furnish no more than an approach to probability, not historical certainty. If we would separate these from history, and present it merely as a connected narrative, new views must not venture to obtrude themselves, and the research must be unprofitable; or else the historian be compelled to repose upon them at every emergency, yet without escaping universal censure, for giving his own hypothesis as historical truth. The history of early Rome can only be a union of narrative with criticism.

CHAPTER XIX.

ON THE SECULAR CYCLE.

THE Roman year, before the Julian Reformation of the calendar, is known to have been a lunar one, which, by the addition of an intercalary month, was brought, or rather intended to be brought, to a coincidence with the solar year. The great Joseph Scaliger, with that acuteness of perception which transforms into sources of truth, testimonies apparently without any bearing, has discovered the original system of this computation with incontrovertible certainty. He has proved that it was a trieterical intercalation in periods of twenty-two years, to each of which was added, ten times, an intercalary month, alternately varying from twentytwo to twenty-three days,-the last trieteris being passed over. As five years constitute a lustrum, so five of these periods form a Sæculum of 110 years 49

The supposition that Italy lay in barbarism, and

49 De Emendat. Temporum, p. 180. et seq.

that Rome derived her science from her connection with Greece, is utterly removed, if we consider how this clear and regular computation had, from the very period of that connection, fallen into such complete oblivion, that Cæsar was compelled to borrow from foreign science an inferior reformation of the calendar; and that the year then differed by an excess of about sixty-seven days from the true time. This confusion may have been very early occasioned by the general ignorance of mathematics and astronomy; the deductions of which, but not the sciences themselves, had been imparted to the Romans by the Etruscans. But it became an habitual and progressive evil, through the shameful dishonesty of the Pontifices, who, taking upon them an arbitrary intercalation, occasionally favoured the consuls and farmers-general by lengthening the year, or vexed them by its curtailment.

It is generally known, that according to the unanimous testimonies of the earliest and most credible Roman Archæologists, the year of Romulus consisted only of ten months, or 304 days. Amongst the mass of proofs, it will be sufficient to refer upon this point to Censorinus, who states the number of days in each month 50. This year which, individually considered, coincides neither with the moon nor the sun, appeared so absurd to those who were accustomed only to the views of the Greeks and of later ages, that Plutarch almost doubts whether it could ever have existed; and what is more surprising,

50 De Die Natali, c. 20.

Scaliger altogether denies it as a fable; and agreeing with Licinius Macer and Fenestella, neither of whom understood the subject, he takes for granted that the Roman year had from the first consisted of twelve months 51. But besides those statements, which, if there be left any foundation for history, cannot by any means be rejected, being narratives superior in accuracy to almost any from the earliest ages, we meet with unquestionable proofs that this year was at one time in actual use; and more than one visible trace of its application in later times, when its origin had already fallen into oblivion. We discover also, from the cyclical relation of this year to the lunar intercalary year, and its sæcular period as explained by Scaliger, in what manner this sæcular year might be used, partly as a continued correction, together with the lunar cycle,-partly as preferable to it for scientific purposes.

The first key to the intelligence of this system, is found in a passage of Censorinus; in which he says, that the lustrum was the great year of the early Romans, or the cycle, in which the beginning of the civil was again brought to coincide with that of the solar year.

It is perfectly evident, that Censorinus confounds the lustrum of his times (the Capitoline Pentaeteris) as the Greeks did their Olympiads with the ancient Lustrum, in respect of its duration; but his misconceiving the sense of the ancient accounts, does not diminish their value, nor their utility, whenever the

VOL. Í.

5 De Emendat. Tempor. p. 173.

P

corruption is so glaringly apparent as in this instance 52

Five Egyptian solar years, of 365 days each, contain 1825 days. Six Romulian years, of 304 days, only 1824. The Roman computation thus lost in five years a whole day, compared with the Ægyptian civil computation, which had no leap year, but with the loss of one in 1461 years, (as the circumnavigators of the globe lose one day), returned again to its original commencement. Compared with the Julian emendation, the loss was within nearly two days and a quarter. This certainly would be so great a variation, that if other divisions of time, clearly of the same system to which the year of ten months belongs, did not offer a systematic intercalation with a clearness and harmony, amounting to intrinsic evidence, its use for the cycle might indeed be rejected as mere hypothesis.

These divisions are, the greatest and the smallest Etruscan periods-the sæculum and the week of eight days. The former was also the measure of the lunar intercalary cycle, the latter was so far retained by the Romans, that every ninth day (Nundina) was a market-day. This day, amongst the Tuscans, was also termed Nonæ; and it was in reference to this division of time, that the ninth day before the Ides continued to retain that name.

But

52 Censorinus de Die Nat. c. 18. Those readers whom Scaliger's demonstration on this subject, and the fact that the duration of the lustrum was five civil years, should fail to convince, I beg leave to refer to more detailed observations on the regulations of the Censor.