James Lind tested six scurvy cures side-by-side on matched sailors

Scurvy killed more British sailors on long voyages than combat, storms and every other cause combined, and had done so for centuries by the time James Lind served as a naval surgeon aboard HMS Salisbury in 1747. Medical authorities disagreed on the cause and cure, and citrus fruit was just one of several remedies with anecdotal backers — cider, seawater, vinegar, diluted sulfuric acid and various tonics all had their own defenders, and no one had ever tested them against each other under matched conditions.

The standard practice was to pick a theory of the disease and treat accordingly, with no controlled comparison possible on land or at sea; a ship's surgeon simply had no way to isolate which of several confounded variables — diet, climate, treatment, individual constitution — was doing the work, so competing remedies persisted indefinitely on reputation rather than evidence.

Lind selected twelve sailors with comparably severe scurvy, matched them into six pairs, and gave each pair one of six remedies then in use — cider, elixir of vitriol, vinegar, seawater, a garlic-and-mustard paste, or two oranges and a lemon daily — while holding their housing and base diet identical, isolating the treatment as close to the only varying factor.

Matched sailors into pairs, held housing and diet identical, gave each pair one remedy at once, so any consistent difference could only be the treatment.

Matching sailors on severity and holding diet/housing constant made treatment the only systematic difference.

Simultaneous testing eliminated time and order effects that plagued sequential trials.

The stark contrast in outcomes made the citrus effect unambiguous despite the small sample.

The design produced evidence strong enough to challenge centuries of anecdotal reputation.

When multiple fixes compete, test them simultaneously on matched subjects to isolate the real effect from confounds.

Within six days the citrus pair had one sailor fit for duty and the other markedly improved; the cider pair improved slightly; none of the other four pairs improved at all — a clear, reproducible result from a trial of just twelve men.