I feel this might be a good point to discuss (read:defend) my choice to use relays. 

Above, some other 3V relays that were physically larger that I also got but didn't use. Inserted for visual interest. 

I desired to keep the consumer products electrically isolated from any control circuitry so as to avoid complications caused by interactions and differing voltages. Both SKELLY and the audio unit were designed to be activated by pushbuttons. The crudest solution would be a solenoid that physically pushed the button with a plunger when activated. Although not impossible, this solution is appallingly inelegant and would require more energy than any other sensible solution. 

Relays are slow, they require significant energy, and the coil requires a snubber circuit to prevent damage from inductive kickback/ back EMF when de-energized.  However, they are cheap, available, reliable, and provide excellent electrical isolation. The switches on a relay may suffer from some mechanical chatter, but they are very low resistance and can handle a wide range of voltages and currents.

The desire for electrical isolation eliminates most transistor- or SCR-type switching with the exception of optoisolators (optocouplers, photocouplers). These devices look like integrated circuits but are really a LED paired with a photodiode, phototransistor, or other photosensitive active component. When one circuit lights the LED, the circuit on the other side is affected by the light striking the photosensitive component. Thus the two circuits are electrically isolated up to hundreds or even thousands of volts, with only light energy passing between them. Most optocouplers do not handle a lot of current and become derated quickly, but could be used to switch a more capable transistor.

So the switches could be replaced by optically-activated transistors.  I abashedly must admit that my transistor knowledge is somewhat lacking, but even a favorite like a 2n2222 or 2n7000 has a significant voltage drop or ohmic resistance when "fully on", which is a bummer. Knowing what sort of switch resistance is tolerated in the consumer circuits requires significant analysis or experimentation. Then again, a 2n2222 driven into saturation only has a voltage drop of .3-1.0V and an "on" 2n7000 has a resistance of no more than 5.3Ω with 4.5V between gate and source.

I used little black relays instead of big blue relays because they were smaller. I also wish to gripe one more time: WHY is it that no relay fits properly onto a breadboard?  I daresay no relay that I have can go into a breadboard without shorting pins or jamming a pin in the middle trough.