The seasons begin at the time of the solstice or equinox.

How it works:
During a local news broadcast, they said that winter begins in December. I prefer to think that the Winter Solstice is in December, and winter itself started a month and half earlier. We will need to have a few definitions here, so hang on!

First off, I am in the Northern hemisphere, so I will make all my descriptions using that reference. I apologize to my Southern readers; just reverse the references to the seasons (i.e., replace "winter" with "summer") and it should straighten things out.

The Earth orbits the Sun in one year (by definition!). The seasons are caused by the tilt of the Earth to the plane of its orbit. The Earth is tilted about 23 degrees, so that sometimes the North pole is tipped toward the Sun, and other times the South pole is pointed more toward the Sun. From here on the Earth, this means the Sun moves north and south about 46 degrees over the course of the year. In the summer, the Sun is very high in the sky, but in the winter it never gets as high; the difference is that very same 46 degrees (about 1/4 of the way around the visible part of the sky!).

Now imagine breaking up the Sun's 46 degree up-and-down yearly swing into 1 day intervals. Start at the top; then 3 months later it is midway between extremes and heading south, then 3 months after that it is at the lowest point (6 months after the peak). After that, it starts heading back up; three months later it is at the midway point on its way up. A full year after we start, it is back at the top of its journey. We have names for these four times; in order, they are the Summer Solstice, the Autumnal (or Fall) Equinox, the Winter Solstice, and the Spring (or Vernal) Equinox).

Now, bear with me: if summer started at the moment the Sun hit its peak, then every day of summer would mean the Sun gets a little lower in the sky. This to me doesn't make sense. You want the Sun to hit its peak at summer's midpoint, meaning summer would start 1.5 months earlier. That way the Sun would be climbing steadily higher until the peak of summer, then start to fall. This would in turn mean that the Sun hits its nadir (lowest point) in the middle of winter, and not at winter's beginning! See?

[Note (added July 31, 1998): As it happens, a lot of countries do actually think of the seasons this way; Japan for example. A lot of European countries do too; they have a Midsummers Day (made famous by William Shakespeare) on June 21 or so, and the days I claim should be the actual season starters are called "cross-quarter" days. I'll note that I didn't know this when I started the page. I'll also note that the email I get complaining about my opinion on this is overwhelmingly from people in the US, who also use weather as an argument. I specifically say here that I am avoiding using weather in my reasoning. Also, the email that supports my idea usually comes from outside the US boundaries.]

The length of the day depends on this effect as well. When the Sun is at the peak of its path, the day is at its longest (think of it this way: it has more distance to travel to get across the sky!). When the Sun is at its lowest, the day is shortest. It makes far more sense to me to have the longest day of the year at mid-summer, and not the first day, much as the shortest day should be mid-winter.

There is one major drawback with this definition of the seasons. Most people like to base definitions on something, well-- definite. That's probably the reason we define the seasons at such specific, easily measured dates. My definition would change that so that the mid-season is defined at these times, and thus the beginnings of the seasons would not have such a definite start and stop.

The point I am trying to make here is that the seasons are defined by an astronomical event-- the crossing of the Sun over the Celestial Equator (see below). This event does not really tie into the weather at all: by December where I live, things have been cold for over a month and will stay that way for at least two more! So I am trying to avoid the whole issue of weather and simply say that the definition of seasons -- while fine in and of itself-- could be tied into our daily lives better by using the length of the day as a guide.

This makes some people uncomfortable, as my email has been telling me! Still, I am not trying to change the world here-- just pointing out that maybe our current definition needs some background to explain why we chose it. Now mind you, I don't have a lot of scientific backing on this claim, just my own personal feelings on symmetry, beauty and common sense.

Bad Addendum (December 11, 1996): I have received quite a bit of mail about this. The messages tend to fall into two groups: one uses weather as a guide to the seasons, and the other talks about definitions. Let me make this clearer now: I am talking about the somewhat arbitrary definition of seasons, not the climatic effects. Temperature changes with seasons are highly latitude-dependent: someone in Ecuador, for example, sees little temperature change year round, while in Washington DC (trust me here!) the change is dramatic (well over 30 degrees Centigrade). So I am avoiding that issue completely. The actual definition of seasons is purely astronomical: the Equinoxes are defined as the two points on the sky where the path of the Earth around the Sun (the ecliptic) intersects the projected equator of the Earth on the sky (the Celestial Equator or CE). These two points are generically called nodes. As the Sun travels around the ecliptic it moves generally in an East/West direction, with some small North/South motion due to the inclination of the ecliptic to the CE. When the Sun hits the Vernal Equinox, it is moving North (on its way up in the sky as I discussed above). Six months later, it is moving South (on its way down) when it reaches the Autumnal Equinox. At the Summer Solstice the Sun has reached its highest point (the farthest North of the CE), and at the Winter Solstice its lowest point (the farthest South of the CE). That is how the seasons are actually defined! (phew!) I know this is confusing and awful without a diagram, but I don't have good software yet to make the diagrams I need for this page. I will include them as soon as I can draw them!