I understand the possible methods for narrowing down the variables, but wasn't sure if Greg was interested in a thesis project
It might be more feasible to take 1'-depth temp-readings from several similar ponds (some dyed, some not), at regular intervals over several weeks and then compare the variances within each pond - instead of directly comparing the temps from one pond to another. If the dyed ponds signifcantly increase in temperature (over their base-lines) more rapidly than the non-dyed ponds, I think the theory would be supported. Still, a lot of work.....
