Contrary to what I have long believed, the spring equinox is Not the time after which days become longer than nights

Sun rising due east: Note*

I have always supposed that on the day of the spring equinox (i.e. yesterday) the name and night both became 12 hours long - or, at least, it was the day in which they were most-equal; so that the next day was the first on which the day was longer than the night.

But this is not true; as I realized when I was checking the sunrise and sunset times. 

It turns-out that on 20 March, the day of the vernal equinox, there were 12 hours and 10 minutes of daylight; and the most equal date wrt. day-night was either the 17th (11:58 h day) or 18th (12:02 h day). 

So I was wrong about the definition of an equinox - which is actually when the sun rises in due-east and sets due-west.

(And also when the sun is positioned exactly above the equator, but that would never have been apparent to people up here at 55 degrees North.) 

When I think about it; I already knew about the east-west definition, but had assumed that this also meant that day and night were equal length - and the day-length aspect of it seemed to be more important. 

Furthermore, I'm pretty sure I have been told that the vernal equinox was the exact time of year when light began to overcome dark... A very poetic, not to say spiritual, conceit.

Well, I guess the timing is "close enough for government work"; but I am rather sad that it is not, after all, exact. 

And it emphasizes my cumulative recognition that exact astronomical measures are not humanly-valid segmentations of time - the reality is more like a phase or season; and we ought not to make as much of specific days as we typically do. 

*Stonehenge is Not aligned on the Spring Equinox