802 



SCIENCE 



[N. S. Vol. XLII. No. 1092 



thermodynamics may not be easy to retain. 

 If this is recognized, then, knowing that me- 

 teorology has never been an exact science, 

 there is no reason why this new departure 

 should not be welcomed and thoroughly tested 

 with whatever modifications and improvements 

 may be found necessary; but warning should 

 be given that the terms are now used in a new 

 sense. The present examples go a long way 

 toward supplying such tests. Bigelow's thermo- 

 dynamics supplements the equally elaborate 

 hydrodynamics of Vilhelm Bjerknes in a much- 

 needed way. As in all other departments of 

 meteorology, however, the application of the 

 principle is more or less handicapped by the 

 imperfection of the record. A single atmos- 

 pheric sounding is never completely syn- 

 chronous, and even if it were, the need remains 

 for connecting the single air column with its 

 surroundings. The importance of a study of 

 cross currents at every level, in both velocity 

 and azimuth, has been shown by Sir W. IST. 



The method introduced in the first part of 

 this work (pp. 262-292) for finding the solar 

 constant, assumes that the reduction of high- 

 sun and low-sun observations by the common, 

 but incorrect, hypothesis (that the transmissive 

 power of the atmosphere has no diurnal varia- 

 tion) will give the solar radiation at the 

 10,000-m. level, but not the radiant value out- 

 side the atmosphere; and that an approxi- 

 mately doubled interchange of thermal quan- 

 tities above the 10,000 m. level (to which the 

 author is led by his thermodynamics) is due 

 to reflection of solar rays from the atmosphere 

 at higher levels; whence the value of the solar 

 radiation outside the atmosphere must be 

 doubled. No one doubts that there is exten- 

 sive reflection of solar rays by the air. The 

 point to be carefully noted is that these rays 

 can not be both reflected and absorbed. Hence, 

 since any rays which are reflected can have no 

 efl^ect whatever on the thermal state of the sub- 

 stances in question from which the reflection 

 proceeds, this hypothesis also is erroneous. 

 The errors made in these two hypotheses have 

 opposite signs, and it is conceivable that the 

 opposite errors very nearly counterbalance each 



other, but the procedure is wholly empirical, 

 and any approximation to the truth obtained 

 in this way is accidental. 



In spite of the foregoing assumption that all 

 losses of solar radiation above the 10,000-meter 

 level are by reflection, the seventh chapter 

 introduces considerations of absorption and 

 transmission of solar rays by the various 1,000- 

 meter layers up to 90 km. Though inconsistent 

 with the preceding hypothesis, this move is in 

 the right direction. Unfortunately, however, 

 both here and in earlier parts of this work, 

 some serious errors occur. An extensive foot- 

 note (pp. 278 to 280) contains the following: 



" Transformation Factors. 



kilogr am ^ Gr. X 10' ^ Gr. 

 '(meter)2" ~ cm^ X 10* ~ cm' 



( M.K.S. \^ 10 X 60 



X lO-i. 

 Gr. cal. 



V mech. units / 41 851 000 cm.' min. ' 

 Factor = 0.000014336." 



The true derivation of the conversion factor 

 for radiant energy from M.K.S. mechanical 

 units into C. G. Min. thermal units is of 

 course : 



1 kilogr. cal. 

 4185.1 



X 



which is equivalent to 



10' gram cal. 

 418571 



60 

 10000^ 



0.0014337 (C. G. Min.). 



The same error is repeated on page 377. On 

 page 415 the author defends his mistake thus: 

 " Since the Erg and the Joule are units of 

 work they must refer to the unit volume and 

 not to the unit area. 



Hence Joule/volume ^ML^T'^-L-^ 



= ML'^T-^ = 103 X 10-2 == 10. 

 While Joule/area = ML^T-^ ■ L-^ 



= il/r-2 = 10^ = 1,000" 



It is quite true that, for example, the radiant 

 energy received from the sun in 1 second on 

 1 cm. 2 of surface represents the energy previ- 

 ously distributed through a volume of 

 3 X 10^° cm.3 ; and that, when absorbed by a 

 pyrheliometer, the temperature-effect is deter- 

 mined by the volume of water (equivalent to 

 that instrument) which is heated. But in 



