Thirdly, I think the frequency of rainfall in the two localities is different. There may not be much difference between the total rainfall at the two points, but I think there is some difference, which might have a considerable influence upon the development of the sugar beet, in the frequency of showers in the two localities.

It would, of course, require a special research to discover what part the soil, either physically or chemically considered, played in this problem. The mechanical analyses of the soils and subsoils at Ithaca and Geneva, made by the Bureau of Soils, are given on page 38 of Bulletin No. 78, report of 1902.

It has been several times mentioned that the lack of sunshine data for Geneva was much regretted, but these observations were not obtainable for any point near enough to be of value. The suggestions made as to the effect of sunshine, length of day (Geneva having a day three minutes longer), and distribution of rainfall might work together to explain, at least in a measure, the inferiority of the Ithaca beets, in accordance with the theories developed by the observations and results at other stations, although no relation between sugar content and sunshine data has been established.

IRRIGATED SECTIONS.

These data are at best only preliminary, since they do not cover the whole period of the investigation. The experience of the last twenty years has shown the probability of a very wide extension of the sugarbeet industry on irrigated lands. The desirability of such an extension rests upon an economic basis. In the first place, the control by irrigation of the distribution of water renders the production of a crop practically certain. The other meteorological data are usually of such a constant nature as not to endanger the production of an average crop. The predominating factor, therefore, in so far as yield is concerned, is the distribution of the water. Thus it happens that the crop of beets that will be harvested in the arid regions may be confidently predicted within a few tons. Such a prediction renders all of the farm operations connected with the production of the crop more certain and more economical. In the second place, it is highly important to secure for irrigated areas a crop which shall have a high money value per acre with a reasonable margin of profit. The cost of bringing lands under irrigation as a rule is considerably greater than that of preparing land in the nonirrigated regions for cultivation. The actual cost, therefore, of the land, other things being equal, is greater in the irrigated than in the nonirrigated areas. This higher cost fastens upon the farmer a fixed charge which must always be provided for in the crop before a margin of profit is possible. The ordinary average crops do not always present the most hopeful avenue of securing this increase of profit. For instance, the amount of Indian corn or other cereals or grass crops, with the possible exception of

alfalfa or other species of clover, does not afford the opportunity of certainly discharging the obligation accruing from the interest on the investment in land. The sugar beet, however, adds further inducements in this direction because of the possible and even certain production of an average crop of not less than 15 tons per acre, having a money value delivered at the factory of from $60 to $75. These ideas are fully borne out by the data from the Utah and Colorado stations. The average yield at Logan is 18.9 tons per acre and at Fort Collins 20.4 tons. The average content of sugar in the beet is quite satisfactory at both places, being very good at Fort Collins. The purities are reasonably high, and the data collectively indicate a probable value at the factory of not less than $5 a ton. The yield at the Pomona (Cal.) Station indicates only half a crop, but the quality of the beets as shown by the data is satisfactory.

At all three of the stations where irrigation was employed the average temperature for the growing season is favorable to the production of beets high in sugar. Pomona, with the highest temperature (68.9° F.), being 1.1° below the maximum mean temperature of 70°. It is interesting to compare these stations with Geneva, N. Y., in this respect. The average temperature at Geneva for the six growing months is 64°; at Logan, 62.4°, and at Fort Collins, 59.4°. For the three principal growing months-June, July, and Augustthe average temperature at Geneva is 69.3°; at Logan, 69.5°, and at Fort Collins, 65.1°. The temperature at Fort Collins is uniformly lower than at the other stations, owing, as is readily seen, largely to its high altitude.

PURITY.

It has been the general observation during this investigation that the purity coefficient always increases with the sugar content of the beet. This general relation is shown very plainly in the tables. Lexington, with the lowest content of sugar, shows the lowest purity, and Geneva, with the highest content of sugar, shows the highest purity. There are, however, irrregularities in the curve representing these data, but these variations only serve to accentuate the general principle illustrated by the data.

PRECIPITATION.

NONIRRIGATED SECTIONS.

A study of the precipitation data for the six growing months shows a close agreement between the nonirrigated stations. The greatest average precipitation during the period observed was at Ames, Iowa, 25 inches. The smallest rainfall is recorded for Lexington, Ky., viz, 14.9 inches. This amount of rainfall is so small as to plainly indicate the chief cause of the small crop produced at that station. At all the

other stations it is seen that the average rainfall is about 20 inches, showing a remarkable uniformity with the exception of Ames and Lexington. The full value of these data can only be determined by studying the distribution of the rainfall in the various annual reports of this investigation. It is evident at once from a study of this kind that the total amount of precipitation for the six months is really greater than the crop required if it could have been distributed evenly and at the proper times. This of course is not possible when dependence is placed upon natural causes alone. In the detailed records. discussed in previous bulletins it has been clearly pointed out that the rainfall does not directly, but only incidentally, affect the sugar content of the beet. The great function of the rainfall is related to the magnitude of the crop. Incidentally the rainfall affects the sugar, content of the beet in the following way: If the distribution of the rainfall or its deficiency is such as to produce a very small crop made up of small or undersized beets, it exerts the incidental tendency of increasing the percentage of sugar in the beet. A very abundant and well-distributed rainfall by supplying the conditions to grow a beet of extraordinary size will have the opposite effect, of diminishing the percentage of sugar in the beet. If after a period of dry weather during which the beets have matured, as shown by the change in color and the falling of the leaves, there comes a period of warm, wet weather, a second growth will be induced in the beets during which to a certain extent the stored sugar is consumed, with the incidental result that the percentage of sugar in the beet will be diminished. The total changes in percentage composition induced by the rainfall may reach. considerable magnitude and may determine whether or not in the process of manufacture favorable economic results will be obtained. Far more important than this effect, however, are the relations of the rainfall to the magnitude of the crop-relations which are well established but which do not form an essential part of this investigation.

IRRIGATED SECTIONS.

The inspection of the table of average precipitation shows that the Fort Collins Station approaches in its average rainfall that of Lexington, Ky., which had the least precipitation of any of the nonirrigated stations. The average precipitation for the six months at Fort Collins is 11 inches, while the average precipitation of Lexington is 14.9 inches. The average precipitation at Logan, Utah, for the six growing months for three years is 5.9 inches. The station at Pomona, Cal., may be regarded as actually arid, since the average precipitation for .two years is only 3.7 inches, making it evident that the natural precipitation at this station is a wholly unimportant factor.

CLEAR DAYS AND SUNSHINE.

In the detailed records representing the work of the different years attention has been called to the fact that the active principle of light, in so far as it affects the sugar content of the beet, is not probably the most luminous element. It is apparent, at least to a certain extent, that the diffused light from a cloudy sky has practically the same effect in producing sugar as the direct sunlight. The study of the environment, therefore, in respect of the number of clear days and the hours of sunshine does not show as close a relation to the production of sugar as was expected before the investigation was begun. Of the nonirrigated stations the one showing the largest average number of clear days during the five years is Ames, Iowa, while Ithaca represents the other extreme. Of the irrigated stations Logan shows the largest number of clear days and Fort Collins the smallest.

In respect of percentage of sunshine at the nonirrigated stations the highest figure occurs at Lexington, viz, 71.6 per cent, and the lowest at Blacksburg, 53.7 per cent. At the irrigated stations the highest percentage of sunshine is found at Logan and the lowest at Fort Collins. It is evident, therefore, from a study of these average data, that it is not possible in the present state of our knowledge to find any direct relation between the content of sugar in the beet and the number of clear days and percentage of sunshine.

COMPOSITION OF SOIL AND YIELD PER ACRE.

The data appearing in the accompanying table show that Lafayette, Ind., had the smallest yield and Madison, Wis., the largest yield during the series of four years. The crop both at Lafayette and Lexington must be regarded as abnormally small. The yield at the other stations is satisfactory. It was not the purpose of the present investigation to study the soil as a factor of the environment affecting the yield, but only as affecting the sugar content. It is evident that there is only one method by which such a study could prove of value, viz, the establishment of the experiment under conditions which would eliminate all the varying factors with the exception of the composition of the soils themselves. This ideal method of studying the effect of soil on sugar content is to bring the soils, in sufficient quantity, from the different stations to one station, thus eliminating all disturbing factors of the environment save those due to soil alone. This test was included in the original scheme for these studies, but opportunity to carry it into effect was not presented. Inasmuch as in all instances the distribution of the rainfall has been shown to be one of the dominant factors in determining yield, it is hardly necessary to continue further any comparison of the soil and yield data which were merely incidental to the principal purposes of the investigation.

a Average for two years.

b Averages for three years except where otherwise noted.
c Data for one year.

The data show that any direct effect that soils may have upon the sugar content is that which has already been alluded to in sufficient detail, viz, the soil diminishes the sugar content when there is a tendency to produce an overgrowth due to an excess of plant food, and has a corresponding tendency to increase to a slight extent the sugar in the beet where the amount of plant food is not sufficient to produce a normal growth.

The study of the use of fertilizers and manures in the growth of beets and the effect on the sugar content naturally would be considered in connection with the study of the effect of the soil. Such a use of manures, including what are known as commercial fertilizers, as would produce an extraordinary tonnage would tend to diminish the percentage of sugar in the beet. In addition to this it has been shown that certain kinds of fertilizers and manures are more effective in this direction than others. This is especially true of the nitrogenous fertilizers, both when used in the form of nitric acid and also in organic compounds. This class of fertilizers tends preeminently to produce increased tonnage, to develop abundant leaf and root growth, and as a result to diminish to a slight extent the percentage of sugar in the beet while greatly increasing the tonnage. On the other hand, phosphoric acid, and to a less extent potash, show a tendency to bring the beet to an early maturity, thus checking undue growth and increasing to a slight extent the percentage of sugar in the beet. When, however, we consider the total yield of sugar per acre rather than the actual percentage of sugar in the beet, it is evident that that system of manuring which would produce a very much larger crop will cause an almost corresponding increase in the total sugar produced per acre. The general conclusion from the study of these experiments and similar ones made by other investigators is that the soil and, to a somewhat less extent, the fertilizers and manures have only a limited influence upon the actual content of sugar in the beet, and that influence is incidental rather to the vigor of growth than to any specific action on the sugar content itself.