APPENDIX. 127 



very small qunnfifj' compnrcd with tlio entire iiiciss of the coral. Some cosy means of 

 completely separating them all, is an ind'spensahle preliminary step in their examination 

 and estimation. 



"I am indebted to n^v friend Dr. J. L. Smith, of Charleston, South Carolina, for sug- 

 gesting to me the use of this test in the analysis of the corals. 



"As the several elements whose presence our researches have determined in corals, have 

 been enumerated in the body of the work (p. 57), it is not necessary to repeat them here; 

 but we may state, in a summary manner, an outline of the general course of analysis pur- 

 sued in determining the constitution of the lime-wnter prccipitalc, which, it will be allowed, 

 contains several elements, whose association has always been considered as offering some 

 of the most difficult problems in the whole range of inorganic analysis. The following 

 plan of analysish as been contrived in part, from the late researches of Von Rammelsbcrg, 

 on the estimation of phosphoric acid, and partly from the labours of Rose and Berzelius, 

 adapting the method to the requirements of the particular problem before us. 



" A. The lime-water precipitate, after ignition, is weighed and then digested in f ne 

 powder in cold chlorohydric acid; it slowly dissolves, leaving a white flocculcnt powder. 

 This collected and washed, will be found to be silica. It is harsh and gritty between the 

 teeth, is not taken up by long digestion in strong acids, dissolves in a solution of caustic 

 potassa, and before the blow-pipe forms a hard colourless glass with carbonate of soda, 

 dissolving in this reagent with eflervescence. 



" B. The solution in chlorohydric acid is supersaturated with caustic ammonia, and boiled; 

 a gelatinous precipitate separates, which is usually coloured by iron, and by its characters 

 indicates the pi-esence of alumina. This precipitate contains the phosphoric and other 

 acids and the bases therewith combined. It is collected and the filtrate therefrom (C) is 

 examined for lime arid magnesia, both of which are usually present. 



" D. The precipitate by ammonia (B) is ne.xt made into a thick paste with strong sul- 

 phuric acid, in a small vessel of piatina. A plate of glass, coated with wax and written on, 

 is placed over the crucible ; and heat being applied, hydrofluoric acid escapes, and attacking 

 the glass, leaves a permanent record of its presence. I have never failed to obtain evidence 

 of the presence of fluorine in any coral which has been subjected to the test. Generally, 

 exposure for one minute will etch the glass most decidedly; and one experiment will 

 suffice to mark distinctly several pieces of glass. Bv this plan of analysis the quantity of 

 fluorine cannot be estimated, and it must be judged of either by the loss or by the defi- 

 ciency of acids to satisfy all the bases formed. The constant association of phosphoric 

 acid and fluorine, renders it advisable, in compounds in nature, where one of these ele- 

 ments is found, to search for the other. 



" E. After the sulphuric acid has been digested on (D), long enough to convert all the 

 bases present into sulphates, a portion of bisulphate of potash or caustic potash is added, 

 and a little water, to dissolve it; to this, a ver}^ large quantity of alcohol of a specific 

 gravity about -SGO, is added, and the whole is allowed to stand for some hours ; during 

 which the double sulphates of potassa, alumina, and iron, crystallize out, while any lime 

 previously combined is separated as sulphate, and in the solution we must look for the 

 phosphoric acid and magnesia, together with a little persalt of iron, held up by the alcohol. 



" F. The mixture (E) being filtered and the precipitate washed quite clean with alcohol, 

 the filtrate is evaporated until all the alcohol is expelled, and then supersaturated with 

 ammonia ; a little trace of alumina and iron separates, which may be added to that to be 

 obtained from the other portion (H). We may now either add an excess of pure chloride 



