How the number is built
The team examined 3,722 Smilodon fatalis vertebrae in the La Brea Tar Pits Museum collection: 849 sacra, 2,130 lumbar vertebrae and 743 seventh cervical vertebrae. In three of them, belonging to three separate individuals from three separate pits, foraminal widening consistent with a nerve tumour was reported. The prevalence of 353 per 100,000 comes from dividing three by 849; the sacrum is the denominator because it is the element used to establish the minimum number of individuals in the collection.[1]
The comparison figure in humans is 0.22 to 0.38 per 100,000. The ratio is roughly a thousandfold, but the two numbers are not of the same kind. The human figure is an incidence in a living, diagnosed population; the fossil figure is a proportion among skeletal elements that were preserved, excavated and catalogued. The paper does not hide this: it states that the prevalence estimate reflects taphonomic bias rather than a true disease rate. Read without that sentence, the thousandfold stops being a number and becomes an impression.[1]
Where the diagnosis stops
Fossils preserve no soft tissue, and so no histological diagnosis. What is in hand is a skeletal marker consistent with a tumour rather than a tumour itself: a widening of the opening through which a nerve passes. The comparison the paper supplies shows that this marker carries different weight from species to species; in domestic dogs foraminal extension is common among spinal cases, while in cats it is very rare. That the same marker means different things in different species is exactly what demands care in the step from marker to disease rate.[1]
The less discussed side of the paper is in fact the stronger one. In the same collection, 32 vertebrae with incomplete arch development, 48 fused vertebrae and 226 transitional vertebrae were counted. These are congenital-type malformations counted against the same denominators, and there are hundreds of them rather than three. This is the leg that could carry a genetic story. The three vertebrae in the headline are the finding standing on the weakest denominator.[1]
How much of the claim stands
The authors' inbreeding suggestion is carefully framed: given the presumed inbreeding late in the species' existence, a similar mutation could have contributed, and they add that the level of inbreeding cannot be measured without ancient DNA. There is a serious alternative to set against it. Tar pits do not sample at random: an animal in pain and restricted in movement is more likely to be trapped. If that holds, the share of pathology in the collection comes out higher than in the living population, and nothing need have changed in the genetics. The team's own reading, that pain and difficulty hunting may have pushed the animals into riskier behaviour, requires the same mechanism to be read this time as a sampling bias.[1]
Two measurements would settle this, and both are reachable. The first is to recover ancient DNA from La Brea Smilodon specimens and estimate the level of inbreeding from runs of homozygosity; the inbreeding claim is testable only there. The second is to count the same malformations separately in early and late pits: since the three affected specimens span roughly 12,000 to 28,000 years, any rise towards extinction should show up between strata. If either comparison is published within two years the claim will stand or fall; until then what we hold is a hypothesis.[1]