•388 Scientific Intelligence. 



freezing point — was verified for moderately dilute solutions by 

 Van't Hoff himself by means of the equation 



t=0'02T 2 /w 

 in which T is the absolute temperature, and w the latent heat of 

 fusion of the solvent. If Bunsen's value for w, 80 calories, be 

 taken, the value of the constant v is 1*87, somewhat smaller than 

 1*878 obtained if 79*6 be taken. Preliminary experiments in 

 1894-5 gave for alcohol and cane sugar the value 1*84 when ob- 

 tained with the 0-001° thermometer. Subsequently 1*89 was ob- 

 tained for cane sugar and resorcinol, 1*87 for urea, 1*86 for milk 

 sugar, 1*85 to 1*86 for dextrose and 1'84 to 1*87 for maltose. In 

 the present investigation, cane-sugar, alcohol, urea, acetone, ani- 

 line, phenol, dextrose, resorcinol, maltose and milk sugar, were 

 the substances used, the solutions were very dilute and the obser- 

 vations for temperature were made with 0*001° and 0*01° ther- 

 mometers, the convergence temperature being both above and 

 below the freezing temperature. The results, which are given in 

 tabular form, show only small deviations from 1*87, the theoretical 

 value. Hence the author concludes that Van't Hoff's thermody- 

 namic equation receives abundant confirmation from the results 

 obtained with dilute solutions, the evidence being even more sat- 

 isfactory than is the case with most of the generalizations estab- 

 lished on the thermodynamic basis. — J. Chem. /Soc, lxxi, pp. 743, 

 796, July, 1897. G. F. b. 



2. On the Spectrum of Silicon. — If a silicate be fused on a 

 platinum plate and a highly condensed spark be allowed to im- 

 pinge upon it,* Arnaud de Gramont has shown that the spec- 

 trum of the spark shows the following lines of silicon : 6969*7 

 strong, 6342*2 very strong, 5978*9 somewhat strong, 5960*3 dis- 

 tinct, 5948*0 doubtful, 5060*0 and 5045*5 very strong, 4575*7 very 

 feeble, 4568*9 somewhat distinct, 4553*7 distinct, 4131*3 and 

 4129*2 somewhat strong but diffuse. These wave lengths are the 

 means of determinations not only with the spark and fused salts 

 but also with the spark and silicon electrodes in very pure hydro- 

 gen. The most characteristic lines are 6969 7 and 6342*2 in the 

 red and 5060*0 and 5045*5 in the green. The silicon lines last 

 mentioned are much more intense than the adjacent lines of plati- 

 num and air. The spectrum obtained by this method is well 

 shown with sodium silicate, not as well by the potassium salt. 

 Both potassium and sodium silico-fluoride show it particularly 

 well, but zinc silicate gives it very imperfectly. Natural silicates 

 if pulverized and fused with sodium carbonate, soon show the 

 pairs of lines in the red and in the green. — C. H., cxxiv, 192-194, 

 January, 1897. G. F. b. 



3. On the Spark Spectrum of Cyanogen. — Several reasons 

 have been offered for supposing cyanogen to have a real spectrum. 

 (1) When nitrogen is present, this substance is actually formed 

 in the arc; (2) without nitrogen, carbon itself does not give the 



* This Journal, IV, iii, 150, February, 1897. 



