You will get the same idea, of course, if you think about the electrons in the lower room. The current in the plate circuit will not stop increasing until the voltage of the grid has become positive enough to make the plate current a maximum. It can't do that until enough electrons have left the upper room and been stored away in the lower. Therefore the charging current will have to flow for a long time if the capacity is large. We have, therefore, the other rule for increasing the frequency of an audion oscillator, that is, decrease the capacity. These rules can be stated the other way around. To decrease the frequency we can either increase the capacity or increase the inductance or do both. But what would happen if we should decrease the capacity and increase the inductance? Decreasing the capacity would make the frequency higher, but increasing the inductance would make it lower. What would be the net effect? That would depend upon how much we decreased the capacity and how much we increased the inductance. It would be possible to decrease the capacity and then if we increased the inductance just the right amount to have no change in the frequency. No matter how large or how small we make the capacity we can always make the inductance such that there isn't any change in frequency. I'll give you a rule for this, after I have told you some more things about capacities and inductances. First as to inductances. A short straight wire has a very small inductance, indeed. The longer the wire the larger will be the inductance but unless the length is hundreds of feet there isn't much inductance anyway. A coiled wire is very different. A coil of wire will have more inductance the more turns there are to it. That isn't the whole story but it's enough for the moment. Let's see why. The reason why a stream of electrons has an opposing conscience when they are started off in a coil of wire is because each electron affects every other electron which can move in a parallel path. Look again at the coils of Figs. 28 and 29 which we discussed in the tenth letter. Those sketches plainly bring out the fact that the electrons in part _cd_ travel in paths which are parallel to those of the electrons in part _ab_. [Illustration: Fig 39] If we should turn these coils as in Fig. 39 so that all the paths in _cd_ are at right angles to those in _ab_ there wouldn't be any effect in _cd_ when a current in _ab_ started or stopped. Look at the circuit of the oscillating audion in Fig. 38. If we should turn these coils at right angles to each other we would stop the oscillation. Electrons only influence other electrons which are in parallel paths.