When you apply two drugs to a system, you can get either the same response as the sum of the two drugs individually (additive), a greater response (synergistic) or a smaller response (if one drug blocks the effects of the other). How do you analyze your data to figure out which of these is the case? Sounds like an easy question.  But in fact it is pretty tricky. This brief article (by Greco and colleagues) gives some perspective.

The term "additive" is, in fact, a slippery term with multiple definitions. For this article, we'll use the definition of Bliss. This is appropriate when both drugs act on the same system (at least down stream) so the maximum response provoked by both drugs is the same. The rule is that the fractional response of a combination of two drugs (assuming Bliss's independence) equals the sum of the two fractional responses minus their product.

Think of it this way. The fractional response to drug A alone at any particular dose is  Fa. Similarly, the fracitonal response of drug B alone is Fb. But what is the additional response of drug B once A is already present? The additional response to drug B is the fraction Fb times the remaining possible response, which is 1-Fa, So the additional response due to drug B, in the presence of drug A equals Fb*(1-Fa). Therefore, the total response to a mixture of the two drugs is Fa+Fb(1-Fa) which equals Fa+Fb-Fa*Fb. This equation assumes that the effects of the two drugs are additive.

We'll assume that you apply the two drugs in 1:1 ratio. Here is a user-defined model, written for Prism, that can fit those three curves:

FA=1/(1+10^((LogEC50A-X)*HillSlope)) 
FB=1/(1+10^((LogEC50B-X)*HillSlope))  
 Y=FA 
Y=FB 
Y=FA + FB - FA*FB   

The first line defines the first dose reponse curve, with its own logEC50 and slope. In this example, all three curves are assumed to have a baseline of zero and a top plateau of 1.0 (the data have been normalized to fractional response). The second row defines the second dose-response curve. The third line tells Prism that for dataset A, Y is defined as Fa (defined in the first row). The next line defines the model for dataset B. The last line, defines the Y (response) for dataset C according to the Bliss independence rule.