the 'instrumental record' goes back to around 1850, this was a world wide standard for temp measurement so im guessing they could in do it 1910
you can have a large scale with a normal mercury thermometer with a big bulb and a thin tube so reading .1 wouldn't be much of a problem.
sounds like they really got their act together in the 1800's and with various mechanical and electrical measuring devices you wouldn't have to check it every 5 mins