( 843 ) 
TABLE 4. 
also Quittner has established the great diversity of samples by 
measuring their variable densities. It is difficult to foretell the influence 
of all this on 'the magneto-optic properties. 
3. Other ferrwnagnetic compounds . The distinct analogy in the 
dispersion for substances of similar chemical structure as e.g. cupri- 
and ferroferrite in contradistinction to iron, nickel, and cobalt 
suggests whether the properties of other ferromagnetic ferrites and 
oxides are not similar. The investigation of cakiumfemte was » 
this respect of importance. This substance is very feebly magnetic 
and brittle. A small piece was surrounded by the easily fusible 
Wood alloy and then thoroughly polished. No KERB-Effect however 
was observed although the mirror was sufficiently good. The effect, 
if it exists, must be smaller than 0,35'. A similar result was 
obtained with ilmmile'). The light was reflected from the base of 
the crystal i as well as from a plane parallel f ,he P™ C,pa l “^ 
but in no case could a rotation be detected. was als0 
impossible to detect any-— , , >. 
without doubt is ferromagnetic. The following alloys were tested. 
4 Nickel-iron with 36% nickel, so called “Invar”, known to 
possess a very small coefficient of expansion, is strongly magne^ and 
distinctly shows the KRBR-Effect. The rotat.on .s exclus.veIy negatm 
in the Lion of the spectrum investigated, and there is onlyashght 
variabon wi.h wave-length. (Table 5, Fig. 6). The dispersmu-curve 
lies considerably below the zero-line; with mcreasmg wavelength 
~.) See B. Bavimk, Map. lnllueuz in KrysUllen, Gottinger Dissertation 1904. 
rotation with chromic oxide Cr t 0„ which 
