We saw above that copper sulphate loses four-fifths of its water at 100° and the remainder at 240°. This means that there are two definite compounds of water with the anhydrous salt. Washing soda or carbonate of sodium, Na_{2}CO_{3} separates out as crystals, Na_{2}CO_{3},10H_{2}O, containing 62·9 p.c. of water by weight, from its solutions at the ordinary temperature. When a solution of the same salt deposits crystals at a low temperature, about -20°, then these crystals contain 71·8 parts of water per 28·2 parts of anhydrous salt. Further, the crystals are obtained together with ice, and are left behind when it melts. If ordinary soda, with 62·9 p.c. of water, be cautiously melted in its own water of crystallisation, there remains a salt, in a solid state, containing only 14·5 p.c. of water, and a liquid is obtained which contains the solution of a salt which separates out crystals at 34°, which contain 46 p.c. of water and do not effloresce in air. Lastly, if a supersaturated solution of soda be prepared, then at temperatures below 8° it deposits crystals containing 54·3 p.c. of water. Thus as many as five compounds of anhydrous soda with water are known; and they are dissimilar in their properties and crystalline form, and even in their solubility. It is to be observed that the greatest amount of water in the crystals corresponds with a temperature of -20°, and the smallest to the highest temperature. There is apparently no relation between the above quantities of water and the salts, but this is only because in each case the amount of water and anhydrous salt was given in percentages; but if it be calculated for one and the same quantity of anhydrous salt, or of water, a great regularity will be observed in the amounts of the component parts in all these compounds. It appears that for 106 parts of anhydrous salt in the crystals separated out at -20° there are 270 parts of water; in the crystals obtained at 15° there are 180 parts of water; in the crystals obtained from a supersaturated solution 126 parts, in the crystals which separate out at 34°, 90 parts, and the crystals with the smallest amount of water, 18 parts. On comparing these quantities of water it may easily be seen that they are in simple proportion to each other, for they are all divisible by 18, and are in the ratio 15 : 10 : 7 : 5 : 1. Naturally, direct experiment, however carefully it be conducted, is hampered with errors, but taking these unavoidable experimental errors into consideration, it will be seen that for a given quantity of an anhydrous substance there occur, in several of its compounds with water, quantities of water which are in very simple multiple proportion. This is observed in, and is common to, all definite chemical compounds. This rule is called _the law of multiple proportions_. It was discovered by Dalton, and will be evolved in further detail subsequently in this work. For the present we will only state that the law of definite composition