THE FREYEZING POINT OF TRIPLE ALLOYS. 65 XI.-The Freezing Point of Triple Alloys. By C. T. HEYCOCK and F. H. NEVILLE. Gold a.nd Cadmium. IN the Chemical Society’s Journal (Trans., 1891, 59, 936), we de- scribed the peculiar fluctuations in the freezing point of mixtures of gold and cadmium dissolved in tin, and we showed that many of the6 6 EEYGOCK AND NEVXLLE: effects could be explained by the assumption that the two metals combine t o form a compound of the formula AuCd. This substance we afterwards obtained in a pure state (Trans., 1892, 61, 914), and we have since found that the beliaviour of these two metals to each other is substantially the same, whether they are dissolved in tin, bismnth, thallium, or lead. From the curves and tables now given, i t will be seen that if we make a solution oE either gold or cadmium in any of the above-men- tioned solrents, and then add doses of the complementary metal, cadmium or gold respectively, the same phenomena repeat themselves.That is to say, the addition of the third metal sometimes causes a small further fall in the freezing point, but, as more of this metal is added, the freezing point begins to rise, and reaches a maximum when an equal number of atomic weights of gold and cadmium have been added. This maximum freezing point is the same whether gold or cadmium is added first, and is independent of the concentration,* so long as we obser-re the condition of maintaining an equal number of atoms of gold and cadmium in the mixture. After the maximum freezing point has been reached, a further addition of either metal causes a fall, but, as we see in the curve with thallium as solvent, the maximum freezing point can be recovered by adding a sufficient aillount of the complementary metal.Whatever be the explanation OE these phenomena, we here see that the reaction between the two metals is remarkably independent of the nature of the solvent. As the new solvents give results so similar to those recorded in o u r previous paper, where tin was the solvent, we feel no doubt that the cause is the same, namely, the formation and partial precipitation of the compound AuCd, and that this sn5stance exists in solution in equilibrium with free gold and cadmium, uhat- ever be the solaent. It is seldom possible ta study the formation of the same compound in so many solvents, and the uniformity of the results appears strongly to bear out the view that solution in itself is a physicai, rather than a chemical, action. Silver and Cadiniurn.The behariour of silver and cadmium, when dissolved together, is recorded in the Tables V-VIII (pp. 71-73) and Figs. 5-6t (p. 68). It will be seen that these two metals produce in the freezing point of the solrent changes very similar to those caused bj- gold and cadmium. In each case, the addition of the t.hird metal causes a rise, which reaches a maximum value and then declines as more of this met'nl is added. * The maximum is not reached in very dilute solutions. t See " Explanation of the Charts," p. 74.TEE FREEZING POINT OF TRIPLE ALLOYS. 67 Although we have not much experimental proof, yet we have no doubt that, as with gold-cadmium, t h i s maximum freezing point will turn out t o be constant for each solvent, independent of the absolute amounts of silver and cadmium present, but reached when the ratio of these two attains a certain value.In tin or lead, this maximum is reached when the ratio of the two metals is exactly 2Ag: Cd. In thallium, the ratio is nearly the same, and we think that experimental errors would account for the discrepancy. I n bismuth, the ratio at the highest point is not far from 4Ag : Cd. The experimental error is probably small in the bismuth series, but, without further trial, we do not feel certain that this ratio would be maintained a t other concentrations. TABLE L-Cadmiitm aid Gold in 400 grams of Lead.Total weight of gold present. 0 Y Y 7 9 7 9 9 , 9 , Y9 0 G85 2 -083 3 '037 5 -089 5 -717 7 -021 9 -353 14 -04.1 18 -70% 21.631 25 -91'4 Total weight of cadmium preaent. Atom of gold per 100 of lead. 0 9, 97 Y ? I 7 99 9 9 0 .$58 0 *546 0.797 1.335 1-500 1 .M2 2 -453 3 -683 4 -906 5 '674 6 -797 Atoms of cadmium per 100 of lead. Freezing point of solution. 327 -60' 327 -24 326 -85 32a -80 323 '86 319 *62 315 '48 308 7ti 310 -04' 312.W 313 -441 314-60 315 -10 316 -06 317 -25 318 -35 318.55 316 *9L 322 -40 * As soon as the first portione of gold were added, a grating of the stirrer was noticed, and other proofs of a precipitate which rapidly increased. Much preci- pitate was found in the crucible at the termination of the experiment. If we compare together the curves for gold-cadmium, and silver- cadmium, in bismuth, and then compare them with all the others, we shall see that these two stand somewhat apart.We do not yet know whether this is due to the specific action of bismuth as a solvent, or to a cause of experimeiital error, which we hace, hitherto, failed to detect. Curve 5 of cadmium added to a saturated solution of silver in tin is singularly like the saturated gold-cadmium curves of onr prerious p p e r : the initial fall caused by the cadmium, followed by what may68 HEYCOCK AND NEVILLE: be a short f l d , is to be seen, also, in the gold-cadmium curves. We haye traced experimentally a 3-atom silver and cadmium cnrve in tin, where, j u s t as in the 3-atom gold-cadmium curves, there is no raaxi- mum freezing point.This curve is not given in the present paper. TABLE 11.-Gold and Cadmium i n 400 grams of Lead. Total weight of gold present. 0 0 '822 0.632 1 -i42 5 -417 9 -451 12 -563 15 231 9 ) 7 9 Y ? Y 7 Y 9 31 Y1 I1 Y Y 9 9 ?? Y 1 Y l 2 9 7) 9 ) Total weight of cadmium present. 0 Y9 Y Y ?f YY Y Y 99 7 9 0 -159 0 '441 0 -730 1 '083 1 -756 3 -345 5 -418 6 -630 7 -6i5 8 -666 11 '841 13.364 16.108 It) *801 21 -463 9 -78% Atoms of gold per 100 of lead. 0 0'085 0 *l66 0 '457 1 -521 2 *479 3.292 4 -0oo > > J Y Y ) Y * 1, I > 9 7 Y Y 9 , 7 ? 9 Y Y 7 9 Y , 1% 7 Y Atoms of cadmium per 100 of lead. 0 9 9 Y Y Y7 9 , Y Y $ 1 1, 0 -073 0 -204 0 -337 0 -5uo 0 -810 1 -643 2 -500 3 -059 3 541 4 -0oo 4 -514 5 -186 6 -165 I ado 8 -675 9 -903 P ..- Freezing point of solution.327 '64' 327 -10 326 '54 324 -TO 318 -45 311 -61 306 -44 301 -84 301 -77 301 -57 302 -53 303 -69 305 -51 309 -62 314 -13 316 -4Q 317 -72 318 '54 318 '78 318 -47 317 -08 315 -46 314 -90 314 -20 A mass of crystalline precipitate was found a t the termination of the experiment. I n Series 6 and 8, silver was again added from the point X, and produced, as we expected a rise. All these curTes contain details that we cannot interpret, but we. are disposed to explain the general character of 5, 6, and 7 by assuming that a sparingly soluble compound, Ag,Cd,* is formed, which call only exist in the presence of free silver and cadmium. Much of what was said in our paper on triple alloys of gold, cadmiurn, and tin (Zoc. cit.) would then be applicable. The only point we will here consider is the position of the summit C.Because the summit C is reached at a concentration of 2Ag: Cd, we are hardly entitled to concliide at once that the compound AaCd * Our attempts to isolate this compound in the same way as SuCd hare so far failed.68 HEYCOCK AND NEVILLE: be a short f l d , is to be seen, also, in the gold-cadmium curves. We haye traced experimentally a 3-atom silver and cadmium cnrve in tin, where, j u s t as in the 3-atom gold-cadmium curves, there is no raaxi- mum freezing point. This curve is not given in the present paper. TABLE 11.-Gold and Cadmium i n 400 grams of Lead. Total weight of gold present. 0 0 '822 0.632 1 -i42 5 -417 9 -451 12 -563 15 231 9 ) 7 9 Y ? Y 7 Y 9 31 Y1 I1 Y Y 9 9 ?? Y 1 Y l 2 9 7) 9 ) Total weight of cadmium present.0 Y9 Y Y ?f YY Y Y 99 7 9 0 -159 0 '441 0 -730 1 '083 1 -756 3 -345 5 -418 6 -630 7 -6i5 8 -666 11 '841 13.364 16.108 It) *801 21 -463 9 -78% Atoms of gold per 100 of lead. 0 0'085 0 *l66 0 '457 1 -521 2 *479 3.292 4 -0oo > > J Y Y ) Y * 1, I > 9 7 Y Y 9 , 7 ? 9 Y Y 7 9 Y , 1% 7 Y Atoms of cadmium per 100 of lead. 0 9 9 Y Y Y7 9 , Y Y $ 1 1, 0 -073 0 -204 0 -337 0 -5uo 0 -810 1 -643 2 -500 3 -059 3 541 4 -0oo 4 -514 5 -186 6 -165 I ado 8 -675 9 -903 P ..- Freezing point of solution. 327 '64' 327 -10 326 '54 324 -TO 318 -45 311 -61 306 -44 301 -84 301 -77 301 -57 302 -53 303 -69 305 -51 309 -62 314 -13 316 -4Q 317 -72 318 '54 318 '78 318 -47 317 -08 315 -46 314 -90 314 -20 A mass of crystalline precipitate was found a t the termination of the experiment.I n Series 6 and 8, silver was again added from the point X, and produced, as we expected a rise. All these curTes contain details that we cannot interpret, but we. are disposed to explain the general character of 5, 6, and 7 by assuming that a sparingly soluble compound, Ag,Cd,* is formed, which call only exist in the presence of free silver and cadmium. Much of what was said in our paper on triple alloys of gold, cadmiurn, and tin (Zoc. cit.) would then be applicable. The only point we will here consider is the position of the summit C. Because the summit C is reached at a concentration of 2Ag: Cd, we are hardly entitled to concliide at once that the compound AaCd * Our attempts to isolate this compound in the same way as SuCd hare so far failed.Jozr rm.C h m . SOC. fib. I894 HEYCOCK & NEVILLE. SILVER AND CADMIUM.THE FREEZING POINT OF TRIPLE ALLOYS. 69 Total weight of bismuth present. 350 365 380 395 410 4 25 Y Y ?I 7 9 Y ? Y ? Y Y 7 7 Y Y Y * 1 7 1 7 Y ? > I JY Y 7 I t l Y Y f Y Y 11 Y 7 Y l Y Y Y Y t t > 9 7 9 9, Y Y Y Y Y Y Y Y Y l ? Y ?I * A has been formed. mately in the following manner. But we can arrive at this conclusion more legiti- ~ Total weight Total weight Of k;,”i”,$ Freezing point of of gold of cadmium present. present. perlOOof per of bismuth. bismuth. -- ----- -- 0 0 0 0 267 a 5 4 O 2 *252 1 9 0 -651 l Y 266 ‘24 4 -826 Y l 1 ’340 9 ) 264 -92 9 -621 9 9 2 ‘569 Y Y 262 ’52 14 -823 l l 3 ’813 I t 260 -03 20 *152 5 ‘001 Y Y 257-88 Y 9 0 ‘227 Y Y 0 ‘10 258 ’39 Y , 0 -4.67 l t 0 ‘2C4 258 -96 Yr 0 -936 19 0 ‘408 260 -00 l t 1 -884 Y l 0.822 261 -67 7 , 2 -917 I ? 1 ‘273 263 -11 Y , 3 -903 l ! 1.704 263 -90 Y 7 4 -913 9 , 2 -145 263 -99 Y, 5 -207 7 7 2 -274 263 ‘98 t l 5 -720 Y Y 2 -498 263 ‘99+ 6 -630 Y l 2 ‘895 264 -16 7 -186 Y l 3 -137 264 *36 97 7 -677 Y Y 3 ‘352 264 -81 > I 8 *079 Y Y 3 -528 265 -01 Y l 8.396 73 3 -666 265 ‘15 Yl 8 -768 7 7 3 -828 265 -39 -360 Y l 4 -087 265 -81 979 -577 Y l 4 -617 266 -30 11 -238 Y Y 11 -801 Y Y I 5.152 266 -29 Y! 12 *332 l .5 ‘385 266 ‘00 t l 13 -199 19 5 -763 265 ‘64 Yt 14 -204 l Y 6 -201 265-10 Y 7 15 *2W l Y 6 -640 264 -30 $1 16 -376 I9 7 -150 263 ‘43 Y Y 17 -628 I 7 -696 262 ‘46 >l 19 -044 Y > 8 -314 261 ‘41 IY 20 -639 Y I 9 -011 260 -35 > Y 22.274 t? 9‘724 259 *Oo I , 23.594 9 ) 10 ‘30 258 ‘05 Yt 25 -622 Y Y 11 -19 256 -66: Y Y 27 -760 Y - 12 -12 255 -20 37 29 -590 Y Y 12 -92 254 -17 l Y 33 -954 Y Y 14 *825 252 -60 Y Y 0 -?la 7 - 0 ‘05 258 -20 l Y Y * Y Y Y ) Y Y 4 ‘357 266 -15 266 ‘36 266 ‘49t Y l Y l Y precipitate here formed.TABLE 111.-Gold and Cadmiwn in Bismuth.70 HEYCOCK AND NEVILLE: TABLE TV.-Citdmiunt and Gold in 6G-34 grams of Thallium. Gei ssler Thermometer. Total weight of gold present. 0 9 ) Y 9 Y 9 Y Y Y Y 0 -%8 0 *408 0 -728 1 '051 1 *409 1 '854 2 -234 2 -514 3 *013 3 -627 7 ) 7Y Y Y Y Y Y , Total weight of cadmium present. 0 0 -104 0 -201 0-517 0 -998 1.328 3 ) Y 7 7Y Y Y 7) Y Y Y Y Y Y Y Y > Y 9 , 1 -999 2 -046 2 -129 2 -229 2 -511 Atoms of gold per 100 of thallium. 0 7 ) Y 9 Y Y Y 7 Y Y 0 "700 1 -250 1'806 2 '419 3'183 3 '835 4 -316 5 -172 6 '226 Y Y Y Y Y 9 9 Y Y Atoms of cadmium per 100 of thallium.0 0 -312 0 -604 1.5657 3 -006 4 -000 6 &3 6 '166 6 '416 6 T I 4 7.568 Freezing point of solution, 301 %7" 300.18 298 '90 294 -86 288 -10 284 '52* 285 '42* 284 .97 287 '28 289 -11 291 -07 293 - 16 294 -68 294 '29 291 -17 285 -69 294 '40 299.53 294 -86 294 -66 293 37 294 '77 * Between these readings, there was an interval of two days, and the block was st,rongly heated during a part of the time. The discrepancy ip probably due t o two causes, a rise in the zero of the thermometer and, perhaps, mme oxidation of the cadmium. Suppose that the compoiind Az,fiCd, is being formed, but that i t can only exist i n the presence of free silver and cadmium.Then, if we have present in solution x atoms of free silver, y at.oms of free cadmium, and z molecules of the compound Ag,,Cd,, the theories of chemical equilibrium accepted at present indicate that 2, y, and z should satisfy the equation xmyn = KZ, where K is a, constant so long as the temperature does not vary much. Let UTJ consider only that part of the curve where the liquid is saturated with the compound ; then, for this part of the curve, z is constant By adding silver or cadmium, we can vary x and y, but only subject to the condition that r%f is constant. Now, a maximum freezing poiiit that is a summit C on the curve corresponds to a minimum number of molecules in solution. That is to say, at C we must have x + y + z a minimurn, or as z is con-THE FREEZE@ POIXT OF TRIPLE ALWYS. 71 stant, we must have a: + y IL minimum snbjeot to the condition x m y * constant. This leads at once to the result that at the summit C the relation between x and 3 is my = ?Ax.Now, if a and c are the numbers of atoms OE silver and cadmium that have been added to the mixture, and P the number of molecules o€ silver-cadmium that has been formed, then a = 7nP t x, c = 71P + y, and maltiplying by n and m and subtracting, we see that n.a = mc a : c = m : n. TABLE V.-f&lver and Cadmium in 300 grpm of Tin. Total weight of eilver present; 0 5 -478 10 -956 13 -695 Y2 Y Y? 2 2 Y 2 ) 7 2 Y Y 22 7 7 7 9 9 1 7 Y 2 > > ;> Y ? Y Y 3 7 7 7 7 7 Y Y Y 3 2 Y 7 ? f 7 2 Total weight of cadmium present. 0 Y Y 39 0 -;85 0 '83.4 1 -424 1 -993 2 -.562 3 -132 3 -701 4 '270 4 -a40 5 -970 7 -118 8 '256 9 -395 10 -534 11-673 13.950 16 -223 19 -075 21 -922 24 -769 25 -908 26 -4'77 27.901 29 -824 30 -748 32 -595 36 -442 Atoms of silver per 100 of bin.- 0 2 4 5 9 1 > Y Y Y > I Y2 Y 7 2 7 7 I Y 2> 7 7 7 7 7, 2 ) 7 9 YY 7 3 ,, 3 9 > 7 x2 27 > 2 9 2 9 9 2 > Y Atoms of cadmium per 100 of tin. 0 J 9 3 ) dfl 0 -3 0 -5 0 -7 0 -9 1 -1 1 -3 1 -5 1 -7 2 -1 2 -5 2 -9 3 -3 3 -7 4 -1 4.9 5 -7 6 -7 7 9 8 '7 9 -1 9 '3 9 -8 10 -8 10 -8 11 B 32 B Freezing point of solution. 23 I -49" 225 '95 221 '12 221 *04 220 -84 220 -68 220 -33 219 -57 219 '56 219 -64 219 -73 219 -86 219 '84 21 9 '93 219 '98 2L9-97 219.93 219 -86 218 77 219 -51 219 '16 218.88 21 8 2 5 218 -00 W 8 '22 218 -12 217 8 9 217 -62 217 -35 216 $7 216 Q272 HEYCOCE AND NEVILLE: Therefore the snmmit will always be reached when the quantities of the two metals added are iu the ratio given by the formula of the compound formed.In the case of silver and cadmium, me find that in t,he solvents tin, thallium, and lead the body formed is AglCd, whilst in bismuth there is, perhaps, the body AgdCd. In the main, me see that the reaction between silver and cadmium is, like that of gold and cadmium, independent of the nature of the solvent. TABLE VI.-Silz.er in 61.18 gmms of Thallium, and then Cadmium added.* Total weight of silver present. --- 0 1 -326 7 7 Y! 77 77 7 1 1 ) Y I 7 7 1 sib3 1 598 1 -681 1 939 1 -821 1 -889 1 -967 2 -060 2 -165 4 -165 Total weight of cadmium present. 0 0 -272 0 424 0 -515 0 569 0 '602 0 -662 0 5'02 0 -969 7 9 17 I Y Y1 7 ) 7 ? 7 7 3 Y l 1 ) Atoms of silver per 100 of thallium.0 4.1 7 9 77 9 9 Y9 Y, 7 7 7 1 9 9 4 &5 4 '9.10 5 '196 5 -375 5 '628 5 -838 6 '080 6 '367 6 '692 12 '875 Atoms of cadmium per LOO of thallium. 0 0 k 7 0 -810 1-263 1 -534 1 '696 1'793 1 -971 2 -09 2.886 Y Y 7 9 9 1 9 1 17 7 7 11 1 9 11 91 Freezing point of solution. 301 -93' 288 *lot 290.19 292 -04 293 -10 293 -2p 293 -27 293 '28 293 -18 293 -08 291 -86 292 -29 292 -4.9 292 *81 292 *76 292 '86 292 '97 292 -99 293 -04, 293 -06 293 '09 * This series, hke all the thallium experiments, was canied out on too smaU a scale for accurate results. f More than saturated. TABLE VII.-Ssa'lver and Cadmium in 400 grams of Lead. Total weight of silver prewnt. - 0 0 -167 0 -687 1 -d91 3 -907 7 -149 Total weight of cadmium present. Atoms of silTer per 100 of lead.Atoms oh cadmium per 100 of lead. 0 0 '08 0 '329 0 '906 1 '872 3 ' a 5 0 1 ) 9 ) 7 9 1 8 7J Freezing point of solution. ~- 327 *so" 327 -18 325 -62 322 -18 316 TO 308 *64THE PREEZlSG POINT OF' TRIPLE ALLOTS. 73 ~ Total weight of silver present. Total weight of cadmium present. 0 - 0 -a62 1 -404 1 -934 2 -4% 3 -502 4 '576 6 -275 f -771 I -951. 8 -492 9 -470 11 -603 14 -688 18 *634 20 "772 24 - 9 i i Atonis of silver per 103 of lead. .5 a 2 (i -002 --- 7 7 > 7 7 Y 7 P 7 1 7 7 7, 7 . 7 - > 7 7 &onis o€ cadmium per 100 OF lead. 0 o .%8 0 -648 0 -892 1.14'7 1 -61-5 2 -112 2 9395 3 -12.5 3 -486 3 *918 4 -370 5 -354 k; -767 8 -395 9 *536 11 '92.5 F r e I? z i n ,D point of solutio 11.TABLE VIII.-Silcer and Cudiiziunz in L'iism u f l ~ Thermometer, Hicks' 12 A. Total weiglit of bismuth present. Total weight of silwr present. 0 3 '838 7 -676 10 -635 -- 9 1 7 J l 97 7 ) J 9 7 9 7 7 '7 7 > r 77 7 Y Y 7 7 77 7 7 14 .k3l Tot a1 weiglit OE cadmium present. 0 --- Y l 7. 0 *553 1 -105 1 -668 2 -210 9 .T63 3 '315 3 *86i 4 -320 4 -9i;j 5 -5:s 6.078 7.193 8 -230 9 -332 3 0 -497 11 -6 13 13 -812 16 -023 20 -443 24 363 9 . 303 -Mi" 303'15 306 -62 308 67 310.32 311 '66 313.55 31 a -99 31.5 -75 315 *7!J 315 '70 315'39 314 98 313 48 311.99 311.13 311 '33 310 ,311 &oms of cadmium bismuth. pel- 100 of Freezing point oi solution. VOL. LXV. G74 HBTCOCK AND XEVILLE: EXPLANATION OF THE CHARTS. The curves gire the freezing points of soliltions of gold nit11 cadmiu~ii.ancl of silver with cadmium in different solvents. Each black dot records the concentration and freezing point of a yarticulal. mixture. In eT-ery case, 100 atomic weights of solvent arc assumed to be present. The horizontal distance of a point measured from the left records the total number of atomic weights of foreign metals, other than the solrent present, each square corresponding to one atomic weight. Tho vertical distance of a point from the ler-el of the first point on the curve records the fall or rise in the freezing point, each square being one 1" centigrade. Along the t h n iine, cadmium is being added : along the thick line, gold or silver- I n Curve 4, we start with a mixture of 4 atoms of cadmium and 100 atoms of thallium; in other cases the curve begins witn the pure solrent.The point A marks where we begin to add the third metal. C is the summit of the curve. The tables are numbered to correspond with the curves. I n the tabIes all weights ore in grams, and temperatures in degrees centigrade. The point X mnrIis where we begin again ta add the second metal. -1lunzinizcnt and Gold dissolced iu Tin. I n striking contrast to the foregoing curves, which owe their character to the partial dissociation of the compound fonned, we non- give curves showing the freezing points of alloys made by dissolving aluminium in tin and then adding gold. Fig. 9 commences by the addition of aluminium to tin. We see that the aluminium produces an atomic fall of 1*4", which is half the normal fall produced by other metals in t'his solvent.This, as we suggested in oiir previous paper (Trans., 1890, 57, 392), may be explained by assuming that the met.al is present, in diatomic molecules as AI,. The addition of aluminium was elided a t A, almost exactly at the point of saturation of the tin. Gold was then added, as recorded by the thick line. It at once produced R rise, the maximum freezing point C, which was identical with that of pure tiiz, being reached with 2 atoms of aluminium and 0.9 atom of gold, or, in other words with a concentration of AI,Au, 9. A further addition of gold produced a fall, which, reckoned from the highest point, is exactly that of gold in pure tin. In Fig. 10, a total quantity of 4 atoms of aluminium mas added up to the point A, saturation occurring at 2 atoms.The first por- tions of gold produced no effect on the freezing point of the mixture until 0.95 atom had been added, to the point B, but on adding more there was a rapid rise ; the highest point, again identical with the freezing point of pure tin, being reached with 1.9 atoms of gold. When we add more gold from this point C, we seem to be starting afresh with a solution of pure tin, and we get the atomic fall 2-86", which is identical with our old numbers. The obvious explanation of these resnlts is, that the gold and0 c3 x! 0 0 0 > W Ir - TEE FREEZIKG POINT OF TRIPLE ALLOTS. 1 .3 Atoms of aluminium per 100 of tin. --- 0 0 . 5 1 -5 0 -2 ? > > 7 7 Y 7 2 7 '3 7 7 aluminium form a stable insoluble compound of the formula Anal,, the gold completely removing the aluminium from solution.There is no curved part to the louiis of the freezing points, inasmuch as free gold and aluminium cannot exist together in solution. That rather less gold is needed to reach C in both figures, or B in Fig. 10, than the amount corresponding to the compound AuAlz, may fairly be put down to the fact that all the loss by oxidation in making the alloys, and probably all the original impurity, falls on the aluminium. The fact that in Fig. 10 the gold from A. to B produced no effect agrees very well with our view that the aluminium flat up to A. is due to saturation of the tin with aluminium. If this be so, then, although the first quantities of gold up to B precipitate aluminium, yet some of the excess mill come into solution and the freezing point will not alter.Bat by the time 1 atom of gold has been added, the excess of aluminium rill have been used up to form the body Al,Sn, and further addition of gold must decrease the aluminium in solution, and so produce a rise in the freezing point. The precipitate AuAl, is, no doubt, the purple alloy discovered and examined by Professor Roberts-Austen. He finds for his alloy the formula A12A~0.95. Our values deduced from the freezing point curves are :- At C, Fig. 9 . . ............... AI,An,.9 ,, ,, 10 ................. L4.12Au,,.L,5 A t B, ,, 10 ................. A12duo.95 TABLE TX.-Alumiwium ctrzd Gold in Tin. Thermometer, Hicks' No. 12. Total weight of tin present. -- 300 3 10 320 330 > Y J > f , 7 I ? 7 f > > Totsl weight of aluminium present -- 0 0-335 1.101 1 -6u5 > 9 7 7 1 . > 2 ) 7 7 ) Total weight present. of gold Atoms of per 100 of tin. goia Freezing point of solution. -- 231 -55" 230 -90 229 4 8 228 67 228 -97 229 -29 229 -90 230 -55 231 - 20 231 -55 231 -34 230 *76 229 -6176 HOOKER AXD CARNELL : THE COXDENSATIOK Tables IX and X give the experimental numbers from which the Total weight of tin present. 250 275 280 300 J 7 9 9 97 Y 9 f $ 9 7 9 > ) Y Y 9 9 alumininm curves were drawn. Total weiglit of aluminium present. 0 1 261 1 -605 2 .is2 7 7 > 1 9 9 Y Y 7 7 Y 9 Y ? 9 2 7 7 9 TABLE X.-Aluminium awd Gold in Tiu. Thermometer, Hicks' No. 11. Total weight of present. gold -- 0 Y Y Y 2 *go6 3 -750 4 260 4'761 5.263 7 -769 9 -022 9 523 10 '024 12 -530 Atoms of aluminium per 100 of tin. -- 0 9 -0 2 ' 5 4 -0 Y ) 9 7 Y i 7 7 7 7 Y ? ? Y 3 , I Atoms of per 3 0 0 of tin. gold 0 3 7 >7 dY5 0 -75 0 -85 0 -95 1 -05 1-55 1 -8 1-9 2 -0 2 -5 Freezing point of solution. -- 231 -65' 228 '72 228 -69 228 -70 228 *'73 228 -7 1 228 '73 228 -74 229 - 07 230 5 7 231 -36 231 -59 231 -29 229 -86 Sidney College, Cambridge.