# Can I sum up my count data instead of performing some repeated measures analysis?

I want to analyze establishment of plants in created gaps (plots): I have 10 treated plots, which are divided into an outer section and an inner one and I have been looking at these plots over a period of time (4 repetitions), noting down how many new plants (per species) get established. After each time, all plants are removed, so it's kind of starting from zero again. My data looks something like this:

date plot_no section    species plants

7   A1  inner   M   1
7   A1  inner   F   1
7   A1  margin  F   4
7   A2  margin  C   6
7   A2  margin  M   1
7   A2  margin  F   1
…


What I want is to show that there is more established plants in the outer section than in the inner for the different species. So I am not interested in the effect of time, and I don't want to look at the differences after each time-step. First I did the analysis simply with a Chi-square test, however, I thought maybe I need to use some repeated measures test, so I've been looking at McNemar's test (which is just 2x2, and I have 2x4), than I detected the Bowker test (which is kxk, so columns and rows need to have the same number, right?), so I still don't really see, what could be applied to my data... so I got back to the start, thinking of whether I can simply sum up my count data (sum up all plants thag got established during the whole time) and do a chi-square on these summed up numbers? And: if not, what test shall I perform??