Thanks DJKeenan for your additional suggestions. Many constants used in the expression "ratio" are products of numbers known to many digits and experimental values known to a few digits, so it would be difficult to specify that kind of constants to say, 16 digits, right? For example, one variable used is sigma*2*Pi*f0*mu where sigma is an experimental value known to 3 digits, f0 is our variable (exactly specifiable?) and 2, Pi and mu could be specified to many digits. I will look up ScientificErrorAnalysis and Tolerances and see how they could be used. Thanks. -Dan

Thanks DJKeenan and JacquesC for your comments. I have re-derived the long expression "ratio" in analytical form and then used evalf[20] for numerical calculations, and found consistency between evalf[20] and plot. (as to whether 0.6666666666e-14 came from ((2/3)*10^(-14)), it probably did. "Ratio" came from electromagnetic wave computation in which there are constants that could originally be of rational form. In the new derivation of "ratio", I kept rational terms in rational form as long as possible before numerical evaluations.) Thanks again for all your help! -Dan

OK thanks everyone for pointing out the reason for the inconsistent numerical answers. Sorry again for the really long expression on my post. I am not too experienced in posting on this format. I will perform calculations using evalf[20] as suggested. Cheers, -Dan

I did try the Select Axes and then Properties. No help there.

Thanks for all your help! I did have in mind the removal of the assumption, so the suggested command assign(d,'d'): unassign('d'): will be quite useful! -Dan