TABLE 1.-Comparison of sodium peroxid fusion and magnesium nitrate methods for total phosphorus.

The referee did not ask that these methods be checked by the carbonate fusion or standard method, as it is sometimes called, his work with the method having shown that it has the disadvantage of a large amount of soluble silica, necessitating dehydration, and in the case of a soil with any considerable amount of organic matter there exist favorable conditions for reduction and, consequently, some loss of phosphorus.

In Table 2 will be found the results of the referee's determination of total phosphorus in the two soils, working as follows: Making a carbonate fusion, then proceeding as for an accurate determination of silica, the silica evaporated with hydrofluoric acid, the residue taken up with hot, strong hydrochloric acid and added to the filtrate from the silica. Iron, alumina, and phosphorus precipitated with ammonium hydroxid, washed, redissolved, and the phosphorus determined volumetrically, as in the magnesium nitrate method.

It will be seen from the table that in Soil I, containing very little organic matter, the results compare very favorably with the other methods, but in Soil II, rich in organic matter, the average is lower.

TABLE 2.-Duplicate determinations of phosphorus by carbonate fusion.

In Table 3 will be found duplicate determinations of total phosphorus in soils by the magnesium nitrate digestion. Of these, Nos. 1127, 1202, and 1204 are presumably Texas soils, and the determinations were made by Mr. E. C. Carlyle, of the Texas station. The others are low phosphate soils of western Kentucky and were made as checks in the general work of this laboratory.

TABLE 3.-Duplicate determinations of phosphorus by magnesium nitrate method.

Table 4 shows the modified method to compare very favorably with the regular J. Lawrence Smith method for total potassium, the former method giving in the general average 0.01 per cent more potassium in both soils. Taking into consideration the fact that the work is done on 1 gram the agreement is as close as could be expected. Referring to the work done by these methods in 1906 and 1907, it will be seen that the results are in the main concordant.

TABLE 4.-Comparison of modified and Smith methods for total potassium.

a Duplicates not reported.
Too late to be included in the general average.
c Not included in the general average.

COMMENTS OF ANALYSTS.

P. E. Brown: The magnesium nitrate method is undoubtedly quicker and easier of manipulation than the peroxid fusion method. It has the advantage that there is not nearly such a large amount of silica to get rid of, as it was found necessary to dehydrate three or four times with the peroxid fusion method. Then, too, in the first operation of the fusion method there seems to be an uncertainty of reaction while avoiding fusion, which is of course eliminated in the other method.

You will notice from the results that the agreement is fair with a tendency for the new method to give slightly higher results. However, if the first determination in Soil I by the fusion method is eliminated, the agreement is much better. On the whole the magnesium nitrate method seems to me to be undoubtedly superior to the other.

W. P. Kelley: I find the magnesium nitrate method as outlined by you to be a very simple and convenient scheme for determining the phosphoric acid in soils; and while I have not had an opportunity to compare this method with others, I have no doubt that the results are reliable.

RECOMMENDATIONS.

The work done this year, while not as extensive as the referee had wished, still warrants in his opinion three conclusions, especially when it is remembered that work along the same line last year and the year before is mainly concordant in the matter of results: First, that the modified J. Lawrence Smith method for total potassium compares very favorably with the regular method and is somewhat shorter; second,

that the sodium peroxid fusion method for total phosphorus gives good results, but the manipulation presents some difficulty, and the time required for making the determinations is a disadvantage; third, that the magnesium nitrate method gives uniformly as good results for total phosphorus as the sodium peroxid fusion, and is quick and easy of manipulation. With these facts in view, the referee would make the following recommendations:

(1) That the modified J. L. Smith method for total potassium be adopted as an optional method of this association.

(2) That the sodium peroxid fusion for total phosphorus be adopted as an official method.

(3) That the magnesium nitrate method for total phosphorus be adopted as a provisional method of this association and be further tested.

REPORT ON THE DETERMINATION OF CALCIUM CARBONATE IN

SOILS.

By JACOB G. LIPMAN, Associate Referee.

Systematic determinations of calcium carbonate in cultivated soils seem highly desirable in view of its important functions in crop production. Unfortunately, there is no unanimity of opinion among chemists as to the methods best adapted for this work. When the proportion of calcium and magnesium carbonates exceeds 1 per cent, fairly accurate determinations may be made by the liberation of carbon dioxid and its absorption and weighing in potash solutions. But when the proportion of carbonate is small, as is true of so many of our soils, the quantity of carbon dioxid which remains in solution in the acid is very large in proportion to its entire amount. This source of error has frequently been commented upon and has led to several more or less successful attempts to correct it.a

The associate referee on soils thought it advisable, therefore, to outline some cooperative work on one or two promising methods for the determination of carbonates in soils. Samples of two different soils were sent to eleven members of the association who had signified their willingness to cooperate in the testing of soil-analytical methods. It was suggested that determinations of carbonates be made in the samples by Knorr's method as described in Wiley's Agricultural Analysis. Where possible, the results secured by Knorr's method were to be checked by the method described by Amos in the Journal of Agricultural Science.c

The samples were sent out early in September, and analyses were made and reported by W. B. Ellett of the Virginia station, by Percy E. Brown of the New Jersey station, and Ernest Van Alstine of the Illinois station. Mr. Van Alstine's data were transmitted to the associate referee by Mr. Hopkins.

Mr. Ellett and Mr. Brown used Knorr's apparatus for the determination of the carbon dioxid. Mr. Van Alstine employed the method regularly used at the Illinois station and consisting of the liberation of the carbon dioxid "by boiling with hydrochloric acid and ascertaining the quantity of carbon dioxid evolved by measuring before and after absorption by a caustic potash solution." The results were as follows:

a See Hall and Russel, A Method for Determining Small Quantities of Carbonates, Transactions, J. Chem. Soc., London, 1902, 81:81.

b Vol. 1, p. 338.

c 1905, 1:322.

The results submitted by Ellett and Brown agree very satisfactorily. Those submitted by Van Alstine are markedly lower, especially in the case of soil No. 2. Apparently the amount of carbon dioxid which remained in solution in the latter work is the cause of the lower results. Evidently Knorr's apparatus is efficient for the determination of comparatively slight amounts of carbonates; however, it is desirable that further work be done along this line, and the associate referee would therefore recommend that it be continued with certain modifications for at least another year.

REPORT ON POTASH.

By B. B. Ross, Referee.

The work on potash for the past year has included cooperative tests of the regular official method in comparison with the phosphomolybdic volumetric method, and, in addition, the referee, associate referee, and some cooperating chemists have made comparative tests with some special methods which will be described in the latter portion of this report.

Twenty laboratories expressed a desire to take part in the cooperative work on potash samples, but reports were received from only eight laboratories.

Two samples were sent out for analysis to each laboratory taking part in the work, sample No. 1 being high-grade commercial sulphate of potash, while sample No. 2 was a mixed fertilizer, the ingredients of which were acid phosphate, cottonseed meal, dried blood, potassium chlorid, and a small amount of magnesium sulphate.

The following instructions with regard to the work were sent out to all cooperating chemists, the details of the volumetric method being those given by the referee for 1906 and 1907, Mr. A. L. Knisely, who had given much time and attention to a study of the phosphomolybdic method.

OUTLINE OF ASSOCIATION POTASH WORK.

Sample No. 1. Commercial sulphate of potash.

Sample No. 2. A complete mixed fertilizer, the nitrogen of which is derived from cottonseed meal and dried blood.

Potash in these samples should be determined both by the official method and the proposed volumetric method involving use of phosphomolybdic acid.