In the second half of the 20th century, astronomers began to develop large computer programs. Scientists were starting to calculate things that they could neither see with instruments, nor understand through theoretical reasoning. The first great breakthroughs came in the 1950s and 1960s with discoveries about the evolution of stars. Common-sense reasoning, backed up by elaborate mathematics, had already given crude insights into how ordinary stars like our Sun changed over billions of years -- evolving into much larger and cooler “red giants,” which in turn evolved into “white dwarfs.” But true understanding came only when the ideas and mathematics were built into models calculated by digital computers. After billions of arithmetic operations, the computers printed out numbers that showed what happened inside a star at each stage of its development. The results were important for cosmology, for they gave reliable values for the ages of stars. Our Sun was about five billion years old, and others were two or perhaps three times as old.

No scientist could get an intuitive grasp of what was happening in these computer models, the way one might picture the workings of a mechanical clock. Gradually scientists realized that computer models offered an entirely new way to look at nature. This Web exhibit is displayed in two parallel streams, “Ideas” and “Tools,” to display the parallel growth of theory and instrumentation. Today a third, independent realm has joined them.

For example, in recent years several groups have built numerical models showing how gravity can pull galaxies together in clusters, and how the clusters themselves can map how matter was distributed at the very beginning. Repeat the calculation with a different number for the density of matter in the universe, and the results will show a tighter or looser pattern of galactic clusters. When the computers' results are compared with observations, we get important information about what kind of universe we actually live in.