The MaKey MaKey's 22MΩ pull-up resistors (which are essential to its main mode of operation of detecting closed circuits through relatively high-impedance real-world objects such as humans and fruit) seem to mean, however, that noise dominates signal when trying to use a banana as voltage divider.
Here is a very clear schematic:
(note that schematic does not correspond to above-linked photo with banana and forks, that's just me measuring the resistance across the banana at roughly 18MΩ)
and here is the data I got from listening on the A0 analog input pin after probing the banana once near the 5V end and once near the ground end:
I'm just sampling A0 in a busy loop and pumping the raw numbers across the serial port and thence through gnuplot. Y-axis is voltage 0 to 5V, X-axis is time.
You can see there's not a huge difference between the two voltage drops there. I imagine doing this whole setup with the Uno, which has no such pull-up resistors built in, might work better. Confession: I don't really understand electrical circuits all that well. Uhhh actually come to think of it, I don't understand why I get a big drop when I probe near the vcc end of the banana.