sigmoidal curve

There are lots of sigmoids -- the arctangent function, the logistics function, the CDF for every distribution, .... Got one in mind?

Or maybe you are thinking of something more periodic/sinusoidal, like the trigonometric sin() and cos() functions? For something like "hours of daylight over time", I would expect a generic sine function would work (one form might be y=A*sin(B*x+C)+D. A is amplitude, B is related to the period, C is the "phase shift" and D is the vertical offset).

http://people.hofstra.edu/Stefan_Waner/trig/trig1.html
http://en.wikipedia.org/wiki/Sine_wave

I think Mr. Shorty is probably correct.

There's a workbook at https://app.box.com/s/ereqysafe9cshbpuu5fg that will fit a sinusoid to your data.

Ok, so i've looked at it and is not what i was asking. I might be not explained myself out. I already have the data but i dont know how to represent a three years cycle of daytime (in hours of sun over that time) in a chart as a sinuosoid smoothed curve, like if it was a trendline, not following the datapoints at the exact value because is a progressive change overtime is not like every week it chances 30 minutes at once.... i like a all week long change. is that ok?
thanks anyway!

Post the workbook and show the curve you're trying to fit.

The hours of sunlight by day over the course of a year (or several years) is a perfect application for Fourier regression.