y^inc 2, 1S87] 



NATURE 



115 



weigh no tons, fire shot 16^ inches in diameter, weighing 1800 

 lbs., and costing £1^ each. The advance thus shortly chronicled 

 is due to several workers, prominent amongst whom may be 

 mentioned Sir Joseph Whiiworth, Sir William Armstrong, and 

 Sir William Anderson. The production of ordnance of such a 

 character has been due to the introduction of steel, and the 

 possit)ility of producing steel in lar^^e masses by means of the 

 open-hearth steel process, with which the name of Sir William 

 Siemens will always be connected. The quick-firing machine 

 guns are known by the names of their inventors, as the Gardner, 

 Nordenfelt, Maxim, Gatling, and Hotchkiss. 



The Prebident also drew attention to the circumstance of the 

 inventive talent of the country having been taken advantage of 

 here, and ignored in France until after the P'ranco-German 

 war ; now, however, there as here, many works have found it to 

 their profit to establish gun factories which supplement the 

 Government factories to a large extent. 



Two papers were read at the meeting on prime movers, the 

 one by Mr. F. Brown, of Montreal, on " The Construction of 

 Canadian Locomotives," and the other, by Major T. English, 

 R.E., detailing experiments on the distribution of heat in a 

 jstationary steam-engine. The former, as its name denotes, 

 [refers to details of construction ; the latter is illustrated by 

 thirty-five figures, mainly of indicator diagrams, and distribution 

 pf heat diagrams showing in one view the applied and wasted 

 [heat. The series of trials extended altogether over fifiy hours' 

 kvjrking of the engine ; but out of this trial, various results, 

 representing in the aggregate twenty-eight hours' working, were 

 rejected, on account of doubtful measurements at some point or 

 pther. The remaining trials are sixteen in number, in two 

 ;ets— one condensing and one non-condensing — each with 

 xiid without the steam-pipe jacketed, and each with a 

 :ut-ofr at approximately one-quarter, one-eighth, and one- 

 ixteenth of the stroke respectively, 'thus making twelve 

 iifTerent combinations. The conclusions drawn by the author 

 ire : that, in order to obtain the best results for any given 

 ange of temperature, there should be a definite relation between 

 he surface of the steam passages, the diameter of the cylinder, 

 ind the length of stroke ; and that in the design of a steam- 

 mginc the adjustment of these proportions is perhaps the most 

 mportant point to be considered as regards economy. The 

 ;alculated results of varying the length of the stroke of the 

 ngine which was experimented on— while the diameter of the 

 ylin ier, the absolute clearance volume, and the clearance sur- 

 ace exposed, remained unaltered — were tabulated for two 

 liflerent points of cut-off, and show that the same number of 

 :xpansions may give widely different results as regards the ratio 

 »f efficiency and the water consumed per indicated horse-power 

 ler hour ; and also that with the same length of stroke these 

 esults are but slightly affected by doubling the number of 

 jcpansions. 



NOTE ON THE SPECTRUM OF DIDYMIUM> 



T is well known that the absorption spectrum usually ascribed 

 ^ to didymium shows six bands in the blue and violet with 

 pproximate wave lengths 482, 476, 469, 462, 444, 428, accord- 

 ng to Lecoq de Boisbaudran. 



The evidence that we at present possess shows, I think, that 

 hese bands belong to at least five different fractions of 

 lidymium. 



Wehbach {Monaishefte, vi. 477) has shown that the band 428 

 ccurs in the absence of all the others mentioned above in the 

 pectrum of the fraction which he names neodymium. On the 

 ther hand, Crookes (Proc. Roy. Soc, l886, 502, Fig. i) has 

 tiown that all the other bands of neodymium can be obtained 

 1 the absence of the band 428. This band, therefore, belongs 

 D a distinct fraction, and should be obtainable quite by itself. 



Crookes has shown that the band 444 varies in strength 

 idependently ot all others, and is therefore distinct. The same 

 anclusion is arrived at by a sUghtly different argimient. 

 ^'elsbach's praseodymium shows the bands 482, 469, and 444, 

 )gether with a faint band in the orange. Crookes (i7;2i/. , Fig. i) 

 as shown that 482 and 469 can be got in a fraction which does 

 3t show 444. It is possible that the faint orange band of 

 raseodymium belongs to the same fraction as 444, since its 

 resenc .■ or absence would make little difference in the appearance 



' Reprinted irom the Chemical Newi, May 20, iZZ-j. 



of the dark orange band of the ordinary didymium spectrum^ 

 one part of which it forms. 



The band 462 is shown to be distinct by a comparison of 

 Crookes's Figs. I and 2, taking into account that 444 and 428- 

 have been shown to be distinct. 



The two bands 482 and 469 seem always to accompany each 

 other. They occur together in Welsbach's praseodymium and 

 in all the spectra of didymium fractions published by Crookes. 

 They are distinct from 476, since they occur in praseodymium in 

 the total absence of 476. They may belong to the same fraction, 

 as the faint orange band of praseodymium. 



The band 476 does not occur in Welsbach's neodymium 

 spectrum. 



In fact the two bands 476 and 462 seen in the didymium 

 spectrum are not accounted for by Welsbach at all in the spectra 

 of praseo- and neodymium. Since 462 is distinct, 476 must also 

 be distinct. 



I have repeated Welsbach's experiments up to a certain point,, 

 and can confirm his results as regards praseodymium in every 

 respect. There is no indication whatever that the three main 

 bands belong to different fraction-. I have not been able to 

 satisfy myself quite that the faint oran;.;e band of praseodymium 

 really belongs to the same fraction as the others, even supposing 

 that the method of fractionation is not changed. In the didymium 

 spectrum the orange band is much darker than the green, and 

 the difficulty of getting a really concentrated praseodymium 

 solution, which does not show a trace of the green band, is 

 extreme. A small remnant of some other fraction of didymium 

 might ihere'ore cause a faint band in the orange some time after 

 the band in the green had disappeared. Nevertheless, there is 

 no doubt that by Welsbach's method the orange didymium band 

 is split up, for the maximum absorption with didymium is not at 

 the point in the orange where the band of pra eodymium occurs. 



I have not yet obtained the neodymium fraction free from 

 praseodymium, but I have no reason to doubt that WeKbach's 

 observations are correct. A study of the intermediate fractions 

 brings out a point which Welsbach does not refer to. As we 

 pass from the praseodymium end the bands 482 and 469 become 

 fainter, whilst 476 and 462 first appear and then grow stronger, 

 till they become distinctly stronger and much broader than 482- 

 and 469. 



It appears then that the absorption spectrum of didymium is 

 splitting up just as the fluorescent spectrum of yttrium is. I have 

 only discussed a few of the bands, but there is no doubt that the 

 other bands will also in time be separated. Indeed, this separa- 

 tion has already been partially efiected by Crookes for some of 

 the bands in the red. 



Perhaps the most surprising result arrive 1 at by Crookes is 

 that the splitting up of the fluorescent yttrium spectrum is 

 unaccompanied by any change in the spark spectrum. On the 

 othci- hand, Welsfjach states that the spark spectra of praseo- and 

 neodymium are- parts of the didymium spectrum, and that, though i 

 similar in general appearance, they are really quite distinct. 

 There does not appear to be any theoretical reason for this 

 difference between yttrium and didymium, and it is to be hoj^ed 

 that the different fractions of didymium will be got pure enough 

 to show whether the spark spectra can be still further split up. 



Claude M. Thompson. 



University College, Cardiff. 



UNIVERSITY AND EDUCATIONAL 

 INTELLIGENCE. 



Cambridge. — The first election to the Harkness Scholarship 

 for Geology and Palaeontology will be made in June. All 

 B.A.'sof Cambridge not beyond M.A. standing are eligible. 

 The Rev. Osmond Fisher is appointed an elector to the scholar- 

 ship. 



The report of the Council of the Senate on the teaching of: 

 geography is to be voted upon on June 9. 



SCIENTIFIC SERIALS. 



American Journal of Science, May. — On red and purple- 

 chloride, bromide, and iodide of silver ; on heliochromy and the 

 latent photographic image, by M. Caiey Lee. To this paper 

 we have already called attention. It is the first of a series- 

 of important papers, the object of which is to show (i) that 

 chlorine, bromine, and iodine may form compounds whh. 



