34 FORMATION OF SHELLS OF ANIMALS, ETC., 
displayed in the act of crystallization, rests partly upon 
experimental and partly upon rational evidence, and as 
the former will supply data for the latter, it will be neces- 
sary to describe first some facts connected with the change 
of the crystalline into the globular form, as well as some 
facts directly connected with the production of crystals— 
facts which have not been before mentioned in this paper, 
but which have formed a part of a communication con- 
tained in the ‘ Transactions of the Microscopical Society,’ 
published in the ‘Quarterly Journal of Microscopical 
Science,’ for January, 1858. It has been already ob- 
served that there is in gum arabic, besides malate of lime, 
triple or ammoniaco-magnesian phosphate. Although 
this latter compound has not to my knowledge been given 
as a constituent of this substance, yet the elementary con- 
stituents entering into its composition are given in the 
analysis of gum by several chemists. (See Turner’s 
‘Chemistry, p. 855.) When the two solutions, of the 
density and composition prescribed in the formula for 
making the artificial calculi, are brought together in the 
manner there directed, a malate of potash and carbonate 
of lime will result; and if the quantity of alkali had been 
only just sufficient to neutralize the vegetable acid in 
combination with the lime, the globular carbonate of lime 
would have been the only compound formed and depo- 
sited, while the triple phosphate would remain in solution. 
But an excess of carbonate of potash is ordered to be put 
into the denser solution, so that, after one portion of 
alkali has precipitated the carbonate, the other may set 
free the triple phosphate, and these combining form the 
largest kinds of artificial calcul. Now, as the carbonate 
is formed first, it will of necessity occupy a position in the 
mixing fluids above that occupied by the triple phosphate, 
