**The most counterintuitive finding is that the species with the smallest flowers, *Iris-setosa*, has the widest sepals.** The per-species summary (from `Iris_by_Species`) shows the pattern: - **Sepal width runs against every other measure.** Setosa averages 3.42 cm sepal width. Virginica averages 2.97 cm and versicolor 2.77 cm. - **Setosa is smallest on the other three measures.** - Sepal length: 5.01 cm, versus 5.94 cm (versicolor) and 6.59 cm (virginica). - Petal length: 1.46 cm, versus 4.26 cm and 5.55 cm. - Petal width: 0.24 cm, versus 1.33 cm and 2.03 cm. - **Sepal width does not rank the species by size.** You would expect a bigger flower to be bigger in every dimension. Sepal width ranks them setosa > virginica > versicolor, so the largest species (virginica) is not the widest. This has a practical consequence for separating the species. The petal ranges show setosa is completely separable on petals alone, because its maximum petal length (1.9 cm) is below versicolor's minimum (3.0 cm). Versicolor and virginica overlap on petal length (versicolor up to 5.1 cm, virginica from 4.5 cm). They also overlap on sepal length (4.9–7.0 cm versus 4.9–7.9 cm). Sepal width is the weakest separator, with its ranges overlapping heavily (versicolor 2.0–3.4 cm, virginica 2.2–3.8 cm). A single petal threshold can isolate setosa, but telling versicolor from virginica needs several features together.