Abstract
The limited purity of available ferromagnetic materials may turn into a pitfall in experimental studies of critical phenomena. In pursuit of asymptotic critical behaviour it is not uncommon to approach the Curie point too closely, violating the condition |ε| > > ΔT/TC, ε = (T − TC)/TC (here ΔT ∼ Δx×103 K is the width of the critical region, Δx being the fractional impurity content). If this happens, the presence of what is normally regarded as a tiny amount of impurity can no longer be neglected. One of the consequences is that the critical exponent β is noticeably over-estimated in comparison with its value in the pure material (which is the quantity sought in such studies). This problem does not occur if the excessive proximity to the Curie point is avoided.