Email from Roger:

Hi Folks:

Just in case the Monte-Carlo gods do not favor giving Qweak more rate or we just want to be "greedy" I would like to 
toss the following detector layout concept into the "to be simulated" pile. The attached pdf file should be self 
explanatory. I note that even 2 inches of quartz does not start a significant shower so a limited overlap of bars should 
do no harm if one remembers to correct the statistics according to the percent of overlap. Also, now the PMT's can be 
very simply end mounted behind the shielding. Don't get excited about the shielding construction as I beleive we can 
make it work too. This arrangement might allow us to significantly increase both the theta and phi acceptance, if we 
need to.

Best Regards

Roger

pdf of Roger's figure