Hi Tal!

I´ve been looking for the same sort of stuff lately, but it seems there are not a lot of bibliography about it. Just an advice: as you want to make repeated measures on the same individuals, care should be taken when modelling growth, because when you make regressions (linear or non-linear) to model growth, residuals will be serially autocorrelated. The independence of the samples inter-days is violated (you measure the same leaf one day and the next one).

I think you have panel data.

You could photograph leaves against a black/white/red background and try to extract the leaf using ImageJ or the LAMINA software or similar, but much care about the non-independence of samples between days!

Best wishes!

I am not sure what you mean by non-destructive sampling.

However, here is a suggestion for a statistical approach that might save replicates: Let's say you observe a steady increase of a phytochemical over time. You will only need one replicate per point in time in order to perform a regression analysis which will tell you whether the increase is significant (p-value of the slope). I would still find it more convincing with more than one sample per time-point but in any case you should need fewer replicates when compared to an ANOVA/t-test approach.

To my experience, regression analysis works well for half-way linear relationships. If there is a non-linear but half-way steady increase or decline it might be useful to transform the data into ranks. I don't know if there is a way to apply regression analysis to more complex relationships that involve maxima or minima within the time line.

Thank you for the replies. I should have been more explicit; non-destructive meaning the plant will be alive and healthy at the end of the study. I plan to take the most recent fully expanded leaf from the trees each month. 

Then, if I understood, you will have pseudo-replicates (you will collect different leaves each month, but from the same trees)?

Yes, I will collect different leaves each month from the same trees (edit: I have 12 trees). Leaf samples will be used to look for changes in foliar compounds as the plant ages.