It's from "The Artificial Production of Carbon Dioxide and its Influence on Temperature," G. S. Callendar, Quarterly Journal of the Royal Meteorological Society v64 Issue 275 pp 223-240 (April 1938). [PDF]
Callendar also mentions the "well known" heat island effect, on page 235.
Note Callendar didn't attribute all the temperature rise to CO2. I don't see in his paper (though I haven't read it carefully) where he considered possible changes in solar irradiance (or, for that matter, a reduction in volcanic cooling that was part of the Little Ice Age). But those would be small, perhaps just a few hundreths of a degree (climate sensitivity to changes in solar radiance is small, about 0.1°C/(W/m2), for solar radiance measured at the top of the atmosphere; see the IPCC 5AR).