The question of how to conduct a chorus of nature's musicians is perplexing because humans aren't in charge --Mother Nature is. The idea of physically "playing" a cricket or other musical animal is compelling, but I'll leave that (cruelty-free) implementation to someone who knows how to play an instrument.

My approach is to let Mother Nature conduct her own, which is to say, allow a circuit to control an electronic chorus. What I want is an autonomous master sequencer that runs indefinitely, controlling the noise of multiple animal circuits in a seemingly random way, but never so abhorrent as to be unbelievable. My control circuit doesn't have to be truly random according to cryptologists, but it does need to be unpredictable for the average observer. Otherwise, listeners will detect a pattern and the effect will be unconvincing if not downright annoying.

The key to my solution is a circuit designed by Charles Platt called "Really, Really Random Number Generator" that was published in Make Magazine in May, 2015. Platt combines a white noise generator (a source of true randomness) with a Schmitt inverter and shift register to produce a never-ending sequence of random bits --eight bits, to be precise, as that is the output of the shift register in his design. I encourage you to read his article and study his schematic, as he does an excellent job of explaining the design principles and I am only standing on his shoulders (as in "standing on the shoulders of giants"). 

If you refer to my schematic "Mother Nature Board (random pulse gen).pdf" included in my project files, you will see that Platt's design makes up the lion's share of what I'm doing to simulate Mother Nature's timing. I made a few changes to his circuit because I am running at 12V instead of 18V, and I use a larger capacitor for the Schmitt oscillator because I want a slower clock for my shift register in order to see the state changes (plus, this is how I get so many bonus fireflies and other LED eye candy). The remainder of my schematic comprises two basic functions: logic gates to generate further random pulses, some of which occur less frequently than the states of the shift register; and a collection of 555 timers in monostable configuration, aka "one-shots" to generate pulses of longer duration. 

Ultimately, it is the outputs of the one-shot timers that I want. These pulses are how I turn animal circuits on and off (or affect some other parameter). Platt's design is a starting point for me to control any number of animal circuits: an output of his random number generator (and logic gates, as desired) acts as a trigger for a one-shot. More frequent triggering equals more pulse out. The duration of a one-shot's pulse out is a factor of the resistor and capacitor on DIS and THR of the 555 timer. You have to breadboard the circuit and mess around with it until you get the frequency of triggers and pulse durations to your satisfaction. It's not hard, but I give you the following hints/ warnings:

-- The circuit is sensitive to noise.

-- Not all 2N3904 transistors are made the same, and you must get the right one for the noise generating part of Platt's circuit. I tested multiples from my collection and some worked, while others did not. If your shift register isn't shifting, it's probably a problem with the noise generator.

-- The trigger into the 555 monostable (one-shot) is active low. You can use leftover inverters in the hex Schmitt IC or build your own, as needed.

-- Careful about impedance. The logic ICs can't drive much of a load, so you might need to add a transistor switch in-between (especially when you want one output to control multiple things). This also applies to the output of the one-shots.

Last but not least, here is a video without sound that is intended to show the different elements of my Mother Nature Board. Pay attention to the LEDs: they reveal how the states are changing at various points in the circuit.

“The Earth has its music for those who will listen.” – George Santayana