over represented adapter in reverse but not in forward reads?

Question

[demoraesdiogo2017]

Hello

I am processing some raw RNAseq reads from illumina. FastQC accused that some reverse files had over representation of adapters. I plan to use trimmomatic to trim those specific sequences as paired-end reads. I assume I should also remove the palindrome in the forward files, but FastQC did not accuse any of the forward files to have any kind of over representation. Is this normal?