The ninth series of electrical researches occupied the autumn of 1834. In it he returns to the study of the magnetic and inductive actions of the current, investigating the self-induced spark at the break of the circuit, to which his attention had been directed by Mr. W. Jenkin. Several points in this research are little known even now to electricians, the laboratory notes being much more detailed than the published paper. He describes an exceedingly neat high-speed break for producing rapid interruptions, using for that purpose stationary ripples on the surface of a pool of mercury. In a wonderful day’s work on 13th November, filling thirty-four pages of the laboratory book, illustrated with numerous unpublished sketches, he tracks out the properties of self-induction. He proves that the spark (on breaking circuit) from a wire coiled up in a helix is far brighter than that from an identical wire laid out straight. He finds that a non-inductive and, therefore, sparkless coil can be made by winding the wire in two opposite helices. “Thus the whole [inductive] effect of the length of wire was neutralised by the reciprocal and contrary action of the two halves which constituted the helices in contrary directions.” The next day he writes: “These effects show that every part of an electric circuit is acting by induction on the neighbouring parts of the same current, even in the _same wire_ and the _same part_ of the wire.” [Sidenote: EFFECTS OF SELF-INDUCTION.] On 22nd November he is trying another set of experiments, also never fully published. They relate to the diminution of self-induction of a straight conductor by dividing it into several parallel strands at a small distance apart from one another. The note in the laboratory book runs thus:-- Copper wire 1/23 of inch in diameter. Six lengths of five feet each, soldered at ends to piece of copper plate so as form terminations, and these amalgamated. When this bundle was used to connect the electro-motor it gave but very feeble spark on breaking contact, but the spark was sensibly better when the wires are held together so as to act laterally than when they were opened out from each other, thus showing lateral action. Made a larger bundle of the same fine copper wire. There were 20 lengths of 18 feet 2 inches each and the thick terminal pieces of copper wire 6 inches long and ⅓ of inch thick. [Illustration: FIG. 13.] This bundle he compared with a length of 19 feet 6 inches of a single copper wire ⅕ inch in diameter, having about equal sectional area. The latter gave decidedly the largest sparks on breaking circuit.