We now come to the question of spark-gaps, and at the very outset one may say that each tube requires different treatment. Sometimes it has been advised to place the balls as near as possible to the tube. This in highly inconvenient, as they are constantly in the way; and further, one never feels quite sure that the tube is not endangered. An arrangement that was in use for some time and answered well, consisted of two pieces of glass tubing about one foot long. These were fixed into two pieces of wood, which were grooved at the sides and fitted to work on a slide. The balls, which were of course mounted on the glass, could easily be made to approach or recede at pleasure. Finally the following arrangement was made, and gives more satisfaction than any other yet tried: AB Fig. 4, is a stout glass pillar 13 1/4 in. long by 1 1/4 in. diameter. One end is fixed into a strong piece of wood 7 in. by 4 1/2 in., through this passes a clamp C which holds it firmly in any position. The other end of the glass pillar is let into a piece of wood at B, and from this a stout mag-nalium rod D springs, which carries the ball. This piece of apparatus is so mounted that the ball is level with the ball of the machine conductor. The wires to the tube can be put on or taken off the rod D in a moment, while by releasing C the whole can be moved out of the way, or the length of the spark-gap varied on either side as is desired.
As regards the size of the balls, experiment will alone decide. We have balls from 5/8 in. up to 8 1/4 in. in diameter. In the case of some tubes the brilliancy is vastly enhanced by the use of large balls. In the next place, we must say a word about connecting up with the tube. The thickness of the wire does not seem to be of much consequence; only if too thick, it will no longer be flexible. It should be passed through india-rubber tubing; but the connecting-up with the tube requires care. Usually with a bianodic tube a tine spiral of wire connects the two anodes. This is best removed. A fine piece of indiarubber tubing, about 8 in. long, should be taken, and a wire passed through it; at each end there should be a second piece of tubing sufficiently large in diameter to pass over the glass projection which carries each anode. The ends of the wire should now be connected with the two anodes, and the second piece of indiarubber slipped over them. The tighter the indiarubber tubings fit over each other the better, and the outer one may be turned back over the inner till the wire is fixed, and then it can be passed over the anode by turning it straight again. We now take the middle of our tubing, connecting the anodes, and make a small hole, and insert the wire connecting with the spark-gap. The junction of the wires should also be protected with a larger piece of tubing to cover it up. By these means the brush discharge from the end of the tube will be greatly diminished if not entirely done away with. Having now described all the different pieces of apparatus, we come to the actual working of the whole. This falls into two divisions (1) with primary circuit alone, (2) with primary and secondary circuit.
Here the bottom of the jars are unconnected, and the tube is excited by the current directly from the conductors. The current from the Wimshurst is, however, feeble, and the larger the number of plates the better. The output depends on the speed of rotation of the plates. In fact, the brilliancy of tube equals numbers of plates, multiplied by the speed of rotation. The brilliancy is also affected by the spark-gap. If this is too large, the tube shows signs of reversal, particularly if a small-sized one; if too small, the glow is faint. One spark-gap of 1/2 in., or two of 1/4 in. generally work the best.
If the bottoms of the jars be joined together, the brilliancy of the tube is, as we should expect, doubled; but as the charge is intermittent, the tube is unsteady and smart sparks occur from time to time which are apt to be disastrous. Leyden jars added still further increase the brilliancy of the tube, but, of course, increase its unsteadiness. For various experiments a large number of jars are
Figs. 5 and 6.
made, and we have two four-pint jars, eight gallon jars, two still larger, and one five-gallon jar. The last is affectionately named " Jumbo," and has a table to himself, and is treated with due respect, since, having been left in the room during one evening's work, though quite away from the machine, he somehow acquired a charge, and when he was touched promptly deposited one of us on the floor. This is said as a word of caution, as it is by no means pleasant to be ""knocked into the middle of next week." Experiments with the secondary soon led to some interesting results. Passing the current from the outsides through a large coil of fine wire gave very bright but intermittent flashes. A series of experiments were then made with various forms of resistances, some of which may be described. Fig. 5 is a glass tube, 6 x 2 in., securely corked and having a small disc of magnalium mounted on a wire carried through the bottom, and outside to a terminal. Through the top passes a piece of glass tubing with a cork at the end, which is pierced with four pins. A copper wire passes down the tube and is connected to the four pins with a drop of mercury. The outer glass tube is filled with oil. The inner tube can be moved up or down till sparks pass freely between the points and the disc. This answers and gives good results.
Figs. 7 and 8.
Fig, 6 is a three-necked retort receiver, also partially filled with oil. Two rods pass through the horizontal corks, and are connected together by wires to a terminal on the stand. The vertical neck has a rod which ends in a magnalium disc. It was thought that as the tube was bianodic the resistance should be as well. It cannot be said that there was any improvement on the original one.
Fig. 7 is a convenient shape and consists of a chloride of calcium tube. The bend is filled with mercury, which extends just up to the bulb. The rest of the tube is filled with oil. The two wire pass through corks at the ends of the tubes, and spark on the mercury. Many other forms and modifications of these were tried with varying results. Generally, the more sparking points there were the better. The principle is the same in all, - viz., a sparking resistance.
Fig. 8 is a resistance of another kind, and gives much superior results. It is a U-tube 8 in. long by 7/8 in., and the liquid used is a solution of copper sulphate. Into this two copper wires plunge. Of course any acid solution will do equally well.
The last form is shown in Fig. 9. This consists of twelve pieces of glass tubing, each 1 ft. long, connected ogether with indiarubber and containing solution of copper sulphate. The mounting is an upright piece of wood carrying two glass rods, and to these the tubes are fixed. A small coil of fine wire, not shown in the figure, is also interposed, and with the spark-gap properly adjusted, the tube is brilliant and steady, and seldom troubled with sparks. The resistance in the secondary act as a balance to the resistance of the primary, which is made up of the tube and the spark-gaps. If the spark-gap is too small the secondary cannot pass, and the tube is as if running from the primary only, and if too large there are sparks at the tube instead of current in the secondary. When the two circuits are properly balanced the brilliancy of the tube is most extraordinary, and the larger the diameter of the tube the better the result. One word of can-tion-don't leave a resistance coupled up at one end only, otherwise it will blow up with an alarming explosion and scatter the liquid contents in all directions. The results of our experiments are communicated in hopes that they will be useful to others.