\i MB U. S. DEPARTMENT OF AGRICULTURE. OFFICE OF EXPERIMENT STATIONS-FARMERS' INSTITUTE LECTURE No. 6. K. C. TRUE, Director. SYLLABUS OF ILLUSTRATED LECTURE OX ESSENTIALS OP SUCCESSFUL FIELD EXPERIMENTATION. BY C. E. THORNE, M. S. A., Director Ohio Agricultural Experiment Station, Wooster, Ohio. ^< \\i WASHINGTON: OOVEBNMENT PRINTING OFFICE 1905. 848 U. S. DEPARTMENT OF AGRICULTURE, OFFICE OF EXPERIMENT STATIONS— FARMERS' INSTITUTE LECTURE No. 6. A. C. TRUE, Director. SYLLABUS OF ILLUSTRATED LECTURE ON ESSENTIALS OF SUCCESSFUL FIELD EXPERIMENTATION. BY C. E. THORNE, M. S. A., Director Ohio Agricultural Experiment Station, Wooster, Ohio. WASHINGTON: GOVERNMENT PRINTING OFFICE, 1905. PREFATORY NOTE This syllabus of a lecture upon Essentials of Successful Field Experi- mentation, by C. E. Thome, M. S. A., director of the Ohio Agricul- tural Experiment Station. La accompanied by 32 views illustrating the methods recommended for conducting agricultural experiment work. The s} T llabus and views have been prepared for the purpose of aiding farmers' institute lecturers in their presentation of this subject before institute audience-. This subject is especially timely because of the rapid increase now taking place in the number of field experiments in which farmers are cooperating with the experiment stations. The numbers in the margins of the pages of the syllabus refer to similar numbers on the lantern slides and to their Legends as given in the Appendix. In order that those using the lecture may have opportunity to fully acquaint themselves with the subject, reference- to its recent literature are given in the Appendix. John Hamilton. Farmer? Institute Specialist. Recommended for publication. A. C. True, Director. Publication authorized. James Wilson. Secretary of Agriculture. Washington, D. C, November J, 1905. (2) ESSENTIALS OF SUCCESSFUL FIELD EXPERIMENTATION. By C. E. Thorne, M. S. A. INTRODUCTION. The experiment station is the distinguishing feature of the new agriculture. It stands for organized research in agriculture, for the concentration of the knowledge and skill and energy of the entire State on the solution of the unsettled problems of the farm. The experiment station is created to do things which are impossible to the ordinary- farmer. It is furnished with costly equipment and is conducted by men trained in the methods of scientific research, who are made free from other cares in order that they may devote their undivided energies to helping the farmer. But while there are some things which can only be done by the aid of such an equipment as that of the experiment station, there are other things which the station can never do, and the farmer who profits most by the work of the experiment station is he who is him- self an experimenter. It is the province of the experiment station to discover and formu- late general principles. The application of these principles must be made by the farmer himself. Even were there an experiment station in every county, there would still be hundreds of farms within each county on which some of the conditions would vary from those of the station, and while no farmer should attempt to duplicate the elaborate work of the experiment station, neither can any farmer afford to blindly accept the conclusions reached at the experiment station with- out subjecting some of them to the test of further investigation on his own farm. Were there no other reason for this than the benumbing effect upon the intellect produced by the unthinking acceptance of the dicta of others, as evidenced by the unnumbered centuries during which the sickle remained unchangingly a chief implement of husbandly, this unchangeableness in implements s} T mbolizing a similar monotony in the farmer's intellectual processes, that alone would be abundantly sufficient. The farmer of to-day must learn to think or he is lost, and nothing is more conducive to exact thinking than scientific experiment. (3) The experiment station should carry its work far enough to demon- strate clearly the lines which practical application must follow, but after it has reached its uttermost limit there will still be much for the fanner to do. Many experiments which farmers attempt, however, are either vtA- ueless or actually misleading, because of failure to observe some of the essentia] conditions of successful experimentation: for investigation in agriculture by experiment is a business by itself entirely distinct from ordinary farming, and many a good farmer will overlook points of vital importance to the success of an experiment until his attention is called to them. It is in the hope of encouraging farmer- to experi- ment more largely and to assist them in making their experimental work more effective that the following suggestions are offered: SELECTION OF SOIL. A matter of first importance in preparation for tield experiment is that the soil shall he as nearly uniform as possible in character. To secure such uniformity the following points should be considered: GEOLOGICAL HISTORY. View. 1 It is important, in experiments with fertilizers especially, to know whether a soil is of sedentary, drift, or alluvial origin; that is, whether it has been formed by the gradual weathering of the underlying rocks; by the action of the great glaciers which once overspread the northern portion of the United States, grind- ing up the rocks over which they passed and rearranging their particles in the moraines and sheets of drift which they left he- hind them as they retreated northward, or by the action of the flood waters of rivers and streams, washing down the soil from the higher lands and depositing it in the alluvial flood plains or "bottom lands" of the streams — a force which is in active opera- tion to-day -as this knowledge may throw considerable light upon different methods of treatment. A sedentary soil is likely to be more uniform than a drift or alluvium, because in the case of these transported soils there is generally a more or less uneven deposit of materials, an excess of gravel and coarse sand appearing in one spot and of silt and finer particles in another. A heavy sheet of drift may some- times become weathered into practically the same condition a- a sedentary soil, while it would seem that some of the great loess deposits would offer especially good conditions to the Held ex- perimenter, the loes^ being the fine-grained, silty soils found in some of the Western States, and whose origin is apparently due to the blowing of the dry surface dust into banks and hillocks, sometimes many feet in depth. PREVIOUS MANAGEMENT. View. It is also important to know something of the previous man- 2 agement of the soil. The great Rothamsted experiments have shown the persistence of the effect of applications of barnyard 3 manure -the yield from a formerly manured plat remaining more than double that of the unmanured land alongside for 4 thirty years after the manuring had been discontinued. Other illustrations arc given to show the effect of leaving fence rows or lanes uncultivated for a period of years and then bringing them into comparison with lands brought under cultivation at an earlier date. Again, a tree which has stood for years after the remainder of the forest has been cleared away, and whose shade has been a resting place for live stock, may cause a con- siderable variation in the productiveness of the soil which will be manifest long after it shall have been cut away. TOPOGRAPHY. Absolutely level land is seldom adapted to field experiment, 5 for the reason that no land is so flat that it does not have slight inequalities of surface, and on such land the minor depressions receive a larger share of the rainfall by surface drainage from the higher portions and thus have an advantage in dry seasons when water may be the controlling factor in producing increase of crop, while they may be at a disadvantage in seasons of excessive rainfall. When the subsoil is porous these depres- sions may merely be a little more fertile than the higher por- tions through increased accumulation of humus, but if the subsoil be impervious to water it may give rise to a semiaquatic growth of vegetation which may radically modify the behavior of the soil under cropping. On the other hand, steep hillsides are not adapted to experiment, because the water from heavy rainfalls in washing over them will cut them into gullies or transport fertilizing materials from plat to plat. The ideal topography for field experiment is a broad, gentle slope of about 1 to 2 per cent, or just enough to permit the surplus water of heavy rains and melting snows to flow off uniformly and completely. ARRANGEMENT OF PLATS. SIZE OF PLATS. Farmers generally have the idea that experiment plats should be made as large as possible, an idea naturally following their observation of the inequalities of most soils; but the practical 6 difficulty in the way of using large plate is the fact that fora comparative experiment the soil most be a- Dearly absolutely uniform as possible, and it Lb extremely difficult to find 1 areas having sufficient uniformity. In almost all cases it will be found better t<> use a large Dumber of small plat- than a -mall Dumber of Large plat-, since by multiplying the plats the variation- <>f the -oil can be more evenly distributed. A field of LO acres, for example, in which it i- desired to make tea comparisons, will yield results of far greater value if cut into LOO plats containing one-tenth acre each. gii ing ten plat-, dis- tributed over the field, to each particular comparison, than if only LO plats are employed. The most convenient size of plat- for computation is one-tenth acre. When the plats are reduced much below one-tenth acre in size another (dement enters into the problem, namely, the individuality of the plant. One-tenth acre of corn, as usually planted, will contain nearly a thousand plants, and this number ought he grown from a single seed ear: hut experiment- have shown that ears taken from the same variety and grown on neighboring stalks may vary 50 per cent in their produce. 6 Hence, even for plat- of one-tenth acre, the greatest care -hould be taken to secure seed which has been drawn from a consider- able number of ears and thoroughly mixed. In the case of wheat or oat- a tenth-acre plat may contain from 100,000 to 150,000 plants, the seed of which has probably been obtained from 3,000 to 5.0o0 heads of grain, representing several hun- dred separate plants. Hence, we may safely use a -mailer plat for the small grains than for corn. SHAPE OF PLATS. 7 It is generally conceded that experiment plats should be long and narrow, both for greater convenience of cultivation and also because tins shape permit- a better arrangement with respect to irregularities of soil. The minor inequalities of the soil are usually caused by the gradual washing of the surface into small ridge- and valleys, which can be bo crossed by a long and narrow plat a- to cover equal portion- of the lower and higher land, whereas this would not be so readily accomplished in square plats. It is well to adapt the width of the plat to that of the machinery in use. A plat l»i feet wide, for exam- ple, will contain four rows of corn 4 feet apart, or live I of potatoes 38 inches apart It may be -own with three '•through-*' of an 8 hoe. 8-inch drill, and cut in three swaths with a machine having a 5j <>r 6 foot cutter bar. The width of the plats should also be adjusted to the drainage. When View - they are more than 16 feet wide it will be necessary on most soils to give each plat a separate drain. SURFACE DRAINAGE. Whatever the natural topography of the land may be, under 8 most circumstances the uniform disposal of excess of surface water ma}' be facilitated by a slight ridging of the plats, such 9 as is accomplished by plowing the plats separately. To attain this end successf \i\\\ the plats should be comparatively narrow, in order that the cross harrowing may so distribute the ridge left by the plow as to leave a gradual slope from the middle of the plat to the sides. On wide plats there will always be a flat space left between the ridge and the furrow. It is not necessary nor advisable to repeat the ridging at every plowing. On the contrary, the effect of ridging will remain apparent for five to ten years, even though subsequent plo wings and cultivations be, as they should be whenever practicable, across the plats. CROSS DRAINAGE. One of the chief objects in separating the plats by dead fur- rows is to prevent cross drainage of surplus rainfall. Where there are no surface drains such water will, of course, follow the natural slopes of the land, and these are practically never suffi- ciently- regular for the purpose of plat experiment. In variety testing and cultural work this may not always be a matter of great importance ; but in fertilizer tests it is essential that the fertilizer applied to each separate plat be confined to that plat, which can not be the case if the surface water is permitted to flow indiscriminately across the plats. Even in such work as variety testing, however, it is desirable that the surface water find its way off the plat as quickly as possible. If it stands in minor depressions over the plat it will cause irregularities in yield ; if the plats are laid off up and down a slope the water will follow down the drill rows, washing out many plants during the spring thaw, while if the plats are laid off across the slopes there will be cross washing which, in fer- tilizer comparisons, will vitiate the results. The best arrange- ment in such work is to so ridge the plats that the surface water will flow from the middle down the sides of the plats into dividing furrows, these being so graded as to carry it gently away. Some gullying of the furrows is almost unavoidable, but it is better there than within the plat itself. CROSS FEEDING. Another prime object of a furrow between the plats is to pre- vent cross feeding- that Lb, the extension of the roots from an unfertilized plat, for instance, to one that has been fertilized — and this is practically accomplished by the ordinary dead furrow. UNDERDRAINAGE. 10 Provision for the removal of an excess of soil water is lately essential to successful field experiment. Such excess may be naturally removed by stratified rocks Lying at Bhallow depths below the surface, or in rare instances deposits of gravel may he found sufficiently uniform to serve this purpose, hut in the vast majority of cases, especially with -oils which have been so long in cultivation that a hardpan has been formed, it will he necessary to aid nature by artificial drainage. In draining for plat experiment the drains should be so lo as to give uniform drainage to each plat When the plats do not exceed a rod in width and the -oil i- reasonably porous, a single drain may serve two plat-, being located under the divis- ion space between them, but in the majority of cases it would be better to give each plat its separate drain, located either under the dividing space or under the middle of the plat. The former location is probably the better one where it is not designed to study the composition of the drainage waters, as the surface water accumulating in the furrow will be more promptly removed if the drain lies immediately beneath the furrow than if it lies under the middle of the plat. DIVISION SPACES. 11 Division spaces 2 feet in width have been found sufficient for the separation of the plats, even in fertilizer tests, provided the land is plowed in ridges with the dead furrows falling at the division .-paces. Such narrow space- necessitate special care in seeding in order that the grain may not be sown too far down in the furrow. The dead furrow is rightly named, and in propor- tion as the seed falls within its influence will the growth of the 12 plant be diminished. When the division space is not more than '1 feet wide and the planting is accurately done, the resulting produce will probably be a more accurate index of what might be expected under ordinary field culture, under the same treat- ment, than where the spaces are wider, as in the latter case the outside rows will have a larger area for root extension, before reaching the dead furrow, than in the former. OLD RIDGES AND DEAD FURROWS. View. It is of prime importance to so arrange the plats as to avoid 13 the errors arising from old ridges and dead furrows, for either a ridge or a furrow, running lengthwise in a plat, may completely reverse the results of the test. Wherever possible, experiment plats should be laid out across the direction in which the field has previously been plowed. CHECK PLATS. A matter of great importance, too often lost sight of in field experiments, is the repetition of check plats. In the most uniform soils there will be some variation in the produce of adjoining plats from season to season. Even were the actual plant food the same, the variations in level which occur on all soils will produce an unequal distribution of moisture, and moisture may often be a more important factor in determining crop yield than plant food. The ideal system of plat experi- ment would leave every alternate plat as a check. Next to this comes the plan of leaving every third plat as a check, thus hav- ing a check plat on one side or the other of every plat under treatment. In fertilizer tests the check plats ma}^ be unfertilized or subjected to uniform dressings with a standard fertilizer or manure, depending upon the object of the experiment. In variety tests the check plats should be planted to a standard variety. The importance of repetition of check plats will be further illustrated under the head of calculating increase. PERMANENT BOUNDARIES. A plat of land designed for experiment should be definitely marked with permanent stakes, especially if tests with fertilizers are to be made. A very convenient marking stake for such purposes is a piece of second-hand gas pipe, about 2, feet long, driven to such a depth that harvesting machinery will pass over it. A block of, say, ten plats may be marked with such stakes, set at the corners of the block, the intervening distances to be measured for each planting. HEADLANDS AND ROADWAYS. Ample headlands should be left across the ends of every block of experiment plats. Not only are these headlands needed to prevent the breaking down, in turning while cultivating, such crops as corn or potatoes, but they are required in the harvest- ing of the small grains, in order that the machine may be driven 11361— No. 6—05 2 10 View. around tin- plats empty, and they arc e\ en more urgently needed in plowing, as the trampling of the soil by the team in turning the j>l<>\\ will materially affect the yield of the following crop. These headlands may be made 18 feet wide, but 20 feet is better. 14 In addition to the headlands there should be roadway- at in- tervals through every large group of plats, to permit the first 15 passage of the machine in harvesting, or the continual pa— 'nil:* where it is necessary to cut the grain l>ut one way. and the pas- 1(> sage of wagons in hauling manure or grain. An experiment plat should never be used as a roadway, for the reason given above, that the additional trampling and packing by the team and wheels, even in the driest weather, will so change the physical condition of the soil as to affect the future yield of crop-. A convenient arrangement, where the plats are not more than 16 feet wide, is to leave a 12-foot roadway between blocks of eight or ten plats. These roadways may be planted in the same crop as that grown on the plats, the roadways to he cut out and the produce set to one side where it will not be mixed with that of the plats before the harvesting of the latter i- begun. The headlands, however, should be kept in grass, in order that the plats may at all times be accessible. Nothing will add so much to the interest in and value of an experiment as to be able to visit it at any time and to see clearly the contrast between plats. It is not waste, but economy, to give a little land to this purpose. LABEL STAKES. IT Every plat should be distinctly marked with a label stake, 18 giving the number of the plat and an indication of its treatment. A very convenient stake for this purpose is a board 4 inches wide and 30 inches long, sharpened at one end. painted with three coats of white-lead paint, and lettered with price markers or with brush. Such stakes, if taken up and sheltered through the winter, will last ten years or more. PREPARATION OF LAND FOR CROPS. UNIFORM PLOWING. It is highly important that the plowing and fitting of the land for a comparative experiment he as uniform as possible in every respect. A difference of a few weeks in date of plowing may cause as great a difference in yield of crop as will he produced by difference in fertilizing, and far greater than is usually observed between varieties. For this reason the plowing should, as a rule, be done across tin 1 plats, and when it becomes 11 necessary to plow the plats separately, in order to ridge them, view - the work should be pushed forward as expeditiously as possi- ble. On this account it is well to have the plats arranged in blocks of ten, planning the test so that each block will have its complete series of checks, independently of the others. MANURING OR FERTILIZING. Both manures and fertilizers may be distributed b} r machinery more uniformly and accurately than by hand. In a compara- tive test of manures or fertilizers it is important that the mate- rial be kept entirely on the plat, leaving the dividing spaces entirely untouched; but where the test is one of varieties or methods of culture the entire surface ma}^ be covered. In applying fertilizers it is a good plan to bring the applications intended for the different plats all to the same bulk, by mixing them with diy sand, so that a single setting of the drill may suffice for the entire experiment. Even then it will not always be possible to exactly gauge the materials, for mixtures which are largely made up of such salts as sodium nitrate and potas- sium chlorid will pass through the drill less readily than those which contain a larger proportion of sand. It is a good plan to test the drill by sowing an equal bulk of material on a plat not designed for experiment before beginning the actual work. The fertilizer will run more slowly as the drill becomes nearly empty; hence it is well to set the drill so that a little will be left in it after the plat has been gone over, and then turn back over the work until all is out. Machines are now made for distributing fertilizers and lime 19 broadcast. They are made wider than the ordinary fertilizer drill and are guaranteed to spread even slaked lime satisfac- 20 torily. Where such a machine is not available, the manure spreader may be used, first spreading a layer of chaff or similar material in the bottom of the machine and spreading the lime on this. Where small quantities of lime are to be applied — 1,000 pounds or less per acre — it will be an advantage to gear the machine so that the apron will travel more slowly than in ordinary work. The best manure spreaders are now fitted with hoods and slow gearing for lime spreading. MACHINES SHOULD BE ACCURATE. While a good machine will do better work than can be done by hand, there are many machines in use which are not fit for experimental work. There are both corn planters and grain drills which do not distribute the seed uniformly, while there are othei's which are unsuited to the experimenter's purpose 12 u never safe to depend upon memory nor upon field stakes for the plan of the experiment; for memory is proverbially treacherous and stakes are liable to he misplaced or lost. A very convenient way of entering the final results of an experiment which is continued from year to year is shown in the accompanying transcript from an actual record. CORRECTING RESULTS. 27 While it is highly desirable that an experiment should be so conducted that its results may be accepted just a< they come from the field, there will sometimes be cases where some cor- rection is needed in order to avoid a statement that would be actually misleading. In the case of corn and potatoes especially it is practically impossible to secure at all times a perfect stand, because, of destruction of occasional plants by insects or from other causes. Where such cases occur it has been found better to count the hills or plants and make the corrections on the basis of the average stand actually attained, rather than upon that of the possible full stand. The former method gives results corresponding to the yield actually obtained, whereas the latter one always gives exaggerated results, since it appears that the destruction of a plant gives opportunity for surround- ing plants to develop more freely and produce a yield larger than could have been attained under normal conditions. CALCULATING INCREASE. 28 An old soil which, from surface indications, is apparently uniform, may nevertheless show considerable variations due to previous treatment. This is shown by the accompanying dia- gram, which shows the average yield of four plats which have received for ten years fertilizers containing the same quantities each of nitrogen, phosphorus, and potassium, together with the yields of three adjoining unfertilized plats, the plats being arranged as indicated by the numbers. It will be seen that 17 where we assume the variation between consecutive unfertilized View - plats to be uniformly progressive we get a practically uniform increase from the different fertilizers, but if we should strike a general average of all the unfertilized plats there would be no consistency in the results. The variation in the unfertilized yield in this case was due to the fact that the land occupied by plat 28 and parts of the adjacent plats was not brought into cultivation until many years after that occupied by the other plats of the series; but about seven years before the test began all had been thrown into the same field, and when the test began there was no superficial indication of previous differences in treatment. CONTINUITY OF WORK. Few of the questions which demand solution by field experi- ment can be definitely answered by a single season's test. The unequal effect of rainfall on soils differently situated with respect to natural drainage may produce wide differences in the apparent results of a test in different seasons. Other climatic variations may produce similar variations in results, so that an experiment must be carried through a number of seasons before its results can be accepted as a reliable guide. This point is illustrated by a test showing the increase or 29 decrease in yield of wheat on several plats in 1894, and in the average for ten years. It will be observed that in the general average acid phosphate and nitrate of soda are the most effective constituents of the fertilizer, but that their effect has varied greatly in different seasons, acid phosphate producing an actual loss of crop in 1894. Had this experiment been limited to a single test in 1894 it is evident that a wholly erroneous opinion of the effect of acid phosphate on this soil would have been formed. Views 29, 30, and 31 show the varying effect of fertilizers 30 over a long period of years on poor soil and on good soil. In most of these seasons phosphorus is unmistakably the chief 31 agent in producing increase of crop; but in 1894, on the poor land, potassium takes the leading place, and in 1903 nitrogen comes to the front, while on the good land there has been still greater variability in the apparent effect of the three fertilizing constituents. Another reason for continuity of work in experiments with 32 fertilizers is the fact that the full effect of such fertilizers is sel- dom realized in the crop to which they are applied. This point is shown by a table giving the average increase from fertilizers in a test which has been in progress for ten years, the figures 18 opposite each year being the average result for the entire period up to and including that season. In this case there was b com- paratively Bteady yield on the unfertilized plat, with a constantly Increasing yield from the fertilizers. SUMMARY. The experiment station Btands for organized research, doe- things impossible to the ordinary farmer, and formulates general principle-. The farmer must make the application. The farmer who profits most by the experiment station must himself be an experimenter. SELECTION OF SOIL. The geological history of a soil is of importance as an index t<> its genera] character. Pn vious managt rru nt may have an important bear- ing upon the result^ of an experiment, as shown by long-continued effect <>f manuring at Rothamsted. In topography an experiment field should not be absolutely level nor very Bteep; a broad, gentle slope of 1 or2 per cent, or just enough to permit uniform surface drainage, has been found best. ARRANGEMENT OF PLATS. The most convenient size of plate, for most purposes, is one-tenth acre. This size is convenient for computation, and holds a sufficient number of plants to eliminate the errors arising from individuality if care be taken in seed selection. The inequalities of soil can be tx eliminated by duplicating the test in small plats than by the m large plats. In ,sj,,. the plats should be long and narrow. It is well to adjust the width to the convenient use of machinery and to the scheme of drainage. Surfod (1 ri)hits is a matter of first impor- tance. Every third or fourth plat should be used a- a check. The boundaries of every block of experiment plats should be permanently marked with stakes that will not be thrown out in plowing. R lands >m. CONTINUITY OF WORK. This is of first importance in field experiment. The results of one season's tests may be the direct opposite 1 of the average outcome of a period of years, and as a general rule the full effect of a fertilizer or manure can only be determined after a series of year-. APPENDIX, TiANTERK SLIDES. • No. of view. 1. Soil formation. Face of a stone quarry in a glaciated district, showing the heavy layers at the bottom, gradually breaking up into thinner layers toward the surface, and finally ending in a finely broken stratum, over which is spread the thin drift sheet with its contained pebbles and small bowlders. In many places in this region this drift sheet is replaced by a thin soil filled with the angular fragments of the underlying rock— such a soil as would be made by a little further disintegration of the upper stratum of rock shown in the picture. 2. Persistence of effect of barnyard manure. Compiled from Memoranda of the Rothamsted Experiments, 1903, pp. 26, 27. Experiments on the growth of barley in Hoosfield. 3. Effect of previous treatment of land. A diagram showing the yield of the unfertilized plats in an experiment in which corn, oats, wheat, clover, and timothy have been grown in a five-year rotation for ten years. There are 30 plats in the series, and every third plat, beginning with No. 1, has been left continuously unfertilized. The average yield of these plats ranges between 1,400 and 1,750 pounds per acre until No. 28 is reached, when it suddenly mounts to nearly 2,600 pounds. This experiment was located in a field which had been in cultivation for fifty years or more, and at the time it was laid out no difference was apparent between the land occupied by plat 28 and other por- tions of the field; but it was afterwards learned that this plat and parts of the adjacent plats had been occupied by a lane until some seven years before the test began. 4. Old fence rows are rich. This point is again brought out in the photograph showing the growth of corn on an old, roadside fence row, in a region where the discovery has recently been made that it is cheaper to fence cattle in than to fence them out, and consequently the fences are being limited to such as are necessary to inclose permanent pastures, and many miles of road are entirely unfenced. 5. An ideal platting field. Part of a field of about 60 acres which is wholly devoted to plat experiments. It slopes gradually toward the north, with secondary slopes toward the east and west, which give suf- ficient fall for the lateral tile drains that are laid between the plats, running east and west and emptying into mains which run to the north— one on each side of the field. 6. A tenth-acre plat of oats. Showing the great difference in number of individual plants. 7. Mowing a plat of grass 16 feet wide in 3 swaths. 8. Effect of ridging on wide and narrow plats. No. 1 shows a plat 2 rods in width, as first plowed, and No. 2 shows the same plat after the harrow has,gone over it, partly leveling the ridge or back furrow. Nos. 3 and 4 show a plat 1 rod wide under similar conditions. It will be observed that In the narrower plat the effect of cross harrowing is to leave a surface having a gradual slope from the crest to the furrow, whereas on the wide plat there will be a flat space left between the two. The prac- tical difference between the two methods of platting is that the surplus water of heavy rains and melting snows will flow off the narrow plat more uniformly and with less cross washing and gullying than off the wider one. 9. Contour of a plat 18 feet wide from center to center of furrow as obtained in actual practice. These plats have just been ridged for the third time in 12 years. Future crosswise plowing will materially reduce the height of the ridges. (21) 22 ■- tO. Natural rmderdrainage. There las gradual transition here bom the heavy layers <>f stratified rock, through the lighter layers and broken strata to tin- thin sheet ol soil, irhieh has been funned bj tfa (feathering of the surface rock. On ttii- soil treei are growing, which send their root* through ti> the rocki below. When the treei arc cut away their decaying roots furnish nat- urul drainage channel* to the rock seams below . but tbeee channels are gradually obliterated by the Mraping of the plow and the trampling of the teams and grazing stock. Eventually i bardpan li formed, and the land require! artificial drainage. From Ohio sta. Bui. no. l i. Two-fool wide dividing space between plate of coin. 12. Two-foot wide dividing >pa< «• between plats of <>ata. 13. Bffed of backfuTTOw. The middle row of corn in the picture la located on the ridge or backfurrow formed by throwing two furrows together, it a ill be Been if one plat has men a backfurrow and others do not. the apparent results of ■ comparative tc^t may be very misleading The effect of »n old dead farrow may be equally fatal to accuracy in n-Milts. 11. Arrangement of |>lat< for t-year rotation. Four ranges or tiers of plats, in plats each, as actually employed in an experiment in which corn, oats, wheat, and clover are grown in rotation, each CTOp being grown every year. 15. Arrangement of plats for 3-year rotation. Six blocks of 10 plats each arranged for a 8-year rotation. 16. Plats arranged for blocks of ten. No. 16 shows 40 plats arranged in blocks of m plats each and being one of three simi'. tions used in a 3-year rotation. The arrangement shown in No. 15 or No. 16 permits thi rate handling of each block of 10 plats, each having its compl- (hecks, and is the better arrangement where circumstances permit its nee. The plats marked X are cheek plats, to be left unfertilized in fertilizer tests or planted with the same variety in variety com- parisons. 17. Label stakes. A convenient label stake 4 inches wide by 30 inches long, painted white and lettered with a brush. These label stakes are useful, but should never be permitted to take the place of a written record on a carefully drawn diagram, as stakes may be easily transposed or lost. 18. A signboard for convenience of visitors. 19. Distributing lime with lime spreader. 20. Spreading lime with the manure spreader. 21. Seed corn. Four grains each from butt, middle, and tips of same ear, showing difference in size. 22. Catting the grass plats. After careful measurement from the permanent iron stakes which mark the corners of each block of 10 plats, small stakes are driven at the proper points on opposite sides of the block. The machine is then set with the tongue over the stake on the opposite side of the block and is driven to a tall stake held by a boy at the end in the foreground. The lx>y has just stepped aside to let the team pass. Returning, the two machines, driven full, will cut the Lrra-s stand- ing on the plat, leaving the 2-foot wide dividing space uncut. This will stand until the hay on the plat is weighed. 2.'3. Harvesting the small grains. This is a simpler matter than harvesting the hay, as no grain stands in the dividing s] but a man or boy should follow the machine to see that every sheaf is placed on its proper plat and that the machine is cleared of grain when it reaches the end of the plat. 24. Harvesting the com plats. In harvesting the corn plats with a harvesting machine an ear will occasionally be knocked off. A man should follow the machine to see that all such ears are placed on their proper plats. 25. The derrick scale. The Bled runners are oak joists 2 inches thick. 10 inches wide, and 8 feet long: the uprights are of any Strong WOOd '2 by 1 inches ami 8 feet long; at tin- top they are bolted to a head- 6 Inches thick. 8 Inches wide, and is indies long. The swivel or rowlock which carries the pole is made by a blacksmith, the ears being i inch by 2 inches, and the tongue 1J inches In diameter by 8 inches Long. The swivel re^ts upon a washer let into tlie headpiece, and the split key which holds the swivel in place works upon a similar washer fitted in the lower side of the headpiece. The pole is 24 feet long and 4 inches in diameter at the swivel. 23 No. of view. 26. Planning a field experiment. 27. Method of recording plat work. 28. Calculating the increase'. When every third plat is a check plat the difference in yield between two successive checks is found, and one-third of this difference Is added t< < t ix- smaller yield of the two checks or subtracted from the larger yield, the result being assumed to be the natural yields of the two plats lying between the cheeks. The difference between this assumed yield and the yield actually obtained is taken as the increase or decrease resulting from treatment. 29. Effect of fertilizers in different seasons. This table shows a decrease in yield on all plats fertilized with acid phosphate in 1894 except one, although this material has increased the yield every year since then. 30. The varying effect of the different fertilizing constituents in successive seasons on poor soil. 31. The varying effect of the different fertilizing constituents in successive seasons on good soil. These show the necessity for continuous work over a long period of time before definite conclusions are drawn. 32. Progressive increase from fertilizers. The cumulative effect of systematic fertilizing is shown by this table, each season's increase being sufficient to raise the average increase for the total period to a higher level. i:i:i 1:1:1. n« E8. Dept. Agr., Office Expt. Stas. Bui. 22. Agricultural Investigations ;it Both- amsted, England, by Bir Joaepfa Henry Gilbert r., Office Expt. Stas. Bui. L42,p. 127. Methods of Conducting Invea- Relating to Maintenance or [ncreaae of Boil Fertility, by C. E. Thome. Dept Alt.. Office Expt Staa. Bpt L904, p. 483. Experiment Station Work in Oorn ( Hilture, by J. I. Schulte. 1 . - Dept Alt. Yearl k. L897, p. 291. Every Farm an Experiment Station, by En in E. Kuril. Connecticut Stoma Sta. Bpt L888, p. 17. Cooperative Field Experimenta with Fertilizers, bj 0. B. Phelps. Georgia Bta. Bui. 65. Corn Culture, by II. J. Bedding. gia Ma. Bul. *'»*''. Cotton Culture, by It. .1. Bedding. Massachusetts Hatch Bta. Bul. 9. Soil rests with Fertilizers, by Wm. P. Brooks. New York Cornel] Bta. Bul. L29. How to Conduct Field Experiments with Fertil- izers, by G. C. < 'aMwell. Pennsylvania Bta. Bul. 35. A Soil Test with Fertilizers, by H. P. Armsby. I'tah Sta. Circ. I. Memoranda of Flans for Arid Farm Investigations. Utah Sta. Circ. 2. Memoranda of Plans ««f Irrigation Inveetigationa. Plans and Summary Tables Arranged for Reference in the Fields Rothamated, 1905, pp. 23. Memoranda of the Origin, Plan, and Results of the Field and Other Experimenta conducted on the Farm and in the Laboratory of Sir J. B. Lawes at Rothamsted, 1900, pp. 119. Review of Agricultural Experiments, by A. P. Aitken. Trans. Highland and Agr. Soc. Scot., 15 (1903), pp. 94-l.M. TheJ Necessity of Practical Field Tests and the Methods of Conducting Them, by Loges and Vibrans. Jahrb. Deut. Landw. Gesell., 13 (1898), pp. 65-74. (Es? f Sta. Rec, 10, p. 955.) T ' Experience Notes on Plat Experiments, by B. D. Halsted. Proc. Soc. Prom. Agr. Sci., 1899, pp. 19-27. (Expt. Sta. Rec, 12, p. 337.) Directions for Making Local Fertilizer Experiments, by K. Hansen. Copenhagen, 0. C. Olsen & Co., 1902, pp. 35. (Expt. Sta. Rec, 14, p. 648.) On the Organization of Local Fertilizer Experiments in European Countries, by A. Rindell. Landbr.Styr. Meddel., 1902, No. 41, pp. 44. (Expt. Sta. Rec, 14, p. 851.) Some Remarks on the Object and Method of Conducting Field Experiments with Fertilizers, by F. W. Dafert. Landw. Jahrb., 32 (1903), No. 1, pp. 149-159. (Expt. Sta. Rec, 15, p. 130.) Exact Methods of Conducting Field Experiments with Fertilizers, and Various I tions Relating to the Use of Nitrate of Soda and Ammonium Salts as Fertilizers, by P. Wagner et al. Mitt. Yer. Deut. Landw. Vers. Stat.. L904, No. 2, pp. IV (Expt. Sta. Rec, 16, p. 859.) The Offering of Premiums for Fertilizer Experiments as a Means of Promoting Rational Fertilizing, by T. Pfeiffer. Mitt. Landw. Inst. Breslau, 1 (1901), No. 5, pp. 1-45. (Expt. Sta. Rec, 13., p. 726.) The So-called Statistical Method of Making Field Experiments with Fertilizers and Caculating the Results, by T. Pfeiffer. Mitt. Landw. Inst. Breslau. 2 H"»4 . No. -l, pp. D47-682. (Expt. Sta. Rec, 16, p. 31.) Experiments with Agricultural Plants, by A. Mitscherlich. Landw. Jahrb I L90S), No. 5-6, pp. 773-818. (Expt. Sta. Rec, 16, p. 558.) Agricultural Experiments and the Presentation of Their Results, by A. Mitscherlich. Landw. Vera. Stat, 61 I L904), No. 1-4, pp. 285-303. (Expt. Sta. Rec, 16, p. 658.) UNIVERSITY OF FLORIDA 3 1262 08929 2048