November 23, 1888.] 



SCIENCE. 



251 



— ' The Old Northwest : with a View of the Thirteen Colonies 

 as constituted by the Royal Charters,' by B. A. Hinsdale, Ph.D., 

 constitutes No. 2 of Mac Coun's Standard Historical Series. ' The 

 Old Northwest ' is a guide to the historical facts of State. Fed- 

 eral, and Inter-State legislation in connection with their formation, 



development, and admission into the Union. P. Blakiston, Son, 



& Co. have just published a second edition of ' Medical Jurispru- 

 dence and Toxicology.' a te.xt-book for medical and legal practi- 

 tioners and students, by John J. Reese, M.D. ; and ' The Physician's 

 Visiting List for 18S9,' being the thirty-eighth year of Lindsay and 

 Blakiston 's ' Physician's Visiting-List.' 



LETTERS TO THE EDITOR. 

 Anemometer Constants. 



Thk last volume of the Rcpcrtoriuiii fiir Metcorologie (Vol. XL 

 No. 7), just received, contains a paper by Dubinsky ("Vergleichende 

 Verification zweier Anemometer in Hamburg, Deutsche Seewarte, 

 und in St. Petersburg, Physical Central-Observatorium '), giving 

 the results of comparative tests of two Robinson anemometers of 

 very small dimensions, and using for this purpose the two whirling- 

 machines respectively at Hamburg and St. Petersburg. 



These experiments are of special interest to the writer, who was 

 himself engaged during the past summer upon similar work for the 

 Signal Service, and used, with the larger anemometers of the ser- 

 vice, a very small one for studying certain parts of the problem. 

 In this work the whirling-machine was very large, having an arm 

 twenty-eight feet long, which in later experiments was increased to 

 thirty-five feet. It is not intended at this time to speak further of 

 this work, but to notice in a few remarks the method (pp. 1 1 et 

 seg.) used by Mr. Dubinsky to ascertain a very important correction, 

 and to compare his results with those obtained by Dohrandt (Rep. 

 fur Met., Vol. IV.-Vol. VI.), who had already used the St. Peter- 

 burg machine in making a large number of experiments upon 

 anemometers of the ordinary sizes. 



The two whirling-machines, which are permanently set up in 

 ■closed rooms, are nearly the same in size ; that at St. Petersburg 

 being much like a letter T in form, and adapted to be revolved about 

 the central stem as an axis, carrying the anemometer to be tested 

 on the outer end of one or the other of its horizontal arms, which 

 are about eleven feet long. In the Hamburg machine one arm is 

 quite short, and carries a counterpoise ; the other is between twelve 

 and thirteen feet long. 



In using such whirlers, there is a tendency of the arm and other 

 moving parts to set up a slow rotation in the air, as a whole, through 

 which they revolve. This movement of the air with the arm is 

 called by the Germans, and aptly so, the Mitwind. The determi- 

 nation of its amount is one of the most serious obstacles to over- 

 come in experiments of this kind. 



Results seem to indicate a pretty close proportionality of this 

 Mitwind to the velocity of the arm : and Dohrandt concluded 

 from his studies that in value it was about 5 per cent of the latter. 

 Dubinsky, working with relatively very much smaller anemometers, 

 though using the same whirling-machine, adopts 7.3 per cent as 

 the correction for the Mitwind. The discrepancy in these results is 

 really larger than it appears, when it is considered how much less 

 the small anemometers would tend to generate Mitwind, as com- 

 pared with those used by Dohrandt. 



A brief description of the method of measuring the Mitwind 

 will aid in understanding the question. For this purpose both ex- 

 perimenters placed close to the path of the whirled anemometer a 

 delicate air-meter, with its axis tangent to the orbit. Its indications 

 during the progress of an experiment give a measure of \.\\t Mit- 

 ■wind, however, being strongly acted upon by the violent disturb- 

 ance of the air which immediately attends and follows just after the 

 passage of the whirled anemometer, and which cannot be consid- 

 ered as a true Mitwind. The velocity given by the air-meter is 

 no doubt, as Dohrandt points out, much greater than that of the 

 true Mitwind. 



The treatment by Dubinsky, of this observed velocity, to reduce 

 it to the .M'it-.uind velocity, is practically the same, at least in in- 

 tent, as the expedient resorted to by Dohrandt (Rep. fiir Met., Vol. 



IV. No. 5. p. 39), who placed on the end of the unoccupied arm 

 of the whirler a small air-meter, which was thus carried in the path 

 of, but diametrically opposite, the whirled anemometer. The 

 whirling-machine is revolved, first with both anemometer and 

 air-meter in position, and then with the air-meter alone. Ow- 

 ing to a decrease in the Mitwind attending the removal of the 

 anemoineter, the whirled air-meter registers a larger number of 

 units in the second case than in the first ; and the difference, 

 in terms of velocity, is considered by Dohrandt as the trzie value oi 

 the difference between the Mitwinds in the two cases. Not ques- 

 tioning the correctness of this assumption, a comparison of the dif- 

 ference thus obtained with that derived from the indications of the 

 stationary air-meter shows the latter to be from two to three times 

 the former or presumed true difference. Finally, it is further as- 

 sumed that the whole observed Mitwind and the true are in the 

 same proportion. Or, if v^ and v^ are the velocities indicated by 

 the air-meter when whirled with and without an anemometer, and 

 Xi and x^ the corresponding velocities of observed Mitwind, we 

 have, the velocity of the arm being the same in both cases, 



and the true Mitwind \% a y. observed Mitwind. 



In applying this method, Dubinsky whirled both of the small 

 anemometers, one on each end of the arm, and then one alone, 

 using the stationary air-meter for observing the Mitwind in each 

 case. This substitution of the small anemometer — an instrument 

 equally influenced by equal winds in a horizontal plane, whatever 

 their direction — for an air-meter not thus influenced, is an impor- 

 tant modification of Dohrandt's method, and may serve to account 

 for a part, at least, of the difference found in their results. Dubin- 

 sky has, apparently without being aware of its peculiar merits, hit 

 upon what is believed to be a more proper method of investigating 

 Mitwind than any heretofore used : that is to say, the Mitwind 

 anemometer must be of the same form as the anemometer being 

 tested, as it is evident the instrument used for measuring the Mit- 

 wind must be influenced thereby in the same manner, and to the 

 same extent, as the instrument whose constants are being deter- 

 mined. 



Throughout the tests upon the small anemometers the Mitwind 

 was carefully observed by means of a stationary air-meter, and 7.3 

 per cent of the arm-velocity was adopted as its value at St. Peters- 

 burg, 7.6 per cent being the value found at Hamburg. A single 

 experiment only is cited, by which the value of the factor a was 

 determined, and is as follows : — 



The paper further states that in the second case, had the velocity 

 of the arm been 62.24 instead of 62.56, the recorded contacts of 

 No. 74 would have been 60.27. Hence we have 



63.34 



60.04 

 60.27 



4.68 



Differcn 



0.33 



1.33 



The author, apparently too hastily, jumps at the conclusion, and 

 places 



0.23 



= I .00 



0.23 

 as the value of the factor a. and in consequence applies all of the 

 7.3 per cent observed Mitwind as the correction for that disturb- 

 ance. It is to be observed that the first 0.23 in the line of differences 

 is in terms of contacts per hour, and is not a velocity. Further 



