Just did this experiment with my Nexus One: Lying on the couch: -75dbm. Held by the top away from body: -70dbm. Held by the top by ear: -75 - -87dbm, depending

Just did this experiment with my Nexus One: Lying on the couch: -75dbm. Held by the top away from body: -70dbm. Held by the top by ear: -75 - -87dbm, depending on which direction I face. Held across the bottom by ear: -75 - -97dbm, depending on direction. So it seems that in the worst case, holding it across the bottom attenuates the signal by about 10dbm, i.e. by a factor of 10, compared to holding it across the top, which is about the same effect as the relative position of your head vs. the tower. It's weird, though, that in the best case, holding it by the bottom has no effect at all on signal. You'd naively think that the best case would be when the phone is on the side of your head facing the tower, in which case the hand across the bottom should make maximum difference. I guess this just goes to show how nonintuitive the results are of loading the antenna with a bunch of conductive goo of strange shape...