18 Messrs. J. Tyndall and H. Knoblauch on the 



A number of thin plates, each about half an inch square, 

 were cut from almond kernels, with an ivory blade, parallel to 

 the cleft which divides the kernel into two lobes. These were 

 aid one upon the other, with an interval of strong gum be- 

 tween, until a cube was obtained. A few minutes in the sun- 

 shine sufficed to render the cube dry enough for experiment. 

 Hung between the poles, with the line perpendicular to the 

 layers horizontal, on exciting the magnet this line turned and 

 set itself parallel to the magnetic resultant passing through 

 the mass. The action here was a maximum. Turning the 

 cube round 90° in the axial plane, there was scarcely any di- 

 rective action. If the word ; crystal - be substituted for ' cube ' 

 in the description of this deportment, every syllable of it is 

 applicable to the case of bismuth ; and if the deportment of 

 the crystal warrant the conclusion that the force is axial, the 

 deportment of the cube warrants the same conclusion. Is the 

 force axial in the case of the cube? Is the position of the 

 line perpendicular to its layers due to the " tendency " of 

 that line to set itself parallel to the magnetic resultant? The 

 kernel is strongly diamagnetic, and the position of the per- 

 pendicular is evidently a secondary result, brought about by 

 the repulsion of the layers. Is it not then possible, that the ap- 

 proach of the magne cry st alii c axis, in bismuth, to the magnetic 

 resultant, is really due to the repulsion of the planes of clea- 

 vage ? 



But here the experiment with the silken axis meets us; 

 which showed that so far from attraction being the cause of 

 action in a magnetic crystal, there was actual recession ; and 

 so far from repulsion being the cause in a diamagnetic crystal, 

 there was actual approach. This objection it is our duty to 

 answer. 



A model was constructed of powdered carbonate of iron, 

 about 0*3 of an inch long and 01 in thickness, and, by atten- 

 tion to compression, it was arranged that the line of elective 

 polarity through the model was perpendicular to its length. 

 Suspending a thread of cocoon silk with a weight at one end 

 a vertical axis was obtained ; a bit of card was then slit and 

 fitted on to the axis, so that when the model was laid on one 

 side, the card stood like a little horizontal table in the middle 

 of the magnetic field. The length of the model extended from 

 the central axis to the edge of the card, so that when the 

 mass swung round, its line of elective polarity was tangent to 

 the circle described. 



When the model was made to stand between the flat-faced 

 poles obliquely, the moment the magnet was excited it moved, 

 tending to set its length equatorial and its line of elective 



