Varying absorbance with varying volume in 384 well plate - (Jan/08/2015 )

A friend of mine is testing some 384 well plates on a spectrophotometer, but she saw something strange: With increasing volume in the well, the absorbance goes down with every sample she tested, even water.

You would expect no absorbance with just water, but in lower volumes (20 µl) there were OD measurements of + 0.5, while higher volumes (100 µl) showed the expected values of almost 0.

Anyone has an explanation?

This could be a result of bubbles in the well (not present at high volume, but maintained by surface tension at low?) too short a pathlength for the plate reader at low volume, or perhaps some interference caused by the meniscus? 

First, I would check the manufacturer specifications in case there is a recommended minimum volume for accurate readings.

Second, I'm not sure but isn't the beampath in plate readers vertical? If so, with such difference in volume you are changing considerably the path length on the sample, which according to the Beer-Lambert law will affect to the absorbance:
 

A=ebc

A is the absorbance of the sample (no units)
e is the molar absortivity with units of L mol-1 cm-1
b is the path length of the sample - that is, the path length of the cuvette in which the sample is contained. We will express this measurement in centimetres.
c is the concentration of the compound in solution, expressed in mol L-1

So, unless the instrument is capable of measuring the path length (height of sample) the volume will really matter and if it can, your volume could be below the instrument operation limits

Thanks for your responses. My friend already contacted the manufacturer and the volume was in the recommended range. The reader automatically adjusts the absorbance for the given volume (and thus path length), it does this too for 96 well plates without any problems. I'll mention her the possible bubble problem.

Thanks a lot!