Rothamsted's Classical Experiments

Guide to the Classical and other Long-term Experiments, Datasets and Sample Archive

Soon after John Lawes started his collaboration with Henry Gilbert, they began a series of field experiments to examine the effects of inorganic fertilisers and organic manures on the nutrition and yield of a number of important crops. Seven of the experiments that were started in the middle of the 19^th century are still going today. Most of them have undergone changes to maintain their scientific and agricultural relevance but they retain many of the original treatments and have an unbroken history that extends over more than 150 years. These experiments are a valuable resource and are used for a wide variety of scientific observations. Soil and crop samples from the experiments are added to the Rothamsted archive every year.

We call these old experiments the 'Classicals' and they are:

Broadbalk - started in 1843
Hoos Barley - started in 1852
Alternate Wheat and Fallow - started in 1856
Exhaustion Land - started in 1856
Park Grass - started in 1856
Barnfield - started in 1843
Garden Clover - started in 1854
Broadbalk & Geescroft Wildernesses 
Other long term experiments 

Harvesting Broadbalk

Broadbalk

The Broadbalk experiment had its first winter-wheat crop sown in autumn 1843, and this crop has been sown and harvested on all, or part, of the field every year since then. The experiment tests the effects of various combinations of inorganic fertiliser (supplying the elements N, P, K, Na and Mg) and farmyard manure on the yield of wheat: a control strip has received no fertiliser or manure since 1843. Originally the weeds were controlled by hand weeding but later by periodically bare-fallowing and cultivating different parts of the field in different years.

From the mid-1950s, herbicides have been used but they are withheld from one part of the field. Two major modifications were made from 1968.One was the introduction of modern, short-strawed cultivars. The second saw crops other than wheat being grown on the experiment, so that yields of wheat grown continuously could be compared to those of wheat grown in rotation.

To accommodate this change, the experiment was divided into 10 sections; four continued in wheat whilst six were used to compare two 3-course rotations. There have since been further modifications and we now have: two sections growing continuous winter wheat; one section growing continuous winter wheat where the straw is chopped and incorporated into the soil (on other sections, the wheat straw is baled and removed); one section in continuous winter wheat where no herbicides have ever been applied (on other sections, herbicides are applied routinely); one section in continuous winter wheat where since 1985 the use of pesticides has been restricted; and five sections testing the rotation oats, forage maize, wheat, wheat, wheat.

Broadbalk

Lawes and Gilbert installed a tile drain under each treatment strip and used these to collect and measure the nutrients in the water that leached through the soil into the drains. After 150 years, many of these drains had collapsed and in 1993 one section was re-drained so that water leaching through the soil could again be collected and analysed.

Uses of Broadbalk

To study plants grown in the field deficient or well supplied with nutrients

To monitor long-term changes in soil organic matter

Soil-borne and foliar fungal diseases

Soil microbial studies, Insect pests and nematodes

Entomopathogenic fungi

Annual variations in yields and relation to weather

Comparison of continuous cropping of wheat and wheat with a 2 year break

Nitrogen cycling and leaching

Ammonia fluxes, Gaseous exchanges at the soil surface, P & K leaching

S, Mg and Ca studies

Soil hydrology

Comparison of old and new varieties of wheat

Cultivation studies using dynamometer and plough

Arable weeds in the absence and presence of herbicides since 1957

Genetic variation of weed species between plots

Nitrogen fixation from leguminous weeds

¹⁵N has been used to look at nitrogen utilisation and residues from crops

Radiometric and chlorophyll meter calibrations

Aerial reflectance measurement to indicate nutrient status of the crop

To observe the regeneration of plants in an area last cropped in 1882 and later divided into untouched, topped annually and grazed areas. (Wilderness)

Aggregate stability and carbon fractionation in Wilderness soils

Archived and modern samples have been used to study atmospheric deposition of NH₄, NO₃, SO_2, cadmium, lead, sulphur, poly-aromatic hydrocarbons, poly-chlorinated bi-phenols and radioactive isotopes from thermonuclear explosions

The roll of glutathione transferases in metabolising herbicides

Incorporation of straw on one section since 1987 has allowed comparisons to be made where straw is removed

Long runs of data used to validate computer simulations of field processes over year-to-century time span

Future work already planned is to monitor carbon cycles using the different carbon profiles of ¹³C from the maize and wheat

Quantify dissolved organic nitrogen and trace elements in the drainage water

XRF analysis of archived soils to measure accumulations from the environment of 60 elements

To assess the contribution of long-term, continued application of organic and inorganic fertilizers on the formation of pedogenic secondary ferrimagnetic minerals

Deuterium and oxygen isotope ratios in plant material for palaeoclimatic reconstruction.

Back to the list of experiments

Hoos Barley

Hoos Barley has tested the effects of different fertilisers (N, P, K and Mg) and farmyard manure on the yields of spring barley since 1852. The experiment offers interesting contrasts to Broadbalk; for example, because it is spring-sown it has only been necessary to fallow it four times to control weeds and it showed that P is a more important nutrient for the spring-sown crop than for winter wheat.

As with Broadbalk, modifications were made to the experiment in 1968. Modern barley cultivars were introduced and a 3-course rotation was started on part of the experiment. This was discontinued in 1979 because the effects of a two-year break were small. One further change in 1968 was the introduction of different rates of fertiliser N; these rates were superimposed on treatments already receiving farmyard manure and different combinations of PKMg fertiliser. Unlike Broadbalk where the same amount of N is applied year-after-year to the same plot, on the Hoos Barley experiment the different rates of N are rotated. In 1862, a test of sodium silicate was introduced, originally to see what effect this had on the strength of the straw, which was used in large quantities in the local hat factories at that time. Although the effects on straw strength were contrary to what was expected i.e. more silicon made the straw more brittle, it was also found that the silicate increases yields on plots not given P, apparently because it increases the availability of the small amounts of P in these plots.

Back to the list of experiments

Alternate Wheat and Fallow

This experiment was set up in 1856, on a site that had been without manures since 1851. The experiment originally had two strips which alternated winter wheat and a bare fallow in successive years. Between 1932 and 1982, a modification was introduced to allow a yearly comparison of a one-year and a three-year fallow but the effects on the yield of the subsequent wheat crop were small and since 1983 the experiment has reverted to the original design. No fertilisers or manures are applied but, like most of the other arable experiments, chalk is applied to maintain the soil at about pH 7.

Back to the list of experiments

Exhaustion Land

This experiment has had several distinct phases since it started in 1856. Between 1856 and 1901, different amounts of P and K were applied, initially as fertilisers to wheat but from 1876 as either fertilisers or farmyard manure to continuous potatoes. From 1902 to 1939 no treatments were applied and, with a few exceptions, cereals (usually spring barley) were grown. From 1940 to 1985, spring barley was grown and nitrogen applied to all plots every year, initially at a single rate but from 1976 the experiment tested four rates. The nitrogen increased yields and also revealed the benefits to the crop of P and K residues remaining in the soil in some of the plots from fertilisers and manures applied prior to 1901. Since 1986, half of the experiment has received annual dressings of N and P and the value of K residues continues to be measured. The other half receives annual dressings of N and K and following recent applications of different rates of fertiliser P to some sub-plots it is now used to study the effect on yield of different concentrations of available P in the soil. Wheat has been grown since 1992.

Back to the list of experiments

Park Grass

The Park Grass experiment has been described as one of the most important in the world in the area of bio-diversity and bio-ecology. This experiment started in 1856, on a field that had been in pasture for at least a century. Various combinations of inorganic fertilisers (P, K, Mg, Na, nitrate-N, ammonium-N and Si) have been tested since the start; organic manures (farmyard manure and fishmeal) have been tested since 1905.

In 1903 most plots were halved and the effects of regular liming tested. This was modified in 1965 with the division of most plots into four sub-plots, three of which are limed to maintain pHs of 5, 6 and 7. The fourth sub-plot receives no lime and the pH of these ranges from 3.5 to 5.7 depending on the fertiliser treatment. The plots are cut each year for hay, usually in June, and a second cut is taken in the autumn.

Dramatically different swards have evolved as a result of the different pHs and nutrient statuses of the soils. There are 50 - 60 species on the unfertilised plots but only 2 or 3 species on some of the fertilised plots. Since 1990, nitrogen fertiliser has been withheld from half of all sub-plots formerly receiving 96 kg N ha^-1 as either ammonium sulphate or sodium nitrate to study processes controlling soil acidification, heavy-metal mobilisation and botanical changes.

150 years of the Park Grass experiment

Click here for more information about the public event

Click here for more information about the scientific meeting

Back to the list of experiments

Harvesting Park Grass in 1941

Barnfield

The oldest of the Classical experiments, treatments were first applied to Barnfield in the spring of 1843 for a crop of turnips. The experiment underwent a number of changes until 1876 when a period of continuous cropping, with mangolds, was started. Because the yields of the root crops grown continuously declined, the cropping was further modified from 1959 to include a range of arable crops and grass. Since 1975 grass and clover has been grown on the site.

Back to the list of experiments

Garden Clover

This is the smallest and simplest of the experiments started by Lawes and Gilbert and until 1956 it was just a single unmanured plot measuring 3m x 2.1m. Although red clover is a perennial crop, Lawes and Gilbert found that on the Rothamsted farm it seldom survived for more than a year or two, for reasons which they did not understand. In 1854 they established this small plot in Lawes' garden. For the first 10 years yields were excellent much to Lawes' and Gilbert's surprise and even for the next 30 years they were respectable. However, yield later declined probably, in part, as a result of clover rot. From 1976, a variety resistant to clover-rot has been grown and, in the period 1980-82, the largest yields in the history of the experiment were obtained.

Back to the list of experiments

Broadbalk and Geescroft Wildernesses

In 1882, about 0.2ha at one end of the Broadbalk experiment was fenced-off and the crop left unharvested. The wheat was allowed to regenerate and left to compete with the weeds. After only four years, the few wheat plants that could be found were stunted and barely recognisable as cultivated wheat. Half of this area has been left completely untouched and has been naturally colonised by trees, mainly ash and sycamore, with a hawthorn understorey and ground ivy in the middle. The other half has had the trees and shrubs removed by grubbing, and, therefore, developed open-ground vegetation containing, for example, coarse grasses, hogweeds and willowherbs. In 1957, this grubbed section was divided into two parts, one remaining grubbed but the other being mown and then grazed by stock to encourage pasture species.

On Geescroft field a much larger area, 1.3ha, was also abandoned in the 1880s after attempts to grow continuous field beans failed. This area was much wetter than Broadbalk and, although trees and shrubs took longer to establish, it is now a mature deciduous woodland dominated by oak and ash with a rapidly expanding understorey of holly leading to otherwise poor ground cover. A major difference between the two wilderness areas is that chalk was applied in large quantities in the 18^th and early 19^th century to the Broadbalk site; its pH is still slightly alkaline. Geescroft had never received large dressings of chalk and, although it was neutral when it was abandoned, it is now very acid.

Back to the list of experiments