stations are made in whole degrees F. The use of the Centigrade would require the use of fractions of degrees for the same accuracy, and this would greatly increase publication costs. Why not combine the advantages of conforming with physicists and the non-English speaking world, on the one hand, and of having reasonably small degrees, on the other hand, by using a half-degree C. scale, with 200 degrees between melting and boiling points, and with the melting point at approximately (2 X 273 =) 546° on an absolute scale. Whatever future determinations of absolute zero might be, the size of the degrees would be the same. For convenience in meteorological observations the first digit, 5, could be left off, giving 46° as the melting point and thus pushing still lower than in the F. scale the point at which minus temperatures would need to be recorded.]

[In the discussion, P. C. Day said that the publication of such figures would cost no more than now. W. J. Humphreys remarked that the Coöperative Observer might object to having a new scale thrust upon him, and that his thermometers might be supplied with two scales. B. C. Kadel mentioned the desirability of World uniformity, but raised the question as to whether countries now using the C. scale would adopt it. Furthermore, the adoption of degrees of F.° size was a recognition of the value of the present Fahrenheit scale. If a new scale is to be adopted the claims of scales made on other principles, e. g., Dr. Aspinwalls's "Homograde” scale (90° from melting point to blood temperature), are worthy of careful consideration.]

The Physics of the Aurora. W. J. Humphreys. Mo. Weather Rev., Abstract, May (?), 1920.

[The rapid changes that occur during auroral displays in the direction and strength of the magnetic force of the earth, the interferences with telegraphing, and the directions of the streamers of light, are some of the things that make it quite certain that the aurora is due to electric discharges in the high, rarefied atmosphere. And as the more pronounced auroral displays occur at the times of large sun-spots it is also evident that they are somehow connected with solar activity. It is known that the lower edges of the auroral arcs, bands and streamers, generally are 60 to 65 miles high; and it is well-nigh certain that all the electric and other phenomena of the aurora are caused by the bombardment of the outer atmosphere by solar particles of great velocity.]

[In reply to a question during the discussion, C. F. Brooks stated that the general westward drift of the elements in an auroral display is due to the eastward turning of the earth under the non-rotating display.]

The auroras of March 22-25, 1920. Herbert Lyman. Mo. Weather Rev., May (?), 1920.

[A popular notion of the nature of an auroral display was illustrated by the question of a local garage man relative to the aurora of March 22: "Who's running it anyway?" The numerous accounts of this display received by the Weather Bureau indicate that the aurora of the night of March 22-23 was visible in every state of the Union [and all the provinces of Canada, and also in Europe]. It was visible even from southern Florida. In the North the display was mostly in the southern sky at first. At about latitude 40° the display was mostly overhead. Far in the South the aurora was visible only in the northern sky. Selected descriptions were read. The predominant characteristics were the general cloud-like background of light, the brilliance, great extent, and long duration.]

[In the discussion, C. F. Brooks mentioned the occurrence of a faint display on the following night, and called attention particularly to a "spot-light" display all night March 24-25. [Note: Observations of these spots from 6 places indicates that the aurora was at a height of about 80 miles and that two spots central over western and southeastern Pennsylvania in the early evening gradually moved southward and spread out, extending in a belt from southwestern Ohio about 15° S. of E. to a few hundred miles out into the Atlantic.] Another mentioned that a similar spot-light display had occurred at the end of a three-day display in the latter part of Sept., 1919.]

The comments took another turn when C. Fitzhugh Talman, Chairman of the Committee on Public Information, read the account of this display as published in a Washington morning newspaper of March 23. This article contained the astonishing statement, alleged to have been made by Rev. F. Tondorf, director of the Georgetown University observatory, that "the flashes were in reality but the reflection of the aurora borealis of far northern latitudes," and an equally startling announcement, said to have emanated from the Naval Observatory, to the effect that the flashes "were caused by the peculiar atmospheric conditions of the last few days." Prof. Talman read letters from Father Tondorf and from the Superintendent of the Naval Observatory, denying that such statements had been issued to the press. The latter wrote that no responsible official of the Naval Observatory, and, so far as he could ascertain, no other person connected with the Observatory, had supplied the paper in question with any information regarding the recent aurora. The speaker remarked that this episode was characteristic of the way in which the leading American newspapers, with three or four exceptions, write up scientific events, and he pointed out some of the really important facts connected with the aurora, such as the passage of a large sun-spot group over the sun's meridian on the same date, which the newspapers completely ignored, and would have been given prominence in their reports if we had, in America, a news service for science comparable, in intelligence and efficiency, with that which exists for politics and baseball. The other morning newspaper of the capital, published a less silly and bungling report of the aurora; but any comfort one might be inclined to take in this fact was dispelled by turning to another page of the same paper, on which was published an article by Frank T. Allen, "Director, Astrological Research Bureau," entitled "What the Stars Say About Governor Cox." It is a humiliating fact that influential newspapers all over the United States are engaged in the nefarious business of keeping alive the gross superstitions of astrology. W. J. Humphreys recalled that one of the newspapers above mentioned some years ago, on the same (?) page that the Wrights were "roasted" for not flying for the edification of a crowd when weather conditions were not suitable, published a long account of how a man in Ohio had solved the problem of flying by nullifying gravity with the ringing of bells. He said that fortunately there is a prospect of combatting such nonsensical stuff by means of a constructive service backed by the National Research Council. J. Warren Smith upheld the newspaper man, stating that in his experience at Columbus, O., they wish to get the facts right and that anticipating demands, he usually had copy ready when it was asked for. In this way, the newspaper

accounts were accurate.

The most intense rainfall on record. B. C. Kadel. Mo. Weather Rev., May, (?) 1920.

[Mr. H. G. Cornthwaite, Acting Chief Hydrographer, Canal Zone, referred in an article on "Panama Rainfall" (Mo. Weather Rev., May, 1919, Vol. 47, pp. 298–302, 4 figs.) to a record of 2.48 inches which fell in 5 minutes at Porto Bello, Nov. 29, 1911. Upon Prof. Alexander McAdie's suggestion the original record was asked for. It shows that 2.47 inches fell in three minutes, at a rate of 0.82 inch (20 mm.) per minute. This exceeds by more than 100 per cent. the next highest record (Curtea-de-Arges). The total fall was 7.60 inches in 24 hours. The 12-inch tipping bucket apparently worked properly. The legible portion was first corrected and the remainder credited to the excessive period. The gage was empty at the beginning of the 24-hour period, the observation was carefully made with the usual stick measurement, no foreign substance was found in the funnel. The heavy rain was remarked: water was standing all round, and boulders were washed down the hillside. In the blurred part of

the record where the downpour was recorded it is possible to count 18 or 19 lines, representing 1.8 to 1.9 inches of rainfall. A similar tipping bucket gage experimented with at the Weather Bureau showed 194 tips (1.94 inch record) when 2.48 inches of water was poured into the funnel all at once. The water required 21 minutes to flow through. Even though there was a 17-minute slack period before the recorded time of the downpour, it seems unlikely that a foreign substance stopped up the hole and suddenly gave way, or that anyone tampered with the gage. The occurrence was at 2 A.M. and the gage was in an inaccessible spot on a slippery hillside.]

[There was considerable discussion of this paper. J. Warren Smith told of how, in a cloudburst in Ohio, in which 7 inches of rain fell in half an hour, people who were out in it said that they were almost drowned and had to hold their hands over their faces to get air. S. P. Fergusson asked if there could be such rapid condensation in the atmosphere. C. F. Brooks thought that the 17minute period of slack rainfall preceding the cloudburst was a necessary accompaniment to such an excessive rainfall, for the rate of rainfall much greater than any possible rate of condensation indicated that there must have been strong upward currents holding the raindrops up in the air, and that therefore there must have been downward current and little rainfall about the region of up-rush. H. C. Hunter called attention to the rainfall at Guinea, Va., in which 9 or more inches fell in less than 45 minutes (Mo. Weather Rev., 1906, Vol. 34, pp. 406-407, 2 figs.). W. J. Humphreys elaborated on the explanation of a cloudburst outlined by Dr. Brooks (see above). "Could hail have occurred?" asked Mr. Kadel. C. F. Talman called attention to the fact that hail is known in the tropics, especially in India, and that it had been reported in the mountains of Haiti and in Jamaica. Dr. Brooks called attention to the fact that most of the Indian hail occurred in the arid and semi-arid parts of sub-tropical northwestern India, and after raising the question as to whether the Jamaica hail was not in the mountains, stated that it appeared extremely unlikely that hail would ever fall at sea-level in Panama. Note: An article by H. G. Cornthwaite, on "Panama Thunderstorms" (Mo. Weather Rev., Oct. 1919, Vol. 47, pp. 722-724) mentions the occurrence of hail in the canal zone on 3 occasions in 12 years.] Mr. Kadel said that where such intense rainfalls occur, the gages should have greater capacity in the tipping bucket.

New aerological apparatus. S. P. Fergusson. Mo. Weather Rev., May(?), 1920. [Judging from the performance of large ballons sondes, with relatively heavy recording apparatus, new ones may easily reach 40 km. The maximum height depends on the quality of the rubber as well as on the weight of the apparatus. new ones may easily reach 40 km. Mr. Fergusson exhibited a new balloon meteorograph with a direct action 4-hr. clock weighing less than 50 grams, including cylinder, and with pens for pressure, temperature, and humidity all on same spindle. The whole instrument with protecting cover and wicker shock absorbing basket weighs but 332 grams, as compared with 4150 g. of the Bosch instrument and 2140 g. of the De Bort. The cost of this new, Fergusson meteorograph is about $100. Balloons only slightly larger than pilot balloons will be needed for carrying it, and these cost about $2.50 per balloon. Considering the fact that the percentage of recoveries of soundingballoon instruments usually exceeds 90, the cost of sounding balloon work has been reduced to a moderate figure indeed.]

[C. F. Marvin in the discussion laid great emphasis on the magnificent progress in instrumentation represented by this new instrument. It is far better than anything that has been produced of this kind before.]

Temperatures Versus Pressures as Determinants of Winds Aloft. W. R. Gregg. Abstract, Mo. Weather Rev., May (?) 1920.

[In this paper were discussed the frequent variations between observed and "gradient" winds, and between the free-air temperatures in anticyclones and cyclones as observed in Europe and in the United States. These differences were shown to be due in large part to horizontal temperature distribution and the resulting changes in free-air density and therefore free-air pressure distribution between adjacent regions. It was pointed out that greater weight should be given to temperature distribution, both horizontal and vertical, in forecasting wind direction and speed in the free air.]

Daily wind charts for stated levels. C. LeRoy Meisinger. Mo. Weather Rev., May (?), 1920.

[The method of determining mean temperatures of the air column from surface wind direction, given in a previous paper, was subjected to a statistical test to determine the pressure error at upper levels due to errors in the mean temperature. Errors as large as 5° C. will occur rarely, since a test of the homogeneity of the data show the error distribution to be fortuitous. Errors of the order of 2° C. have a chance of occurrence of about 1 in 37 times, yielding pressure errors of the 2 kilometer level of about 1.4 mb., and at the 1 kilometer level, about 0.8 mb. This small degree of error in computing pressures for stated levels indicates that the actual construction of free-air pressure charts and their use for computing wind directions and velocities awaits only further aerological observations to establish the necessary constants.]

[In discussing this paper, C. F. Marvin made it clear that for the purposes of constructing such free-air pressure maps aerological observations at more places rather than more observations at the three places for which the factors have already been computed are needed. W. R. Gregg, however, showed that this did not mean that present stations should be abandoned, for many more observations are needed to establish values for each wind direction and each season at each place.]

Cloud base altitudes as shown by disappearance of balloons and kites. O. L. Lewis. Mo. Weather Rev., May (?), 1920.

[6000 observations, made in the eastern half of the United States, when tabulated by altitudes without respect to the name given to the cloud, exhibit an uninterrupted decrease in the frequency of occurrence of cloud bases with increased altitude. The maximum frequency according to measurements with balloons and kites occurs at 400 meters above the surface. Owing to difficulty of observation, however, the actual level of maximum frequency may be lower. These results lend weight to the question raised against the usually accepted idea that there are 5 or more levels of maximum condensation in the atmospherean idea based on the fact that when cloud heights are tabulated according to the names arbitrarily assigned to them there are, of necessity, levels of maximum frequency for each form.]

B. C. Kadel announced that a new nephoscope is being prepared and that 100 will soon be delivered to selected Weather Bureau stations. At that time a new program of cloud observations is to be embarked on, especially to obtain better observations of directions and strengths of the wind at different heights. Cloud nomenclature. Charles F. Brooks. Mo. Weather Rev., May (?), 1920. [Cloud appearance is the only safe and universal basis for cloud classification. Form, apparent size, and density are the elements in appearance, and these criteria determine the differentiations between cloud forms according to the International Classification. Height is only a secondary basis, for the apparent height is derived from the apparent size and density. The present lack of detail and comparability of cloud observations shows that there is need for a uniform method both for recording cloud appearance and, if desired, for translating the recorded appearance into terms of the International Classification. To this end, it was suggested that a form with space for recording, when present, the following aspects for each cloud layer be used: Form-fibrous, sheet-like, rounded-top, downward, rounded, flocculent, ragged, and lenticuloid (more or less lentil-shaped); coarseness (i. e., angular size)-coarse, medium, fine; and density semi-transparent, medium, dense, very dense. After an observation has been recorded in such terms the use of a tabular guide, such as that presented, would give uniformity in naming the clouds at each layer in accordance with the International definitions. The unsatisfactory character of the International Classification was dwelt upon, but it was recognized that it would be inadvisable to embark on any changes in it without international action.]

[In the discussion S. P. Fergusson emphasized the need for modification of the International system and suggested that a sufficiently detailed nomencla

ture with brief symbols for records could be built upon the present International' system after some modifications. In reply to the criticism of the International Classification, W. J. Humphreys stated it was his impression that the wording of the definitions in the International Cloud Atlas did not express accurately the intention of the Commission. C. F. Brooks held, however, that we must follow the wording as printed.]

Some meteorological observations of a bombing pilot in France. Thomas R. Reed. Mo. Weather Rev., Apr., 1920.

[Smooth flying in summer is generally found above cloud summits, and bumpy air beneath them. Exceptions to this rule were so rare in the speaker's experience on the Western Front during the summer of 1918 as to be worthy of note, and one such exception was made the subject of memoranda on the occasion of its occurrence. This was on the afternoon raid of August 29, 1918, in connection with the Oise-Aisne operations, when roughness was found at all levels. The cumuli that occupied the lower levels were left far below, but roughness was still encountered, although an altitude of 5000 m. was reached before the lines were crossed. Phenomena, associated by the speaker with the disturbed conditions in the upper air, were noted later in the appearance of cirro-stratus clouds overhead, the remainder of the journey to the objective being made between two cloud-strata; cirro-stratus above, cumulus and strato-cumulus beneath.]

[Discussing this paper, C. F. Brooks suggested that, as is frequently the case in the forward part of a cyclone, the "bumpiness" above the cumulus-forming con-vection was due to convection between a colder over-running wind aloft and the relatively warm southerly wind below, and mentioned similar experiences recorded by J. C. Edgerton in the air mail service in the summer of 1918 between Washington and Philadelphia. C. F. Marvin called attention to how convection from the surface works up higher and higher as the lower air is heated; and W. J. Humphreys remarked on how cooling above will lower the underboundary of convection going on aloft. The junction of a layer of convection aloft with that near the surface will produce "bumpy" conditions to great heights.

Project for local forecast studies. R. H. Weightman. (By title.) Mo. Weather Rev., Mar., 1920, Vol. 48, pp. 154-155.

[An attempt is being made to find an adequate basis for classifying local weather observations as a basis for investigating the prognostic worth of different combinations and sequences of values of the various meteorological elements. and for applying the results.]

Climatic conditions in a greenhouse as measured by plant growth. Earl S. Johnston. Abstr. Mo. Weather. Rev., Apr., 1920. [The growth of successive plantings of buckwheat seedlings was used to obtain monthly indices on the effectiveness of greenhouse climate in different months at College Park, Md. Stem height, dry weight, leaf area, and transpiration were carefully measured. Although evaporation and temperature changed but little the rate of growth showed seasonal and irregular changes.]

Modifying factors in effective temperature. Andrew D. Hopkins. Mo. Weather Rev., Apr., 1920.

[Dr. Hopkins mentioned the effectiveness of applying the Bioclimatic Law, of altitude, latitude and longitude to determining for any locality the proper time for applying remedies to control insects, or to plant crops. This law, which postulates that periodic events progress 1 degree of latitude, 400 feet of altitude, and 5 degrees of longitude every four days, northward, upward, and eastward in spring, and southward, downward, and westward in autumn, needs considerable regional correction in parts of the United States. These corrections are retarding, in general for low-land regions and accelerating for highland regions as a general rule in spring. The degree of departure has been worked out from phenological observations.]