In my opinion (1881) these data should be expressed with exactitude by a straight line, 67·5 + 0·87_t_, which entirely agrees with the results of experiment. According to this the figure expressing the solubility of salt at 0° exactly coincides with the composition of a definite chemical compound--NaNO_{3},7H_{2}O. The experiments made by Ditte showed that all saturated solutions between 0° and -15·7° have such a composition, and that at the latter temperature the solution completely solidifies into one homogeneous whole. Between 0° and -15·7° the solution NaNO_{3},7H_{2}O does not deposit either salt or ice. Thus the solubility of sodium nitrate is expressed by a broken straight line. In recent times (1888) Étard discovered a similar phenomenon in many of the sulphates. Brandes, in 1830, shows a diminution in solubility below 100° for manganese sulphate. The percentage by weight (_i.e._ per 100 parts of the solution, and not of water) of saturation for ferrous sulphate, FeSO_{4}, from -2° to +65° = 13·5 + 0·3784_t_--that is, the solubility of the salt increases. The solubility remains constant from 65° to 98° (according to Brandes the solubility then increases; this divergence of opinion requires proof), and from 98° to 150° it falls as = 104·35 - 0·6685_t_. Hence, at about +156° the solubility should = 0, and this has been confirmed by experiment. I observe, on my part, that Étard's formula gives 38·1 p.c. of salt at 65° and 38·8 p.c. at 92°, and this maximum amount of salt in the solution very nearly corresponds with the composition FeSO_{4},14H_{2}O, which requires 37·6 p.c. From what has been said, it is evident that the data concerning solubility require a new method of investigation, which should have in view the entire scale of solubility--from the formation of completely solidified solutions (cryohydrates, which we shall speak of presently) to the separation of salts from their solutions, if this is accomplished at a higher temperature (for manganese and cadmium sulphates there is an entire separation, according to Étard), or to the formation of a constant solubility (for potassium sulphate the solubility, according to Étard, remains constant from 163° to 220° and equals 24·9 p.c.) (See Chapter XIV., note 50, solubility of CaCl_{2}.)