Everything moved except the night
2026-08-26
Across two years of one child's hand-logged sleep, bedtime swung three hours down and two and a half back, naps went from five a day to one, and day sleep fell from six hours to two. Through all of it, the night barely moved: from month five onward, every monthly median of actual night sleep — wakings subtracted — landed between 9h17 and 10h57. Twenty-two months inside a band of an hour and a half, while everything a parent actually watches was reorganizing around it. I spent yesterday charting why bedtime falls to 19:00 and stopped one question short. The night was the other half, and I nearly didn't ask.
Where the night's ten hours came from
At one month, day and night were nearly equal — 5h58 against 7h57. By month five the night had climbed to over ten hours, and it gained almost all of that at the front: bedtime fell three hours in those months while the morning wake-up barely moved, from 7:00 to 6:23. The famous early-bedtime descent didn't relocate sleep, it converted it — day sleep collapsing at one end, night sleep absorbing about two of those hours at the other.
Then the direction reversed and the night still didn't budge. From month six to today, bedtime climbed back from 19:00 to 21:37 — and instead of shrinking, the night slid: the morning gave way at the same rate, 6:23 drifting to 8:10. A ten-hour block, moving through the clock as one piece. Weissbluth's claim that an earlier bedtime buys the same or a later morning, not an earlier one, is exactly what this looks like from the outside — though with everything changing along with age, my data can agree with him without being able to prove him.
The staircase underneath
The nap count is how the day paid for all this, and it moves in steps, not slides. Some steps are rigid to a degree I didn't expect from a small child: months eleven and twelve held exactly two naps on every one of their 61 days; from month twenty it has been exactly one on every day but a single napless afternoon just this week — a preview of the last transition of all, which the books park at three to five, mostly three to four. The transitions between steps are the opposite — month eight split perfectly in half, fifteen days on three naps, fifteen on two, and the two-to-one move took two full months of mixed days.
That last transition is also the only event in two years that shows up in the night itself: the single dip in the whole plateau — 9h17, the band's floor — falls exactly on the month the second nap started dying. One month later both had recovered. And the nap that survived has anchored: it starts at 13:21 on the median of the last sixty days and ends at 15:10, drifting about four minutes a month later as the morning stretch grows. The same force that is pushing bedtime up is pushing the nap, slowly, in front of it.
What this argues against
The reflex — mine included, and the entire nap-rescue industry's business model — is to treat every daytime change as a threat to the night: a dropped nap, a shrinking day total, a bedtime sliding somewhere new. In this data the night survived a five-to-one nap collapse, a four-hour day-sleep decline, and a three-hour bedtime round trip without giving up more than its transition-month wobble. The day is the negotiation. The night is the settlement. Nothing needed defending.
The caveats are the usual ones, stated plainly: one child; naps counted from a hand-kept journal, so an unlogged car nap is missing from the staircase and the count starts where the journal does (the earliest weeks exist only in a smart bassinet's record, which knows the nights but not the naps); and "night" here means logged wakings subtracted, which is why my ten hours reads about an hour under the Zurich in-bed percentiles for the same ages. Same direction as the measurement problem — if anything, the real night is steadier than my floor.