that this infectious principle grows and multiplies in the soil or upon the herbage, and that the land is never free of their presence. Where these favorable climatic and telluric conditions do not exist, the disease organisms are not found unless carried there by Southern cattle that have grazed on infected land, and the conditions being such that the organism can live only for a certain time, the lands do not become permanently infected. (In Northern latitudes cold destroys the infection.)

From the foregoing it will be seen that we have territory which is permanently infected, and territory which is not. As before mentioned, most of the territory in the States bordering on the Gulf of Mexico, and on the Atlantic as far north as Virginia, including a part of Virginia, Tennessee, Arkansas, and Indian Territory, are permanently infected. If animals north of this infected territory are brought South and placed at pasture, they contract Southern cattle fever in severe form and usually die. If cattle from the infected territory are taken North to non-infected lands, the lands become temporarily infected, and cattle not immune contract the disease the same as when brought South. Cattle moved from one infected place to another remain comparatively healthy and do not communicate the disease.

As far as I have been able to learn, this State (S. C.), with the exception of a small portion along the foot of the mountains, is permanently infected. Cattle brought from near the mountains to the lower portion of the State contract the disease the same as cattle brought from the Northern States. As in other Southern States, here the disease (in other than a latent form) is observed usually in animals brought here for breeding purposes. A large majority of the animals purchased at the North for this purpose take the disease sooner or later and die, unless kept on dry food and not allowed to run at pasture. This disease has, consequently, proved a great detriment to the improvement of the breeds of cattle in the permanently infected States.

Last October, through the kindness of Mr. L. D. Childs, I had the opportunity of examing cases of Southern cattle fever in some fullblooded Herefords brought by him the fall before from New York. These animals were pastured on the Congaree River bottom and did exceedingly well during the first of the season-became very fat. In October a majority of the herd became affected with Southern cattle plague, and showed the characteristic symptoms of the disease, the most important of which are as follows: Very high fever, temperature ranging from 105 F. to 107 F., frequent pulse, increased respirations, loss of appetite, fæces usually hard and coated with blood,

and large quantities of blood-colored urine is passed. On post mortem examination the blood is found thin and watery, the spleen enor mously enlarged, weighing often five or six pounds when the normal weight would not exceed two; the liver is also much increased in size, being about twice the normal size and of a yellowish tint; the gall sac is enormously distended with a black granular-looking mass, and the bladder contains large quantities of wine-colored fluid. In addition to the above constant symptoms there is usually a congested appearance of the mucous membrane of the fourth stomach with occasional erosions. After the first symptoms appear the animal usually lives but three or four days.

Much time has been devoted to the study of this disease by the Department of Agriculture at Washington, and others, with a view of finding out the exciting cause. It has for some time been supposed to be a germ disease, i. e., to belong to that class of diseases caused by living organisms, but the Department at Washington so far as the reports show, have failed to find the specific germ. Dr. F. S. Billings, however, during his connection with the University of Nebraska, met with cases of Southern cattle plague, and succeeded in isolating an organism which he considers as the exciting cause of the disease. This organism he describes as resembling in many particulars the hog cholera bacillus of Salmon.

For description of organism, etc., the reader is referred to the Bulletin of the Experiment Station of Nebraska, containing Nos. 7, 8, 9, and 10.

Autopsies were made on two of the Herefords owned by Mr. Childs, previously mentioned, and cultures were obtained in the regular way. From these cultures but one organism was met with which would seem likely to be pathogenic; this was obtained from both apimals, and will be seen from the following description to very closely resemble the one described by Billings in his report, and also to a certain extent to resemble the hog cholera bacillus :

This organism is a medium-sized bacillus with distinctly rounded ends, 1.4 to 2.8 micro-millimetres in length, 0.7 to 0.9 in thickness. Upon cultivating this organism in peptone gelatine it forms a continuous white line along the line of inoculation. On the surface of the gelatin, which never becomes liquid, is formed a white growth that gradually covers the surface. The development on agar-agar is similar to that on gelatine.

The organism grows very vigorously upon potatoes and has a delcate straw-colored appearance.

The organism did not prove pathogenic in rabbits, and not always in mice. On no other animal was it tested, for want of time and

other necessary conveniences. It is to be hoped, however, that this work can soon be resumed, and the fact as to whether the organism in question has any part in the causation of the disease determined. It is also hoped that some work can be done toward determining whether or not the disease can be prevented by inoculation. As the disease often exists in a latent form without causing any interference in the general health of the animal, it seems probable that this can be done.

Treatment of this disease is seldom followed by good results. The changes in the body caused by the infectious principle (or germ) are so great and take place so rapidly that drugs fail to save the life of the animal. In some cass, when the attack is less severe, proper care and treatment may enable the natural forces of the animal to overcome the disease. In these cases soft food and stimulating fever medicines are indicated.

Nitrate of potash in two drachm doses, sulphate of quinine in drachm doses, and alcohol in ounce doses may be given three times daily. After recovery commences tonics to increase the appetite may be required.

The portion of this report on hog cholera, first published as a bulletin in July, was the result of the combined labors of Dr. B. M. Bolton, the former Bacteriologist of the Station, and myself. For assistance in the subsequent investigations, I am much indebted to Mr. F. W. Pickel, the present Bacteriologist.

W. B. NILES, D. V. M.

FIELD EXPERIMENTS.

The field tests begun in 1888 were intended to cover a period of years. They were, therefore, carefully repeated in 1889. Their plan. and scope were fully set forth in last year's report (1888), and the reader is referred to it for explanation of details. This is necessary in order to avoid repetition. Some of the series of tests were enlarged by the addition of new experiments. These will be explained in their proper place. Among the most important questions attacked by the tests were: The requirements of the soils of our three experimental farms; the requirements of our leading crops; and the relative values of different nitrogenous, phosphatic, and potassic manures. The order of discussion adopted in last year's report will be followed in this, namely: I. Oats; II. Wheat; III. Corn; IV. Cotton; V. Miscellaneous Crops; VI. Feeding Experiments.

Before entering upon the consideration of the results reached in the following sections, a few words of a general nature as to the unsatisfactory character of the work of field experimentation and of explanation of what we mean by "the probable error," and our method of determining it, are necessary. In Bulletin No. 1 of our New Series, published April 1st, 1888, giving the results of our tests of varieties of cotton, occurs the following:

"The whole subject of plat experimentation is now being warmly discussed. So conflicting and unsatisfactory are the results too often obtained that many are hastily inclined to hold that this method is a failure. No one who has given the subject attention will hesitate to admit that it is one of exceeding difficulty. The assumption is usually made that it is impossible to get plats of uniform productive value, and that differences in the results are due to differences in the fertility of the plats used. This appears to us to be simply begging the question. And to urge that because the results vary, the system should be forthwith abandoned, is to disregard all scientific experience and precedent. In earlier days chemical analysis gave very dif ferent results in the hands of different chemists, and astronomical observations, of assumed equal weight, differed widely. But the abandonment of these methods was not the remedy suggested. The chemists and the astronomers went to work to correct and improve them, and by repeated tests and experiments to determine the amount of error involved in the method employed. It was found

that the average of a number of analyses or observations would furnish accurate and reliable information. Our efforts, therefore, should be directed towards the determination of the amount of error incident to this system of experimentation. When this is ascertained, and the necessary allowances made, the results obtained can be fairly compared and estimated. Large plats (fromto 1 acre each) will give, generally speaking, the most accurate results, for any error is not multiplied when the yields are given per acre. But the use of such plats involves a large outlay and abundant resources [in order to insure the uniform treatment of the several plats, i. e., their planting, cultivation, harvesting, &c., at the same time]. Hence small plats must be used. Where these are tenths of acres, an error of 1b. in the product is multiplied tenfold when stated per acre. The greatest accuracy and care are, therefore, necessary, and, in our opinion, the differences in the yield of plats subjected to so-called uniform conditions are as often due to other causes, such as slight differences in the mode of planting, tillage, harvesting, weighing, seed, or fertilizer, loss from disease, accident, or insects, as to differences in the soil. With scrupulous care and minute attention to every detail the employment of duplicate plats can be made to give valuable results, provided it is continued through a series of years in order to determine the amount of error incident to it. Our tests of varieties have been in progress since 1883, and, therefore, cover a period of five years. Each year we have carefully noted the sources of error and endeavored to remove them. As stated above, every test was duplicated, and in the preceding and all the following tables only the averages of the duplicate plats of each variety are given and dealt with. Now, by taking the differences between the average and the separate yield of each of the two plats of [each of] the 7 varieties tested in 1883, the 13 tested in 1884, the 14 in 1885, and the 27 in 1886 (1887 must be thrown out, as the duplicates were not in the same field), we can determine the amount of variation. There is a steady and marked improvement noticeable each year. For the five years, the variation or probable error amounted to 95 lbs. of seed cotton per acre, for the four last years to 83 lbs., for the three last to 80 Ibs., and for the two last to 75 lbs. It may be assumed, since even greater care was used, that the difference for 1887, had the duplicate plats been in the same field, would have been about the same as for 1886, or 75 lbs. of seed cotton, or 24 lbs. of lint per acre. By the employment of the well known formula for the determination of the probable error in the mean of several observations or tests of equal weight, or made under the same conditions