Hi,

fluorescence anisootropy can be negatine, if there is a large angle (like 90%, to be precise, > 54 deg) between transition dipole moments of absorption and fluorescence. This can happen if the compound is excited to the higher excited energy level (e.g. S2), then falls down nonradiatively to the first excited level (S1), and then fluoresces (S1 --> S0). If there is an angle between S0 --> S1 and S0 --> S2 absorption transitions, then (given that S0 --> S1 and S1 --> S0 are almost always parallel) you will have negative anisotropy. You can read more in famous book by Lakowicz.

To illustrate Mykhaylo's point look at this paper.

Another case of rising anisotropy is that of a mixture composed of a fast decaying, fast rotating major component mixed with a slow decaying, slow rotating minor component.

At short times, the anisotropy is dominated by the major component and decays. Then it rises again when the fluorescence from the slow component starts to dominate.

Remember that =(a1(t)I1(t)+a2(t)I2(t))/(I1(t)+I2(t))

Article Homo-FRET Microscopy in Living Cells to Measure Monomer-Dime...

In the frequency domain this effect is called the "Chip Dip" while in the time domain it is call "dip and rise".

Look near the end of the attached lecture.