Ok, this is for an "ideal" situation, in other words a box without cigars or a humidifier, the only thing inside it is air and a tiny bit of water vapor.
Ken, the reason your humidity went up when it got warm isn't because this graph or physics is wrong, it's because you have a practically sealed container with a source of humidity inside it. So as the temperature went up, your humidifier became more "active" and pumped out more moisture than it did when the A/C was running.
Lou, this all depends on if your $10 hygrometer is temperature compensated. At this point, it's say that if you're on the cool side of things, try to keep the humidity a bit higher than normal, but I wouldn't go too far above 70% so you don't promote mold growth. The basic idea is that as it gets warmer, you can drop the humidity a bit, and as it get cooler you can afford to raise the humidity a bit.
------------------------------------------------
Now then, the scientist in me is going to go buy a "Lock-n-Lock" container and perform a little experiment. These containers, if you've never seen them at the grocery store or your local jumbo-mart, have a gasket seal and four locking tabs (one on each side) and are nothing more than a really fancy Rubbermaid container.
Then, I'll take the container into my lab where I have an inert atmosphere glovebox with 99.9999% pure argon sealed air-tight inside it (less than 1 part per million of water) and it has constant HEPA filtration as well as active oxygen and water traps... yeah, it costs a hundred thousand dollars, it's fancy, LOL.
Here's a picture of what the glovebox looks like, but this is at another university (in China):
http://www.pcoss.org/manager/images/mbraun.jpg
This glovebox has two vacuum antechambers attached to it, so I'll take the Lock-n-Lock container, a micro-pipette of water (yes I know taking water into a drybox is a no-no, but a few drops is no problem for it), and a digital hygrometer and load them into the vacuum chamber, remove all the air, and refill it with the pure (dry) argon so we know it's at 0%RH. Then I'll add one drop of water so the total amount of moisture is fixed (as is the case for the theoretical curve shown above) and I'll seal up the box with the hygrometer inside it. The one drop of water should be enough to give the container some measurable RH value once it's sealed up. If that one drop of water doesn't make an appreciable RH register on the hygrometer, I'll add more, drop-wise, until it's looking good (around 50% or so) and no drops are visible so that it's all in the vapor phase. Then I'll take it out of the glovebox, keeping it perfectly sealed, and I'll use a thermostated water bath to vary the temperature up and down by 10 degrees (or more depending on how much time I can spare) and I'll record the RH inside the box at each temperature. This should replicate the theoretical "ideal" conditions as close as possible.
After this is done, I'll repeat this experiment, but this time I'll use a Humi-Gel humidifier inside the Lock-n-Lock box rather than a drop or so of water.
Hopefully I can do this over the weekend and get some results, but I have other work to do first... sigh. LOL
In the meantime, light up a good one guys.
-JT