The blood should be examined for the various anemias. If the thyroid gland is deficient or altered in activity, thyroid extract may be indicated--this acts also as a diuretic. Uterine malpositions should be corrected. Treatment should be given where there is any evidence of toxemia, as headache, altered secretion and excretion, neuralgia, mental eccentricity, increased vasomotor stimulation, high tension, disturbance in the sensory apparatus, obstinate constipation and jaundice. Toxemia is not necessarily renal in origin. In any of these conditions the proteids should be kept low in the diet, so that the kidneys may not be overtaxed. To throw off toxins, the emunctories should be stimulated by laxatives, water for diuresis, tepid bathing. If the symptoms grow threatening, and the kidneys are involved, the woman should be put to bed, on water alone. After three days an absolute milk diet should be begun. As she improves, starches are added, then the vegetables containing proteid, vegetable oils, and butter. As the improvement goes on, the diet may be vegetables, fruit easy of digestion, and one egg a day. Later fish and chicken are used, but never a full meat diet. Beef, mutton, veal, and similar heavy meats are not to be eaten. The drink is to be water, buttermilk, or koumiss. When the eclampsia is inevitable the question of inducing labor arises. If the child is not viable, abortion is out of the question, as has been proved in the chapter on Abortion and the general chapter on Homicide. If the child is viable, there are three opinions: one, that the premature delivery should be effected as soon as possible; a second, that this delivery should be delayed as long as possible; and a third, that it should not be attempted at all. Those who hold that the uterus should be emptied as soon as possible, induce labor at the first convulsion, rapidly and under deep narcosis. Chloroform is dangerous to the heart in such cases for full anesthesia; ether is better. Braun first observed that the convulsions cease or are lessened after delivery. Dührssen found these results in 93.72 per cent., Olshausen in 85 per cent., Zweifel in 66 per cent. Peterson said that in 615 cases of early delivery--as soon as possible after the first convulsion--the maternal mortality was 15.9 per cent., but 28.9 per cent. in the same maternities under the expectant method. Olshausen was not in favor of forced delivery. Charpentier[122] held that forced delivery is dangerous and should be absolutely proscribed. His statistics of mortality are: after spontaneous labor, 18.96; after artificial labor, 30.04; after forced delivery, 40.74. [122] _Nouvelle Archives d'Obstétrique et de Gynécologie_, 1893.