The body is made up of about 60 per cent of water, about 20 per cent of fat, and 40 per cent of dry substances. Water is, therefore, an essential constituent, and must be constantly supplied to replace that which is being lost. The chemical changes taking place in the tissues are reactions between dissociated molecules in a weak saline solution. Hence any considerable drying of the body would arrest all vital processes. When a man is deprived of food he is able to live for a considerable period upon his own tissues; this is not the case with water, total deprivation of which can only be survived for a short time. Atwater and Benedict found that, upon an ordinary diet, an individual not doing muscular work took on an average about 4 pints of water a day in a 49 days' experiment; this includes that taken in the solid food as well as in liquids.

The elimination of water takes place in the urine and faeces, and from the skin and lungs. The tissues retain approximately the same quantity in health, so that with a sufficient supply the amount of water lost is determined by the amount taken in : if enough water is not available to make up the minimal loss thirst arises. Variations in the loss of water are chiefly due to evaporation from the skin, for this is the means by which any excess of heat is dissipated, and the temperature kept constant. The amount of water in the urine, if the diet be the same, will vary inversely as the loss from the skin, and therefore in hot weather the urine is scanty, and in winter abundant. The total loss of water from all sources is rather more than the quantity taken in the food, for the oxidation of the hydrogen of the solid food produces a half pint or more which must be eliminated. The average quantity evaporated from the skin and lungs in a day is about a pint and a half, two-thirds of which is given off from the skin; this total increases very much with active movement and may be over half a pint in an hour, i.e. ten times as much as is lost in that time during rest. The fluid taken must therefore be greater when much work is being performed; thus, in an individual working daily for 66 days the average intake of fluid (including that in the solid food) was found to be six and a half pints. The drinking of an increased quantity of water does not, apart from muscular exertion, lead to a greater evaporation from the skin, but the extra quantity appears in the urine, that is to say, other things being equal, water has a direct diuretic action. The evaporation from the skin is determined during rest by the needs of the regulation of heat and not by the amount of water taken in.

The amount of water vapour in the atmosphere has an important bearing upon the loss of water from the skin and lungs, for it is obvious that in moist air evaporation is hindered or suppressed, and the loss of heat from the skin therefore limited. This might cause a rise of temperature, and must be remembered in prescribing a steam kettle for a febrile patient, especially upon a damp day.

The nature of the diet has a considerable influence upon the amount of water passed out in the urine. If much protein is contained in the food the quantity of water excreted is increased and correspondingly more must be drunk. This is because the nitrogen of the protein is excreted in the urine in the form of urea, uric acid, and other bodies, and it involves less work to the kidney for these bodies to be passed out in a weak solution. On the contrary, a diet which is poor in protein will involve but little excretion of nitrogen, and the amount of urine will be considerably diminished. This may be exemplified by the following case. A patient in bed under the care of the writer, taking a mixed diet containing about 15 grammes of nitrogen, passed out 1,100-1,200 c.c. of urine a day. When placed upon a diet containing practically no protein, the quantity of urine sank on the first day to 530 c.c, and on the second day to 350 c.c. and on the three following days was about 500 c.c. On a vegetarian diet containing but little nitrogenous food less water will therefore be required than upon a diet containing plenty of meat.

We may now consider what experimental evidence we have as to the effect of an increase or diminution of the amount of water supplied in health and in disease.

The drinking of a large quantity of water by an ordinary man, not doing excessive work or living in a tropical climate, leads, as we have seen, to a considerable increase in the quantity of urine. This urine will be of low specific gravity, that is to say, the greater flow of water from the kidney is not accompanied by a corresponding excretion of solids. There is, however, some increase, for the total amount of nitrogen in the urine is raised. The amount of extra nitrogen in one day rarely exceeds 1 gramme, and in a number of careful experiments has not been found to be greater than 2.8 grammes. After a day or two, or at most a few days, the excretion of nitrogen returns to its ordinary level. Neumann found that about 6 grammes of nitrogen could be washed out in 9 days, but on returning to an ordinary quantity of fluid nitrogen was rapidly retained equivalent to that which had been eliminated. It appears, therefore, that although the flushing of the body with a large quantity of water washes out temporarily a certain limited amount of nitrogenous matter, the metabolism of protein on the whole is unaffected. The drinking of much water does not affect the oxidation processes of the body, as shown by the excretion of carbon dioxide. The same is true of waters containing small quantities of salts. Hence the statements made in the advertisements of various watering places that their waters increase metabolism is not supported by exact observation.

In fevers water should be supplied freely, with the object of replacing that in the perspiration or with the hope of promoting that secretion if it is deficient. Matzkevich observed a number of patients with typhoid fever to whom large quantities of water were given for four days at a time. During this period the amount of nitrogen in the urine was slightly increased, but not more than we have seen is the case in normal people. The copious draughts did not appear to have any influence on the temperature, pulse, or respiration. Gruzdiev also found an increase in the nitrogen in the urine when much water was drunk, in some cases amounting to several grammes. His experiments, again, only lasted four or five days at a time, and do not show whether it would be possible to continue to wash out nitrogenous material from the tissues in fever if the treatment were persisted in. Sollmann and Hoffmann's results indicate that the effect would fail after a few days as it does in health. They gave patients with typhoid fever 4 oz. of distilled water every quarter of an hour for a great part of the day, the average for the twenty-four hours being nearly 9 pints, and found that the excretion of nitrogen and of chlorides was unaffected; sweating was favoured, but the treatment was without obvious influence on the progress of the disease. In fevers there is reason to believe that bacterial toxins cause a breakdown of protein, and a good supply of water may assist in the elimination of these products. It is not, however, proved that there is any advantage in supplying such quantities of fluid as were given in these cases amounting sometimes to 10 pints in a day. In phthisis and pneumonia, Gruzdiev found only a slight increase in the nitrogen of the urine when much water was given, similar to that which occurs in health. In all the above experiments, the absorption of food was improved rather than diminished.