J 1885 .
] Absorption-spectra Thermograms .
77 being made for temperature , the determination of the currents by this method was abundantly accurate and very simple .
The results are grouped in three series , of which the first two were considered in the Preliminary Note .
In both of them the same tube was used , the principal difference being that in the first the light traversed the tube three times , and in the second but once .
In the third series another tube was employed , and some improvements in respect to thermal insulation were introduced .
The readings were taken with a double image prism in place of the ordinary analysing nicol , a substitution by which it is believed some advantages were obtained .
From the fifteen sets of observations of Series I , we find as the rotation of sodium light in bisulphide of carbon at 18 ' corresponding to a difference of potential equal to unity C.G.S. the value *04208 minute .
From the four observations of Series II we get in like manner *04198 minute , and from the seven observations of Series III *04202 minute .
The last value is adopted as the most probable .
In an appendix some remarks are made upon polarimetry in general , especially in relation to the half-shade method .
A device proposed by M. Becquerel for augmenting the precision with which rotations can be determined with the aid of a half-wave plate is .
considered , and the conclusion is arrived at that no advantage can thus be obtained .
III .
" Absorption-spectra Thermograms .
" By Captain Abney , R.E. , F.R.S. , and Colonel Festing , R.E. Received December 31 , 1884 .
When employing a grating for visual work , the choice of a medium which shall absorb the overlapping parts of orders of the spectrum other than that under examination is a comparatively simple affair .
When working photographically , however , it becomes necessary to know what invisible rays the medium will cut off .
For instance , in investigations in the infra-red of the spectrum , we have found it necessary to ascertain whether media which absorb the blue and allow the red rays to pass would also transmit the above-mentioned parts of the spectrum .
Photography would seem to be the simplest method of experimenting in this direction ; but the results not being quantitative , as are those obtained by a thermopile , we found it better to work with that instrument .
It should be remembered that we were dealing with the infra-red part of the spectrum , in which the energy is so great as to be well shown by the thermopile ; in the more refrangible part the indications 78 Capt. Abney and Col. Festing .
[ Jan. 15 , of the thermopile are so feeble that photography would be more serviceable .
The absorbing media usually used were :\#151 ; 1 .
A solution of potassium dichromate .
2 .
Ruby glass combined with orange glass .
3 .
Deep orange glass alone ( often called " stained red " ) .
4 .
Iodine dissolved in an aqueous solution of potassium iodide .
5 .
Iodine in alcohol .
6 .
Iodine in carbon disulphide ( the violet rays being absorbed by canary glass ) . .
' 7 .
Dyes of different kinds , combined with coloured glasses .
8 ; Deep coloured cobalt glass and orange glass .
Of these the 1st and 3rd were employed when the green rays had not to be cut off ; the 2nd , 4th , and 5th when the whole of the more refrangible end of the spectrum down to the orange had to be absorbed , while the 6th cut off very nearly the whole of the visible spectrum .
The dyes employed were soluble in either alcohol or water .
With No. 8 ( deep cobalt and orange glasses ) a very long exposure to the dark rays was found necessary to obtain a satisfactory impression on a sensitive plate .
In these experiments we used the thermopile described in our paper in the " Proc. Roy .
Soc. " ( No. 232 , 1884 ) .
The source of radiation was an incandescence lamp maintained by a Grove 's battery of 40 cells , the filament being approximately at the temperature of 1650 ' .
An image of the filament was thrown by a lens on the slit of a collimator , behind which was the dispersing prism .
The rays emerging from this were collected by a second lens into a spectrum at a definite distance from the prism , and the thermopile was passed along this spectrum , as described in our paper above mentioned .
The absorbing media were placed in front of the slit of the collimator , glass cells one-eighth of an inch of internal thickness being used for containing the liquids .
The thermogram of the radiation unabsorbed , except by the prism and lenses , was first made ( see Curve I , figs. I and II ) .
An empty cell was next placed in front of the slit ; the thermogram ( Curve VI , fig. II ) shows that no special , but a general absorption has been produced.* A comparison of Curves II , III , and V with Curve I , in fig. I , shows the absorption of radiant energy by orange glass , ruby glass , and cobalt glass respectively .
Through the kindness of Messrs. Chance Brothers we can state that the orange colour is given by the oxides of silver and antimony combined , whilst the ruby colour is due to cuprous oxide .
The * The empty cell-curve has been corrected for reflection from the two inner surfaces of the glass , so that all comparisons are made under similar conditions .
1885 .
] A bsorption-spectra Thermograms .
Fig. 1 .
Curve T. Naked incandescence lamp .
Curve II .
Absorption of orange glass .
Curve III .
" ruby " Curve IY .
" green " Curve Y. " dark cobalt glass .
cobalt colour , as is well known , is given by cobalt .
A glance at the curves shows why any combination in which the last is employed would necessitate a prolonged exposure , as it will be seen that it entirely absorbs the infra-red rays about the part of maximum energy .
The use of this glass should therefore be avoided , if possible , when the infra-red rays are to be used in photography .
In a paper by Captain Abney , which appeared in the " Philosophical Magazine " ( vol. x , 1880 ) , it was shown that reversal of the photographic image was caused by rays of low refrangibility .
If a photographic plate which has been exposed to white light in the camera be subsequently exposed to radiations of low refrangibility , as for instance to light passing through a combination of ruby and orange glass , we should expect that there would be a gradual extinction of the effect of the white light .
This is true in practice , though it rarely happens that the after exposure to such radiation is sufficiently prolonged to be hurtful .
A case , however , might arise when a knowledge\#151 ; -or perhaps it Capt. Abney and Col. Festing .
[ Jan. 15 should rather be said an appreciation\#151 ; of this fact would be of practical use .
In such a case the visible radiation coming through a combination of cobalt and orange glass might be successfully employed .
Curve IV is the absorption-spectrum thermogram of a green glass , which obtains its colour from a combination of cupric oxide and iron oxides .
An examination of the prismatic spectrum photographed through such glass , has led us to suspect that green glass is transparent to certain rays of very low refrangibility. .
The thermogram fully confirms this .
Some four years ago one of us proposed to the Meteorological Committee of the Royal Society the use of green glass for thermometer bulbs , in the belief that it totally absorbed the infra-redrays. .
This is now found to be a mistake , and some other colouring matter must be sought for for this purpose .
Fig. II illustrates some interesting facts in connexion with absorption .
Curves I and VI , as already stated , are the thermograms of the lamp without and with an empty glass cell in front of the slit .
Curve VII is the absorption thermogram of benzene .
As far as Scale No. 6 there is no appreciable absorption , but from that point there is a drop in the curve , which proves to be coincident with the radical band of the group of benzene compounds , which we mapped photographically ( see " Phil. Trans. , " 1882 ) .
At a point between 7 and 8 on the scale absorption again commences , and although of some small extent , is appreciable as far as the end of the thermogram , thermal action being traceable nearly down to Scale No. 20 , or wavelength 27,000 .
This point we presume to consider of some importance , for we have shown that there is reason to believe the existence of some compound of the aromatic series between us and the sun , as the radiant band is coincident with a very strong line in the solar spectrum .
The solar spectrum as determined by the thermopile being coterminous in this direction with that of the incandescence lamp , had it been found that benzene cut off the lower part of the latter , our belief as to the indications of benzene in the solar spectrum would have been much shaken .
Curve VIII is the thermogram of a deep-coloured solution of iodine in alcohol .
This terminates near Scale No. 16 .
We found a slight dip near Scale No. 6 ; this was too slight to be conveniently shown on the diagram .
It is coincident in position with the alcohol radical band .
The thermogram of uncoloured alcohol is the same as that of Curve VIII , with the exception that some further thermal action in the visible part of the spectrum has to be registered .
Curve IX is the thermogram of the absorption-spectrum of a saturated solution of common salt in water .
Rock salt , as is well known , allows nearly all radiation to pass , and it seemed desirable to ascertain what effect it would have in solution .
The effect of the salt appears to have been a mitigation of the absorptive power of the 1885 .
] A bsorption-spectra Thermograms .
Fig. 2 .
Curve I. Naked incandescence lamp .
Curve VI .
Absorption of empty cell .
Curve VII .
99 of cell and i-inch of benzene .
Curve Till .
99 99 J-incli iodine solution in alcobol .
Curve IX .
99 99 " sodium chloride saturated solution in water .
Curve X. 99 99 " water .
Curve XI .
99 99 " iodine in an aqueous solution of potassium iodide .
Curve XII .
99 99 " iodine in carbon disulphide .
water , as the curve shows the same qualitative absorption as Curve No. 10 , which is that due to -|-inch of water , the absorption being equivalent to what would have been produced by a diminished thickness of water .
This is a fact of some importance theoretically , and has been confirmed by experiments with other solutions .
Curve XI is that due to the absorption of iodine in an aqueous solution of potassium iodide .
Compared with Curve X it shows that the molecular structure of potassium iodide is such that it intensifies the absorption due to water alone , no trace of the spectrum being found beyond Scale No. 14 .
Curve XII is the thermogram of the absorption-spectrum of a soluVOL .
XXXVIII .
00 to I II Ill IY Y YI YII Till IX X XI XII Curves .
Scale 00 00 C3 Dark cobalt glass .
00 00 rS r3o \#169 ; .0*4 || rt 3 3 6 .a .is ApproxiNumber .
Naked light .
\#169 ; .
hA ao ( H on g p-H 6 60 r3o .a Green glass .
1 Empty | cell .
S N \#169 ; m H 0 S ^ S S o'a NaCI ir water .
Water .
Iodine i potassi iodide water .
Iodine i carbon disulp !
mate wavelengths .
-i 0 0 0 0 0 0 0 0 0 0 0 0 5500 0 1 0 0 0 0 Of ; 0 !
0 0 !
0 !
0 0 5850 2 3 I2 0 O2 0 2 !
2 !
0 2 !
2 !
0 !
0 6500 3 , , 1 0 , , , .
3 , , 2 !
, .
0 6850 4 7 4 !
u 0 Of 5f 5f 3 5f 5f 4 0 !
7250 5 11 7 4 0 2 8 8 6 8 8 7 1 !
7750 6 17 11 8 0 3 12 !
11 !
9 12 !
12 !
rH](N rH 6 8400 6 !
, , 0 , , .
, 11 , t , , .
, . . . .
8800 7 27 17 15 0 4 17 !
17 !
17 !
17 !
17 !
17 14 9200 8 37 !
25 22 0 3f 24 22 24 24 24 21 24 10100 9 45 35 29 I !
2 29 28 29 27 !
26 !
22 29 11250 10 49 42 34 4 1 32 30 29 19 18 13 32 12500 10| 48 !
43 35 !
( , 0 31 !
30 . .
, .
, . . .
31 !
13200 11 46 !
42 35 !
10 1 !
30 !
29 18 8 4 !
3 !
30 !
13900 11 !
# # , , , , 13 9 !
5 4 . .
14600 12 35 !
32 25 !
12 1 !
26 21 12 !
10 !
6 4 26 15300 13 25 21 17 12 !
2 17 12 8 5 3 .
1 !
17 16800 14 15 !
12 10 8 2 !
9 6 !
4 !
0 !
0 0 9 18300 15 9 7 5 5 !
2 !
5 !
4 2 0 0 0 5 !
19800 16 4 4 2 !
3 !
1 !
4 2 0 !
0 0 0 4 21300 18 2 14 1 14 Of 0 !
0 0 0 0 14 24500 20 0 !
0 !
0 Of 0 0 !
0 0 0 0 0 04 27000 isorption-spectra , 1885 .
] The Upper Partial Tones of a Pianoforte String .
tion of iodine in carbon disulphide .
It differs from the curve due to the empty cell only in the visible and shortest wave-length portion of the invisible spectrum , as might be expected .
We have also examined plain carbon disulphide and carbon tetrachloride , but have found only small traces of absorption with these liquids .
We have found that alum solution gives a curve which differs but little from Ho. XI in the least refrangible part of the spectrum ; the alum seems to intensify the absorptive power of the water in which it is dissolved .
It has been so often stated that an alum solution cuts off all rays of low refrangibility- ( or as it is incorrectly and commonly said , all " heat rays " ) that we were not prepared for the comparatively small effect that it produces .
It may be said that , roughly , one thickness of a saturated solution of alum in water is equivalent to a double thickness of water , and not more .
Judging by the thermograms , even this would be an exaggeration of the truth ; but the use of glass in the cells , prisms , and lenses diminishes the effect as found when the total radiation is taken directly .
We may add that dyes seem only to absorb in the visible spectrum , and to have but little , if any , action in the invisible regions .
The positions which we assign to the maxima of energy in the different absorption spectra of glasses do not agree with those that have been published ; but as ours are the result , not of one set of experiments , but , in some cases , of dozens , we feel fairly confident as to their correctness .
IV .
" Observations on the Upper Partial Tones of a Pianoforte String , struck at one-eighth of its Length .
" By Alfred James Hipkins ( of John Broad wood and Sons , London ) .
Communicated by Alexander J. Ellis , F.R.S. Received January 7 , 1885 .
' This is a postscript to my paper on the harmonics of such a string , read on the 20th of November , 1884 .
According to Professor Helmholtz 's theories , the tone of a struck string is compounded of a number of simple partial tones , with the ratios of their frequencies as 1 , 2 , 3 , 4 , \amp ; c. The harmonics are themselves also compound tones of which the primes or lowest partials are the partials of the original tone .
These are produced on damping the other partials by touching them at a node .
How , in my former paper , I showed that by so touching I could bring out twenty different harmonics of the string , and among these the 8th and 16th .
Young 's law , however , makes an harmonic impossible to produce , if its node is the point struck .
Hence the string being struck at one-eighth of its length , no 8th or 16th

