Lower courts supposedly follow Supreme Court precedent — but they often don’t. Instead of adhering to the most persuasive interpretations of the Court’s opinions, lower courts often adopt narrower readings. For example, recent court-of-appeals decisions have narrowly interpreted the Court’s rulings on police searches, gun control, and campaign finance. This practice — which I call “narrowing from below” — challenges the authority of higher courts and can generate legal disuniformity. But it is also beneficial. Narrowing from below allows lower courts to update obsolete precedents, mitigate the harmful consequences of the Court’s errors, and enhance the transparency of their decision-making process. This Article contends that narrowing from below is generally legitimate when lower courts adopt reasonable readings of higher-court precedent, even though those readings are not the most persuasive ones available. This conclusion holds true — with some significant modifications — under multiple scholarly models of vertical stare decisis, including models that view higher-court rulings as legally authoritative, comparatively proficient, or usefully predictive. Understanding narrowing from below as a legitimate activity also points toward a new “signals” model of vertical stare decisis. Under this model, lower courts follow the Court’s relatively informal cues to resolve ambiguity in conventional precedent, including by narrowing from below.