First, a comment to T. Miyata. I think it is interesting and important to acknowledge the contribution of aeronautical engineers to early developments in aeroelasticity and non-steady-state aerodynamics. Here, two names come to mind: one is Prof. Wagner who was the first to introduce the convolution-type integrals presented here by Prof. Kareem, and the second name is Prof. Theodorsen, who applied the frequency domain to flutter analysis of shaped wings. There are a number of publications by NACA (the precursor to NASA) on the topics of non-steady aerodynamics and aeroelasticity. Second, a question to F. Ricciardelli. You mentioned that a better understanding of the aerodynamic forces leads to the optimization of the design. The question is how do you optimize the design? Is there a method for optimizing the design of the cross section or is this done by a trial and error approach? F. Ricciardelli, University of Reggio Calabria, Italy: Of course, it is a trial and error procedure; you do not have a closed form way to do that. What we did was test just one section and that gave us some results that lead us to the understanding of the behavior. Of course, if you test more sections not very different from one another, with just minor modifications, then you can understand what's going on and what is the effect of that minor modification to the aeroelastic and aerodynamic behavior of that section. So, for sure, we use trial and error, but the understanding of the mechanisms of the excitation may help us in finding an easier way to get a good solution. M.P. Bieniek (to N. Jones): In your equations you showed us 18 coefficients of derivatives. How many did you actually use in your calculations and how many did you actually determine in your section tests? N. Jones, Johns Hopkins University, USA (now University of Illinois at Urbana-Champaign, USA): We used nine in the calculations and we determined eight directly from dynamic wind tunnel tests and one indirectly from the static test using the pseudo-steady approximation for the P1 ร . So even though we have the whole 18 in the formulation, we are still not at the point where we routinely use all 18 of those. We actually do not think that they are all necessarily important. So having gone all the way up to 18, I think we are now in the process of coming back to fewer than 18, to the ones that are the most significant and influential in our analysis process. M.P. Bieniek (to A. Kareem): In your presentation of the time domain approach, you mentioned approximation to steady-state aerodynamics, am I correct? Are the error functions, functions of frequency?