30 



SCIENCE. 



[N. S. Vol. XII. No. 



of these and other compounds of this group 

 that we may hope to gain an insight into 

 the constitution of these interesting com- 

 pounds of which so little is known, and 

 further extend our knowledge regarding 

 valence, for it is just at this point that the 

 generally accepted theory of valence begins 

 to break down. 



Before alluding to the ammonia bases, 

 which are so well developed in this group, 

 and would naturally follow these complex 

 salts we have just considered, a brief di- 

 gression may be made to refer to three 

 classes of anomalous compounds, which 

 should not be passed without reference. 

 The first of these is the nitroso compounds. 

 It is only recently that, largely through the 

 efforts of Joly, the nature of the so-called 

 nitro-prussids was discovered, — double cy- 

 anids in which one cyanogen group is re- 

 placed by one nitroso group, NO. Joly 

 then found that the old osmiamic acid of 

 Fritzsche and Struve is also to be consid- 

 ered as a nitroso compound, and that the 

 supposed tetrachlorid of ruthenium is, in 

 reality, a nitroso-chlorid. But while there 

 appear to be no representatives of the 

 nitroso compounds in the cobalt or the 

 nickel groups, several other compounds of 

 iron are known which contain this group, as 

 the potassium iron tetra- and heptanitroso- 

 sulfonates, K,Fe,(NO),S, and KFe,(IvrO),S„ 

 and the iron nitroso-thiocarbonate and thio- 

 antimonate of Low. There seems also to 

 be a nitroso-cyanid of ruthenium, corres- 

 ponding to the nitroprussids, but it has not 

 been isolated. In none of these cases has 

 the interesting question been brought out 

 as to whether the nitroso group remains 

 attached to the metal when in solution, or 

 whether it is electrolytically dissociated 

 and acts the part of an acid radical. 



In some respects yet more remarkable 

 are the compounds formed with carbon 

 monoxid and with phosphorus trichlorid. 

 The best known compound of this class is 



the nickel carbonyl, 1^1(00)4 of Ludwig 

 Mond. The nature of this volatile liquid 

 is yet unknown, but it is by no means 

 unique. Immediately after its discovery it 

 was found that iron formed similar com- 

 pounds, Fe(CO)- and Fe,(CO),. That a 

 volatile compound of iron exists had been 

 very apparent on the lime of the Drummond 

 light, when water gas, compressed in iron 

 cylinders, was used instead of hydrogen, 

 and also in the clogging of gas burner tips 

 with an oxid of iron, especially when a car- 

 buretted water gas is used as an illumi- 

 nant. The volatile iron carbonyl seems to 

 be formed at ordinary temperatures bj' the 

 passage of carbon monoxid through iron 

 pipes. But it is not alone with metals 

 that carbon monoxid combines to form 

 volatile compounds. As early as 1868 

 Schiitzenberger * discovered that platinous 

 chlorid PtCl^ would combine directly with 

 carbon monoxid, with the formation of 

 three distinct compounds, containing re- 

 spectively one, two and three molecules 

 of CO to one of PtCl,. A compound is 

 also known in which one CO group replaces 

 one cyanogen group in potassium ferricya- 

 nid, that is K3Fe(CISr),C0. This reminds 

 us naturally of the nitroso-ferrocyanid, the 

 so-called nitroprussid. Again in 1870 Ca- 

 hours and Gal f discovered a series of com- 

 pounds containing platinous chlorid united 

 with phosphorus trichlorid, and also with 

 some of the organic phosphins. These 

 compounds are not of the nature of double 

 chlorids, for they can be hydrolyzed with 

 the formation of chlorplatophosphorous 

 acid. An analogous class of compounds of 

 iridium has been made by Geisenheimer,J 

 which are also capable of hydrolysis, giv- 

 ing chloriridophosphorous acid. Geisen- 

 heimer has formed similar compounds con- 



^Compf. rend., 66, 666,747 (1868). 



t Comp. rend., 70, 897, 1380 (1870); 71, 208 (1870). 



t Comp. rend., 110, 1004 (1890). 



