Poor "global" coverage of land weather stations pre-1920 -- ocean temperature measurements were even worse -- a bucket and thermometer in Northern Hemisphere sea lanes, never in the same place, with few measurements elsewhere = not fit for real science