CAPILLARY CONSTANTS 439 



therefore if the constant of the tip be known, the surface-tension 

 may be determined. Of late years the practical details of the 

 method — the most compact form of the apparatus, the proper 

 shape of the tip, the accurate control of temperature and of the 

 conditions governing the fall of the drop — have been most care- 

 fully studied and worked out. 1 It possesses several advantages 

 over the capillary-rise method — for, if no counterbalancing dis- 

 advantages exist, a weighing method is more sensitive than one 

 which depends on the estimation of a length, and at high 

 temperatures the percentage error is not increased, inasmuch 

 as one can allow more drops to fall, so compensating for the 

 diminished weight of each drop. 



But it is not quite certain that the method is independent 

 of the contact-angle, for, in the words of the experimenter who 

 has made a special study of the subject : " the weight of a 

 falling drop of liquid is thus found to be strictly proportional to 

 its surface-tension, determined by capillary-rise or any other accurate 

 method" 2 



If this be so, then, equally with the capillary-rise method, the 

 drop-weight method depends on contact-angles, and its value 

 as an instrument of research is correspondingly minimised ; it 

 is true that the elementary theory of the falling drop gives an 

 equation which is independent of the angle of contact, but it 

 is well known that the theory as usually given is only a rough 

 approximation to the truth, and the point, which is one of some 

 importance, would be best settled by experimental evidence. 



The elementary theory of this method is so uniformly mis- 

 stated in treatises and text-books on physical chemistry that no 

 apology is needed for discussing it here in detail. In the great 

 majority of text-books 3 it is stated that the weight {ing) of 

 the falling drop is related to the surface-tension of the liquid 

 by the equation 



mg = 277-rT ..... (iv). 



As a matter of fact any experimenter who used this equation 

 would soon be convinced of his error by rinding that his results 

 for T would have about half their true value. The error arises 



1 See various papers by Morgan and colleagues in the Journal oj the American 

 Chemical Society for 1909 and later years. 



1 Morgan, I.e. xxxii, 349, 191 1. (Italics mine.) 



8 Dolus latet in universalibus, but where so many books go wrong, it would be 

 invidious to select any particular treatise for special mention. 



