The approach to magentic saturation in the cases of iron, nickel, and a number of iron-cobalt alloys at various temperatures has been studied in the range of field from a few hundred to 6,000 oersteds. It has been found that for annealed and moderately cold-worked specimens the differential susceptibility can be represented fairly accurately by a formula of the form ((?I)/(?H))T=A/H2+(2B)/H2+ C/H1/2+D where I is the magnetization, H the applied field, and T the temperature. Near room temperature, the last two terms taken together are so small compared to the first two that they can be approximated by a single constant. For a specimen starting in the annealed state both the coefficients A and B increase at first tremendously with the amount of cold-work the specimen receives, but the formula fails to apply when the specimen is severely cold-worked For annealed specimens these coefficients decrease with increasing temperature, vanishing at considerable distances below the Curie point. In general, D cannot be evaluated with certainty from the present data. However, the coefficient C evaluated by neglecting D is in satisfactory agreement with the theoretical results of Holstein and Prima-koff in so far as its order of magnitude and the way in which it varies withthe temperature are concerned. B is undoubtedly proportional to K12/I5 where K1 is the first magnetocrystalline anisotropy coefficient and I5 thesaturation magnetization. However, values of K1 obtained by comparing theory and experiment are satisfactory only in regard to the order of magnitude.