When to use ANOVA or ANCOVA?

I could use some help here..
I am trying to see if Island Area (sq. km) has a significant effect on Island Population Size across multiple species sampled across multiple islands. I have sampled 8 species that are co-distributed among 3 islands (24 populations in total).

ANOVA tells me that island means are not significantly different from one another, nor are species means. However, larger populations tend to be found on the larger islands (5 out of 8 species have largest populations on largest islands). 

If I want to test the effect of Island Area on Population Size, do I need ANCOVA? If so, what would be the best way to set it up?
Thanks!