At his Bad Astronomy blog on July 10, 2018 Phil Plait posted about Science, humanity, and the necessity of making mistakes.
That post was based on his TEDx Boulder talk, Failing upwards: science learns by making mistakes. It’s a great 12-minute presentation on how scientific research is a very human activity.
One example from metallurgy is the allotropes of iron found on the iron-carbon phase diagram. At room temperature pure iron is alpha (ferrite), with a body centered cubic crystal structure. At higher temperature it changes to gamma (austenite) with a face centered cubic structure. And at even higher temperatures it changes to delta (ferrite), again with a body centered cubic crystal structure. But beta is missing.
Why isn’t there a beta iron anymore? The phase behavior or iron was first investigated by thermal analysis. That was back before we knew how to use x-ray diffraction to determine crystal structures, or about ferromagnetism and that it disappears at the Curie temperature, and can only be understood using quantum physics (as discussed in a chapter of the Feynman lectures). Beta iron just was the paramagnetic form of alpha iron, above the Curie temperature. It’s not really a different phase, so it got removed from the diagram.