Samples of all the experimental crops are brought to the laboratory. Weighed portions of each are partially dried and preserved for future reference or analysis. Duplicate weighed portions of each are dried at 100° C., the dry matter determined, and then burnt to ash. The quantities of ash are determined and recorded, the ashes themselves being preserved for reference or analysis. In a large proportion of the samples the total nitrogen is determined, and in some the amount existing as albuminoids, amides, and nitric acid. In selected cases, illustrating the influence of season, manures, exhaustion, etc., complete ash-analyses have been made, numbering in all more than 700. Also in selected cases, illustrating the influence of season and manuring, quantities of the experimentally grown wheat-grain have been sent to the mill, and the proportion and composition of the different mill-products has been determined In the sugar-beet, mangel-wurzel, turnips, and potatoes, the sugar in the juice has, in many cases, been determined by polariscope or by copper, or both. In the case of the experiments on the mixed herbage of permanent grass-land, besides the samples taken for the determination of the chemical composition (dry matter, ash, nitrogen, woody fiber, fatty matter, and composition of ash), carefully averaged samples have frequently been taken for the determination of the botanical composition.

Samples of the soils of most of the experimental plots have been taken from time to time, generally to the depth of nine, eighteen, and twenty-seven inches, and sometimes even to four times this depth. In this way more than fifteen hundred samples have been taken, submitted to partial mechanical separation, and portions of the sifted soil have been carefully prepared and preserved for analysis. In a large proportion of the samples the loss on drying at different temperatures, and at ignition, has been determined. In most, the nitrogen

1 Including 1 year fallow.

2 Including 1 year wheat and 5 years fallow.

3 Including 4 years fallow.

4 Including 2 years fallow.

5 Clover, 12 times sown, 8 yielding crops, but 4 of them very small, 1 year wheat, 5 years barley, 12 years fallow.

Including barley without manure 3 years (11th, 12th and 13th seasons).

determinable by burning with soda-lime has been estimated. In many, the carbon, and in some the nitrogen, as nitric acid, and the chlorine, have been determined.

Almost from the commencement of the experiments the rain-fall has been measured-for more than thirty years in a gauge of one-thousandth of an acre area, as well as in an ordinary small funnel-gauge of five inches diameter. From time to time the nitrogen as ammonia (and sometimes as nitric acid) has been determined in the rain-waters, also chlorine in many samples.

Three drain-gauges, for the determination of the quantity and composition of the water percolating, respectively, through twenty inches, forty inches, and sixty inches depth of soil (with its subsoil in natural state of consolidation), have also been constructed. Each of the differently manured plots of the permanent experimental wheat-field having a separate pipe-drain, the drainage waters have been, and are frequently, collected and analyzed.

For several years in succession experiments were made to determine the amount of water given off by plants during their growth. In this way various plants, including representatives of the gramineous, the leguminous, and other families, have been experimented upon; also evergreen and deciduous trees.

Experiments upon the feeding of animals were commenced in 1847, and have been continued at intervals up to the present time. The following points have been investigated:

1. The amount of food, and its several constituents, consumed in relation to a given live weight of animal within a given time.

2. The amount of food, and of its several constituents, consumed to produce a given amount of increase in live weight.

3. The proportion, and relative development, of the different organs or parts of different animals.

4. The proximate and ultimate composition of the animals, in different conditions as to age and fatness, and the probable composition of their increase in live weight during the fattening process.

5. The composition of the solid and liquid excreta (the manure) in relation to that of the food consumed.

6. The loss or expenditure of constituents by respiration and the cutaneous exhalations-that is, in the mere sustenance of the living meat-and-manuremaking machine.

Several hundred animals-oxen, sheep, and pigs-have been submitted to experiment. The amount, and the relative development, of the different organs and parts were determined in two calves, two heifers, fourteen bullocks, one lamb, two hundred and forty-nine sheep, and fifty-nine pigs. The percentages of water, mineral matter, fat, and nitrogenous substances were determined in certain separated parts, and in the entire bodies, of ten animals, namely: One calf, two oxen, one lamb, four sheep, and two pigs. Complete analyses of the ashes, respectively, of the entire carcasses, of the mixed internal and other "offal" parts, and of the entire bodies, of each of these ten animals, have also been made.

From the data provided as just described, as to the chemical composition of the different descriptions of animal, in different conditions as to age and fatness, the composition of the increase whilst fattening, and the relation of the constituents stored up in increase to those consumed in food, have been estimated. To ascertain the composition of the manure in relation to that of the food consumed, oxen, sheep, and pigs have been experimented upon. The

loss or expenditure of constituents, by respiration and the cutaneous exhalations, has not been determined directly, but only by difference-that is, by calculation, founded on the amounts of dry matter, ash, nitrogen, etc., in the food, and in the (increase) fæces and urine.

Independently of the points here enumerated, the results obtained have supplied data for the consideration of the following questions:

1. The characteristic demands of the animal body, for nitrogenous or nonnitrogenous constituents of food, in the exercise of muscular power.

2. The sources in the food of the fat produced in the animal body.

3. The comparative characters of animal and vegetable food in human dietaries.

Having given a brief outline of the scope and plan of the investigations that have been in progress at Rothamsted for so many years, I propose to draw my illustrations as to the character and significance of the results obtained, mainly from those relating to the growth of wheat for more than forty years in succession on the same land:

1. Without manure.

2. With farm-yard manure.

3. With a great variety of chemical manures, both individual constituents and mixtures.

Table I. gives the number of bushels of dressed grain per acre without manure, and with farm-yard manure, in each of the forty years, 1844 to 1883 inclusive; and on some of the artificially manured plots, mainly selected to illustrate the effects of exhaustion and of manure-residue. In most cases in this table, and in all cases in the subsequent tables, the results obtained on the artificially manured plots are only given for the last thirty-two of the forty years, as, during the first eight years the manures were not the same year after year on the same plot as they were subsequently.

TABLE I.-Wheat grown for forty years in succession on the same land, Broadbalk Field, Rothamsted. Results showing the effects of exhaustion, and of manure-residue. Quantities per acre. Produce-Dressed Grain in bushels.

FIRST.-WITHOUT MANure.

After a five-course rotation since manuring (turnips, barley, peas, wheat, oats), the first experimental wheat crop was harvested in 1844. The highest yield of the series was 234 bushels in 1845, and the lowest was 4 bushels in 1879. Other yields have been 214 bushels in 1854, 20 in 1857, only 53 in 1853, and only 8-9 bushels in 1867, 1875, 1876, and 1877.

In the lower division of the table (I.) the average produce is given for each four years, each eight years, each sixteen years, and for the thirty-two years from 1852 to 1883 inclusive; also for the whole period of forty years. Without manure, the average annual produce over the four-year periods was 14, 17, 148, 12, 13, 10§, 84, and 12 bushels; over the eight-year periods, 161, 131, 121, and 10; over the sixteen-year periods, 143 and 113; over the thirtytwo years, 131, and over the forty years, 14 bushels. With such wide variations due to season, it is very difficult to estimate the rate of decline due to exhaustion. Excluding the very bad seasons, the decline due to gradual exhaustion is reckoned at from one-fourth to one-third of a bushel per acre per annum. It is estimated that over a period of thirty years the unmanured plot yielded an average of 18.6 lbs. of nitrogen per acre per annum in the crop, and lost a minimum of 10.3 lbs. in drainage, in all 28.9 lbs. ; whilst on the mixed mineral manure plot (5), it is estimated that the crop removed an average of 20.3 lbs. of nitrogen, and that at least 12 lbs. were lost by drainage, or in total 32.3 lbs. Further it is estimated that the soils lost to the depth of twenty-seven inches about two-thirds of these amounts, leaving, say, 10 lbs. more or less to be otherwise accounted for. Of this, the rain, etc., would supply 5 lbs, or perhaps rather more, and the seed about 2 lbs., so that there is but little to be provided from all other sources. Lastly, as at the commencement the soil was, agriculturally speaking, exhausted, the nitrogen supplied by it would be largely due to old accumulations.

SECOND.-FARM-YARD MANURE EVERY YEAR.

In the application of farm-yard manure every constituent is supplied in excess. The highest yields of the series of years were 44 bushels in 1863, 414 in 1868, 41 in 1857, and 41 in 1854. The lowest yields were 16 bushels in 1879, 19 in 1853, 20 in 1844, 237 in 1876, and 244 in 1877.

The average produce per acre per annum over each of the five eight-year periods was 28,34, 35, 351, and 283 bushels. Excluding the first eight years, and several of the recent very bad seasons, the average produce is about 35 bushels per acre per annum.

On the farm-yard manure plot, the first nine inches of soil show a great accumulation; it is nearly twice as rich in nitrogen as any other plot, yet this richness is not proof against bad seasons, nor are the highest amounts of produce in the field obtained on this plot.

Thus, without manure, or with mineral manure alone, there is a gradual decline in yield, and with this a marked reduction in the nitrogen of the soil. With farm-yard manure, on the other hand, there is great accumulation, and yet not the fullest crops, a large proportion of the constituents becoming very slowly available.

The next question is, which constituents of farm-yard manure are the most effective for wheat in this agriculturally exhausted rather heavy soil, with a raw clay subsoil. The first illustrations on this point will be drawn from Table II.