When did the lunar-inspired subdivisions of months - the kalends, nones, and ides - and the system of counting down to them, fall into disuse, to be replaced with the modern system of counting up across the period of the whole month?
When did the system of holding the "bisextile" day for 48 hours fall into total disuse, to be fully replaced by the insertion of an additional day numbered 29 at the end of the Feb month?
I've been researching over the past few years the evolution of the practices of keeping a civil calendar, from the "Roman Republican" calendar, to the Julian, to the Gregorian.
Needless to say, the pre-Julian calendar is difficult to fully characterise.
One clear point appears to be that the calendrical "month" has its origin in tracking the phase of the moon, but the only hard evidence remaining of Roman calendrical practices dates from later years of the Republic when any correspondence with the observable lunar phase must have been lost (at least anything beyond a 'stopped clock being right twice a day' sort of correspondence).
The evidence appears to suggest that the original length of the Roman Republican calendrical year was not 365 days, but generally somewhat less, and so the correspondence between the months of the calendar and the seasons was originally maintained on an ad-hoc basis by the intercalation of whole months, according to the judgment of Roman clergy.
But indeed by time of the Julian reform, any correspondence with the solar seasons had also been lost - firstly on account of political meddling in the maintenance of the calendar (as political offices were renewed annually, as reckoned by the calendar), and finally a large amount of foreign war which led to intercalations being neglected.
This led to the "years of confusion" which culminated in the Julian rationalisation.
One could speculate that lunar calendars are of most practical use to sea-faring cultures, as the moon regulates tidal waters and night-time visibility, whereas solar calendars are of primary interest to agrarian cultures as the sun regulates temperatures and growing conditions.
Since the pre-Julian calendar had lost relation to both moon and finally also the solar seasons, one can speculate that by its final stages it was existing purely to coordinate civil and military activity taking place predominantly on dry land.
This is reinforced by accounts which suggest it was seen as "unlucky" to intercalate the calendar in years of major wars - probably on account of the difficulty of communicating the change to the active soldiery located at a distance from Rome, and the confusion and incoordination that would ensue from intercalation during military campaigns (or only shortly before they were expected).
The Julian reform commencing in January 45BC introduced a fixed 365-day calendar year, and restored the traditional correspondence between particular months and the solar seasons. But it also retained lunar trappings, not only in its retention of the "month" period itself, but also in the continued subdivision of its months into three periods of kalends, nones, and ides.
The last day of each of these periods represented in principle the new phase, waxing quarter phase, and full phase of moon, respectively. But even following the Julian reform, the lengths of the Julian months, and their subdivisions, meant there was no intended correspondence to the lunar reality.
Each individual day of the calendar was reckoned as a count down to the final day in each of these lunar-inspired periods.
I would assume, speculating again, that the primary practical reason for the retention of these lunar-inspired periods, was to work within the calendrical concepts that were already traditional, and maintain (at least as much as possible) the placement and reckoning of birthdays, annual feasts and holidays, and other civil activities which the pre-Julian calendar was regulating.
Another innovation of the Julian reform was the introduction of a "bisextile" day every four years. This is the analogue of what nowadays is the Feb 29 leap day.
But originally this "bisextile" day occurred on Mar Kal 6 (the equivalent to the modern Feb 24), and was reckoned not by the insertion of an additional day into the calendar as such, but by holding the calendar at Mar Kal 6 for 48 hours (or two 'natural' days). Hence the name "bisextile" - loosely, the "doubled sixth" day.
It should be noticed that the "first" day, in this down-counting scheme, falls later, in this case being Mar Kal 1 (equivalent to the modern Mar 1).
I'm also not exactly representing above how the Romans spoke of these days - they apparently spoke of "ante diem 3 Kal Mar" (3rd day before, or Mar Kal 3), "pridie Kal Mar" (the day before, or Mar Kal 2), and "Kal Mar" (the day of the kalends, or Mar Kal 1) - but using a conceptually-consistent representation of the sequence which I at least find more consistent and less confusing.
My understanding is that at some later point, it became conventional to explicitly distinguish between the first and second natural day of the bisextile calendar day, although when this system was still in use, the two days were still reckoned as the first and second days of Feb 24 (or Mar Kal 6).
If it's pertinent, I'm primarily interested in when these changes became established in England (or later, the UK), but I'm equally interested in where the changed practices may have first originated and spread.