ii8 



SCIENCE. 



[Vol. X. No. 239 



LETTERS TO THE EDITOR. 



*♦* The attention of scientific men is catted to the advantages of the correspondence 

 ^otumns of ?'C\-e.ftz^/or placing promptly on record brief preliminary notices of 

 their investigations. Twenty copies of the nujnber containing his communication 

 'will be furnished free to any correspondent on request. 



The editor will be glad to publish any queries consonant with the character of 

 the journal. 



Correspondents are requested to be as brief as possible. The writer's name is 

 in all cases required as proof of good faith. 



Wind Pressure and Velocity. 



The importance of an accurate determination of the relation be- 

 t\vee» the pressure and velocity of the wind will be readily recog- 

 nized. This relation is especially needed by the architect and 

 bridge-builder, since most instrumental determinations are of the 

 wind's velocity. The problem is much more intricate than is ordi- 

 narily supposed, and the diverse results obtained by experimenters 

 of great ability show how the determination of the movements and 

 behavior of gaseous media are hedged about with difficulty, and, as 

 already pointed out in Science for July 8, the absolute necessity of 

 building up the science of meteorology on a firm foundation of fact 

 rather than theory. 



There have been two methods of experimentation : the earliest, 

 with plates rotated upon an arm seldom exceeding lo feet in length, 

 and, later, by the exposure of plates to direct air-motion. Borda, 

 in 1763, with plates ranging from 16 to 85 square inches area, ob- 

 tained the following relation, 



p = (.0031 + .00035 1^) Sv", 

 in which p = pressure in pounds on the plate, c = contour of plate 

 in feet, 5 = surface in square feet, and v = velocity in miles per 

 hour (this notation will be maintained throughout). 



In 1874, Hagen tried most careful experiments with an arm of 

 8 feet. The velocity ranged from i to 3 miles per hour, but the 

 room was so small that at the latter velocity the air was set in feeble 

 rotation. The plates ranged in size from 4 to 40 square inches. He 

 found, as did Borda, that the pressure per square foot increased with 

 the size of the plate. The following is the relation established by 

 him, 



p — (.002894 -1- .0001403c) S V-. 

 This forrtlula for this relation has repeatedly appeared in print, and 

 each time it has been changed. This is believed to be correct. 



Singular as it may seem, these experiments have been almost the 

 only ones quoted in discussions of this question, and yet it is easy to 

 see that they are utterly useless for determining the pressure of a 50- 

 mile wind on the side of a building. 



In November, 1886, a few experiments in Washington with an 

 arm of 4 feet, and plates from 16 to 576 square inches area, gave 

 the relation, 



p = (.0032 + .00034 c) Sv". 



The agreement with Borda's results is very interesting. 



Afterward, with the same style of apparatus and an arm of 16 

 feet, the relation found was 



/ = .0034 S v''. 

 The velocity of the larger plates was only 4 miles per hour, so that 

 this formula does not help us for greater velocities. It was certain- 

 ly established that there was no difference in pressure per square 

 foot depending on the size of the plate. Turning to experiments 

 of the second class, we find that Thibault obtains, with plates from 

 I to 1.5 square feet area exposed to the wind, the relation, 



p = .00475 S V-. 

 In France, with plates exposed on a locomotive running 44 miles 

 per hour, the relation established was 



/ = . 00535 S-d^. 

 In this case, probably, a slight allowance must be made for the 

 wind with the train. 



In the ' Encyclopcedia Britannica,' article ' Hydromechanics,' the 

 mean of all the better determinations is 



p = .00496 5 V ^. 



We may conclude, ist, that experiments with whirling arms of 

 less than 16 feet are very untrustworthy ; 2d, that we need determi- 

 nations with rapid, straight-line motion, best obtained, perhaps, by 

 pushing two or three platform-cars loaded with iron in front of a 

 locomotive, exposing the plates on the front car ; 3d, the relation 



p = .005 S v'' is the most satisfactory yet determined, and does 

 not differ by more than four or five per cent from the truth. 



While there has been this great difficulty in determining the 

 above relation, there has been just as much, if not more, in connec- 

 tion with the relation between the velocity of the wind and that of 

 the cups of Robinson's anemometer. Some confusion has arisen from 

 the fact that the standard anemometer in England has 9-inch cups 

 and 24-inch arms, while in our country we have 4-inch cups and 

 7-inch arms. 



It has been determined, by careful experiment in England, that, 

 if the large type of anemometer has a factor of 2.5, then the smaller 

 should certainly have 3.00. Dr. Robinson, after a long research 

 with a whirling machine, decided that the factor (of the smaller 

 instrument probabl)') should be about 2.5. After tiying a few ex- 

 periments in the open air, however, he changed his view, and de- 

 cided that the factor should be 3.00. In Washington, with an arm 

 of 16 feet, and a velocity of 12 miles per hour, the factor was found 

 to be 3.00. 



Quite recently the Chief Signal Officer, through the kindness of 

 the officials, as a preliminary to carrying on experiments on plat- 

 form-cars, as suggested above, has had an anemometer placed 

 upon a locomotive of the Baltimore & Ohio Railroad running 

 from this city to Baltimore. Only one round trip has been tried 

 thus far : in the outward trip the velocity of the train was about 20, 

 and returning it was about 46 miles per hour. Allowing for the 

 actual wind, we find the anemometer indication 46 miles going, and 

 47 returning. The distance was 40 miles, and we may consider 

 that the excess of about 6 miles was due to the heaping up and 

 flowing over of the air in front of the locomotive. All things con- 

 sidered, it seems probable that the factor 3.00 now used in our 

 anemometers of 4-inch cups and 7-inch arms is entirely correct : 

 certainly no change in the present factor can be thought of for an 

 instant. A complete discussion of this question has already been 

 prepared by me, and will appear in October. The other side of this 

 question has been recently presented by Professor Ferrel in the 

 August American Meteorological Journal. H. ALLEN Hazen. 

 Washington, Aug. 22. 



The Formation and Dissipation of Sea-Water Ice. 



Mr. W. a. Ashe's opinion on the freezing-point of sea-water, 

 and the conclusions he draws from his experiments, cannot be ac- 

 cepted. The arrangement of the experiment described in No. 228 

 of Science seems to be insufficient. A hole was cut through ice 87 

 centimetres (2.85 feet) thick. The water within was thoroughly 

 agitated by stirring from below, and during the actual observation 

 slightly agitated. The thermometer was held nearly horizontally, 

 the bulb slighty lower than the rest of the instrument, just below 

 the surface of the water. When the ice-film began to form, the 

 reading of the thermometer was — 2°.9 C. (26°. 7 F.), the tempera- 

 ture of the air being — 24°.8 C. ( — I2".6 F.). The greatness of the 

 difference between the freezing-point of the sea-water and the 

 teinperature of the air detracts from the value of these observations. 

 The ice is forming so rapidly that brine is included among the crys- 

 tals : it is even probable that cryohydrates are formed at the sur- 

 face. On the other hand, the freezing-point of sea-water was not 

 only found by melting sea-water ice, as Ashe assumes, but also by 

 freezing sea-water, and was always found to be between — i°.6 C. 

 and — i°.8 C. (29°.: and 28^.8 F.), according to the concentration 

 of the solution. Mr. Ashe's second remark on this subject in No. 

 232 of Science does not agree with Buchanan's interesting researches 

 on the melting of fresh-water ice in solutions of salts. He has 

 shown by an excellent series of experiments {Nature, April 28 and 

 May 5, 1887), that, when sea-water is frozen to the extent of fifteen 

 per cent of its mass, and the crystals so formed are allowed to melt 

 in the liquid in which they have been produced, they melt exactly 

 as they have been formed. If snow or pure ice be immersed in the 

 brine formed by partially freezing sea-water, it melts at the same 

 temperature as the ice which had been formed by freezing the sea- 

 water, so long as the chemical composition is the same in each case. 



In a third letter to Science (No. 237), Mr. Ashe makes some re- 

 marks on the formation and character of Arctic ice. He says, that, 

 as the density of sea-water increases till the freezing-point is reached, 

 ice is not formed at the surface, but at a certain depth. In fact, the 



