A fundamental question in biology is whether large-scale evolutionary patterns arise from the same processes that drive change within populations. Fly wing shape offers a striking test case: The directions of genetic variation and species divergence are closely aligned, yet wing shape evolves far more slowly than expected—a rate paradox under a simple constraint hypothesis. I show that this paradox can be explained by single-axis selection, which constrains other nonselected traits through pleiotropy. More broadly, the results suggest that multivariate constraint can emerge from selection on a single trait whenever mutations are pleiotropic within developmental modules. This implies that the alignment between micro- and macroevolution may often reflect simple, low-dimensional selection rather than complex adaptive landscapes. The alignment among mutational variance (M), standing genetic variance (G), and macroevolutionary divergence (R) in Drosophila wing shape poses a rate paradox under a simple constraint hypothesis: Evolution follows mutational lines of least resistance, yet proceeds orders of magnitude slower than the abundant genetic variation would permit. This is difficult to reconcile with a simple constraint view in which long-term evolution merely tracks the amount of available variation in each direction. Previous explanations invoke deleterious pleiotropy on unmeasured traits or correlational selection on trait combinations, but recent empirical work finds little evidence of fitness costs beyond flight performance. Here, by reanalyzing published data, I show that wing size shows the hallmark of the primary selection target: Among all wing traits, size exhibits the lowest ratio of standing genetic to mutational variance, indicating the strongest selective depletion. Based on this empirical observation, I develop a single-axis selection model in which natural selection targets only a single trait while all other traits evolve as correlated byproducts via within-module pleiotropy. This minimal model reproduces both the observed M–G–R alignment and slower-than-neutral divergence rates, explaining micro- and macroevolutionary patterns in fly wings without invoking complex adaptive landscapes.