second. To facilitate reading when the direction is very variable this instrument should have two time-scales, one of about 10 mm and one of 40 to 60 mm an hour, the former for use during the period before the eclipse when the local characteristics of the station are determined, and the latter only during the eclipse. If recording instruments are not available an Assmann ventilated psychrometer can be submitted for the thermograph, a wide-scale aneroid for the barograph, and the direction of the wind can be determined from observations of a vane reflected in a nephoscope, as was done during the eclipse of 1918; but, to allow for errors and accidental variations, observations must be made very frequently-every two or three minutes if possible and a full program will require at least two, and preferably three observers, one of whom should give his entire time to clouds, which have received little attention heretofore.

Concerning exposures of instruments, stations should be placed where there are no hills, trees, buildings, etc., likely to obstruct or deflect the normal wind in any direction; obviously, the more first-class stations, well-distributed, the better. Since temperatures in closed shelters lag appreciably, particularly during calms, a small, open shed having a double roof and suitable boards on three sides to exclude direct sunlight and heat radiated or reflected from surrounding objects, will be best for the thermometers and can easily be improvised from inexpensive materials. It should be erected over grass in an open space freely exposed to the wind.

To summarize, adequate observations and records during solar eclipses are no more difficult to secure than are data of uncertain quality and do not necessarily require elaborate or very costly apparatus; the chief requisites are instruments with wide scales permitting frequent readings, proper exposures and care, information concerning which is available in recent pupblications.

The meteorology of solar eclipses may not be of the highest importance, but, viewing a program of study simply as an experiment or an exercise in the measurement of very small quantities, the results of which may have an incidental value in astronomy or physics, it is believed that the small effort required is worth while.

Washington, D. C.,

12th August, 1925.

Cause of Shadow Bands

Plates exposed here (Middletown, Conn.) during the period of shadow bands in the total eclipse of the sun have been developed at Wesleyan University and prove beyond any doubt, it is said, that the bands were caused by irregularities in the density of the atmosphere. This is the theory that has been held for years, but is the first time that it has been proved by actual photographs. The plates were forwarded by Dr. A. E. Douglass, director of the Steward Observatory at Tucson, Ariz., who designed the camera. The conditions under which the photographs were taken were perfect. Dr. Douglass says that the same effects witnessed

during the period of shadow bands can be seen from distant electric lights or on stars viewed with a powerful telescope.-Boston Evening Transcript, May 2, 1925.

The shadow bands, rippling alternations of light and shade that chase over the landscape just before and after total solar eclipses, are due to disturbances in the atmosphere, according to Dr. Charles Clayton Wylie of the State University of Iowa.

These phenomena, which have long been a puzzle to astronomers, were especially pronounced during the most recent of total eclipses, which darkened the populous northeastern part of the United States last January. Dr. Wylie states that this fits in well with his theory, inasmuch as the atmosphere above cities is much disturbed, especially in winter, by rising currents of warm air.

Dr. Wylie has tested his theory on a small scale, with the image of a bright star in a darkened room. "If the light from a star such as Sirius is allowed to fall on a white surface in a room otherwise dark," he said, "a person of keen eyesight may see a pattern of light and dark mottlings, because the source of light is a point. Ordinarily in sunlight these mottlings are not seen, because the patterns overlap, but at the time of an eclipse, just before and just after the moon covers the sun, a narrow sliver of light remains, which is practically a line, and so the overlapping is in one direction only, and the effect may resemble the stripes in a flag.' -James Stokley, of Science Service, Sept. 11, 1925.

DROUGHT AND HEAT IN THE SOUTHEAST

Weather history has been made "with a vengeance" in the southeastern states during the past crop season. Persistent low rainfalls and high temperatures for months blighted the crops and pastures. "Many localities had ten, fifteen, or even twenty inches less rainfall than usual. In Georgia, crop losses were particularly widespread. At Nashville, Tennessee, many wells went dry for the first time, the river was the lowest on record, and the use of water in the city had to be restricted. Montgomery, Alabama, and Norfolk, Virginia, experienced the driest August in the fifty-three years of record.” (Why the Weather? No. 773, from Weekly Weather and Crop Bulletin, Oct. 6, 1925). The drought began in February and lasted into October. A few contributions received from different parts of the drought stricken area give samples of the extraordinary meteorological situation.

South Carolina

(Excerpts from article by RICHARD H. SULLIVAN, in The State, Columbia, S. C., Sept. 7, 1925.)

"The present drought is the most widespread, most severe and most disastrous in the history of this state, as far as climatological records are available.

A tentative computation for August, developed from reports of 25 selected stations, shows an average of 1.78 inches, or 4.22 inches below normal, and the lowest August state average of record. By districts. the total average precipitation for the five principal cropgrowing months, is as follows: The coastal plain, 16.46 inches as against a normal of 23.76 inches; upper coastal plain and lower Piedmont, 14.10 inches as against a normal of 19.80 inches, and the Piedmont, 13.85 inches as against a normal of 21.65 inches. These data show a shortage

5.70 to 7.80 inches at times when most needed for crop development and to replenish springs, wells and stream-flow.

"In the case of the current season, abnormally heavy rains up-state in January were followed by floods of considerable volume in the various streams. These conditions were followed by cumulative monthly deficiencies in rainfall from that time to the present, ranging from 0.92 inch to 4.22 inches below normal per month. The total deficiency of 4.84 inches in February and March practically offset the excess of 4.95 inches in January, so that the crop season opened, say on April 1, with the rainfall going into the usual normal spring minimum; and the succeeding deficiencies simply aggravated the drought situation.

"Under crop-growing conditions, it has quite frequently happened that drought, or, more properly, lack of serviceable growing rains, would cover a period of a month, possibly two months and rather rarely, three months; but the growing season has seldom ended without good rains in July or in August. It also happens that droughts are most frequent during the periods centering around the time of minimum normal rainfall in spring and autumn.

"Conditions like these carry in their trails the evils of heat and drought curtailment of crop production and its reflection in commercial activities, reduction in the volume of water in springs, wells and streams necessary for the welfare of man and beast, and a general disarrangement of the power situation, now a most important factor in this state.

"Whenever a serious drought or heated term occurs, the writer has frequently heard statements that it does not rain or snow as it did formerly, that it does not get as hot or cold as in former times, etc. While the records of South Carolina are not of sufficient length satisfactorily to demonstrate the Brückner theory of climatic oscillations and periodicities, and fragmentary records are not always acceptable, the general averages are ample in number to show certain changes with which all meteorologists are familiar-changes that are occasionally detected by observant sages.

Human memory is defective, and we are prone to enlarge upon certain peculiarities or vagaries in climate until they become magnified beyond their just values. But there can be no doubt of the length and strength of the drought of the spring and summer of 1925."

Mr. C. G. Andrus, of the Weather Bureau's kite station at Due West, South Carolina, has compiled some interesting statistics concerning the weather there since April 1. He calls attention to a deficit of 20 inches of rainfall and a surplus of 1000 day degrees of temperature in his vicinity since January 1. In discussing general pressure and wind conditions he says in a letter dated October 16:

"One of the notions prevalent in the local college and with casual observers is that the strength of the Bermuda high directs the temperature here, both rising and falling together. I noticed, however, that on most of the 100° days a high would overlie the southeast with a lobe or even a separate crest over the mid-Gulf coast, and, therefore, considered the changes in Bermuda pressure might not play the greatest part. Up per winds here have been consistently out of the northwest eighth of the compass since February, and except for that moisture extracted from ground sources, the upper air carried no "makins" for rain, and of course we never received much. The requisites for extreme convection were present, however, and kites and balloons showed this plainly.

"The summer was one of extraordinary discomfort, due partly to the excessive heat and partly to the unusual clearness of the sky.

"In all of the 1925 months beginning with February and including September, the mean daily range was the highest on record for the several months, due to abnormally clear skies."

The cloudiness each month from February to September was the lowest in the five years of record. The total rainfall from April 20 to September 22 inclusive, was 7.47 inches, the normal for the same period being approximately 21.80. On 18 days when the temperature was 90 or over a rainfall of .01 inch or more was recorded. The first 90 temperature was recorded April 20, and the last September 22, both the earliest and latest of record.

ALABAMA

September, 1925, in Alabama will long be remembered for its intense heat. It was the hottest month since the beginning of State-wide records in 1884, a fact that is most remarkable, especially when it is considered that most of the hot weather came two months after the normal peak of summer. All State records for high temperatures were broken on the 5th when the thermometer at Centerville reached 112°. The highest temperature previously recorded was 109° in 1902 and 1914. Records are not numerous prior to 1884, but there are a few complete for more than 50 years, and fragmentary ones extending back about 100 years; these indicate that in all probability no higher temperatures have been experienced in Alabama within the last century. At three-fourths of the reporting stations the absolute maximum temperature record was equalled or exceeded, while at nine stations this record was broken by as much or more than 3°. The lowest monthly maximum temperature at any station was 103°, and at more than half the stations the maximum temperature was above 105°. The heat was prolonged as well as intense. Two successive heat waves, occupying nearly the whole month, brought the largest number of hot days ever recorded. In places the maximum temperatures were above 90° on nearly every day, while temperatures of 100° or above occurred on an average of 10 days, and at some stations on half the days of the month. The mean temperature was more than 2° higher than for September, 1921, also a remarkable month, and hitherto the warmest September of record.

Notwithstanding the long spell of extremely hot weather, there was little interruption of activities. Work, outdoors and indoors, went on about as usual. There were few heat prostrations, and very few resulting deaths.-Alabama Section, Climatological Data, U. S. Weather Bureau, Patrick H. Smyth, Meteorologist.

Another Summer in Saint Louis

Some "man in the street" mimeographed the following as his impressions of heat in St. Louis:

On the top of the Railway Exchange Building, 265 feet above the street, the U. S. Weather Bureau keeps the instruments that tell us what sort of summer we have really had. From their records we learn that during the first half of July the temperature rose to 90° or above on every day but one; but that during the latter half there was only one day when it did reach that mark. The average of all the highest temperatures for the month was 87.4°, which is exactly normal for July. In