which has been generally cold in the East, though still mild in the West. The accumulations of snow in the Northeast became unusual, Quebec having 54 inches on the ground, and much of Maine and New Brunswick over 40 inches on March 8.-C. F. Brooks, in "Why the Weather." (Science Service).

The Winter's Snowfall

With the middle of March at hand, and the date of the "blizzard" of 1888 past, we may talk of the snowfall of the winter of 1925-26 as largely a closed incident. Much of the winter's snow fell in October; in fact, the snowy period began in September. Over wide regions, snow came at dates earlier than any previously known record and fell in unprecedented quantity: 6 feet in parts of the central and northern Rockies, 2 feet on the piedmont in Montana, and up to a foot or more in the Northeast. Snow occurred almost throughout the country south to the 35th parallel. This did not happen again till late January. November was generally much less snowy, except in the central plateaus and southern Rockies and from the middle plains to Wisconsin. Heavy snows also fell on these regions in December. Moderate depths covered virtually all of Texas and part of Mexico late in the month, and again four weeks later.

The first important snowstorm did not affect the East till January was well begun. Some more a fortnight later still left January's total far below the average. Though there had been 4 feet of snowfall in western New York in December, the snow in northern regions to the end of January was generally inadequate for the usual winter logging. But the first half of February buried the Northeast under 2 to 4 feet of new snow, and in the northern portions more and more snow accumulated. March opened with 38 inches of snow on the ground at Quebec; then it snowed much more, and in western New York brought on impassable conditions.-C. F. Brooks, in "Why the Weather." (Science Service),

Heavy snow fell over a broad belt from south New Mexico to Lake Michigan and beyond at the close of March.

A 1925 Record From the Middle West

Mr. A. M. Hamrick sends us the following notes, which were printed in the Davenport, Iowa, Daily Times, Dec. 31, 1925:

SIXTEEN NEW RECORDS

1. A maximum temperature of 87 on the twenty-second of April, 1925, is the highest temperature on record for Davenport that early in the spring.

2. The snowfall on the twenty-eighth, twenty-ninth, 0.7 inch, is the heaviest snow that ever fell in Davenport that late in the season.

3. The average daily range in temperature, from minimum to maximum, 23.7 degrees, in May, is the highest on record in Davenport for any month.

4. A maximum temperature of 94 degrees on the twenty-second of May, 1925, was the highest ever recorded in this section in any May during the last 53 years.

5. The Mississippi river fell to 1.3 feet above zero on May 30, the lowest stage on record for the month of May at this place.

6. On June 2 the stage of the Mississippi river was 1.2 feet above zero, establishing a new low record for June.

7. A total rainfall of 8.16 inches in September makes 1925 the wettest September in the history of the weather bureau at this place.

8. On five days in September the rainfalls amounted to one inch or more a new record.

9. There was more cloudiness in September, 1925, than in any other September on record. The average cloudiness was 6.4, on a scale of 0 for a clear month and 10 for a completely clouded month.

10. A maximum temperature of 92 on the nineteenth of September, 1925, set a new record for highest temperature so late in the season.

11. October, 1925, was the coldest October in this section, during the last 53 years.

12. A minimum temperature of 11 degrees on the twenty-eighth of October set a new record which will probably stand for a long time.

13. October 29 was the coldest day ever before experienced in October in this section, with a mean temperature of 18 degrees.

14. The temperature remained below the freezing point for 72 consecutive hours in October, a new record.

15. The sun shone on Davenport only 34 per cent of the possible time in October, 1925, making it the gloomiest October on record here.

16. The total snowfall, 10.8 inches, in November, 1925, establishes a new record for that month, and the 24-hour fall of 5.7 inches on the seventh is another record for November 24-hour snowfall.

October, 1925, will go down in the records as the year without an "Indian Summer." The mean temperature for the month was 42.6, the coldest October in this section during the last 53 years. The daily temperatures were below normal on all but the first three days of the month. Minimum temperature of 11 on the twenty-eighth was the lowest on record for October; and the coldest day on record for October was established on the twenty-ninth with a mean temperature of 18. The previous coldest day in October was the twenty-fifth, 1887, with a mean temperature of 24, therefore the 1925 record will probably stand for a long time to come. The maximum temperature did not go above 72 during the month, and with the exception of October 30, 1917, it is the only October on record in which the temperature remained below the freezing point for an entire day. The temperature remained below the freezing point for 72 consecutive hours, from 2 P. M., October 27, to 3 P. M., October 30. The sun shone only 34 per cent of the possible time, 26 per cent below normal, and 6 per cent below the previous low record for October, established in 1911. Rain fell on an average of every other day during the month, although the total was only 3.01 inches.

Tornadoes and Other Severe Winds of 1925 (Written by H. C. Hunter for the U. S. D. A. Clip Sheet, and published

there in abbreviated form)

To anyone with a general idea of the total losses of life and property caused by tornadoes in this country, the newspapers of last March 19 and 20 indicated that the year 1925 would surely find place at or near the top in the matter of such losses.

The last of the 1925 monthly tables of storms, prepared for the Monthly Weather Review, has been completed. These tables are compiled for early publication, so the figures are subject to some correction by way of later, more careful study, especially to increases.

It now appears that 1925 must stand very high not only in the aggregate losses, but in the total number of individual tornadoes reported to the Weather Bureau. No less than 115 are indicated, compared with 106 listed in the 1924 monthly tables. By careful compiling at leisure, the total count of the 1924 tornadoes was found to be 130, or 24 more than the monthly tables showed; these additional tornadoes were chiefly storms of minor importance, occurring in sparsely settled districts. However, it is likely that additional tornadoes in 1925 will similarly come to notice, raising the 115 to some extent, and very likely to more than 130, which is the largest count yet found among recent years.

The 1925 loss of life by tornadoes was very large, 853 being the present estimate; this is unlikely to be changed as much as three per cent by later study. It is more than twice the life toll of the 1924 tornadoes, which was 376. The years 1896 and 1917 each slightly exceeded 500 in deaths due to tornadoes, and it seems probable that 1884 also reached a large figure, because of the dozens of serious tornadoes on February 19, though the information now readily accessible indicates that that day was by no means so tragic as March 18, 1925. The reports about this latter day indicate that the seven tornadoes caused 792 deaths, or more than nine-tenths of all tornadoe deaths during 1925.

After March, October is indicated as the worst month of 1925 for tornado fatalities, 26 being reported. This is an unusual period of the year for severe tornadoes, the month very often passing without any such storms coming to notice in the whole country. The four months, April to July, inclusive, usually about the worst during the year, were in 1925 less marked by important tornadoes than ordinarily, and this was particularly true of May.

It is more difficult to judge of the total property losses from the monthly storm tables, but the present indication is that the tornadoes of 1925 caused damage aggregating considerably more than twenty million dollars, probably about twenty-five million. The 1924 tornadoes set the highest mark, over $26,100,000.

There are very few years when either California or Oregon reports a tornado, but in 1925 each seems to have had one; the Oregon storm on November 11, was of small importance, and the California one not of great note, as tornadoes average the country over.

There have now been 25 months in succession, from January, 1924, to January, 1926, inclusive, with at least one tornado reported in each month. This is an unusual record, as ordinarily at least one fall or winter month of the year passes with no news of any tornado.

In regard to the losses of property and life by severe winds, not tornadoes, the year 1925 seems to have fallen somewhat short of the 1924 figures; these latter indicated 78 deaths and between 13 and 14 millions of losses, the deaths being fewer and the property destruction slightly greater than in the average recent year by such winds. The losses by windstorms other than tornadoes, especially the losses of life, are not likely to total much more per year than the figures just indicated unless the year happens to include an exceptionally severe hurricane of tropical origin striking some portion of the South Atlantic or Gulf coast.

A POSSIBLE WEATHER AND OCEAN CONNECTION

Temperatures reported by ship masters for October, November, and December, 1925, showed the highest temperatures of the Gulf Stream in the Straits of Florida for this season for at least 6 years. The departures were, respectively, 1.4, 1.5, and 0.5 degrees F. above the October to December averages for 1920 to 1924.

Since warm waters favor warm air and high humidities and, therefore, lower pressures and greater storminess, it is very easy to guess that these warm waters south, southeast, and now perhaps east of us had something to do with the following extraordinary features of our weather during the past few months: (1) stormiest, cloudiest, coldest and snowiest October known generally over the United States from the Rocky Mountains to the Atlantic; (2) storminess and coldness of the southern plains and Mexico east to the south Atlantic coast, November to January, including a very late tropical hurricane; (3) unprecedented disturbances of the North Atlantic and extraordinary snowiness of the North Atlantic states and Maritime Provinces from late January into March, 1926. Pressures in the mid-North Atlantic through February were extraordinarily low.

Of course, we cannot yet prove that these weather abnormalities were actually due to the warm waters, even in part; but if we were to specify ocean temperatures which would favor such weather, we should describe those observed.-C. F. Brooks, in "Why the Weather." (Science Service).

INTERNATIONAL MONTH FOR AEROLOGICAL OBSERVATIONS

Meteorologists generally will welcome the announcement that investigations with sounding balloons are to be resumed in this country during May of this year. In years past several series of soundings to great heights have been made, notably those at St. Louis, Mo., by the Blue Hill Observatory and later those at Omaha, Neb.; Huron, S. D.; and Avalon, Cal., by the Weather Bureau. During the war this type of investigation was temporarily suspended, partly owing to the difficulty of procuring suitable balloons, but largely because it was deemed wise to

concentrate on the more practical work of obtaining data in the lower strata of the atmosphere for the benefit of aviation. Since the war this survey of conditions at flying levels has been continued, the data serving the double purpose of furnishing current information for flying weather bulletins and forecasts and of forming the basis for special studies and summaries which set forth in considerable detail the characteristics of the atmosphere up to 8 or 10 kilometers.

The need for similarly detailed information at greater heights, well up into the stratosphere, is urgent. The data thus far obtained are not sufficiently numerous and are not suitably distributed through the year and over the country to show the variations in the temperature of the stratosphere and in the height of the tropopause with season, latitude and pressure types. For procuring the desired data several extended series of observations are necessary, some of them carried on simultaneously at widely separated points, and including in some instances successive soundings at short intervals through day and night. The coming series in May is the first of several which, it is hoped, may be carried out during the coming two or three years. This first one may properly be described as experimental or preliminary, in some respects. New meteorographs, designed by S. P. Fergusson, will be given their first thorough test, new types of balloons will be used and some other changes in incidental equipment and in methods will be made.

The month of May has been chosen because it has been designated as "international month" for 1926, by the International Commission for the Exploration of the Upper Air. Heretofore, this Commission has selected certain days, usually one a month, but in some cases 3 to 6 in a group as "international days," and effort was made to have all countries send up sounding balloons on those days. At the meeting of the Commission in London, April 16-22, 1925, a proposal was presented by the U. S. Weather Bureau to substitute a month during which soundings would be made daily or oftener, thus securing observations bearing a close relationship one with the other, for the scattered days whose data would represent only isolated samplings. The Commission compromised by agreeing to the selection of an international month while retaining also, for the present, the custom of designating international days in the other months. May, 1926, was selected as the first international month and the meteorological services of all countries were urged to make daily soundings during that month. The Weather Bureau will send up sounding balloons at Royal Center, Indiana, and there will also be available, as its part in the international campaign the kite, airplane (furnished by the Navy Department) and pilot balloon observations, (including those by the War and Navy Departments) regularly made at several stations widely scattered over the country.

The meteorological services of Canada and Mexico will also participate, the former with sounding and pilot balloons, and the latter with pilot balloons and airplanes. Thus, it is believed that North America will offer a very real contribution to this first effort to study internationally the upper regions of the atmosphere for a period of some 30 consecutive days.-W. R. Gregg.