April, 1906.] 



KNOWLEDGE & SCIENTIFIC NEWS. 



409 



evidently hc.sn shot." We join with him in his expressions 

 of regret, and agree further that " It is possible that these 

 two were a pair and might have remained to breed." 



King Eider in Orkney. 



Mr. 11. W. Robin-on nrunls | /•';,/,/, MaiL-h 17) the fact 

 that an adult female of the king cider ISiDiuitci ia spectabilix) 

 was shot off the west of Graemsay, Orkney, on February 21. 



PHYSICAL^ 



I?V Al-FKEI) W. PuKTEK, 13. Sc. 



Resolving Power with Wide Slits. 



In tiie ordinary use of a diffraction grating or prism, it is 

 found to be necessary to use a very narrow slit if two very 

 close spectral lines are to be seen separate from one another ; 

 because each line as seen is simply an image of the slit 

 cmplo\ed. The grating is used with the light falling 

 nearly normal to it ; while a prism is usually set in the 

 position of minimum deviation (i.e., the position for which 

 the incident and emergent rays make equal angles with the 

 normals to incident and emergent face respectively). 



Mr. Morris-Airey, in a paper in the Pliilosophical 

 Maqazine for March, shows that quite wide slits, or even 

 no slit at all, can be employed without any sacrifice of re- 

 .'■olving power, if the gratmg or prism be placed so that the 

 light strikes it at nearly grazing incidence. Schuster had 

 previously shown that the purity of a spectrum (which is a 

 measure of the smallness of overlapping of images arising 

 from different wave-lengths), is given by 



_l (5?v) _ 2X+ Am 



Purity ~ IT ~ XK 



where (SX) is the smallest difference in wave-length vi'hich 

 will give distinctly separable images. A is the angle sub- 

 tended at the slit by the iens of the collimator ; w is width of 

 the slit ; and R is the resolving power corresponding to an 

 infinitely narrow slit. 



Now, Morris-.Virey points out that A should not refer 

 really to the angle subtended by the collimator lens, because 

 only part of the complete beam of light passing through 

 this lens may fall upon the grating. The etfcctivc portion 

 of the lens is that which corresponds to the width of beam 

 received by the grating or prism, and this is very small when 

 Ihe grating is placed in a very oblique position. But as A 

 becomes small, w may be proportionately increased without 

 any loss ol purity ; and, in fact, slits half a centimetre wide 

 may still allow separation of the two yellow lines of 

 sodium with a 1300 line grating. Owing to the magnifi- 

 cent casts of Rowland's grating, made by Mr. Thorp, of 

 Manchester, which are on the market, good gratings are 

 much more conmion possessions than once they were. All 

 who own any such grating can obtain much more power 

 out of them than has commonly been thought. For it is 

 clear that if the slit be not widened when the grating is 

 placed oDuquely, (jrcatcr purity will be obtained. A limit to 

 what can be gained in this way comes in owing to the 

 great loss of light for great angles of incidence. But it 

 would .seem that it is only transmission spectra which will 

 be affected in this way. For reflection spectra, the increased 

 reflecting power is an additional advantage which will 

 enable a narrower slit to be employed. .Anyone who wishes 

 to show spectra on a screen to an audience will find it an 

 immense advantage to make use of the above f.icts. The 

 present writer has jusj; set up a Rowland reflection grating 

 with a slit 2 cms. wide, and with the grating very oblique. 

 The lines of the " Swan spectrum " stand out as quite 

 narrow bands. The spectra obtained are so brilliant that 

 10 or 12 can easily be obt.iined. The increasing elTect of 

 overlapping as the oider of spectrum increa.ses is very in- 

 structive. .\fcer the first few, the remainiler appear as a 

 succession of reds and greens, very similar in general 

 elTect to the tints of Newton's rings of high order. Of 

 course, in the case of metal gratings not much advantage is 



to be expected because at normal incidence the reflecting 

 power is at least 90 per cent., but in the case of glass ones 

 very consider.'ible gain may be obtained in exhibiting the 

 reflection spectra. 



Secondary Spectra. 



VA'hile discussing a dilTraction grating it may be inter- 

 esting to point out that, besides the spectra usually seen, 

 theory shows that there are a large number of very much 

 feebler ones between each pair of principal spectra. These 

 are the secondary spectra. They must be distinguished 

 from " ghosts," which arise simply from irregularity in the 

 ruling of the grating. Their number between two princi- 

 pal maxima is very great — practically the same as the total 

 number of lines in the grating; thus there will be as many 

 as 80,000 on a 14,000 line grating six inches wide. In 

 practice, they are invisible, and it is usuallv asserted that 

 this is to be expected when the number of lines is large. 

 It is at any rate certain that when the number is small they 

 are very visible. They may be obtained as follows : Cut 

 four parallel slits in a piece of thin brass or aluminium sheet, 

 each slit a millimetre wide and with an interval of i mm. 

 between each. Now by means of a long focus lens focus a 

 narrow and very bright source of light (suitably obtained by 

 putting a very narrow slit in front of the condenser of a 

 lantern, with the lens removed) upon the focal plane of an 

 observing eye-piece, and interpose the four opening grating 

 just in front of the lens so that the light can only pass 

 through the four openings. Take care that the sides of 

 the openings are parallel to the slit. On looking through 

 the eye-piece, a diffraction pattern will be seen consisting, if 

 a piece of red glass be interposed, of a central principal, 

 maximum (the direct image), followed by two feeble second- 

 arv maxima; then a second principal maximum (the " first 

 order spectrum,") ; then four feeble secondary ; then another 

 principal one (the " third order spectrum " — the second 

 order is absent in this case). Everything is repeated in the 

 same way on the other side of the central image. 



Now it is easily calculated that the relative intensity of 

 any principal maximum and the secondar)- maximum next 

 to it is not very much greater when there are So, 000 lines 

 than what it is when there are only a few. The reason that 

 thev are not visible in practice is that even the principal 

 ones are always comparatively feeble, so that their feeble 

 satellites are bejond detection. If, however, the " direct 

 image " formed by any grating be examined, the satellites 

 are quite conspicuous. With a 20,000 line Rowland grat- 

 ing about six can easily be .seen on each side of the direct 

 image — those of the six being more and more evanescent 

 as they aie more removed from the centre. This simple 

 observation is sufficient to show that there is considerable 

 .agreement between theory and practice even in minor de- 

 tails, although an examination restricted to the lateral 

 spectra .seemed to indicate that there was some disagree- 

 ment. This is by no means the only ca.se in which a pecu- 

 liarity in the effect seen is due simply to imperfection in our 

 sensations of vision. 



The Joule-Thomson Effect. 



When a gas is forced through a porous plug, it is usually 

 at a dilfcrent temperature at the two sides of the plug. 

 Considerable interest has centred round this fact, because it 

 can be shown that if a gas were a perfect gas (i.e., one 

 satisfying accurately both Boyle's Law and Charles' Law) 

 it should undergo no change in temperature in such an ex- 

 periment. In reality, it is found that most gases cool at 

 ordinary temperatures in passing through ; hydrogen, how- 

 ever, is an exception, it gets slightly warmer. 



Olszewski, some years ago, showed that when hydrogen 

 is at a temperature of about 80 degrees C. below zero, it 

 does not change temperature when forced through a plug. 

 This temperature is known as the " inversion temperature," 

 because, for temperatures below it, cooling takes place in- 

 stead of heating. 



It now appears, however, that most probably this point 

 is not a unique one, and that it all depends upon the 

 pressure of the gas as to what the inversion temperature 

 will be. There is no direct determination except the one 

 quoted above. But calculations based on equations of state, 

 which fairly satisfactorily represent the observed connections 

 between the pressure, volume, and temperature of a given 



