Stats-in-R | Introduction to Basic Statistics with R | Analytics library

As I found later somewhere else, the Chi² test is probably not appropriate for my data here or rather does not test what I want to find out. Therefore, I conducted a generalised linear model (glm) with a Poisson distribution on my data which worked out quite nicely. So bear this in mind...

.

After consulting various websites on this problem (like this, this or this) and of course the official documentation of the chisq.test function, I still cannot figure out a solution to my problem.

I want to conduct a Chi² Test of Independence on my data via the chisq.test function in R. My data is composed of 4 epiphyte species found on 4 host tree species (that means: plants of 4 species growing ON these 4 tree species). Now, I want to find out if the epiphytes are equally distributed among those trees or if maybe one tree species tends to host more epiphyte individuals as the others. The standard Chi² test I can conduct quite easily (see below). But this would then as well test if epiphyte species were equally distributed, which I don't want to be tested. So, how can I submit different probabilities for my contingency table in the cisq.test function? Namely, I want the expected matrix to be according to the number of epiphyte individuals per species while expecting them to be equally distributed among the tree species. This sounds complicated, so just have a look at my example data:

(I edited the data format as suggested by @paoloeusebi)