522 Mixter — Thermochemistry and the Periodic Laic. 



heat effects of one gram-atom of oxygen and two gram-atoms 

 of chlorine combining with the different elements. 



Fig. 1 shows the periodic characteristics mentioned by 

 Mendeleeff. It always shows that the quantities of heat fall in 

 a series from the maximum with increasing atomic weights 

 from the metals to the non-metals. The irregularities or varia- 

 tions from linear are due in part to the different molecular 

 conditions of the elements combining, to the crystalline or 

 amorphous form of the compounds resulting, and in part to the 

 number of atoms of oxygen or chlorine in the compounds from 

 which the values are obtained. Thus in series 2, Table I, we 

 have lithium which probably has one atom in the molecule, 

 and boron and carbon with doubtless complex molecules. The 

 value taken for 16 grams of oxygen combining with boron is 1/3 

 the heat effect of 2B + 30. We do not know what B + O is 

 In some cases, but not all, the heat effect of the first atom of 

 oxygen is greater than the second. Hence there is some un- 

 certainty attached to the values derived from the heats of 

 formation of sesqui-, di- or trioxides. 



The values in series 3 are the heats of formation of Na 2 0, 

 MgO, Cl.O, 1/3 of A1 2 3 , 1/2 of Si0 2 and S0 2 , and 1/5 of 

 P 2 6 . The elements with even valence, magnesium, silicon and 

 sulphur, fall nearly on a line as regards their heat equivalents, 

 while aluminum, phosphorus and chlorine, having odd valences, 

 fall on another line. Aluminum and phosphorus oxides have 

 higher heats of formation than indicated by their atomic 

 weights and position in the series. Potassium oxide falls on a 

 line between chlorine of series 3 and calcium of series 4. The 

 values of the oxides in series 4 are for Cr 2 3 (1/3), FeO and 

 MnO, and crystalline Ti0 2 (l/2), MO and CoO. Omitting 

 MnO, the curve is fairly regular from CaO to CuO. The heat 

 equivalents of the chlorides of the series fall quite regularly. 

 The values in series 5 are for amorphous CuO and ZnO, and 

 crystalline As 2 3 (l/3) and Se0 2 (l/2). In series 6 strontium 

 chloride falls on the line between rubidium chloride and silver 

 chloride of series 7. The values for one atom of oxygen com- 

 bining with iodine, csesium, barium and lanthanum are nearly 

 linear. For tellurium it is high and the heat of formation of 

 Te0 2 is greater than that of SeO„. In series 11 the values are 

 relatively low while in series 12 the quantity of heat evolved 

 by 16 grams of oxygen uniting with thorium exceeds that of a 

 combination of oxygen with any other element. 



Thermal data to complete the diagram fig. 1 are lacking for 

 a number of the rare elements and fluorine does not combine 

 with oxygen or chlorine. In some of the series the quantities 

 of heat appear to be nearly linear functions of the atomic 

 weights. It should be observed, however, as the increase in 



