Now, if we do introduce a word or two between “and” and “in,” and such words do not suggest relation to what precedes, we may still omit the comma before “and.” This reasoning, with that above, suggests the best punctuation of No. 21, which is as follows:
21-2. The deliberate good sense with which Franklin treated matters of religion and morality, he displayed equally in his scientific writings and, a little later, in his public documents and correspondence, which made him as eminent in diplomacy and statecraft as he had earlier been in science and local affairs.
Thus we find no need of any mark before “and” in this sentence, where a painstaking writer uses a semicolon.
As “and,” even in No. 21-1, does not connect “writings” with “documents,” but connects the two groups of words beginning with “in,” the adjective clause in No. 21-2 beginning with “which” can hardly go over into the first group, and there find a noun which it may seem to modify.
To illustrate how puzzling a rule may be, and how wrong, we quote the following rule from Mr. Wilson’s work (page 113):
When a sentence consists of three or more clauses, united by a conjunction, none of which are susceptible of division, a semicolon should be put between those which are least connected in sense, and a comma only between the others.
To illustrate this rule, the following sentence is given in Mr. Wilson’s book (the italics are ours):
22. The woods may disappear, but the spirit of them never will now; for it has been felt by a poet, and we can feel for ever[5] what he felt.
Does the punctuation indicate the real sense relations in this sentence? We think not. The semicolon before “for” divides the sentence into two parts; but what follows the semicolon is clearly not in the for relation with all that precedes it, as it should be if the sentence is divided into two parts at this point. The sense relation expressed by “for” is unmistakably between all that follows it and what precedes it back to “but.”
The real division of the sentence into two parts is made at “but,” as shown in the following: