From Conover‘s History of the Blue Hill Meteorological Observatory: The First 100 Years—1885-1985.
Okay so what I need you to understand about this photo of the inner workings of Blue Hill Observatory’s 1935 Model F radio-meteorograph1 is that the barber-stripe pole in the center in fact did rotate, that’s what all the clockwork underneath is for, and the “stripe” is actually a strip of metal that, as the pole rotated, would briefly touch the contact for each instrument in sequence, conducting power to the transmitter about once a minute. The weird bellows-looking thing on the left is the pressure sensor, the curved piece of metal all the way to the right is the bimetallic temperature sensor, and the little hammock-y thing between the pole and the thermometer is in fact suspended by human hair as a humidity sensor. The idea here is that each of these sensors moves their contact up or down slightly based on changes in what they measure—so the timing of each signal pulse is how this thing transmitted information to the ground observer. My understanding is that by the end of the ‘30s, this design had been more or less entirely supplanted in the U.S. by the Diamond-Hinman models which sadly drop the pole and instead make the wild decision to not bother timing the signals at all, instead having the device transmit whenever it hit specified pressure levels, as the arm hooked to the barometer swings across a whole array of contacts.2
You do make weird choices when you need to build something that works in the stratosphere, I guess!
1 The term “radiosonde” would later become standard, though they are often simply referred to “weather balloons.”
2 Diamond and Hinman credit Lange—of the Blue Hill Observatory—for this mechanism.