All of the observations described previously make most sense if mitochondria evolved from free-living bacteria. But if the alternative — that mitochondria originated step by step inside the cell — is true, then there’s no reason to expect mitochondria to be passed on to offspring, to have DNA that codes for unique traits, and to have close bacterial relatives. To give the alternative a fair hearing, Margulis tried to imagine what expectations it generated — to see if it had any evidence supporting it. She reasoned that …
If mitochondria evolved step-by-step inside the cell, then we’d expect that:
There are organisms that preserve early stages of mitochondrial evolution — that contain “proto” mitochondria
Biologists know of many examples of transitional forms, living or extinct organisms that have “intermediate” structures that help us understand how major changes in the history of life happened. It seemed reasonable to suppose that if mitochondria evolved from another structure in the cell, we might be able to find some organism with “transitional” mitochondria — early evolutionary forms of mitochondria. However, try as they might, no scientist had (or has yet today) observed anything like this. All the cells known to science either contained full-blown mitochondria or none at all. This makes perfect sense if mitochondria evolved by endosymbiosis, but not if they evolved from another part of the cell.
Based on the available evidence, the accepted hypothesis did not look very compelling, and Margulis’ hypothesis looked reasonable, but there still wasn’t any smoking gun. Here’s a summary of all the lines of evidence discussed so far: