The restriction of fluid is often used as a therapeutic means, especially in aneurism, nephritis, heart disease and obesity. In health, if fluid is entirely withdrawn, the body receives only that contained in the solid food, about four-fifths of a pint, and that derived from the oxidation of the hydrogen of the food or tissues, about half a pint; the total may reach a pint and a half. The evaporation of water from the skin is controlled but that from the lungs goes on; the quantity of urine falls to half a pint or less; but in spite of the restriction of the loss of water from the skin and the kidneys, the total output is greater than the intake, and water is lost from the body. In " cures " by deprivation of water the tissue fluids are reduced by 6 to 8 per cent, and the blood serum has been found to lose over 10 per cent of its water. If the reduction became greater than this, we should expect harm to follow, for in animals a loss of much over 10 per cent of the total water is incompatible with health. In such restriction cures the appetite is lessened: the absorption of food is normal. The oxidation processes as shown by the respiratory exchange are not affected; nor, in animals, is the amount of fat in the body reduced. The experimental evidence is therefore against the statement that the oxidation of fat is favoured by the restriction of water. In obesity cures with limitation of fluid the reduction in weight is largely due to the loss of water. There is, however, also a loss of nitrogen in the urine, and this may persist for a time after the intake of fluid is increased again to the normal. We may sum up by saying that the restriction of fluid causes a diminution in the total water in the body, with a corresponding "thickening" of the body fluids, a lessening of the appetite, and a loss of protein.

The quantity of water which should be allowed in nephritis requires careful consideration, and has been the subject of many experiments. Dropsy may be due either to an inability of the kidney to pass out water, or to the retention of water by the tissues; in kidney diseases the blood appears to contain an excess of non-protein nitrogenous and saline molecules, and, supposing the same to be the case in the tissues, water would be retained by osmotic attraction in greater quantity than normal.

In acute nephritis drinking much water does not usually lead to diuresis, and the quantity taken should therefore be limited to that necessary to satisfy thirst.

In parenchymatous nephritis experiments show varying results corresponding, no doubt, to different types of the disease. It has sometimes been found that an increase in the fluid is followed by a similar increase in the urine : in such cases it may be assumed that it is advisable not to restrict fluid. If, on the other hand, the quantity of urine is unaffected, any extra water taken in is likely, in the absence of free sweating, to increase oedema.

In cases of chronic granular kidney there is usually a free secretion of urine. Bradford, and Ribbert, found the same to occur when the kidney substance was artificially reduced. We may regard this natural diuresis as advantageous to the kidney, for less work is involved in passing out the solids of the urine in a weak solution than in a strong one, and this being so it is undesirable to restrict fluid in this form of the disease. v. Noorden believes that the large amount of fluid dealt with is harmful to the circulation, and he limits the intake of water on this account. Dealing with much greater quantities of fluid does not, however, affect the circulation in diabetes, and it is probably more important to bear in mind the tendency to the accumulation of the non-protein nitrogenous and saline molecules in the blood in nephritis. We have already seen that restriction of fluid, unless very moderate, is likely to lead to a still greater increase in the solids of the blood and tissues, and that in animals there is a very definite limit to the degree of concentration which can be sustained without harm. This appears to be an indication that whenever the kidney can excrete water it should be encouraged to do so, with the object of washing out this material. Such a conclusion is supported by the observations of Frey, that in diuresis produced by water, the osmotic concentration of the urine may sink below that of the blood. This would greatly favour excretion by the kidney. The estimation of the molecular bodies in the tissues and blood in nephritis offers considerable difficulties, which are not, however, insurmountable, and it is desirable that a larger number of analyses should be made.

In the dropsy of heart disease the blood appears to be more watery than normal, but regains its proper composition when compensation is re-established. Although in circulatory failure the excretion of water from the kidney is diminished, this is not the cause of the oedema, which must be looked for in the conditions obtaining in the vessels of the affected tissues. Changes in the capillary walls probably have an important influence, especially in the peripheral parts of the circulation, for, the blood tending to stagnate in these regions, its oxygen is soon used up and the cells lining the capillary walls become insufficiently nourished, and fail to regulate the passage of both fluids and solids from blood to tissues and vice versa. It must be remembered that water is constantly being produced in the tissues, and it is only necessary to suppose that this water is not taken up by the blood vessels in order to explain the production of oedema. With an improvement in the heart's action, oedema is relieved from both sides; the abnormally high venous blood pressure sinks to its proper level, whilst on the arterial side the pressure rises and the circulation through the kidney is increased. We should not expect these conditions to be closely related to the amount of water taken in heart-disease, and in experiments dealing with the supply of water it is often difficult to separate the results of the restriction of that supply from those of other treatment employed at the same time, such as rest and digitalis. The cause of the whole condition is the failure of the heart, and treatment must be directed primarily to that organ. The restriction of water will do no good of itself. Experiment shows that a moderate limitation aids the heart, but, according to Minkowski, it should be very gradually carried out, 2 1/2 pints being prescribed at first and reduced by a small amount daily, but not below a pint and a half. Thirst is lessened if no salt be allowed.

In the treatment of aneurism, the restriction of fluid is intended to be carried to such a degree that the blood becomes inspissated. We have seen that the water in the blood may be reduced by as much as 10 per cent, but that any greater reduction than this is attended with risk. The sensation of thirst may usually be taken as a guide. Most patients will not suffer extreme deprivation for long, but in the case of a very determined subject the danger of causing an increase in the breakdown of body tissues and of general failure must be kept in mind.