Long-term experiment:
Several agricultural field experiments have run for more than 100 years, but much shorter experiments may qualify as "long-term" in other disciplines. An experiment is "a set of actions and observations", implying that one or more treatments (fertilizer, subsidized school lunches, etc.) is imposed on the system under study. Long-term experiments therefore contrast with nonexperimental long-term studies in which manipulation of the system studied is impossible (Jupiter's Great Red Spot) or undesirable (field observations of chimpanzee behavior).
[edit] Physics
Visible sunspot measurements have been made since the 18th century.
The Oxford Electric Bell has been ringing at Oxford University since 1840, although there is some reason to believe it may be 15 years older.
The Beverly Clock at the University of Otago has been running since 1864.
The pitch drop experiment has been running at the University of Queensland since 1927.
Arthur Covington's microwave solar observatory has been running since 1947, currently at the Dominion Astrophysical Observatory.
[edit] Long-term agricultural field experiments
Long-term experiments test the sustainability of different farming practices, as measured by yield trends over decades. Examples include the Rothamsted Experimental Station (1843-present), the Morrow Plots (1876-present), Alabama's Old Rotation (1896-present), and the Haughley Experiment (1939-1972?).
Experiments at Rothamsted showed that "grain yields can be sustained (and even increased) for almost 150 years in monocultures of wheat and barley given organic or inorganic fertilizer annually" (data in Agronomy Journal 83:2-10). These results show that practices considered unsustainable by some advocates of sustainable agriculture may preserve "the ability of a farm to produce perpetually", at least under some circumstances. But even if crop diversity in space or time (crop rotation) and organic inputs are not always essential to sustainability, there is abundant evidence from Rothamsted and elsewhere that they are often beneficial.
The Haughley Experiment was noteworthy as a rare example of a long-term experiment in organic farming without external inputs of nutrients. After about 30 years, however, it was decided to start importing manure. There is some disagreement whether a "decline in relative yields from the organic section" was due to a depletion of soil nutrients (Agriculture Ecosystems and Environment 30:1-26).
Various short-term experiments have used legumes (in symbiosis with nitrogen-fixing rhizobia) as a nitrogen source, but good short-term yields do not prove the system is sustainable. The problem is that release of nitrogen from soil organic matter can make up any shortfall of nitrogen from legumes for a decade or more. The Old Rotation showed that nitrogen from legumes can balance N removed in a harvested crop over the long term. A key point is that the N in the legumes was not removed, as it would be with a soybean crop, but was plowed under as a green manure. In the Old Rotation, the green manure was grown during the winter to supply N to a summer crop (cotton); this would be less practical in colder climates.
Long-term agricultural experiments that have been started more recently include the Long-Term Research on Agricultural Systems experiments at UC Davis, started in 1993.
[edit] Long-term ecological experiments
The US National Science Foundation supports a number of long-term ecological experiments, mostly in ecosystems that are less directly impacted by humans than most agricultural ecosystems are. See LTER.
[edit] Long-term experiments in evolutionary biology
The experiments of Richard Lenski on evolution of E. coli have been underway for more than 20,000 generations. Experiments with the evolution of maize under artificial selection for oil and protein content (Crop Science 9:179-181) represent more years but far fewer generations (only 65).
| 
