1865.] Sorby on Spectrum-Analysis applied to the Microscope. 203 



inches or more, closed at the ends with plate glass. These answer 

 very well, and give a much longer spectrum than might be expected, 

 on account of the light being reflected down the sides of the tube. 



On placing a crystal on the stage of the microscope as at g in the 

 figure, and the tube containing the solution in front of the slit as at a, 

 it is easy so to arrange the focus of the condenser and the position of 

 the tube and of the crystal, that the spectra of the salt and of the 

 solution may be seen side by side, or with a narrow band of natural 

 spectrum between them. 



Since the width of the spectrum is modified by the thickness and 

 refractive power of the substance placed between the condenser and 

 object-glass, it is desirable to mount the crystal in Canada balsam, or 

 else to fix a piece of glass of the same thickness alongside of it, other- 

 wise the spectra seen through the crystal and outside it are not at the 

 same focus or of the same width. It is also best to use a condenser 

 and object-glass of as long focal lengths as circumstances will permit, 

 raising, if it be necessary, the glass on which the crystal is mounted 

 by means of a small stand fitting on the stage of the microscope. 



Proceeding in this manner, I find that, as a general rule, crystal- 

 line salts have the same action on the spectrum as an equal amount 

 dissolved in water ; but there are many interesting exceptions due to 

 special optical properties of the crystals, to the dissolved salt being 

 probably combined with a different amount of water, or, in some cases, 

 to isomeric changes. At all events, the facts throw much light on the 

 state in which substances exist when dissolved. 



4. Determination of the relative amount of Per- and Prot- oxide of Iron 



in Minerals. 



When I first thought of applying a prism to the microscope, one 

 of my chief objects was to employ it in studying the minerals seen in 

 thin sections of rocks. So far, however, I have only been able to 

 apply it to determine whether the iron exists as protoxide or per- 

 oxide, and, employing polarized light for the spectrum, to learn 

 certain facts connected with the double refractive power of the 

 minerals. 



The action of persalts of iron depends partly on the amount of 

 combined water. When anhydrous, the point of maximum trans- 

 parency is situated about halfway between D and the extreme red, so 

 that when the blue end of the spectrum is cut off to D, all the red is 

 still transmitted. When hydrated, this maximum point is farther and 

 farther from the red end, according to the amount of combined water. 

 For example, in pentahydrated perchloride of iron, the maximum 

 point is f the distance from it towards D, whilst in the dodecahydrated 

 it is situated near D, and the crystals are therefore yellow, but melt 

 when heated into a solution of the orange-red pentahydrated. Very 

 dilute solutions of the persalts absorb the blue end of the spectrum, 

 and, on increasing the strength, they absorb more and more of the 

 green. On the contrary, a strong solution of protochloride a foot 



VOL. II. Q 



