1. The surface tension of pure serum, heated at 55°C., decreases progressively and regularly until the serum coagulates. A drop of 8.2 dynes was observed in 168 hours. The mean drop was 5.7 dynes in 120 hours and 4 dynes in 96 hours.
2. The initial surface tension of solutions of serum at the concentrations 10–1 up to 10–6 is practically not affected by heat, but the time-drop in 2 hours is modified.
3. Each serum seems to react in its own particular way as far as the time-drop is concerned. However, there is a general tendency for the solution to show an increase of time-drop at the concentrations 10–1, 10–2, 10–3 and 10–5, and a decrease at 10–4, when heated at 55°C. a decrease of the time-drop at the concentrations 10–1, 10–2 and 10–3, and an increase at 10–4 when heated at 100°C.
4. Nevertheless, only the changes observed below 10–3 are constant in sign, within 75 or 80 per cent of the cases. The concentration 10–4 seems to correspond to a state of greater instability. This confirms the hypothesis of the existence of a monomolecular layer at that concentration, which was assumed on the basis of the existence of a maximum drop at 10–4. Should this be true, and provided the principle of Gibbs could be transformed so as to be applicable to mixed solutions of colloids and crystalloids, an idea of the size, or at least of one of the dimensions of the molecules or group molecules composing the serum could probably be obtained.
5. 1 drop of the solutions at 1:10 evaporated on glass and examined under the microscope shows the marked differences brought about by heating at different temperatures.