Bias in a chi-square contingency table?


I am setting up a 7 x 7 contingency table for a chi-square test of independence of forest tree species (below).


I want to see where the associations are. But note, for example, that if I want to see the association between red oak and white oak (I suspect there is an important one), I cannot, because they are both on the same side of the table.

So I thought maybe I should find the six most frequent species encountered in the forest plots (listed at right), and arrange them on opposite sides of the table to generate lots of "hits". But then would not this introduce bias in the results re: the chi square stat and z critical value? Then I thought maybe I should randomize the placements of the species on the axes.

Any thoughts?