In July, I promised to redo the measles arithmetic when the new kindergarten numbers came in. They came in on August 17, and I'll give them to you straight before getting to the stranger story underneath them. For the 2025-2026 school year, CDC data show coverage for the measles, mumps, and rubella vaccine slipped again, to 92.4 percent, while vaccine exemptions reached a record 4.2 percent, rising in 41 states and Washington, D.C. Only ten states now sit above the 95 percent threshold for MMR, down from 19 in 2019-2020. About 280,000 kindergartners started school without documentation of both measles doses. Meanwhile the CDC had counted 3,471 confirmed cases this year as of September 17, the most since 1991, against 2,289 for all of last year.
So the pockets I described in July are still growing. But there's a question hiding inside those numbers that arithmetic can't answer, and it's the one that will decide whether the United States officially loses its measles elimination status. The question is: are all these outbreaks one fire, or many?
What "endemic" actually means
The Pan American Health Organization's verification commission, which will review the US status in November, uses a precise definition. Endemic transmission means uninterrupted circulation of the same virus genotype and lineage for 12 months or more. Notice what that rules in and out. A thousand cases from fifty separate sparks imported from abroad is a terrible year, but it isn't endemic measles. A single chain that keeps burning for a year, hopping from state to state, is endemic, even if it's smaller.
The large Texas outbreak, which PAHO dates to January 20, 2025, is the obvious place to look. If that chain never truly ended, and the virus simply traveled onward into later outbreaks, then the clock ran out long ago. If instead the later outbreaks were sparked by fresh importations, the clock reset each time. Both stories produce the same pins on a map. To tell them apart, you have to read the virus itself.
Reading a family tree
Here's the mechanism. Every time measles copies itself, there's a small chance a letter in its genetic code changes. The measles genome is relatively stable, so these changes are rare, but they're inherited. A virus passes its quirks down to every infection it causes, the way a misspelling copied into one manuscript gets copied into every manuscript made from it. Compare enough samples and you can draw a family tree: which viruses are close cousins, which ones branched apart long ago.
For decades, public health labs didn't need the whole genome to do this. According to the CDC, they sequenced a small stretch called N450, about 450 letters out of a much longer code, and that was enough to separate one strain from another. Then something happened that I didn't expect when I started reading. In 1990, 18 genotypes of measles circulated worldwide. Since 2021, only two have been detected anywhere: D8 and B3. Vaccination worked so well that it pruned the family tree down to almost nothing. That's wonderful news, with a strange side effect. When nearly every case looks identical in that small stretch, the CDC explains, it becomes extremely difficult to tell whether outbreaks are connected through domestic spread or are separate importations. Success made the detective work harder.
The fix is whole genome sequencing, which reads every letter instead of a sample. Tiny differences that the short stretch misses can show up elsewhere in the code. But the CDC adds a caveat I want to underline, because it's the kind of caution this column respects: the genome alone can show with certainty that two cases are not related, but it cannot prove that two cases are linked. Genes can break a family tree apart. To join two branches, you still need old-fashioned epidemiology: who traveled where, who sat in which waiting room.
The first big chapter
On September 21, the CDC's journal Emerging Infectious Diseases published one of the most detailed looks yet. Scientists from the Texas Department of State Health Services and the Children's Hospital of Philadelphia sequenced 491 samples from the 2025 Texas outbreak and got usable whole genomes from 359 people. Within Texas, the genomes were close kin: pairs differed by between zero and 15 letters, with a median of just three. Compared with the rest of the world, the Texas viruses clustered with genomes from Utah and the Netherlands collected around the same time. Arizona and Utah genomes were related to Texas but sat in mixed positions on the tree. South Carolina's genomes formed what the authors called a distinct sublineage, related to the others but distinguishable from them.
What the paper did not do is announce a verdict. The authors wrote that the analysis "prevented definitive inference of geographic origin or transmission direction," and they noted that multiple introductions couldn't be ruled out. That is not a dodge. It's exactly what the method can and can't support. The study documents the expansion of a closely related D8 lineage within the broader North American family. Whether that family is one unbroken chain on American soil is the question the November review has to weigh, alongside the epidemiology.
The tree and the soil
I find two things about this genuinely moving. The first is how much patient work hides behind one line on a news ticker: hundreds of samples, each one a person with a fever, each one turned into a string of letters and placed on a tree. The second is that the genes and the kindergarten numbers are telling the same story from opposite ends of time. The family tree is a record of the past, of where the virus has already been. The coverage figures are a forecast, of where it can go next. A lineage only stays unbroken if it keeps finding unprotected people, and 280,000 kindergartners without documented protection is a lot of room for a family tree to grow.
So whatever the panel concludes in November, the decision that matters most is the one that comes after. A ruling can tell us whether measles has been living here for a year. It can't stop the next branch. Only the soil can do that, meaning the immunity gaps the virus needs to keep growing. As always, thresholds don't negotiate. I'll be back when the verdict is in.
