…I just thought I’d let you know so it wouldn’t be unexpected…
The first such machine is the glass maze way back from Theodosius Dobrzhansky‘s lab at Columbia. This resembles nothing so much as a plinko machine for flies: flies start out at one end, and face series of up-or-down branching tubes that are fitted with angled bristles that only let the flies move forward in the maze. After every branch, the individual fly’s position is the number of upward moves minus the number of downward moves. After several stages, the glass pipes outlet into bottles. So you put the flies in the glass maze, you go home for the night, and next morning you come in to see that the flies have sorted themselves out: the ones who enjoy going up are in the top bottles, the ones who enjoy going down are in the bottom bottles and the goldilocks flies are in the middle. Well, I mean, there’s some randomness involved obviously, but, though it seems bizarre at first glance, there’s actually a genetic component. If you breed only the top-going flies and put that generation through the maze, their distribution will actually be skewed towards the top. Keep doing this for several generations, and you breed a race of veritable fly alpinists. Not that that’s helpful for the fly: the best fly is a well-adjusted one that doesn’t have a strong preference for up or down, and so if you let natural selection work its magic, a few generations is enough to delete that hard-bred alpinism.
Next, we have a machine called the dorsoventrometer that was made by Kenneth Weber at University of Southern Maine. It consists, from what I can gather, of a glass floor and a sticky plate on a very precise translation stage. Flies are put to sleep on the glass floor, and the sticky plate is lowered microns at a time. The fattest flies are the ones that stand out from the floor the most, and so the sticky plate catches them earlier in its descent than the slender flies. By raising the plate and collecting the stuck flies at several increments during its descent, you can separate out groups of different sizes. In this way you can select flies for girth. This does seem a bit macabre – I wouldn’t want to be a test subject in a human version of such a machine, for instance. But remember, it’s okay to artificially select flies ’cause they don’t have any feelings.
Finally, another machine by Kenneth Weber which is whimsically named the inebriometer. It consists of a large glass jar with several funnels on the inside (called baffles). The flies enjoy themselves in this jar, and all the while, alcoholic fumes are being pumped in. At some point, the flies can’t take it any more, and slip and fall and layabout on the floor of this jar like the miserable drunks that they are. What they don’t know is that the glass jar actually has a trap floor, and the flies which fall in during a given timespan are shunted via the baffles to a given bottle. As time goes on, the bottles move along, and you have an automatic means of separating the flies from the cheapest drunks to the Finns of the fly-world. And then, given several generations of artificial selection… (this also makes for interesting thinking about what effect natural selection has had on our tolerance for alcohol).
Felsenstein also mentioned a third Weber invention called an “icarometer” that measured phototaxis (the flies’ attraction to light) – presumably based on having them soar towards a lightbulb and then fall, mini-Icarus-like. Unfortunately I can’t find anything on the internet about this machine. Anyone have any information? Clearly, Weber is a master of the fly-selection machine. I can’t help feeling that we have our lucky stars to thank that history has not placed us, Weber and the Spanish Inquisition (see! Told you!) at the same space and time, or else he could have been a formidable torture instrument maker and we would have been really screwed, probably.