4.2. The Role of Simulations

Our view of cosmic evolution is, like Darwinism, a compelling general scheme. As with Darwinism, how the whole process got started is still a mystery. But cosmology is simpler because, once the starting point is given, the gross features are predictable. The whole course of evolution isn't, as in biology, sensitive to ``accidents''. All large patches that start off the same way, end up statistically similar.

That's why simulations of structure formation are so important. These have achieved higher resolution, and incorporate gas dynamics and radiative effects as well as gravity. They show how density contrasts grow from small-amplitude beginnings; these lead, eventually, to bound gas clouds and to internal dissipation.

Things are then more problematical. We're baffled by the details of star formation now, even in the Orion Nebula. What chance is there, then, of understanding the first generation of stars, and the associated feedback effects? In CDM-type models, the very first stars form at redshifts of 10-20 when the background radiation provides a heat bath of up to 50 degrees, and there are no heavy elements. There may be no magnetic fields, and this also may affect the initial mass function. We also need to know the efficiency of star formation, and how it depends on the depth of the potential wells in the first structures.

Because these problems are too daunting to simulate ab initio, we depend on parameter-fitting guided by observations. And the spectacular recent progress from 10-metre class ground based telescopes and the HST has been extraordinarily important here.