> I was told by our school electrician that long ago the clocks were adjusted by the power company. So around 2am DST they would either send power at 59Hz* to fall back or 61Hz* to spring forward. This was from the era when power grids were separate and analog.

That wouldn't have worked.

Grid-tied synchronous clocks (https://en.wikipedia.org/wiki/Electric_clock) were once ubiquitous and did derive their timekeeping stability from the grid frequency, but that is only suitable for keeping time once set. It works because electric grids are frequency-controlled and stabilized, where deviations from the reference frequency are corrected over long intervals.

In principle you could send power at 59 or 61Hz to adjust these clocks, but there are two problems:

1. The grid frequency isn't arbitrary, it's also directly coupled to the rotary inertia of every turbine that powers the grid. This is why power fluctuations cause frequency changes: temporary oversupply or undersupply causes these turbines to slow down or speed up respectively, using their inertia as power storage. Potentially screwing up every other piece of mechanical equipment tied to the grid for the sake of clock changes is... a poor optimization.

2. A one-Hertz difference isn't enough to change the clocks in a timely way. To change time by an hour you need to gain or lose 3600 * 60 = 216k cycles, but a 1Hz change gains or loses one cycle per second. It'd take about 2.5 days to complete a one-hour change via this mechanism.

I don’t know about that, but I was told by someone who worked at a power generation plant in the 1970s that they had a clock in the plant, that they would compare to the time broadcast on the radio frequency that carried the official time, and any time it diverged they would slightly adjust the frequency of the alternating current until it matched.