Now buy some more apparatus. You will need a grid condenser of about 0.0002 mf. The grid leaks to go with it you can make for yourself. I would use a piece of brown wrapping paper and two little metal eyelets. The eyelets can be punched into the paper. Between them coat the paper with carbon ink, or with lead pencil marks. A line about an inch long ought to serve nicely. You will probably wish to make several grid leaks to try. When you get satisfactory operation in receiving by the grid-condenser method the leak will probably be somewhere between a megohm and two megohms. For this method you will not want a C-battery, but you will wish to operate the detector with about as high a voltage as the manufacturers will recommend for the plate circuit. In this way the incoming signal, which decreases the plate current, can produce the largest decrease. It is also possible to start with the grid slightly positive instead of being as negative as it is when connected to the negative terminal of the A-battery. There will then be possible a greater change in grid voltage. To do so connect the grid as in Fig. 115 to the positive terminal of the A-battery. [Illustration: Fig 113] About this time I would shop around for two or three small double-pole double-throw switches. Those of the 5-ampere size will do. With these you can arrange to make comparisons between different methods of receiving. Suppose, for example, you connect the switches as shown in Fig. 113 so that by throwing them to the left you are using the audion and to the right the crystal as a detector. You can listen for a minute in one position and then switch and listen for a minute in the other position, and so on back and forth. That way you can tell whether or not you really are getting better results.