Hello everybody.

Broadly, I'm studying the anatomical drivers of wood density in a set of tree species from a lowland tropical forest. So I'm comparing 22 species, I sampled between 2-5 individuals per species, and within each tree I have measurements at different ontogenetic stages (i.e., positions from pith to bark). I'm especially interested on what's happening at the individual level. To do this, I'm working on mixed effect models with individuals and ontogenetic positions as random effects. In the case of big trees, there are several ontogenetic positions and the model runs well. However, for small trees, there are usually 3 or 4 ontogenetic measurements and here the model fails to predict interactions between my response and predictive variables.

Does anyone could give me some advice to analyze my data (considering the few observations for some individuals)?

Thanks in advance.