Correlation/LM of residuals with different length


I am currently working with two different datasets. In the first one, I have the mass of 30 individuals. In the second one, I have the force of 90 individuals. I took the residuals from a linear regression to get rid of the size effect. Still, I have 30 and 90 residuals, respectively.

What I would like to do, is testing if there is a correlation between the residuals of the mass, and the force. Using either a correlation test or a linear model. But, as it should be paired data, R doesn't like that...

I could use the mean for the correlation test, but I would lose a lot of information.
I could design two models and extract 50 data for each model and then do the correlation/LM, but I would prefer keep my actual data.

Have you any idea how I could do that?

Thanks a lot!