In parkhomov translated paper, there is curve power vs. temperature, and it has been fitted to polynomial with terms for temp difference and also for absolute temp for T^4. Compared to Stefan-Bolzmann law, there is rather big difference in the coefficient for T^4, is that error or what should the curve look like theoretically? (or actually the area of the object must be used to scale, then t^4 coeff makes more sense).
Tried decomposing the curves: