Phage biology has a clean binary. Temperate phages integrate into the bacterial genome and wait — the lysogenic cycle. Lytic phages replicate inside the cell and kill it — the lytic cycle. Integration or destruction. Patience or violence. Every microbiology textbook draws the same forking diagram.
Two independent groups scanning bacterial genome assemblies found the diagram was wrong (Nature Microbiology, 2026). One team analyzed 3.6 million bacterial genomes from 1,226 species and found 119,510 lytic phage genomes inside bacteria that were alive when sequenced. The other focused on E. coli and found 373 virulent phage genomes persisting without killing their hosts. These are phages classified as obligately lytic — they lack integrases, they lack the molecular machinery for lysogeny. By every criterion in the textbook, they should have killed the cell. They didn't.
The mechanism is a third state: the carrier state, sometimes called pseudolysogeny. The phage genome exists inside the bacterium as an extrachromosomal element — not integrated into the chromosome (that would be lysogeny), not replicating to burst the cell (that would be lysis). It just sits there. The predator inside the prey, doing nothing. The association is especially common with jumbo phages — the Chimalliviridae family, phages so large they build a protein shell around their DNA inside the host cell, a structure called a phage nucleus.
The 119,510 genomes represent a five-fold increase in known phage-host associations. This isn't a rare curiosity. It's a major fraction of phage biology that was invisible because the detection methods assumed the binary. If you sequence a bacterial genome and find a phage, you check for integrases. If the phage has an integrase, it's temperate — expected. If it doesn't, it shouldn't be there. So you assume contamination, assembly error, or a recent infection caught mid-lysis. You explain it away because the framework has no category for it.
The binary was a classification artifact. The lytic-temperate distinction describes two endpoints of a continuum. Between them is a gradient of persistence strategies — carrier states, pseudolysogeny, chronic infection — where the phage maintains itself without committing to either integration or destruction. The binary persisted not because evidence contradicted it, but because evidence that contradicted it was systematically discarded as noise.
The general principle: when a classification has only two categories, observations that fit neither get assigned to whichever category they least violate — or discarded as error. The third option is invisible not because it's rare but because the framework has no name for it. Naming it — carrier state, pseudolysogeny — doesn't create it. It makes it countable. And once counted, it turns out to be everywhere.