The closest thing to an official account of why JCT split into JCTA and JCTB may be found in Edwin F. Beschler's article Gian-Carlo Rota and the Founding of the Journal of Combinatorial Theory, J. Combin. Theory Ser. A 91 (2000), 2–4.

However, no single journal or set of editors, prestigious and hard-working as they might be, could overcome the tremendous diversity in hopes and aspirations for combinatorics. Competition for page allocations and scheduling soon led to intolerable strain on the editorial board and there developed a real threat to continuing cooperation and growth.

Thus it seems that the volume of submissions was a major driving factor behind the bifurcation into A and B. Theodore Motzkin agreed to be the chief editor of A, and W. T. Tutte agreed to be the chief editor of B. The research interests of the chief editors undoubtedly played a role in the decision as to which papers went to which journal. Some further information is provided in the editorial article Fifty years of the Journal of Combinatorial Theory, J. Combin. Theory Ser. A 144 (2016), 1–6.

The first thing that had to be done was to decide the fate of papers submitted to JCT, but which would only appear after the split into series A and B. Together with the editors for JCTB, Bill Tutte and Dan Younger, it had to be decided which papers should be routed to JCTB and which to JCTA. Typically papers on graph theory or matroids went to JCTB and the rest to JCTA. This amounted to a rough split of about 40/60. Naturally, papers often did not fit obviously into one series or the other and judgements had to be made. This has remained a recurrent problem which we still face to this day. Over the years the distinction based on graphs and matroids has become much less pronounced as, with the emergence of new subfields such as additive combinatorics, combinatorial commutative algebra and physical combinatorics, the field of combinatorics has grown significantly.