I think there will be lots of answers, or more accurately responses, to this question but substaintially only speculation and antecdote. Ironically, it's a thing farily easy to test accurately.
So, here's my speculation and antecdote. Wood with big pores are the most vulerable, ring porous. Woods with natural resins tend to react less quickly to changes. I have witnessed antecdotally, but not measured scientifically, swings of 5# to 7#, or about 10%, on woods like elms and hickories. I don't personally use oaks but I'd expect they qualify for the same reason. This with Rh swings of 25%, say from 35% to 60% with acclimation periods on the order of weeks, versus days or hours. Osage and yew enjoy a reputation for being not as much effected, but this may be due simply to the relative slowness of acclimation or their tendancy to be inherently overbuilt due to their relatively uncommon capaicity for compression. But there may be inherent qualities which mitigate the effect in these species. Bamboo is much effected, ipe closer to elm than osage. I'd presume the area of most bending, under the most stress, would be most effected in terms of tiller, so inner and mid limbs would tend to set and weaken as they tend to carry more of the compression load.
I wish somebody would do some testing, as this aspect of natural bowyery is in my view perhaps the most important. Another less investigated area, but frequently speculated on is different species capacity for repeated bending, or work fatigue. We know for sure some metals, spring steel for example (albeit engineered by virtue of grain structure and carbon content) enjoy much higher capacity than others.