
Bruce P. answered 08/05/17
Tutor
5.0
(310)
University genetics teacher and tutor
You can't... what you have listed is the OBSERVATION; that's half the battle. chi-square is a measure of how closely the observed data match those that were expected. So you need a hypothesis that generates predictions about the expected numbers. These then become the Obs and Exp values in the chi-square equation.
Example--if you had anticipated EQUAL numbers of all the different sorts, you have (10+15+20+22+12+25)=94 total plants. Six kinds, so if equally distributed, you would expect 94/6= 15.67 of each.
You then take the difference between each count vs. the expected count, square the value, divide by expectation, sum all... and you're off to the races! (or to a p-value table :-)