I would not expect a linearity of response with sample size. We are not in a transparent solution of substrate and fluorescein. The interactions may even be quite different from substrate to substrate. The theory behind may be very complex and cannot be generalized for any substrate. Trust your established calibration curve and use it only in the region where the calibration was established and tested.

Pragmatically I would say it's better to use an experimentally established calibration line even then when we do not understand the theory for its shape, than a straight line your empirical system does not like..... Use the test and share the results with colleagues. One day you may find out that your samples behaved correctly and NATURE was right and you learnt something about of how to bend an expected straight line.

Reto J. Strasser

Prof.Emeritus Geneva  Switzerland

Chair Prof. Nanjing China

Prof. extraordinarius Potchefstroom South Africa                                                                                                  

Thanks Dr Strasser for taking time to answer me.

My calibration curve is nothing else than a colorimetric scale done with fluorescein solution (not FDA)... is it the good way to do (as in Adam and Duncan 2001)?

Thanks,