XXI .
On the dispersive power of the atmosphere , and its ffeect on astronomical observations .
By Stephen Lee , Clerk and Librarian to the Royal Society .
Communicated by W. Ho Wollastor , M D. Sec. R. S. Read June 15 , 1815 .
NOTWITHSTANDING the pains which astronomers have taken to determine accurately the refraction of mixed light , nothing , I believe , has ever been done towards ascertaining the dispersive power of common ' air , or comparative degree of refrangibility of the differently coloured rays in their passage through our atmosphere .
The importance of such an inquiry , however , must be obvious to every one who duly considers the effect which the , different degrees of .
refrangibility of the variously coloured lights must necessarily produce in the apparent situations of differently coloured objects .
Stars of different colours must be differently refracted , and the apparent altitude of the sun must vary according to the colour of the dark glass through which he is viewed .
Perhaps this cause alone is sufficient to explain the disagreement which is found to exist between the latitude of a place deduced from observations of circumpolar stars , and that deduced from observations of the sun during the solstices , which has so long occupied the attention of astronomers , and has never yet been satisfactorily accounted for.* The dispersive power of the atmosphere will also show why Aldebaran and the red stars are sometimes seen projected on the moon 's disk in occultations by that planet , especially when the immersion or emersion happens to be near her upper limb .
For the light of the moon being white , is more refracted than that of the star , and consequently her limb more elevated , which would occasion the star to appear within her disk a few seconds before or after contact.t The great disagreement which is found to exist in the declination of several of the fixed stars , as given by different observers , may probably be traced to the same cause , stars being more or less refracted according to the predominant colour of which their light is composed .
That the fixed stars differ from each other in respect to the composition of their light , must be obvious to any one who will only take the trouble of comparing them on a fine night .
They present a striking variety of colour even to the naked eye .
But this difference becomes still more perceptible when they are viewed through a prism properly adapted to the eyepiece of a reflecting telescope .
A star viewed in this manner is converted into a prismatic spectrum .
Sirius and the brilliant white stars exhibit a large brush of beautiful violet , and the most refrangible colours in great abundance .
Aldebaran , X Orionis , and the red stars show only a small proportion of those colours , whilst the dull white stars exhibit a great quantity of intense green light .
t Vide Philosophical Transactions , Vol. LXXXIV .
p. 345 .
Histoire Celeste Fran9oise , Tome I. p. 393 , 403 , 413 , 425 , 428 , 467 , and Connoissance des Temps for 181x7 $76 The planets also differ much from each other in this .
respect .
The moon , Venus , and Jupiter , seem to possess every colour ; but the green is very pale in all of them .
Mercury and Mars appear deficient in the middle and most refrangible rays , whilst the light of Saturn seems to be composed principally of the mean rays with a very small proportion of the extreme colours of the prism.* The different refrangibility of the differently coloured rays is very visible in stars near the horizon .
If viewed on a fine night with a power of 200 and upwards , they appear expanded into a prismatic spectrum .
Sirius , when within a few degrees of the horizon , presents a most beautiful object .
Having remarked the very oblong figure which the spectrum assumes when near the horizon , and found from repeated observations of different stars that the separation of light begins to be visible as high as 40 ?
or 5o ?
of altitude , I was led to believe that the dispersive power of the atmosphere must be sufficient , in many cases , to produce considerable effect on astronomical observations ; and , consequently , to suppose that it would be desirable to ascertain , if possible , the exact degree of separation of the several rays.t With this view , therefore , I began a series of observations ; the result of which , and the manner of conducting them , I shall now take the liberty of laying before the Society .
The first instrument employed for the purpose .
with any degree of satisfaction , was the two feet reflector made by Mr. SHORT , and which belongs to the Royal Society .
In the compound focus of the eye piece of this telescope , I fixed horizontally a narrow slipof ivory .
With the instrument thus prepared , I observed Capella , and other low stars near the meridian .
By carefully noticing the intervals of time between the first contact and total immersion , and between the first appearance and complete emersion of the star from behind the slip of ivory , I obtained data from which it was easy to calculate its vertical breadth , which , compared by estimation with its horizontal breadth , gave the separation of the extreme rays of light It was impossible , however , to remain long satisfied with such coarse measures , and not finding it convenient to go to much expense on this account , I applied to my friend Mr. RENNIE for the loan of his seven feet reflector made by Dr. HERSCHEL , to which I adapted a very excellent wire micrometer made by Mr. TROUGHTON ; and thus , by the kind assistance of my friends , I obtained instruments capable of measuring small angles to the fraction of a second of space .
With this apparatus I repeatedly measured the diameter of Mars during his opposition in 1813 .
The Society 's apartments being well situated for the purpose , I observed the planet as soon as he became visible over the buildings , until he attained his meridian altitude , which never exceeded 15 ? .
With a power of 170 and upwards , the disk of the planet appeared much elongated , especially when near the horizon ; the upper limb was of a fine blue , the lower limb of a deep red .
By carefully measuring the diameter of Mars and the breadth of the coloured edges , I endeavoured to ascertain , as exactly as possible , the degree of separation of the differently coloured images of the planet .
But after all it was no easy matter to measure the coloured edges exactly , for the light which was necessary to illuminate the wires , rendered the colours so very faint as to make it extremely difficult to distinguish their precise boundaries .
For this reason , and because I wished to apply higher powers than could be used with the micrometer , I adopted the following method , which I found far more convenient , and is , I believe , quite as accurate .
I drew on a sheet of paper several figures of two equal circles cutting each other , placing the centres of the circles in the first figure To of their radius from each other ; in the second figure .
; in the third 3- ; and so on .
The upper crescent of these figures I painted blue , the lower crescent red , and the part common to both circles of a reddish yellow , softening the colours into each other as they appeared in the planet .
For I considered that , in fact , it was not a single image of Mars that was seen , but a number of differently coloured images , lying in the same direction , though lifted one above another , as represented in the annexed figures .
X23 N &gt; joLo _ , 379 9Ca Having prepared a number of these drawings , I repeatedly compared them with the planet viewed through the telescope with different magnifying powers , carefully noting which figure he most resembled , and the time of observation .
This being done , it was easy to calculate the exact altitude from the time of observation , and to make a very near estimate of the separation of the images from the figure referred to , compared with the diameter of the planet found by the micrometer .
From a great number of observations on Mars , Venus , and the fixed stars , taken in all these different ways , I found the deviation of the extreme rays of light to be between o and ~ part of the total refraction .
It has already been observed , that the disagreement which is found to exist between the latitude of a place deduced from observations of circumpolar stars , and that from observations of the sun , may perhaps be traced to the use of dark glasses .
But this will appear more evidently from a reference to the method employed by Dr. BRADLEY for determining the quantity of refraction , which method is very clearly described by Dr. MASKELYNE in the seventy-seventh volume of the Philosophical Transactions .
He says , That Dr. BRADLEY got the height of the pole from observations of the circumpolar stars , and the height of the equator from observations of the sun at the two equinoxes ; that he found these two altitudes together amounted to 89 , 68 ' , ^ " , which being subtracted from 9o ?
, leaves l ' , 57 " , for the sum of the refractions at the pole and equator ; and that of this quantity he assigned 45x " to the former , and 71"/ to the latter .
But Dr. BRADLEY undoubtedily made use of dark glasses for observing the sun , probably smoked glasses , which would give him a pale orange coloured image , or one of less than mean refrangibility ; consequently , the quantity of refraction as found by Dr. BRADLEY must be too small for white light .
This alone is sufficient to produce a small difference between the results of our observations of the sun and of the stars .
I shall now mention two other circumstances which appear to me to have produced a still greater apparent disagreement .
The publication of the Nautical Almanac in 1767 , led to the general use of HADLEY'S sextant .
In the construction of this instrument , coloured glasses were indispensibly necessary ; and the , great convenience in the use of them over smoked glasses , soon occasioned the application of them to all other instruments .
These glasses generally give a deep red image , or one of less refrangibility than smoked glass .
The effect of this alteration , therefore , should have been , that arising from too great correction for refraction in every thing depending on observations of the sun .
The introduction of achromatic object glasses * produced an error of a different kind ; and one which , in certain cases , -tends to correct the other .
In the single object glass telescope ( and there were no others in BRADLEY'S time ) the differently coloured images are formed at different focal distances , which , in a manner , compels the observer to adjust his instrument to the most intense light ; that is to say , to the orange coloured f image ; by this means the fainter colours , which occupy the greatest space in the spectrum , * are dissipated , and lost among the more powerful rays .
In good achromatic telescopes the case is very different , for all the rays being collected by them into one point , every colour is seen in its proper place , so that the observer , in bisecting the spectrum , takes the altitude of the mean , or the upper extremity of the green image .
But if the upper extremity of the green image be taken in observations of circumpolar stars , a greater correction than Dr. BRADLEY'S ought to be applied , in order to get the true height of the pole .
It may not be amiss to observe here , that the observations of Mr. LALANDE at Paris , show a greater disagreement than those at Greenwich ; and the observations of Mr. PIAZZI at Palermo , a still greater than those of Mr. LALANDE .
This , I apprehend , must arise partly from the lesser elevation of the pole in those places , and partly from the fainter colours in the stellar spectra being more distinctly visible in the clear atmospheres of France and Italy than in England .
It should seem then , that in order to get a perfect knowledge of astronomical refraction , we ought to employ at least three different methods of investigation .
ist .
By observations of the fixed stars during the night , when all the prismatic colours are visible .
2dly .
By observations of the stars ( luring the day , when none but the orange coloured rays are to be seen .
And 3rdly , by observations of the sun with different coloured glasses .
By these means we might hope to obtain such an accurate knowledge of atmospheric refraction as would enable us to form tables adapted to every possible circumstance .
But I must not take up the time of the Society by any additional observations .
It is in vain for me to pursue the subject any farther , in a situation so ill adapted to astronomical observations as Somerset Place ; I shall therefore resign the task to those who are more favourably placed in this respect , and who possess instruments better clIculated for an investi , gation which requires so much acc racy .

