After Hisserdude answered my blog survey (it's still open!) about insects in captivity, I decided to elaborate on his idea.
With many poorly-known insects, it is possible to assume husbandry requirements are similar to those of their relatives, or gather a limited amount of info from reading research papers. Unfortunately, this does not provide fine-tuned information, so we end up having to figure some things out ourselves (like unusual dietary requirements).
Right now I am using a "cautious trial-and-error" approach, meaning that I experiment on wild beetles so they are not killed or hurt by mistakes in captivity. This does not always work, due to some insects being difficult to watch for long periods. Anyone have suggestions on how to improve this using non-risky methods? Maybe some sort of "gradient cage" where the insect can self-regulate its humidity, diet, environment by moving to different areas if it senses something we do not know about?