Poetry and Science
In the tradition of Romanticism generally (a pivotal early document being Wordsworth's "Preface to Lyrical Ballads"), Barfield was much interested in the relationship between the seemingly strange bedfellows of poetry and science.

As early as Poetic Diction, he had concluded that

Really there is no distinction between Poetry and Science, as kinds of knowing, at all. There is only the distinction between bad poetry and bad science. That the two or three experimental sciences, and the two or three hundred specialized lines of inquiry which ape their methods, should have developed the rational out of all proportion to the poetic is indeed an historical fact--and a fact of great importance to a consideration of the last four hundred years of European history. But to imagine that this tells us anything about the nature of knowledge instead of a way of testing, this is--instead of looking dispassionately at the historical facts--to wear it like a pair of blinkers. (PD 139)

In Romanticism Comes of Age, he would insist with conviction that "Science must itself become an art, and art a science, either they must mingle, or Western civilization, as we know it, must perish, to make room for one that may have spirit enough to learn how to know God's earth as He actually made it" (RCA 65).
See in particular Poetic Diction, passim.