03/18/04: I have learned from Dr. Noe that the beam size issue has been resolved. It was a matter of selecting the correct Gaussian [exp(-2r^2/w^2) vs. exp (-r^2/2*w^2)] and using the appropriate width parameter, which is the full width at 60.6%, not the full width at half max, to match the size of the microscope objective. So the upshot is that the unexpanded beam (i.e., not going through the pair of spherical lenses) is the correct size, and thus it makes sense that an unexpanded beam would do a better job trapping. Trapping efficiency declines markedly once you go far from the size of peak efficiency in either direction, although it's worse for beams that are too small. Refer to Peter Amendola's site
for more details, or see his poster in the basement (S level) of the Physics building. Note: this is a classic problem with Gaussian optics, and I should have been more careful to keep track of factors of sqrt(2), e, etc.
03/21/04: After discussion with my husband, who has a background in biology, I learned that yeast cells have flagella at some stages of their life cycle. This might explain why they are so much more prone to moving - not only are they affected by Brownian motion, they are also self-propelled. I need to learn more about this to determine whether they have any mobility within a few minutes-a few hours of being stirred into lukewarm water (lukewarm because it's been heated briefly by the heat gun as part of sample preparation).