fixed in the seed and the skin at the ends removed, thus saving the sweetest part of the fruit. The illustration on page 326 shows a special mango fork secured in Mexico by Dr J. N. Rose. The long slender tine in the center easily penetrates the seed, and the shorter outer tines need only to touch the seed to prevent it from turning.

The best varieties of mango have hardly any fiber and the pulp is sliced with a knife, or sometimes is so soft that it is eaten with a spoon.

Porto Rico seems very well adapted to the production of mangoes and, as the plant is strictly tropical and very susceptible to cold, would seem to have a decided advantage over Florida, where good varieties are already successfully

grown, but where, except in the extreme southern part, the danger of injury from cold is very great. A really high-grade mango is unknown in Porto Rico, and the first steps toward making their exportation profitable is the introduction from the other islands, or from Florida, Mexico, or the East Indies,, of grafted stock of the best varieties.

The season of ripe mangoes in Porto Rico is from May to August. By selecting proper varieties this might be prolonged, since in some parts of India it extends over a period of six months. This would be a great advantage in shipping the fruit to temperate regions, as at present the season coincides with the season of temperate fruits, which places the mango at a decided disadvantage.

M

ANY people think that the rainfall, although differently distributed through the year, averages about the same one year as another, or if there is a deficiency one year, it will be made up the next. With this erroneous notion in mind, those concerned with navigation on the Great Lakes have naturally looked for some other explanation of changing water levels, for from 1888-1901 they witnessed a period of low water so long it seemed unreasonable to expect it ever to attain its former level. However, a comparison of the level of Lake Erie, as shown by the gage at Cleveland, with the record of rainfall along the Great Lakes shows a complete correspondence.

The high water in Lake Erie in 1902 and the heavy rainfall of that year are fresh in the minds of those who live near it.

The Weather Bureau established a number of stations on the Great Lakes in 1870. The first marked deviation from normal level in Lake Erie after this

was in 1872, when the water was lower than for many years before or after. The rainfall that year was below the normal at every station on the Great Lakes. (I have taken no account of stations on Lake Ontario.) In 1876 the water was higher than for many years before and higher than any year since. The rainfall was above normal at all stations except Marquette, where it was nearly an inch below. At Milwaukee the excess was 18.28 inches; at Grand Haven, 11.52; at Detroit, 8.07.

In 1878 the lake was considerably higher than the preceding or following year. The rainfall was a little below normal at Duluth and Grand Haven, but above at all other stations, being 60.24 inches at Buffalo, where the normal is only 38.04, and 53.51 at Cleveland, where the normal is 36.29.

In 1882 the lake was higher on an average than in any other year since 1876. The rainfall was below normal at Buffalo, Detroit, and Milwaukee, but

above normal at the ten other stations, the excess at most of them being greater than the deficiency at any of these three. In 1890 the water was higher than in the years immediately preceding or following. The rainfall was not far from normal on the upper lakes, but above normal at all Lake Erie stations.

In 1895 the water was the lowest for half a century, and the rainfall on the Great Lakes probably the least, certainly the least recorded at the Weather Bureau stations since their establishment.

These include all the years that differ in any marked degree from those that precede and follow.

If we consider parts of years we find also a close agreement between rainfall and lake level. Examination of the monthly record of lake level at Cleveland led me to think the rainfall at the different stations must have been below normal for the first half of 1888 and the last three months of 1887. On consulting the record I found it so. To exactly account for the stage of the water during brief periods, of course several things must be considered-surplus or deficiency in the different lakes at the beginning of the period, time required for water to flow from the upper lakes, evaporation, melting of snow on the watershed, whether ground is frozen, whether rain falls gradually or so fast that a larger portion passes quickly into the streams.

Most of the time since 1887 Lake Erie has been lower than for many years before. The rainfall has also been less,

as the table shows. If any one could tell us when this dry cycle will give place to a wet one, the information would be highly appreciated. Perhaps

the wet one has already begun. At any rate, those concerned need not fear any appreciable lowering of Lake Erie below its level in the past decade from any other cause than drouth. I believe that people now at Cleveland, Toledo, Detroit, Milwaukee, and Chicago will live to see the water higher than their fathers ever saw it. The same cannot be said of places on Lake Huron or the northern part of Lakes Superior and Michigan, for the slow tilting of the earth's crust is such as gradually to lower the water in those regions.

MEAN ANNUAL RAINFALL ON THE GREAT LAKES

THE RAILROADS AND FORESTRY

TH

HE Bureau of Forestry has continued this year on a far larger scale the experiments in timber seasoning and preservation for the railroads.

which it began last year under Dr Hermann von Schrenk. This work will be done for the New York Central, the Erie, the Baltimore and Ohio, and the Pennsylvania railroads in the East, and for the Illinois Central, the Santa Fé, the

St Louis and San Francisco, the Missouri, Kansas and Texas, the Northern Pacific, and the Burlington in the South and West.

The scarcity of valuable timbers is felt by no class of consumers more keenly than by the railroads, which use every year 110,000,000 ties merely to renew those worn out and decayed. The price of timbers has risen in some instances to a figure which makes their use prohibitive; in other cases the supply is so nearly exhausted that the roads have been compelled to look about for new timbers.

The Bureau of Forestry has been called on to assist in solving the difficulty, and has come forward with the very practical and simple suggestion that the railroads, instead of continuing to use expensive, high-grade timbers for such a low-grade purpose as that of railroad ties, shall use the cheaper woods. For example, to the complaint of the New York Central that it finds it more and more difficult to secure longleaf pine ties from Georgia at the price it can afford to pay, the Bureau suggests that the road use the beech, maple, and birch of the Adirondacks. The complaint that the timbers rot very quickly when laid in the ground is answered by the suggestion that they should be seasoned and preserved, just as beech is seasoned and preserved in France. The Great Eastern Railroad of France has succeeded in making beech ties last 35 years by impregnating them with tar oils. The unseasoned longleaf pine ties used by the New York Central last only five years; and the beech, if laid green, without seasoning or preserving, would in many cases last no more than three years. substance of the proposal which the Bureau has made to the railroads, and which the railroads have thought so well of as to adopt, is that experiments be made to determine whether cheaper timbers may be treated with preserva

The

tives at a cost so low and be made to last such a long time that it will pay to substitute them for the more expensive timbers now employed.

The railroads have thought so well of these ideas that they will not only carry on under the Bureau's direction the necessary experiments in seasoning and preserving, but have engaged the Bureau's help in learning where cheap timbers for ties may be obtained. In other words, the railroads have decided that if they can be convinced that it will pay to season and preserve cheap timbers for ties, they will acquire large areas of timber lands on which they will grow their own trees, cut their own ties, and thus be assured of a steady supply. This means that some of the great railroads of the country are in a fair way to practice forestry on a very large scale, and to employ a great many foresters.

Work of a similar nature to the railroad experiments is being carried on for the American Telephone and Telegraph Company, which used last year 150,000 telephone poles and 3,000,000 feet of timber in cross-arms. Seasoning experiments are being conducted on chestnut telephone poles near Harrisburg, Pa., and on cedar poles near Wilmington, N. C.

Important and valuable as this work is to the railroad and telegraph companies, it is of far greater importance and value to the country at large. The use of cheaper timbers for railroad ties is in several ways an economic saving; it relieves the high-grade timbers of a part of the heavy demand that is being made upon them, opens a market for timbers for which there is now little sale, and affords splendid opportunities for conservative management of timber lands. The work is being prosecuted according to the regular coöperative system of the Bureau, by which the field and traveling expenses of the Bureau's agents are paid by those for whom the work is done.