Estimation of random versus fixed effect size in mixed models in r

I am analyzing an effect of food deprivation on bird chicks' calls. An idea is to show that food deprivations (experiment) contributes much more to acoustic parameter changes than species or individual identity. We have 3 species, 8 individuals of each species. A reviewer of our ms said that we should use mixed models with experiment and species as fixed and individual as random variables. I would like to rank a contribution of all 3 variables. How do I do it? I get a variance which random intercept adds to a population intercept, but it's not quite what I need. Also, anova(model_name) gives me F-values for both fixed effects but not for a random one. A paper by Nakagawa & Schielzeth (2013) is talking about R-sq. for random vs fixed effects, but I have two fixed variables and would like to rank all three. Any thoughts? Thank you! Here is the code (a model with a random intercept works better than intercept and slope).

m_Fmax <- lme(F_max ~ Experiment + Species +  Experiment*Species, random=~1|chicks, method = 'REML')