Jan. 1G. 1885.] 



KNOWLEDGE ♦ 



49 



tion of silver from a sp.>ilt solution is one in ■which the 

 silver is precipitate^l by the additiou of acid. The prccipi- 

 t.ite havins; been dried, is then melted, and a quantity of 

 silver obtiiiued. Some of the metal, however, remains in 

 the residue, which ought to be removed. It in^y be re- 

 covered by a further <;radual fusiug, or by dissolving and 

 adding a piece or two of zinc and a small quantity of 

 hydrochloric acid. The precipitated silver and chloride of 

 silver is then well washed, dried, and heated with potash 

 in a crucible, wheu the metal septrates as previously indi- 

 cated when dealing with the recovery of silver from strip- 

 ping solutions. This process is, however, not commendable, 

 more particularly to amateurs, as it is accompanied v?ilh 

 the evolution of prussic acid and other almost equally 

 pleasant (or unpleasant) gases. 



Another process is to evaporate the solution to dryness, 

 or nearly so, and mixing the residue with small quantities 

 of nitrate of potash and common salt, fuse it till the silver 

 is reduced to the metallic state and appears as a button at 

 the bottom of the crucible. The molten metal may be 

 obtained in the granulated form by pouring it into a quan- 

 tity of cold water. The metal so obtained is not often pure, 

 and, therefore, not tit to be used in preparing a fresh solu- 

 tion, or in making up into an anode. It has, however, 

 a marketable value. 



The best plan for amateurs, if not for trade purposes, is 

 that suggested by one to whom all electrical students owe a 

 debt of gratitude. The process is, to some extent, a modi- 

 fication of the one first mentioned ; but the gases, which 

 would otherwise prove a fruitful source of trouble and 

 danger, are utilised in the preparation of a new plating 

 solution. These gases are, therefore, not only prevented 

 from acting injuriously, but are a cheapening element in 

 subsequent processes. " Place the solution is a large tlask, 

 titted with a safety-funnel and delivery-tube,* joined by 

 an india-rubber pipe, to a wide glass tube, the end of 

 which dips half an inch into a solution of silver 

 nitrate in another vessel. Xow add sulphuric acid 

 gradually by the safety-funnel, allowing the efferves- 

 cence to subside, and shaking the fla^-k occasionally ; 

 continue adding acid as long as it produces any fresh 

 precipitate. Then, by means of a sand bath, t heat the 

 flask, and keep the solution boiling as long as a precipitate 

 continues to form in the other vessel. The precipitate is 

 pure cyanide of silver, and only needs dissolving in cyanide 

 of potassium to form the fresh solution. The precipitate in 

 the flask is also cyanide of silver, but not pure, though 

 sufficiently so for use in moib ca^es. If it is preferred, it 

 can be reduced by zinc and hydrochloric acid, or dried and 

 fused." The saving above referred to is effected on the con- 

 version of the nitrate of silver into the cyanide by means of 

 the decomposition produced by boiling the bad solution. 



Having, I think, said all that is necessary on the subject 

 of silver-plating, I feel tempted to turn next to electro- 

 gilding, but the practical difficulties are so many and so 

 serious, while the expense is prohibitive to the majority of 

 experimentalists, that I am compelled to defer, for a time 

 at least, an account of the processes necessary or available 

 to secure a good coating of the qusen of metals. There 

 ■are, however, a number of other metals almost if not 

 cjuite as beautiful, and much cheaper to work. Some of 

 them, too, involve questions of great interest, and will well 

 repay one for the time and labour expended upon them. 



* The safety-funnel shonid dip below the level of the solution, 

 but the delivery-tube should only dip fairly into the flask. 



t Consisting of a tin saucer containing a quantity of fine sand, 

 in which the bottom of the flask is embedded. The object is, on 

 applying the flame to the tin, to heat the flask nuiformly, or as 

 nearly so as possible. 



THE WEATIll'Ml FOllECASTH OF 1884. 



By John W. ST.VNironTii. 



AT the beginning of 1884, T submitted to the readers 

 of Knowlkdiik the results of an analysis of the fore- 

 casts issued by the Meteorological Ullico during the previous 

 year. In doing so, it was necessary to show how those 

 results were obtained, and to explain the system by which 

 the forecasts were analysed, but having tested the predic- 

 tions of the past year on the sanu! principle, I shall now 

 only state the results, referring those who care to go further 

 into the subject, to my previous articles.* 



1. — The Wind. — (re) lis Direction. — During last year 301 

 predictions were made of the direction of the wind, as 

 compared with .'JOO in 1883. The following table exhibits 

 the proportional correctness of these forecasts for each 

 month, tlioso under A' beiiis; quite correct, under A'' nearer 

 right than wrong, under A- nearer wrong than right, and 

 those under A' being hopelessly iucorrec^. 



A' .\' A^ A' 



January IG ... 5 ... 2 ... 2 



February 17 ... ... ... 2 



March 2:2 ... ... ... 2 



April 18 ... 2 ... 1 ... 2 



May 19 ... :i ... 2 ... 1 



Juno IG ... 2 ... ;{ ... 4 



July 2.3 ... 1 ... ... 2 



August 14 ... 4 ... ■!■ ... 2 



September 20 ... 2 ... ... 4 



October 19 ... 4 ... ... 4 . 



November 18 ... i- ... 1 ... 1 



December l(i ... 7 ... '<i ... 1 



Totals 218 ... 40 ... IG ... 27 



In other words, rather more than 72 per cent, of the 

 forecasts were correct, 18 i)er cent, were doubtful, and 

 less than 10 per cent, were wrong. In 1883 the figures 

 were C6 per cent., 21 per cent, and 10 per cent, so that 

 we may say there has been an ap[)reciable improvement 

 in the forecasts of the wind's direction. 



(b.) Its Force. — 291 times has the strength of the wind 

 been predicted, and how far these predictions have proved 

 reliable is shown in the following table, where the B's 

 have the same values as the A's in the preceding list : — 



B' B' B- B' 



January 11 ... 3 ... 7 ... 4 



February 5 ... 2 ... 3 ... 14 



March 10 ... 2 ... G ... 6 



April 12 ... 2 ... 5 ... 3 



May 10 ... S ... 5 ... 3 



June 14 ... 7 ... 2 ... 



July 17 ... 4 ... 2 ... 2 



August 15 ... 7 ... 1 ... 1 



September 15 ... 4' ... 2 ... 1 



October 13 ... 5 ... 1 ... 6 



Xovember 15 ... 4 ... 1 ... 4 



December 15 ... 4 ... 2 ... (j 



Totals 152 ... 52 ... 37 ... 50 



To summarise and compare these figures, we may say of 

 the forecasts of wind-force that 52 per cent, were correct, 

 31 per cent, doubtful, and 17 per cent, incorrect, as com- 

 pared with CO per cent., 2") per cent., and 1.5 per cent, ia 

 1883. This is a distinct falling off, for, althougli, as I have 

 previously said, the table by which the.'^e fnrecasts are 

 tested is of too arbitrary a character to do them full justice, 

 the comparative results are in no way affected since both 

 series were tested by the same table. In nearly all those 

 forecasts which are classed under B', B', and B', the error 

 has lain in too great a strength being predicted ; very 



rarely, indeed, is the Meteorological Office uufJcr the mark. 

 ^ -^* 



* Vol. v., pp. 112 and 159. 



