The alkalinity of the blood seems not to be changed, although Chotin and Bouvier noticed a slight increase. The saline constituents do not {175} vary greatly from the normal standard. The average amount in the cases of Becquerel and Rodier and Busk was 8.1 per 1000, the smallest being 5.5 parts and the largest 11.5. In Ritchie's two analyses the proportion of saline matters is given as 6.44 and 6.82 parts per 1000. Opitz and Schneider have found less than the physiological mean. In Frick's case the amount was 8.8, the iron being 0.721 parts per 1000, and 0.782 to 127 parts of globules; lime 0.110, chlorides 6.846, and phosphates 1.116 parts per 1000. The iron was in excess of that in the normal blood, but in Becquerel's cases the mean was 0.381--less than the normal. The proportion of iron in Duchet's cases was respectively 0.393, 0.402, and 0.476 parts, giving a mean of 0.423 parts per 1000, which nearly approximates the normal. Garrod in one analysis of the blood found a deficiency of the potassium salts, upon which he erected his well-known theory of the etiology of the disease. It is an interesting fact that in the physiological state the quantity of sodium chloride is not subject to variation, any excess introduced with the food being thrown off by the kidneys. The quantity in the urine bears a relation to the amount introduced as food, but the proportion in the blood is constant. The quantity of water in the blood has been found to be increased in all the analyses which have been made. Chotin and Bouvier estimated water and loss at 831.1; in Frick's case it was 791.69 parts per 1000; and in Becquerel's five cases it was put at 807.7, 810.9, 811, 813.7, and 854.0 parts per 1000, respectively. In Busk's three cases the lowest amount was 835.9 and the highest 849.9 parts per 1000. The specific gravity of the defibrinated blood was in all cases low in comparison with the normal standard (1057), the average in Becquerel and Rodier's cases being 1047.2, the lowest 1083.3, and the highest 1051.7. In the single observation of Chotin and Bouvier it was 1060. The specific gravity of the serum was also less than normal (1027), the average of four of Becquerel's analyses giving 1023.8, the lowest 1020.8, and the highest 1025.5. Busk gives 1025 in one case and 1028 in another. The results of the most recent analyses, those of Chalvet, are shown in the following table, in which scorbutic blood is contrasted with that of a healthy, robust female: