OCTOBEE 9, 1908] 



SCIENCE 



465 



and land, are matters of real solicitude. In 

 sueh circumstances there might be reason 

 for despondency if one were dependent 

 merely upon one's own ventures and the 

 results achieved thereby. But when one 

 has the advantage of the gradual develop- 

 ment of investigations of long standing, it 

 is possible to maintain a show of cheerful- 

 ness. When Shylock demands his pound 

 of flesh in the form of an annual report, it 

 is not at all uncommon to find that some 

 argosy that started on its voyage long ago 

 "hath richly come to harbor suddenly." 

 There have been quite a number of such 

 happy arrivals within the last few years. 



I will refer quite briefly to the interest- 

 ing relations between the yield of barley 

 and cool summers, or the yield of wheat and 

 dry autumns, and the antecedent yield of 

 eleven years before, which fell out of the 

 body of statistics collected in the "Weekly 

 Weather Report" since 1878. The accom- 

 plished statisticians of the Board of Agri- 

 culture have made this work the starting- 

 point for a general investigation of the re- 

 lation between the weather and the crops, 

 which can not fail to have important prac- 

 tical bearings. 



Let me take another example. For more 

 than a full generation meteorological work 

 has been hampered by the want of a definite 

 understanding as to the real meaning in 

 velocity, or force, of the various points of 

 the scale of wind estimates laid down in 

 1805 by Admiral Beaufort for use at sea, 

 and still handed on as an oral tradition. 

 The prolonged inquiry, which goes back 

 really to the report upon the Beckley ane- 

 mograph already referred to, issued quite 

 unexpectedly in the simple result that the 

 curve 



^ = .0105^" 



(where p is the force in pounds per square 

 foot, and B the arbitrary Beaufort num- 

 ber) runs practically through nine out of 

 the eleven points on a diagram represent- 



ing the empirical results of a very elaborate 

 investigation. The empirical determina- 

 tions upon which it is based are certainly 

 not of the highest order of accuracy; they 

 rely upon two separate investigations be- 

 sides the statistical comparison, viz., the 

 constant of an anemometer and the relation 

 of wind velocity to wind pressure, but no 

 subsequent adjustment of these determina- 

 tions is at all likely to be outside the limits 

 of an error of an estimate of wind force; 

 and the equation can be used, quite reason- 

 ably, as a substitute for the original speci- 

 fication of the Beaufort scale, a specifica- 

 tion that has vanished with the passing of 

 ships of the type by which it was defined. 

 This result, combined with the equation 

 p = .003V^, which has been in use in the 

 office for many years, and has recently been 

 confirmed as sufficiently accurate for all 

 practical purposes by Dr. Stanton at the 

 National Physical Laboratory and Monsieur 

 Eifiiel at the Eiffel Tower, places us upon 

 a new plane with regard to the whole sub- 

 ject of wind measurement and wind esti- 

 mation. 



Results equally remarkable appear in 

 other lines of investigation. Let me take 

 the relation of observed wind velocity to 

 barometric gradient. You may be aware 

 that in actual experience the observed 

 direction of the wind is more or less along 

 the isobars, with the low pressure on the 

 left of the moving air in the northern hemi- 

 sphere; and that crowded isobars mean 

 strong winds. Investigations upon this 

 matter go back to the earliest days of the 

 office. 



There can be no doiibt that the relation, 

 vague as it sometimes appears to be upon 

 a weather chart, is attributable to the effect 

 of the earth's rotation. In order to bring 

 the observed wind velocity into numerical 

 relation with the pressure gradient Guld- 

 berg and Mohn assumed a coefficient of sur- 

 face ' ' friction, ' ' interfering with the steady 



