The Park Grass Experiment is a long-term field experiment at Rothamsted Research in Harpenden, Hertfordshire, England. Established in 1856 by John Bennet Lawes and Joseph Henry Gilbert, it investigates how fertilizer treatments affect the yield and composition of permanent grassland. Originally designed to improve understanding of hay production, it became an important experiment in ecology, demonstrating how nutrient supply, soil acidity, and management can produce sharply different plant communities within a single field. Its continuing treatments, botanical records, and archived samples make it one of the world's longest-running ecological experiments.
Origins and purpose
Park Grass was established on existing grassland rather than on a newly sown mixture. Before the experiment began, the field had long been managed as pasture and meadow. The researchers divided it into plots receiving different combinations of nutrients, while retaining plots without fertilizer for comparison.
The initial agricultural questions concerned which nutrients limited hay yield and whether different fertilizers produced equivalent results. The experiment soon revealed another effect: treatments changed not only the amount of vegetation harvested but also the identities and relative abundance of the plants producing it.
This distinction became central to the experiment's scientific importance. An increase in agricultural output could coincide with a loss of biodiversity, while similar yields could arise from communities with quite different species composition.
Experimental design and management
Nutrient treatments
The experiment compares long-established fertilizer regimes involving nitrogen, phosphorus, potassium, and other mineral nutrients. Treatments differ in nutrient combinations, application rates, and, for nitrogen, chemical form. Some plots receive no nitrogen fertilizer, and some receive no fertilizer at all.
These comparisons help distinguish nutrient limitation from the wider ecological consequences of fertilization. A nutrient can increase plant growth directly while also altering competition, soil chemistry, and the cycling of other elements.
Not every treatment has remained unchanged since 1856. Additional divisions and deliberate changes have created further comparisons. Consequently, the treatment history of an individual plot or subplot is essential to interpreting its results.
Liming and soil acidity
Liming became an important part of the experiment because fertilizer regimes produced substantial differences in soil acidity. Ammonium-based nitrogen fertilizers, in particular, can acidify soil through processes associated with the nitrogen cycle, including nitrification and the loss of nitrate with accompanying positively charged ions.
Systematic lime treatments were introduced during the experiment's history. In 1965, many plots were subdivided to establish a more extensive comparison of liming regimes, including subplots managed toward different soil pH levels and subplots left unlimed.
Liming is therefore an experimental factor rather than simply routine maintenance. It allows researchers to investigate how much of a fertilizer's effect operates through nutrient enrichment and how much through acidification.
Harvesting and observations
The grassland is managed by cutting and removing vegetation, with a principal hay harvest and a later cut. Removing cut material exports nutrients and prevents the site from developing into scrub or woodland through ecological succession.
Measurements include harvested dry matter, the botanical composition of the vegetation, and soil and plant chemistry. The available records differ in frequency and method across the experiment's long history; not every variable has been measured annually since its establishment.
Principal ecological findings
Nutrient enrichment changes community composition
The plots demonstrate that plant communities are strongly shaped by nutrient availability. Fertilization can favor productive grasses that form a dense canopy, reducing the light available to shorter plants. Many broad-leaved herbs and smaller species become less abundant or disappear from heavily enriched vegetation.
This is not simply a direct toxic effect of fertilizer. It can result from altered competition: supplying a limiting nutrient increases growth, after which light or another resource becomes more limiting. Park Grass consequently provides evidence relevant to the competitive exclusion principle and species coexistence theory.
Different nutrient combinations produce different outcomes. Phosphorus and potassium supplied without nitrogen can favor legumes, which obtain nitrogen through associations with nitrogen-fixing microorganisms. Adding nitrogen can reduce this advantage and favor other plants, including members of the grass family.
Fertility and acidity have interacting effects
Some of the strongest differences among plots reflect the combined effects of nutrient supply and soil acidification. Acidic conditions alter nutrient availability and the chemical environment experienced by roots. Species differ in their tolerance of these conditions, so acidification acts as an environmental filter.
Limed and unlimed subplots show why fertilizer effects cannot always be attributed to nutrient supply alone. Adding lime can permit a broader range of plants to persist in a fertilized plot, but it does not necessarily eliminate competitive effects caused by high productivity.
The experiment therefore distinguishes two processes that often occur together: enrichment can change competition, while acidification changes which species can tolerate the soil.
Productivity and diversity are not interchangeable
Park Grass has contributed extensively to research on the relationship between productivity and species richness. Highly productive vegetation need not contain many species, and a species-rich community need not provide the largest hay yield.
The relationship is nevertheless conditional. Nutrient balance, acidity, management, and the species available to colonize a plot all affect the outcome. Park Grass does not establish a universal numerical relationship between yield and diversity.
Harvested hay yield must also be distinguished from total net primary production. Hay measurements record removed above-ground material, not all plant production, which also includes roots and material that dies or is consumed before harvesting.
Recovery and environmental change
Changes to selected treatments allow researchers to examine whether vegetation recovers after nutrient inputs are reduced or stopped. Recovery need not mirror the original response to fertilization. Residual nutrients, persistent acidity, established dominant plants, and limited seed arrival can delay the return of species.
These observations are relevant to ecological restoration because removing a pressure does not necessarily recreate the conditions that existed before it began. Restoring soil conditions and restoring a plant community are related but distinct processes.
The experiment also records responses to environmental influences shared across the site. Even plots receiving no fertilizer are exposed to atmospheric nutrient inputs, changing climate, and year-to-year weather variation. “Unfertilized” therefore does not mean isolated from human influence.
Research using Park Grass records has linked recent increases in plant diversity with declining atmospheric nitrogen deposition, while showing that responses depend on treatment history. Such findings complement the deliberate plot treatments by examining environmental changes that affect the entire experiment.
Evolutionary research
Park Grass has also been used to investigate evolution over very short distances. Adjacent plots can present sharply contrasting soil conditions, creating different selective environments despite their geographical proximity.
Research on sweet vernal grass, Anthoxanthum odoratum, has documented genetic differentiation associated with contrasting plot conditions, including tolerance of acidic soils and differences in flowering time. This makes the site important for studying local adaptation in the presence of gene flow.
The evolutionary significance is not that every plot contains a separate species. Rather, neighboring populations can become differentiated when environmental selection is strong, even where opportunities for pollen and seed movement remain.
Archives and scientific value
Rothamsted maintains historical records and archives of soil and plant material associated with its long-term experiments. Preserved material can be analysed using methods unavailable when it was collected, extending the questions that the experiment can address.
Park Grass combines several kinds of evidence:
- Persistent interventions, which reveal effects that may take decades to develop.
- Comparisons within one field, which reduce some differences in climate and geographical setting.
- Long observational records, which distinguish persistent changes from short-term fluctuations.
- Archived samples, which support retrospective chemical and biological analyses.
Its value lies partly in duration: soil properties, nutrient stores, and plant populations can respond too slowly for short experiments to capture their eventual trajectories.
Interpretation and limitations
Park Grass predates modern randomized, replicated experimental design. Many historical treatments were not independently replicated in the manner expected of a contemporary field trial. Repeated measurements through time add information, but they are not equivalent to independent replicate plots. Spatial variation, treatment history, and temporal dependence must therefore be considered in statistical analysis.
The site also represents a particular grassland, soil, climate, and cutting regime. Results cannot be transferred unchanged to grazed pasture, tropical grassland, or other ecosystems. Some long-running fertilizer treatments create conditions more extreme than those found in many ordinary grasslands.
Finally, harvesting is itself a continuing intervention. It removes nutrients, changes competition, and constrains vegetation development. Park Grass is best interpreted as an experiment on managed grassland, not as an untouched ecosystem or a complete model of every form of nutrient enrichment.