Looks like South America makes up for Africa. Not a lot of excess warming, but just a little.
Mostly just a slow drift to a colder past. 1/4 C in the 1920s, 1/2 C by the 1880s. Nothing dramatic. Some 1850s “lift” to the data (at the far right) that gives “offsetting upward” adjustments that don’t change the trend. A little pruning of some very old very hot data. About the late ’60s just a tiny bit of ‘orange on top’ increase in more recent temperatures.
It is worth noting that once again we see the “Ironed flat” character to recent temperatures. About 1/2 C of range, while in the past we have ranges of 1C and more from year to year. Something is removing the natural variability from the more recent measurements. (Be that tarmac holding heat into the night, airports being in flat more stable valley environments, or thermometers just being pulled out of the mountains and put near the shore – as we saw in prior GHCN analysis where we saw the movement of thermometer locations from volatile mountains to stable waterside sea level.)
Whoever manages the South American data did a very nice job of putting more “trend” into it while not being ham handed about it. Just a tiny bit of change in the upper vs lower dT ranges (at least until prior to 1890 when the blue vs yellow dT lines start to show offset.) But it DOES have trend… so I guess we’ll have to rescind 1/2 the Climate Indulgences for South America. Brazil can pay the Climate Sin Tax and we can let the others have a pass, since Brazil is more industrialized and has all that offshore oil. ;-)

Here are the reports from which the South American graphs were made:
First up, the v1 report:
Then the v3 report: