230 DR. ORD, 



only crystals left were three-sided prisms with shallow pyra- 

 miciic ends (PI. XVI, fig. 5). 



The series constructive of the large octohedra requires 

 some special description. The little spicules before men- 

 tioned as having no very definite shape were first gathered 

 into radiating tufts ; the tufts became cruciform, with very 

 irregularly outlined arms ; then six-rayed, with arrangement 

 of rays corresponding to the angle-joining lines of an octo- 

 hedron; by the filling up of the inter-radial spaces, and 

 simultaneous smoothing of surfaces and pointing of angles, 

 the perfect crystal Avas obtained. 



It was constantly to be observed that the cruciform bodies 

 were joined in j^airs face to face by a bar, so as to resemble 

 the amphidisci of Spongilla ; the two halves then correspond 

 to the two pyramids into which an octohedron can be divided, 

 and the bar joining them to the line joining the apices of the 

 pyramids, or third axis of the octohedron. 



{b) Chloride of calcium was used in large excess of the 

 oxalic acid. 



The line of greatest deposit was here removed to the upper 

 end of the plug. The constituents of the deposit were the 

 same as in the preceding experiments, but above it very few 

 of the forms seen in the corresponding part of the plug in 

 tiie preceding experiment had been deposited ; there were no 

 octohedra, their place being taken by little rhombs. It was 

 found that all the oxalic acid had been withdraAvn from the 

 solution, and the predominance of the calcic solution seemed 

 to have driven away all the forms characteristic of the oxalic 

 side. 



3. Solutions of ammonium oxalate carefully purified by 

 repeated crystallisations, and of chloride of calcium of known 

 strength, were prepared. 



The formula of ammonium oxalate is given in Miller's 

 ' Elements of Chemistry ' as follows : — 



(H4N)oC20„H,,0 = 142; 

 of calcic chloride in the fusible form as follows : — 

 CaCl2,6H20 = 219. 



from which it can be calculated that 100 parts of calcic 

 chloride Avill be decomposed by 65 parts of ammonium oxa- 

 late. The solutions were therefore made to contain respec- 

 tively 100 grains of calcic chloride and 65 grains of ammonium 

 oxalate in four ounces of distilled Avater. 



Experiments were now carried on with more exactness : 



1st, with equivalents of the two salts ; 



2nd, with 4 equivalents of oxalate to 1 of chloride ; 



