Comparing groups - repeated measures or n-way anova?

Hello. I'm new to this site as well as to statistics. So, please forgive me if I have posted in the wrong place. I have a project where I measured a single variable over the course of 120 days. I have four treatments and five replicates. I want to know if I have a significant treatment effect on each day of analysis. I've tried the repeated measure anova in R, and I get results for significance of between days (for all treatments) and between treatments (for all days), but not for between treatments on each day. I was advised to use the repeated measures anova by a postdoc in my lab. I have read up on it, but I'm not sure if it is appropriate. I applied a treatment at the beginning of the study and then measured over 120 days, and the variance is very different as the days progress. Someone else told me I should do a one-way anova on each day because the variance changes so much from day to day. I've tried log transformation and inverse square root transformation (indicated by boxcox function in R), but I am still not meeting homoscedacity. I hope this is clear and understandable. Can anyone give me any advice, please? Thanks so much.
Last edited: