ZOOLOGY AND BOTANY, MICROSCOPY, ETC. 1107 



titles, such as iodine, irou, and manganese, microscopic tests depending 

 on crystal formation will only be occasionally employed. The like 

 holds good for substances wliicli may be distinguished by spectro- 

 scopic appearances, as iridium, thallium, lithium, &c. This branch of 

 petrology, though of comparatively recent date, has lately received 

 greater attention, so that now quite a series of petrographic re- 

 searches are known, and which may compare in exactitude with the 

 most accurate analytical methods. The microscopic crystals which 

 serve for proof of the existence of certain substances are formed 

 either as precipitates after definite reactions or on evaporation of 

 solutions. In practice the former method is usually found . to be the 

 more speedy in the end, for the slower the process the more perfect 

 is the crystal. When, however, crystallization is defective from any 

 cause, the crystals become skeletal, malformed, or jumbled together in 

 masses. These aggregate malformed or skeletal forms are for many 

 substances very characteristic, and certain combinations can only be 

 obtained in such forms as, for example, copper nitrate, thorium nitrate, 

 thallium chloride, &c. The formation of normal crystals is favoured 

 by the employment of very dilute solutions. Very insoluble sub- 

 stances, such as barium sulphate, lead sulphate, silver chloride, are 

 little suitable for the microscopic test applied directly. 



The general method of examination when only small quantities 

 are available, is to place a drop of the solution to be tested on a slide 

 on the stage of the Microscope, and then add a drojD of the precipi- 

 tation-reagent. Cover-glasses are not needed unless any development 

 of gas occurs, or when observations are made on fluorine and its com- 

 pounds. In the latter case it is necessary to protect the objective by 

 fixing a cover-glass in front of the face of the anterior lens. It is 

 also necessary when hydrofluoric acid is given off during the reactions 

 to cover the slides with a thin layer of Canada balsam, and to con- 

 duct preliminary operations in platinum vessels. 



In the majority of the examinations carried out by these methods 

 perfectly trustworthy results are obtained with 1-2 mgrm. of sub- 

 stance. Thus in a drop of gypsum solution which weighs only 

 10 mgrm. is contained merely • 03 mgrm. gypsum or • 01 mgrm. calcium 

 oxide, and tliey can be recognized with certainty under the Micru- 

 scopo as sulphate or oxalate. 



For the examination of the composite silicates the methods of 

 Boricky and Behrens are reconimended. Boricky's method is 

 founded on the property of hydrofluosilicic acid to develope hydro- 

 fluoric acid on evaporation, and thereby to set free silicates even 

 without the aid of heat. A nunuto fragment of tlie size of a i)in'8 

 head is placed on a slide protected by a layer of Canada balsam and 

 a drop ui a 3-4 per cent, hydrofluosilicic acid is added. After acting 

 from two to six hours, the decomposition is so far advanced tliat tlio 

 crystallized double salts of fluorine permit the recognition of the basic 

 constituents of silicates. This method, although simple, is not free 

 from defects. 



Behrens proceeds by completely decomposing the mineral to bo 

 tested with hydrofluoric acid, and by removing any fluorobilicona by 



