utilized because of the danger of introducing ammonia into the compound and thus increasing the nitrogen content. Inasmuch as the writer's analyses now show the approximate nitrogen content, it is extremely desirable that some one should attempt to isolate the pigment by the ammonia process. Grandeau (3) found that when an ammoniacal solution of humus containing 53 per cent of ash was dialyzed against distilled water for 36 hours, the resulting residue in the dialyzer had an ash content of only 8 per cent. Perhaps this method could be employed in freeing the soil pigment from ash.

Pigment Preparation 2. This sample of pigment, weighing approximately 7 gm. was prepared from Pigment Solutions 4, 5 and 6, following much the procedure as was used for Pigment Preparation 1, with the exceptions that the solution of the pigment in dilute alkali was not precipitated by adding NaOH up to 4 per cent concentration and that hydrochloric acid was used for acidification instead of sulfuric. The final extractions were the same for both preparations. The following analytical data were obtained on the sample:

Ash determinations indicated 50.77 and 51.57 per cent, or an average of 51.17 per cent of ash.

Organic carbon (by CuO in oxygen) determinations gave 55.82 and 56.23 per cent C, or 56.03 per cent of organic carbon calculated to an ash-free basis.

Hydrogen determinations gave 5.89 and 6.36 per cent, or 6.13 per cent calculated ash-free.

Nitrogen determinations gave 2.94 and 3.02 per cent N, or an average of 2.98 per cent calculated to an ash-free basis.

These analyses would indicate an approximate content of

It will be noted that this preparation contains appreciably more hydrogen and less carbon than Preparation 1. This is exactly what would be expected from a sample having a higher ash content, if the ash consisted of clay. Apparently the clay has retained enough water of constitution to vitiate entirely the hydrogen analysis and to lower the carbon precentage appreciably.

SUMMARY

A sample of soil of the Fargo silt loam type was leached with 1 per cent hydrochloric acid and then extracted 9 consecutive times with fresh portions of a 4 per cent sodium hydroxide solution. A sample of the residual soil was removed at this point and analyzed, while the remain

(ii-38)

ing portions were further extracted for 6 consecutive times with very dilute (approximately 0.15 per cent concentration) sodium hydroxide solution. Following these extractions the remaining soil residue was dried and analyzed. The following conclusions are evident from the analytical data:

1. The writer's earlier observation that 4 per cent sodium hydroxide solution does not dissolve the black soil pigment is confirmed.

2. The soil pigment is soluble in very dilute sodium hydroxide solutions but precipitates from solution on the addition of sodium hydroxide in sufficient amount to make a 4 per cent solution.

3. The soil pigment is also precipitated from solution by salts of the heavy metals and by acidification; is not dialyzable and forms a stable water-soluble compound with ammonia when an ammoniacal solution is evaporated to dryness.

4. Two attempts were made to prepare the soil pigment in pure form but the resulting products contained such a high content of ash (37.47 per cent and 51.17 per cent, respectively) as probably to render the ultimate analysis unreliable. The analysis of the product with the least ash content gave C., 61.3 per cent; H., 4.3 per cent; N., 2.8 per cent; O., 31.6 per cent; all calculated to an ash-free basis.

5. The first six extractions with NaOH removed relatively more nitrogen than carbon from the soil, but the remaining three NaOH extractions as well as the six pigment solutions contained relatively more carbon than nitrogen. The final soil residue had a ratio of C/N very much higher than that of the original soil.

(1) DETMER, W.

LITERATURE CITED

1871. Die natürlichen Humuskörper des Bodens und ihre landwirthschaftliche Bedeutung. In Landw. Vers. Stat., Bd. 14, p. 248. Reprinted in Jahresber. Agr. Chem., 1870-72, p. 68-72.

(2) GORTNER, R. A.

1916. The organic matter of the soil: I. Some data on humus, humuscarbon and humus-nitrogen. In Soil Sci., v. 2, p. 395-442.

(3) GRANDEAU, Louis.

1872. Recherches sur le rôle des matières organiques du sol dans les phénomènes de la nutrition des végétaux. In Compt. Rend. Acad. Sci. [Paris], t. 74, p. 988-991.

(4) MARBUT, C. F., BENNETT, H. H., LAPHAM, J. E., and LAPHAM, M. H. 1913. Soils of the United States. U. S. Dept. Agr. Bur. Soils Bul. 96, 791 p. 13 fig., 2 maps.

(5) RATHER, J. B.

1911. An improved method for the estimation of humus. In Jour. Indus. Engin. Chem., v. 3, p. 660-662.

FERRIFICATION IN SOILS'

BY

P. E. BROWN, Chief in Soil Chemistry and Bacteriology and G. E. CORSON, Fellow in Soil Bacteriology, Iowa Agricultural

Experiment Station

INTRODUCTORY

Iron has long been known as an essential plant-food constituent. The experiences of the early plant physiologists indicated that the formation of chlorophyl was dependent upon the presence of the element iron. It was recognized, of course, that other constituents were likewise necessary, but in the absence of iron plants soon took on a yellowish, or blanched appearance known as chlorosis, and died. Analyses failed to show the presence of iron in chlorophyl and hence its importance was concluded to be due to a stimulative catalytic, or some other unknown action.

Molisch (8), Raulin (11) and Benecke (1) believed that iron was necessary for the formation of protoplasm and that the chlorosis brought about by its absence was a secondary effect.

Many experimentalists have confirmed the work of the early investigators in showing that iron is essential for the proper development of plants, but little light has been thrown on the question of its exact function in the plant cell. Experiments along such lines would certainly prove extremely interesting and valuable.

While, therefore, the importance of iron to plants has long been recognized, it has always been believed that there was sufficient present in the proper condition in all soils to keep plants supplied for an indefinite period. Certain experiments have shown, however, that applications of iron salts to some soils may prove beneficial. Griffith (4) secured large increases in beans when ferrous sulfate was applied to the soil. Increases were likewise secured when the same salt was applied to turnips, meadow hay, mangel-wurzels, potatoes, wheat, tobacco, onions and cabbage. While there is some question regarding the exact action and cause of the value of the ferrous sulfate used in these experiments, Griffith reports the soils used as extremely low in iron content and the assumption is that the addition of iron benefited the crops. Soils, in various localities, have been found deficient in the element iron, and plants grown on them show this deficiency by their typical chlorotic appearance. It seems therefore, that iron fertilization of some soils may be rather important.

On the other hand, Halligan (6) believes that enough iron is soluble in the weak soil acids to supply the needs of plants so that this material 1 Received for publication October 2, 1916.

need not be considered when applying fertilizers. Russell and Hutchinson (13) assume likewise that iron salts are unnecessary as manures. Hall (5) is even more emphatic, in declaring the use of iron sulfate either for farm or garden crops as unnecessary. He insists that no direct evidence has yet been adduced for a beneficial effect of iron salts either on color or yield, and that experiments have never been conducted in a manner to raise the supposed increase due to the iron, beyond the range of experimental error.

While it must be admitted that the experiments along this line are rather unsatisfactory and broad conclusions should not be drawn from them, it seems that the importance of iron is too great and the possibility of its absence, at least of its absence in the proper form, is too real, to dismiss the question of iron fertilization with an unqualified statement that it is never necessary. Future experiments may demonstrate its value on certain soils, and in any case the value of iron and the importance of insuring its presence in the proper condition should be kept in mind.

Iron occurs in soils in inorganic and organic forms, largely unavailable, and it is necessary therefore that iron oxidation or ferrification as Lipman (7) calls it, occur in soils if plants are to secure the iron they need.

Regardless, therefore, of the great question of the need of iron fertilization, there is much of interest and value attached to the problem of ferrification in soils.

Does this process occur in soils? Is it different in different soils? What conditions govern its occurrence? How may it be controlled? These are some of the questions which immediately arise in a consideration of the problem. Then too, there is the fundamental question of the measurement of the process by laboratory methods. Can accurate methods be devised? If so, how? Broad practical applications of such a problem follow as a matter of course upon the more technical beginnings. Assuming definite, positive results, the whole question of iron fertilization, soil treatment to make iron available, and other problems, would be opened up. In short, the value of experiments on iron oxidation in soils seemed eminently sufficient to warrant considerable attention. No other work along such a line has been undertaken so far as the authors are aware and while the results are far from definite, it is felt that sufficient data were secured to warrant more or less general conclusions and to open up the subject for further study. If nothing else is accomplished by this work, it is clearly shown that chemical and bacterial processes occur so intimately associated in soils that differentiation is often difficult, if not quite impossible. Ferrification, like many other processes occurring in soils, is shown to be partly a chemical process and partly due to a bacterial action. There is some question which is the more

important but from these results the chemical oxidation of iron appears to be much greater than the bacterial action. Later results on a wider variety of soils may modify this conclusion. No attempt has been made yet to isolate and describe iron-oxidizing soil bacteria. Several organisms have been isolated from water supplies and described by Rullmann (12), Ellis (23), Molisch (9), Schorler (14) and others but it is regarded as extremely doubtful if these organisms occur in soils.

Probably an entirely different group of organisms is responsible for iron oxidation in soils. An interesting study is indicated here. The present work, however, considers only the ferrifying power of soils in general without regard to specific organisms. It is an attempt to answer, in a general way at least, the questions which have been mentioned above. The first problem in connection with the work was, of course, the devising of chemical methods which would allow of accurate determinations of ferrous or ferric iron in soils. Just as has been the case in the study of other bacterial processes in soils, it was found that the chemical methods which must be employed were entirely unsatisfactory for application to soils. The first section of the work considers briefly, therefore, the tests that were carried out in order to secure a satisfactory means of determining either ferrous or ferric iron in soils. Such a method was absolutely essential to a measurement of ferrification.

METHODS FOR THE DETERMINATION OF IRON IN SOILS

As the work was originally planned it required a method for the determination of ferrous iron in the presence of ferric, and it seemed advisable to test the methods available.

Many questions arose in this connection. Does ferrous iron occur in the soil? Would the methods available detect all the ferrous iron present? What effect would the strength and kind of acid have upon the results? If the ferrous compounds were all extracted from the soil, would they not be readily oxidized to the ferric form? Does the organic matter of the soil interfere in the determinations? The following preliminary tests were planned to answer some of these questions, if possible.

The colorimetric methods for the determination of total and ferrous iron as described by Schreiner and Failyer (15), Wiley (16), and in the "Standard Methods of Water Analysis," all involve numerous reagents and a difficult procedure. Since several of these series required sixty to seventy determinations, these colorimetric methods were not considered.

The following method for determining the total iron was tested and found to give quite satisfactory results.

Weigh out 5 gm. of soil and transfer to a nickel crucible. Thoroughly