Abstract We study historic observations of solar activity from the 20th-century rise towards the peak of the Modern Grand Solar Maximum (MGSM) and compare with observations of the decline that has occurred since. The major difference in available solar observations of the rise and of the fall are accurate magnetograms from solar magnetographs: we here use synthetic magnetograms to interpret the rise and employ historic observations of Polar Crown Filaments to test them and verify their use. We show that eclipse images at sunspot minimum reveal the long-term variation of open flux deduced from geomagnetic observations in Paper 1 (Lockwood et al., 2022). We also make use of polar coronal hole fluxes derived from historic white light images of polar faculae, but have to consider the implications of the fact that these facular images do not tell us the polarity of the field. Given this caveat, the agreement between the polar coronal hole fluxes and the values derived from open flux continuity modelling based on sunspot numbers is extremely good. This comparison indicates that one possible solution to the “open flux problem” is open flux within the streamer belt that potential-based modelling of coronal fields from photospheric fields is not capturing. We take a detailed look at the solar cycle at the peak of the MGSM, cycle 19, and show the variation of the polar coronal hole fluxes and the inferred poleward flux surges are predictable from the asymmetries in flux emergence in the two hemispheres with implied transequatorial flux transfer and/or “anti-Hale” (or more general “rogue” active region flux) emergence late in the sunspot cycle.