50 PROFESSOR GREGORY ON AN ANOMALOUS 



36.67 per cent, of magnesia. In three analyses, the per-centage of magnesia in 

 the new salt was found to be 10.78, 16.92, and 15.94. In the last of these I 

 pushed the washing further than in the first two, and I did the same in three 

 more analyses, which yielded 15.12, 15.34, and 15.12, per cent, of magnesia. The 

 mean of the six analyses, all made with portions of salt prepared at different 

 times, is 15.87 per cent, of magnesia. As the salt lost no weight by ignition, it 

 contained no water, and therefore Avas composed of magnesia and phosphoric 

 acid alone : the absence of lime being previously ascertained. Its composition is 

 therefore. 



Magnesia, 15.87 



Phosphoric acid, 84.13 



The only formula which at all approaches to this composition is 3 P, Og + 2 M^ 0, 

 according to Avhich, it is an acid sesquiphosphate of magnesia. The composition, 

 calculated according to this formula, is, 



Magnesia, 16.18 



Phosphoric acid, 83.82 



Considering the imperfection of the means for determining the amount of 

 magnesia in analyses witli jirecision, there can, I think, be no doubt, that the 

 above formula expresses, empirically, the composition of the salt in question. 



But what is the rational formula of this salt ? As far as I know, there are 

 no known sesquiphosphates of protoxides. Indeed, the very characteristic of 

 the three known classes of phosphates, is their tendency to form salts Avith 1, 2, 

 and 3 equivalents of base for 1 of phosphoric acid. 



Metaphosphoric, or monobasic phosphoric acid, forms salts of the general for- 

 mula. P., (),, M ; and if our salt is to be classed as a metaphosphate, it must be 

 one with the very singular formula 2 (P., 0., M^ 0") + P 0^; in which 2 eq. of 

 metaphosphate of magnesia are combined with 1 eq. of anhydrous phosphoric acid, 

 which has either entered the radical of the acid, without increasing its neutralising 

 or assumed a neutral character. 



Pyrophosphoric acid, or bibasic phosphoric acid, forms salts of the general 



formula, P^ 0.,2 MO, or P^ ^5 j fr o" ^^ °^^^ ^^^* ^^ ^ pyrophosphate, and the tem- 

 perature appears favourable to its being so, its fonnula must be (P„ Oj., 2 M_$r 0) + 

 2 P., 0., in which 1 eq. pyrophosphate of magnesia is combined with 2 eq. of 

 anhydrous phosphoric acid, in one of the characters alluded to above. 



Lastly, common or tribasic phosphoric acid forms salts of the general for- 



mula, P^ Og, 3 M ; P^ O5 p ^ J, or P^ ^5l2^I^o- If our salt belong to this class, 



its fonnula will be P., 0, <]''p ^. + Pg 0^. Here, 1 eq. of phosphoric acid, act- 



