Relation of Reason
…a relation of reason subsists in virtue of two facts, one only of which would disappear on the annihilation of either of the relates. Such are all resemblances: for any two objects in nature resemble each other, and indeed in themselves just as much as any other two; it is only with reference to our senses and needs that one resemblance counts for more than another.
By logical relations, I mean those in respect to which all pairs [of] objects in the universe are alike; by hemilogical relations those in respect to which there is in reference to each object in the universe only one object (perhaps itself) or some definite multitude of objects which are different from others; while the alogical relations include all other cases. The logical and hemilogical relations belong to the old class of relations of reason, while relations in re are alogical. But there are a few not unimportant relations of reason which are likewise alogical. In my paper of 1867, I committed the error of identifying those relations constituted by non-relative characters with relations of equiparance, that is, with necessarily mutual relations, and the dynamical relations with relations of disquiparance, or possibly non-mutual relations. Subsequently, falling out of one error into another, I identified the two classes respectively with relations of reason and relations in re.
The "paper of 1867" is 'On a New List of Categories'
…the only kind of relation which could be veritably described to a person who had no experience of it is a relation of reason. A relation of reason is not purely dyadic: it is a relation through a sign: that is why it is dicible. Consequently the relation involved in duality is not dicible, but surd; and duality must contain as an ingredient of it a surd disquiparance.
…what is true by virtue of a relation of reason is representative, that is, is of the nature of a sign.