You should spike the sample. Your protein is found only in the sample. So your spiked control will be most meaningful when you put it there. Spiking the tracer will likely work too, but why bother with the "likely"? You do this to study what's in the sample, so that's where it goes for best comparability. You don't need to care about what would happen if it went anywhere else.

It's easiest to use a highly concentrated stock, so that the volume change of the sample becomes negligible. For example, a 1% volume change is usually going to be smaller than the spec'ed inaccuracy of your pipettes. And the accuracy of the entire ELISA process is going to be much worse than that. So anything smaller than 1% is certainly negligible.

Alternatively, you can mathematically correct for the volume increase after the fact, by multiplying your final result in ng/ul by the (spiked / non-spiked) volume ratio.

Hi John Schloendorn

Thanks for your answer. I will try to measure the histamine levels in plasma samples from rats. To validate the ELISA assay I will dilute my plasma 5x, 10x, 20x and so on. In addition to the diluted samples I will spike 25 ng/ml (final conc of Histamine). I had also thought about spiking blanks with the same amount to see if buffer composition has an matrix effect.

According to your answer I think I will spike my sample with 50 ng/ml so the final concentration will be 25 ng/ml in the well.

The concentration of my stock is 4000 ng/ml so I need to dilute this 80x to get a concentration of 50 ng/ml (this will mean I need to add 1,25 ul to my sample). When I add 50 ul of tracer and 50 ul of antibody my final concentration will be 25 ng/ml right?