The considerable diversity of Conus peptides in the I(1)-superfamily provides a rare opportunity to define parameters important for the post-translational l- to d-isomerization of amino acids. This subtlest of post-translational modifications is not readily detectable by most techniques, and it would be a considerable advance if one could predict its potential occurrence purely from gene sequences. We previously described three I(1)-conotoxins, iota-RXIA (formerly designated r11a), r11b and r11c, each containing a d-amino acid at the third position from the C-terminus. In this work, we investigated two novel I(1)-superfamily members, r11d and ar11a, which we show have only l-amino acids. Based on these observations and an analysis of cDNA sequences of other group members, we suggest that there is a rule to predict d-amino acids in I(1)-superfamily peptides. Two factors are important: the residue to be modified should be three amino acids from the C-terminus of the precursor sequence, and it should be in a suitable sequence context. We apply the rule to other members of the I(1)-superfamily, to determine a priori which are probably modified.