the whole calculation is deduced from the time of the reduction of Egypt, unless that reduction took place at the very end of a year, neither could the first of Xerxes have begun at the very end of a year. It is much more probable that the reverse was the case; that his years bore date from the spring or summer quarter of some year: and therefore his first bearing date from that time B. C. 486, his sixth did so from the same time B. C. 481; and his seventh from the same time B. C. 480. The same conclusion follows if we reckon backward from another indubitable date, that of the battle of Marathon, the sixth of the on ". The arrival at Thermopyle would not be earlier than the beginning or the middle of July; about which time the Olympic festival, B. C. 480, must have been actually celebrated. The heralds, therefore, met Xerxes on their return about the end of May or the beginning of June: and they could not have been less than four or five months absent on their mission. They could not have been sent then later than the end of January or the beginning of February. If so, neither could Xerxes have arrived at Sardis Attic month Boëdromion*, later than the end of January or the beginning of February; nor consequently have set out from Susa later than the end of September or the beginning of October. * Plutarch, Operum ix. 420. De Herodoti Malignitate: vii. 378, 379. De gloria Atheniensium: Camillus, 19: dates the battle of Marathon on the sixth of Boëdromion, the battle of Salamis on the twentieth of Boëdromion. Cf. Polyænus, Strategematum iii. xi. 2. Yet Plutarch, De Gloria Atheniensium, loco citato, and Lysander, 15: dates the battle of n Herodotus, viii. 26. B. C. 490. Three years of renewed preparation followed on that defeat; in the fourth year Egypt revolted"; and in the fifth year Darius died. This fifth year being deduced from the close of the summer quarter B. C. 490, would begin with the same time B. C. 486: and Darius might die as soon after its beginning as we please. The beginning of the reign of Darius is necessarily connected with the length of the reign of Cyrus: and this is differently represented by ancient authorities; some putting it at twenty-nine years, others at thirty: which statements would obviously be consistent, if the one were understood of complete years, and the other of current. One thing is certain; according to Herodotus he died in the summer quarter of some year"; and according to Xenophon, in the spring 9. The reign of Cambyses too is reckoned by Herodotus at seven years complete, and five months of an eighth. It seems therefore a reasonable inference that both together they ought to be computed at thirty years plus seven, or twenty-nine plus eight; that is, thirty-seven years complete: whence if the one began at a certain time, Olympiad 55. 1. B. C. 559, the other expired at the same time, Olympiad 64. 2. B. C. 522. The reign of Cambyses was followed by that of Smerdis; a reign of seven months in durations; which beginning some time in the first half of B. C. 522, would not expire until the same time in the second. On this principle the reign of Darius might truly begin some time between the first and the sixth sacred month; and nearer perhaps to the latter than to the former; a conclusion which squares exactly with what has been already shewn : especially from Haggai and Zechariah. The order or succession of subsequent kings is not disturbed by this supposition that the reign of Xerxes began B. C. 486. To Darius and to Xerxes in conjunction the canon assigns fifty-seven years complete; which beginning B. C. 522, would expire B. C. 465. But it says nothing of the odd months of the reign of Artabanus, after the death of Xerxes; which yet, there is no more reason to suppose included in the last year of Xerxes, than in the first year of Artaxerxes. In this case the nominal Thoth of Artaxerxes might be really the Thoth of Artabanus; and coincident with the demise of Xerxes. But this Thoth is December 18, B. C. 465; seven months from which would bring us to June or July B. C. 464, as about the true Thoth of Artaxerxes which also will agree with what was established elsewhere t. The first of Darius, then, bearing date in, or not long before, the sixth month in the sacred year, that is, Elul, B. C. 522, his second would bear date from the same time, B. C. 521. The allusion therefore to seedtime, or to the future harvest, in the ninth month of this year, would be both possible and probable; for it would be just after the beginning of the sixth year of the sabbatic cycle-a year which the providence of God was already pledged to bless in a triple proportion to s Herodotus, iii. 67. t Dissertation xv. vol. ii. 16, 17. any other: and the appositeness of this very coincidence to the language of the prophet is no slight confirmation of the conclusion itself, that the second of Darius was such a year in particular. The very difficulty, then, suggested by these words, turns out, upon examination, to be in harmony with every other case; and a strong corroboration of the truth of our previous deductions. Nor is it the least satisfactory result of its solution, that it has been the means of establishing so clear an agreement between Herodotus, the oldest of profane historians, and the testimony of two contemporaries, Haggai and Zechariah; an agreement which, if it were necessary, ought to outweigh even the authority of the canon of Ptolemy. But the authority of this canon is in no danger from any such consent; since it does not profess to be minutely exact; and pays no attention to parts of years as such. The first Thoth of Xerxes, according to this canon, is made to have synchronised with December 23, B. C. 486"; between which, and the close of the thirty-sixth of Darius, if that was the Jewish Elul in the same year, the difference is only three or four months *. * The authority of the canon of Ptolemy is undoubtedly very great, and its general accuracy is undeniable. But with regard to the question at issue, no authority, which it can possess, ought to be considered greater than that of the testimony of contemporaries, Haggai and Zechariah: according to whom, if the second of Darius coincided with B. C. 520, B. C. 520 was no part of a sabbatic year. But if any of the years before it, B. C. 709-708, B. C. 590589, and still more, if any of the VOL. IV. years after it, B. C. 163—162, B. C. 135-134, B. C. 37-36, were sabbatic years, then B. C. 520-519, must have been so too: and these facts are so certain, that they may be assumed with confidence. We are reduced then to this dilemma; either of supposing that no sabbatic years were celebrated in the time of Haggai and Zechariah, as they were before and after it; or not by the same rule in their time, by which they had been celebrated before, and by which they were celebrated u Fasti Hellenici, cap. 5. 247. B. C. 521, there was an eclipse of the moon, for the meridian of Jerusalem, June 24, at 4. 27. in the morn after-or of supposing that the second of Darius did not coincide with B. C. 520, and if so, that neither did his first with B.C. 521. Which of these suppositions in itself is the more probable, no one can hesitate to decide a supposition too, supported by the evidence of the oldest of profane historians, and next to contemporaries, Haggai and Zechariah, the nearest to the times in question. pays A document, like the canon of Ptolemy, which follows an artificial rule in computing the lengths of its reigns; which no regard to fractions of years as such; which refers the beginnings and the ends of reigns to the Thoth of a year which was perpetually shifting backwards; it might be presumed, a priori, would be liable at least to trifling errors. The very principle of its reckoning supposes that the Thoth which it assigns to a particular reign is but an approximation to the truth; and that it can never be critically exact unless the Thoth of the year of Nabonassar, and the day of the king's accession, both fell out together. There might, then, upon its reckoning, be as much as a year's difference between the nominal Thoth and the true: which, in cases where strict exactness was requisite, would obviously be a source of mistake. advocates of the canon may say this difference can never exceed a year; but a year will often be the utmost wanted to reconcile things together which would otherwise be incongruous-as The the very subject under discussion is sufficient to prove. They may say also that it is always a year at the utmost in excess, and never in defect; that a certain king's reign might begin on, or after, but never could before, the nominal Thoth assigned to it: whereas we are supposing the true Thoth of a reign to have fallen out in B. C. 522, the nominal Thoth of which is placed in B. C. 521. It may be said, too, that these errors of excess, whether greater or less, are all admitted by the canon knowingly-whereas an error of defect, of whatsoever nature, would not be admitted except unknowingly. And is it impossible that an involuntary error might be committed? is it impossible that, at the distance of time when Ptolemy was compiling the canon, the exact day and month, when the reign of Cyrus, of Cambyses, or of Darius, actually began, might be unknown-and not within the power of any sagacity to determine? Now the precise truth, with respect to these points, would be necessary, in order to the solution of the problem whether the ninth Jewish month, in the second of Darius, was the ninth Jewish month B. C. 520, or not. The determination of the year in general, and that of the month, and much more of the day in the year in particular, when a given reign began, would be very different things; and the former might be sufficiently possible when the latter would be absolutely impracticable. The prin |