Observations on the blood picture of Johne's disease in sheep and cattle with special reference to the magnesium content of the blood