**The most counterintuitive finding is that the smallest-flowered species, *Iris-setosa*, has the widest sepals.** This comes from the `Iris_by_Species` gold table. - **Sepal width runs opposite to every other size measure.** - *Setosa* has the lowest average sepal length (5.01 cm) and the smallest petals (length 1.46 cm, width 0.24 cm). - It also has the highest average sepal width (3.42 cm). - *Versicolor* has the narrowest sepals (2.77 cm), and *virginica* sits in between (2.97 cm). - Sepal length, petal length and petal width all rise from setosa to versicolor to virginica. Sepal width does not, so "bigger flower means bigger in every dimension" is false here. - **One petal measurement separates setosa completely, while sepals barely separate anything.** - Setosa's petal length tops out at 1.9 cm, while versicolor's bottom is 3.0 cm. That is a clear gap, so a single petal cutoff identifies setosa perfectly. - Sepal measurements overlap heavily. Sepal length ranges are 4.3–5.8 for setosa, 4.9–7.0 for versicolor and 4.9–7.9 for virginica. Sepal width ranges are 2.3–4.4, 2.0–3.4 and 2.2–3.8, which overlap almost completely. - Versicolor and virginica do overlap on petals (versicolor petal length maxes at 5.1 cm against virginica's minimum of 4.5 cm, and petal width 1.8 against 1.4). Telling those two apart needs a multivariate model rather than one threshold. **Practical takeaway:** sepal width is almost useless as a "size" proxy. It carries signal only alongside the other measurements, and petal dimensions do most of the discriminating work.