**The most counterintuitive finding: setosa has the widest sepals but by far the smallest petals, so sepal size ranks the species in a different order than petal size does.** This comes from the per-species aggregates (gold table `Iris_by_Species`). - **Sepal width runs opposite to everything else.** Setosa averages 3.42 cm, versus 2.97 for virginica and 2.77 for versicolor. On every other measurement, setosa is the smallest of the three. - **Sepal length, petal length and petal width all rank the species the same way.** The order is setosa < versicolor < virginica. For example, average petal length is 1.46, 4.26 and 5.55 cm. - **Setosa is perfectly separable on petals.** Its largest petal length (1.9 cm) is below the smallest versicolor petal length (3.0 cm). Its largest petal width (0.6 cm) is below the smallest versicolor petal width (1.0 cm). A single petal threshold therefore isolates setosa with no errors. - **Versicolor and virginica overlap on every feature.** Petal length overlaps between 4.5 and 5.1 cm. Petal width overlaps between 1.4 and 1.8 cm. Sepal length overlaps between 4.9 and 7.0 cm. The versicolor sepal-width range (2.0–3.4 cm) sits almost entirely inside the virginica range (2.2–3.8 cm). Sepal dimensions, which look like the obvious features to measure, are the weakest separators. Petal dimensions carry most of the signal. A simple petal rule is enough to pick out setosa, but separating versicolor from virginica needs a multivariate model. The overlap ranges are min/max only. They show that the species overlap, not how many of the 50 specimens per species fall inside the overlap.