In order to produce a current in the direction of the arrows shown in Fig. 4, the spiral and axle must revolve from right to left. In this case the rubber, M, occupies the position shown in the same figure, the brushes embracing an arc of a little less than 180 deg.. As soon as the lower touch comes in contact with the brush, R, when the revolution is being effected from left to right, the rubber, M, establishes a communication between the two halves that have until now been isolated, and the current is no longer interrupted. The second touch during this time is at any point whatever of the arc, W N _o_, and the spirals corresponding to the latter arc outside of the circuit. In short, thanks to the rubber, M, we have an ordinary Gramme collector in that portion of the circuit comprised between the brushes, and a collector with a breakage of the circuit in the portion to the right. [Illustration: FIG. 5.] This type of machine is entirely theoretical. In the apparatus used for Prof. Pfaundler's experiments in 1870, the armature revolved with the solenoid. The core and armature were of soft iron, and the core was arranged in a manner analogous to the preceding, and remained in place under the action of its weight, and the shell, forming a complete circle, revolved with poles fixed in space. Practically, the machine that we have just described would prove inconvenient to realize, and would present serious inconveniences. In the first place, it seems to us quite difficult to transmit the motion of the solenoid to the axle, supposing the former to revolve within the armature. In the second place, considerable friction would surely occur between the spirals and core, and the axle, being submitted to a lateral stress, would be placed in a poor condition for work. It is even allowable to doubt whether such a type could be practically got up. At all events, no trial has as yet been made of it. Compared with the Gramme machine, from an absolutely theoretical point of view, the Pfaundler apparatus presents undoubted advantages. A theoretically perfect dynamo electric machine would be one in which there was a complete reciprocity between the magnetizing action of the current and the inductive action of the magnetic field. Now, such is not the case in the Gramme machine. In this apparatus the soft iron core is at the same time a magnet through favorable induction and a disadvantageous electro-magnet. This double polarization is only remedied to a certain extent by the adjustment of the brushes. In the Pfaundler machine, on the contrary, the electro-magnetism and magnetism through induction act in the same direction, and concur in effecting a polarization that favors the production of the current. Looked at it in this light, the latter machine more nearly approaches the type of perfection than does that of Gramme.