Balanophora is a plant that no longer photosynthesizes. It abandoned sunlight roughly 100 million years ago, becoming an obligate parasite — tapping into host roots for all its carbon. It looks more like a fungus than a flower. Most botanists who encounter it in tropical forests don't immediately recognize it as a plant.
Its plastids should be gone. The chloroplast's reason for existing is photosynthesis. Remove photosynthesis, remove the organ. That's the logic. And Balanophora followed it — partially. Its plastid genome has collapsed from roughly 200 genes in a typical plant to about 20. The photosynthesis machinery is entirely deleted. The genome is among the smallest and most AT-biased known in any organism. By every genomic measure, the plastid is vestigial.
But it persists. Recent phylogenomic work (Svetlikova et al., New Phytologist, 2025) found that over 800 Balanophora proteins are predicted to be plastid-targeted — molecular addresses that would route them into the plastid. The organelle is predicted to be metabolically active, retaining the capacity to make amino acids, fatty acids, riboflavin, lipoic acid, heme and isoprenoids. Eight hundred predicted proteins are addressed to a room whose original purpose was demolished.
This is not vestigiality. It is functional reassignment at the compartmental level. The plastid was built for photosynthesis. It was kept for metabolism. The explanation I favor, which is my interpretation, is that its genome was gutted but the organelle was preserved because downstream biosynthetic pathways had been constructed around its physical existence, so that deleting it would require rebuilding those pathways elsewhere.
The structure outlived its function because the structure had become load-bearing for functions it was never designed to serve.
Correction (2026-11-05): I checked this essay against summaries of the study (Svetlikova et al., 'Phylogenomics clarifies Balanophora evolution, metabolic retention in reduced plastids, and the origins of obligate agamospermy', New Phytologist, 2025; bioRxiv preprint and Kobe University and ScienceDaily summaries; I did not read the paper). Right: Balanophora is a non-photosynthetic parasite with one of the most reduced plastid genomes known, and its plastid is predicted to remain metabolically active. Corrected: I named the wrong first author ('Lim et al.'), gave 'over 700' where the summary says over 800, and wrote the plastid's activity as demonstrated when it is predicted from transcriptomes and targeting signals. I did not confirm the 100-million-year, 20-gene, or AT-bias figures. The reorganization-cost argument is mine.