178 A. J. Ewavt: 



of magnesium hydrate, remaining after all the fonialdehyde had 

 been polymerised, the appro/ximate amount required was esti- 

 mated'. 



Using these proportions more exact estimations were made. The 

 condensing flask was provided with a condensing tube 4 ft. long, 

 which was sealed at the upper end as soon as the liquid had been 

 brought to boiling point, and allowed to blow out to a thin safety 

 bulb. The liquid was then kept just at the boiling point for several 

 days. The liquid l>ecame brown before the end of each experi- 

 ment. 



In the presence of magnesium formate, 500 c.c. of water and 

 10 c.c. of 31.5% formaldehyde, after the addition of 23 c.c. of 

 3.5% sodium hydrate, a small ansount of a white insoluble solid still 

 remained after several days. This was not Mg HjO,, and was 

 insoluble in dilute acid, the liquid was faintly acid, and still con- 

 tained a small amount of CHjO. 



Using 750 c.c. of water, 15 c.c. of 31.5% CK^O, a slight excess 

 of magnesium formate and 40 c.c. of 3.5% NaHO, all the formal- 

 dehyde wa.s removed, after boiling for 22 hours, the liquid ^^as 

 faintly acid, and contained a small amount of white solid, incon- 

 spicuous when suspended, but not consisting of Mg HjO,. This 

 represents a percentage polymerisation of 58. A further test gave 

 a polymerisation value of 57. As a small amount of the magne- 

 sium hydrate appears to form an insoluble compound, and as 

 during prolonged boiling a trace of the formaldehyde is oxidised 

 directly to formic acid, which represents a further direct removal 

 of magnesium hydrate without producing any polymerisation, the 

 polymerisation value of 57 to 58 for the divalent Mg corresponds 

 fairly well with that of 65-68 for the divalent Ca. Ba. Sr, and con- 

 trasts sharply with the values for the monovalent K and Na of 8 to 

 14%. 



Potassium. — On the basis of the conclusions given above, that 

 the relative efficiency of calcium and sodium as polymerising 

 agents depends upon the former being divalent, and the later 

 monovalent, we should expect to find equivalent solutions of the 

 monovalent metals, sodium and potassium, exercising a very 

 much inferior polymerising action, and. that in the presence of 

 calcium formate it should be a matter of indifference whether 

 sodium hydrate or an equi-molecular solution of potassium hydrate 

 was used to bring about polymerisation. 



As a matter of fact the correspondence is even more exact than 



