In the early 1900s, scientists were puzzled by the behavior of bacteria. They seemed to pick up the traits of their fallen comrades, but that wasn't possible, right? In a simple experiment, Frederick Griffith showed that it was, through something he dubbed the "transforming principle."
To prove that bacteria could change themselves if exposed to better-adapted, if dead, compatriots, Griffith grabbed a bunch of mice and two different strains of bacteria. Both were of the pneumococcus strain — Griffith studied bacterial pneumonia — but only one had a smooth casing that left it untouched by its host's immune system. The other was rough, and could be recognized and taken out by the immune system. The first was deadly. The second was not.
Griffith used heat to kill off the deadly strain, and injected it into a mouse; the mouse went on with its mousy business, perfectly healthy. He injected the nondeadly strain into the mouse, so that the mouse was infected, but its immune system fought the infection off. He then mixed the heat-killed deadly strain and the live nondeadly strain, and injected the combination into the mouse. The mouse keeled over and died. Afterwards, Griffiths isolated the nondeadly strain, but also the deadly one, from the mouse's body. The deadly bacteria hadn't resurrected itself — the nondeadly bacteria had somehow taken on the properties of the deadly one, acquiring both the casing and the deadliness.
Griffith pointed to the experiment as evidence of some mysterious "transforming" or "transformative" property. Griffith's experiment was performed in 1928, and so it was only much later that that transforming property was shown to be DNA. Although the bacteria was killed, the DNA stayed alive. The nondeadly strain of pneumococcus acquired the gene for the casing from its fallen brethren, and essentially became the more deadly bacteria.