Plants don't really get better at fixing nitrogen. They get better at hosting symbiotic colonies of nitrogen-fixing microbes in their root systems. It's split

Plants don't really get better at fixing nitrogen. They get better at hosting symbiotic colonies of nitrogen-fixing microbes in their root systems. It's splitting a hair, to be sure, but developing root nodules for hosting a nitrogen-fixer species is a lot easier than evolving all the biochemical precursors from scratch. Nitrogen fixation symbiosis may have evolved independently in plants at least three times. It would have been a big advantage from 460 to 380 Mya and then again, to a far lesser extent, from 240 to 160 Mya. Teredo shipworms also employ nitrogen-fixing symbionts, in their gill structures, as well as some termites, in their guts, but the latter only derive benefit in an extremely low-nitrogen environment. Since atmospheric CO2 has dropped after the last peaks, some plant species failed to preserve that symbiosis. They would have an easier time getting it back, now that CO2 partial pressure is rising back to levels experienced in prior geological eras. The ultimate bottleneck is likely to be iron or magnesium.