]\gt ; IX .
On ( Eclipse of Angust ) L. DARWIN , R.E. , ARTHUR .
D. , , clnd E. WALTER Reccived January 28 , \mdash ; Reffi ebruary 1 , 10 .
CONTENTS .
I. Origin of the Expedition and General .
By Ca A. , and .
W. MAUNDER . . . .
21 11 .
Preparations the Eclipse at Prickly Point .
By and A. ) ; 1II .
Totality at Prickly Point .
By Captain and A. SclIUblER . . . . .
) .
On the Accol.acy required in ' an EqnatoriaI for rposes a Total olar Eclipse .
By ARTHUR STLR . . . . .
ResuIts phic C Point .
By VI .
The Coronagl.aph .
By Caplain . . . .
The Prismatic Came ] .
By tain D . . . . .
VIII .
The Spectroscopic Cameras at ickly Point .
By cHUbl . . . . .
IX .
Photographic Results obtained at Carriacou Island .
By .
W. MAUNDLR . . . .
) X. escription of the Eclipse and awing of the Corona .
By Captain C. MALING . . . . . . . . . .
XI .
On the Pbotographs of the Corona obtained at Point and Carriacou Islalld .
By W. H. . . . . . . . . . . . . .
347 I. ORIGIN OF THE EXPEDITION AND GENERAL CAPTAIN DARWIN , ARTHUR SCHUSTER , AND E. W. expedition for observing the clipse of the Sun of August , 1886 , was organised and sent out by the Royul Society , the necessary funds being obtained partly by a special vote from the Treasury , partly from the annual grant to the Society , and partly from the Society 's private funds .
A Committee appointed by the Council of the Royal Society discussed the principal questions to which observers were to direct their attention , and distributed the the1n .
It was also decided that , as far as the scientific part of the work was concerned , ) observers should be independent of each other , and reporc separately to the Society ; but that they should elect one of their members as chief , to represent them in all 10 .
12 .
89 L. DARWIN , DB .
A. SCHUSTER , AND MR. ] .
W. MAUNDER with the authorities in the West Indies .
Mr. NORMAN LOCKYER was , Iy chosen to be this representative .
escnt report only deals with the photographic results obtained by its authors .
was the only other observer who took out photographic ; most , the weather proved so bad at the station he selected was unable to see anything of the eclipse .
ABNEY was nnfortunately not able to take part in the expedition , but he invaluable ] to the observers in their preparations , and in this way tributed most to whatever success the photographic part of the expedition obtained .
The photographic plates used by Dr. SCHUSTER and Mr. bIAUN 1 were prepared by him , and we wish to offer him our best thanks for the assistance he has rendered us .
The expcdition left on the 29th of July , 1886 , and arrived at St. George , , on the 12th of A gust .
A letter had kindly been sent by the Colonial Office to the ] oxly , what the requirements of the expedition would be .
The members sequently found on their arrival that the Governor , Mr. W. J. SENDALL , had made every inquiry as to the best sites for the observatories , taking int , considerationl the weather ties as well as their personal comfort ; and they have to ) the greatest courtesy and consideration during the whole of their stay in the .
The protection of the instruments having been mentioned in the Colonial Office lecter , Captain I. C. , the Colonial Secretary , very kindly prepared models of ] which could be cheaply and readily constructed on the spot .
When the expedition arriyed at Barbadoes on their outward journey , they these models awaiting them .
A telegram was despatched to Grenada approving erally of the design , and thus work was actually commenced before the arrival of the expedition in the island .
Before the observers left England , the President of the Society had written to the esti the co-operation of any men-of-war that might be on the .
As a result of this communication , three of Her Majesty 's ships\mdash ; the " " \ldquo ; Commander H. ARCHER , R.N. ; the " " Bullfrog Lieutenant J. MAS .
N. ; and the " " Sparl.owhawk , \ldquo ; LieLltenant C. F. OLDHAM , R.N.\mdash ; were found ready and prepared to render every ce .
Every member of the expedition felt fur the willing way in which the able assistance of both the officers and men of these ships was given ; and the President of the Royal Society , in a letter to the on the return of the expedition , expressed the value to science of such ready co-operation .
After the arrival of the expedition at Grenada , two or three days were occupied in selecting stations and preparations for conyeying the observers to their destitions .
St. George itself was not favourably situated , it was moreover considered to scattcr the observatories as much as possible , so as to avoid the chance of a single of cloud ving fatal to the whole expedition .
The observers were therefore ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29 , 1886 .
divided into four groups , and , with the aid of the aboye-mentioned men-of-war , they soon found themselves distributed their various stations .
) who accompanied the Rey .
S. J. , was conveyed in the " " \ldquo ; to Carriacou Island , where a station selected near the southern extremity of the island , close to a small house called " " the Hermitage belonging to .
PETER DRUMMON gentleman of Jersey , who happened most fortunately to be Carriacou at the ti1ne , who , in the most generous mannel ; gave up the use of his premises to the observers .
Mr. cHUSTER and Captain DARWIN were dropped at Prickly Point by the " " Fantome and there they found excellent quarters in a house which .
F. M. CHADWICK , the Colonial Treasurer , kindly placed at their disposal .
II .
PARATIONS F0R THE EcLlPSE AT ' POINT .
By DARWIN AND ARTHUR SCHUSTER .
The instrunents used totality at Prickly Point were mounted on two equatorial stands , which were placed in separate huts at a distance of about 20 yards from each other .
On the of Thursday , August 19 , the polar axes of both stands were adjusted in the usual way .
Finders had been attached for this to the phic cameras , but more attention should be given in future to have these finders in convenient positions for observation , and of not too small an aperture .
One instrument , which was under Captain DARWIN'S charge , was placed on a solid foundation , and the first adjustment was therefore considered sufficient , especially as extreme accuracy was not required for the purpose for which it chiefly employed .
The foundation for the stand which carried rument did not , unfortunately , prove sufficiently iirm .
As there was reason to fear the heavy rains during the week the eclipse might have altered the position of the polar axis sufficiently to interfere with the sharpness of the irnage , and as the camera could not be reversed .
adjustment as long as the spectroscopes were attached to it , these were once more removed on Friday , August 27 , two days before the eclipse , and the routine of djustment once through .
From obsel .
Vations taken the day after the eclipse it appeared that the altitude of the axis was about 3 ' too low ; the error therefore sufficiently small not to produce a detrimental effect during the time of exposure .
The clocks of the were frequently adjusted , by the rotation of the hour circle the instrument was going with the time on a chronometer .
In nei her case were the instruments and their stands designed for each other , and it was found possible to balance the instrtlments properly without a too gre increase of weight .
Thus the work which the clocks to perforltl was difl'erent in different positions of the irJstrument .
To remedy this L. DARWIN , ) .
A. SCHUSTEH , AN1 ) MR. E. W. ossible , the clock adjustment was carried on in the position which the equatorials to totality .
Owing to the unsettled state of the there , the for the eclipse c.arl.ied on under reat difficult.ies , the time at disposal was found barely sufficient .
We ived the observatory on a , and the remainder of that week was taken ) with the erection of the equatorials and the adjustments of the various instruments .
the week preceding the eclipse much time was lost .
to the -requent terruption of the work by heavy tropical showers .
Tuesday and Wednesday were wet and stormy and no direct Sun light was available , much required , to get the instruments into working order .
fine , in the good progress was made ; but Friday was , at was followed by a light .
turday , the day before the eclipse , cloudy in the morning , and the only appeared at intervals .
Our experience thus us that a fortnight 's time for preparation is hardly sufficient when two obseryers have to look after five rent instruments , all requiring careful treatment .
After the days had beeu spent in continuous work , the evenings were taken up with the preparation of photographic chelnicals and occasional star observations .
were without intelligent help except the two days when Mr. to Prickly Point .
We had taken him out , jointly with Professor THORPE , as private assistant , anticipating the difficulties we should have to encounter .
His time , however , was chiefly taken up at Hog Island , where Professor THORPE was observing , but our are due to him for the assistance he gave us during these two days .
As for help , we engaged for the whole time one negro servant , but now consider we should have done better to have had two .
, with its steady ture , day and night only between and , proved very exhausting , nd the work could not be carried on as well and as quickly as it might have been under more favourable circumstances .
The richness of aniInal life proved a source of great annoyance .
Mosquitos abounded in our residence .
Wasps built lests and spiders spun their webs with remarkable rapidity .
The photographic roorn and even one of the stands bad to be cleared of wasps ' nests , and a dense spider 's web found an hour before stretched across the slit of ) of the spectroscopes ; the slit having been erfectly clear the before .
It is difficult to realise at home the special difficulties of temporary observatories , but we venture to suggest that more skilled assistance should be provided at future eclipses .
at Prickly Point we were enabled to carry out our programme with hardly any mishaps , .
we feel obliged to point out the difficulties under which we worked and which might easily have led to serious accidents .
For example , five utes before totality , Captain clock stopped altogether , alt'nough it is ON THE TOTAL SOLAR OF AUGUST 20 , 1886 .
2 5 believed every po , ssible ) had been taken in its regulation .
It only by a alteration in balance at the last moment that it was indnced to go on at all .
Captain kindly sent us two ilors to render help during totality , BROWETT , signalman , and , petty officer .
One of these two place(l between the two huts to call bhe time of a } while the other was to Captain DARWIN in the manipulation of his insfrtiment .
They both did what was required of .
totality Captain required no other assistance , his was not ; but with Dr. .
whose whole observations were inch1ded in that time , the case different .
He ccepted therefore the to him by Dr. P. F. , the officer of health in ' by RRAY , who accompanied the expedition naturalist .
Captain MALING undertaken to a of the outer of the According to a stion of Mr. LOCKYER'S , a disc had been prepared inches in diameter .
When placed at proper distance the Moon.and the inner parts of the corona could be screened off by means of this disc .
T-t placed on a wooden support at the top of an incline which down from.the ry to the sea .
A post was driven into ground some distance from ) disc .
The observer was to 1ook a small hole in the post .
arrangement , however , ants very careful adjustment , and we had not time to spare , the ncipal objects of the expedition monopolised all our attention .
from want of adjustment or from other causes , we decide , drawing includes the whole of the corona ) to the prominen ces .
Captain statement is included in this The following Table the ition of the obse ) ' vatory and various connected with the eclipse : \mdash ; III .
TOTALITY AT PRICKLY POINT .
BY CAPTAIN DARWIN AND ARTHUR SCHUSTER .
It has been mentioned that during the week preceding the eclipse , which took place on the Sunday , the weather ] ) been very unfavourable ; but the clouds deared away on aturday afternoon , the sunset was fine , and experience had taught us that a fine evening was generally followed by fine morning .
Our prospects , therefore , were very good on night , but early on Sun the wind rose , which .
At five o'clock , however , when we got up , the sky was still perfectly tlJe clouds came from ) half-past five , at first in the form of detached cumuli , but before the whole sky the East was oyercast .
The Sun rose behind the clouds and was still hidden when the time for first contact arrived .
It only twenty minutes before totality that the crescent of the Sun appeared , and then the wind soon made a clearance round him .
Ten minutes before the sky wa clear .
The last tions were given , the put in position , and everybody took his appointed place .
Five , as the darkness increased visibly , the weather seemed safe .
Another minute passed $)nd danger once more threatened ; a small cloud rising from the South-East was driven by the wind right towards the Sun .
The Moon from above descended over the ] disc , and it seemed a race between the Moon and the cloud which should cover it first .
Totality began , and the corona became distinctly visible , but was obscured again almost instantaneously by the cloud .
For about three-quarters of a minute the corona was only seen ough a film of cloud as a narrow hazy ring .
The corona finally appeared , and remained clear as long as totality lasted .
Lieutenant R. J. KIDD , then Private Secretary to the Governor of Grenada , took temperature observations every five minutes for one quarter of an hour before and after totality .
His nbers are as follows:\mdash ; h. ahr .
As the encement of totality took place at it will be seen that during the ometer only fell one quarter of a degree Fahrenheit , and reached its lowest point ten inuntes after totality .
The whole change in temperature was exception small .
ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29 , 1886 .
PURPOSES DUR1NG TOTAL ECLIPSES .
By ARTHUR SCHUSTER .
Observers preparing for a total Solar eclipse have in general only a very moderate time at their disposal , and it is of great importance to them to settle beforehand to what degree of accul.acy the adjustment of their instruments is to be carried .
The time which is } ) over adjustment in any one direction must necessarily be taken away from other more important matters , as there is never any lack of work on these lf , for instance , a photographic picture of the Solar corona is to be obtained , it would be clearly waste of time to refine on the adjustments of the or the rate of the clock beyond the point at which the Sun 's change in declination would produce a visible effect .
We shall see that this consideration limits the time of exposure which he advantage of the aperture of the lens can be realised , and this again will give us a limit beyond which it would be unnecessary to adjust the orial .
We might , indeed , take the Sun 's apparent motion into , and point the in strumental axis , not to the pole , but to some point near it , which easily be determined by calculation .
We shall see , however , that the ayailable tinle of exposure of a 4-inch lens is quite sufficient for our present requirements , and that it is therefore unnecessary to make allowance for the Sun 's change of tion .
When the instrument is nearly adjusted , the relation between the true declination of a star 8 and the apparent declination S ' is given by\mdash ; @ , where is the angle between the true pole and the instrumental pole , is the hour angle of the star , and S the hour angle of the instrumental pole .
If S is constant the change of apparent declination from bad adjustment in a short time is found from the aboye equation to be , which will , numerically , always be smaller than .
If the time , measured in minutes of time , is , we can write for this maximum change\mdash ; We find in this way that a change of apparent declination of one second of arc per minute could be produced if Now , the change of the Sun 's declination may be , and during the last eclipse was , nearly one second of rc per minnte .
It will be unnecessary , therefore , to spend any * During the total phase of late , owing to the low altitude of tho apparenl change of albitude due to of refraction was about seconds of but the in declination due to refraction was smaII , and generally the effects of refraction may be neglected .
time in making the angle smaller than that given by the above value , as a closer adjustment may in some cases make matters worse instead of better .
Special considerations may of course induce observers to adjust their ruments more accurately .
If , for instance , the equatorial adjustment is to be done once for abouC a fortnight or a week before the eclipse , and if the conditions of the foundation on which the pillar rests give grounds for fear that the position may slightly change , it would be wise to aim at a greater accuracy in the first instance , so as to give some play for change .
So far the adjustments do not depend on the size of the object-glass , but , if we want to make the best use we can of the aperture , an apparent shifting of the Sun 's image during the time of exposure must be confined within small limits , which we have now to fix .
The central disc of a star , or of a small object at a great distance , has an angular radius in the focal plane of the telescope , given by when is the diameter of the objective .
Experience shows that small objects can , under favourable circumstances and with strong illumination , be resolved when the centre of one diffraction disc coincides with the first dark ring ; but for objects having no well-defined boundaries we must give wider limits , and I therefore take the point of fair resolution to be reached when the first two dark rings touch , so that the star discs stand perfectly clear of each other .
In order to find out experim to what extent a small shifting of the images , such as might be produced by defective adjustment of the equatorial , might influence the possibility of resolution , I drew on a piece of paper two circles touching other , corresponding to the two star discs , and another similar set above them slightly displaced in the line joining the centres .
It was found that a displacement of of the distance between the centres of the disc could only affect the lesolution to a slight degree .
We may , therefore , consider a displacement through one-fifth of the radius of each disc allowable without impairing definition .
There is always something arbitrary in the fixing of these limits , it must be done in every investigation in which we want to determine the accuracy to be aimed at .
As object-glasses of different makers would probably differ 10 per cent. in their resolving powers , the limit chosen seems a .
and reasonable one .
We may allow then an angular shifting of if we substitute for its mean value of about cm .
This gives a value for ( which varies inversely with the ) , and is equal to seconds of arc for one centimeter aperture .
If the exposure takes place during minutes , and if ON TH TOTAL SOLAR OF , 1886 .
the or the Sun 's declination in one minute is seconds of , the displacement will be , and we for the longest time of compatible with fnll definition\mdash ; In the West Indian Eclipse was 10 and was 9 , which gives about 17 seconds time of exposure .
This is sufficient for obtaining a good image ; and we therefore , that an ad.iustment of the polar axis to about sufficient .
full definition .
Photographs taken with longer exposures would chiefly serve to fix the fainter parts of the outer corona .
By the ordinary means of an equatorial it is not difficult to obtain the necessary accuracy ; nevertheless , the calculation shows that we can by no means neglect to go through the fnll routine of adjustment , no equatorial should be sent out for ] ipse purposes hout complete appliances in the way of good finders , circles which are easily read , \amp ; c. The adjustment of the rate of the clock to the necessary accnracy is , as will appear , much more difficult .
We have seen that the angular displacement reach , but should not exceed , during the time of exposure .
Reduced to seconds of time , this is equal to , or , if the exposure is to be seconds of time , the error of the clock should not exceed , which for an exposure of 17 secon ds and an aperture of 10 .
is one part in a thousand .
The instrument we hod at our disposal did not 911ow adjustment as accurate as this .
At Prickly Point it found difficult to go beyond an accuracy of one per cent. No record of the rate was taken at Carriacou , but it was probably no better than at Prickly Point .
This would show that full nition would not be possible with an exposure of more than two seconds.ae We finally calculate by how much the efficient aperture is reduced if the errors due to adjustment are greater than those we have hitherto allowed .
If the centres of two star discs are separated b.y a distance , a displacement equal to would just bring them into contact .
A telescope which to the eye resolve two stationary stars at a distance will , if during the time of photographic exposure the displacement is , have its effective aperture reduced in the ratio of : Hence , if be the effective aperture *Some of the statements made here do not agree with those made in the Account ( ' Roy .
Soc. Proc. ' vol. 42 , p. 180 ) .
The discrepancy is accounted for by the fact that I previonsly took wider limits for the allowable shifting due to the motion of the telescope , and that ]lave since the photograpbs to a closer invefi } igation as regards fineness odetail actnally shown .
Hitherto we have put ; but supposing the rate of the clock is such that during the exposure , which was very nearly the case during the late eclipse , the above equation shows that the effective aperture was reduced to half , or that the photographs could not show anything which an eye-observer with a telescope of 2-inch aperture might not have seen .
Let us now turn our attention to the resolving powers which have hitherto been actually obtained in photographs taken of the corona .
the Eclipse of 1871 two prominences were sepalated in a photograph , which were at a distance of 15 seconds , and the corona itself gives no evidence of a finer structure .
The aperture used was 4 inches .
* In the corona photograph of 1882 the diameter of one prominence subtends an apparent angle of about 15 seconds .
As far as the photograph is concerned , we have no reason to suppose that an aperture two stars at a distance of 15 seconds could not have shown everything that is seen on the photograph , for no detail of the corona gives evidence of smaller structure .
This gives an aperture of 2 inches as sufficient to resolve all the detail shown in this photograph .
An examination of the drawing made by Mr. WESLEY of the present eclipse gives substantially the same result .
The 4-inch glasses used in the eclipses which I mentioned all give , therefore , a resolution equal to that obtainable with a 2-inch aperture on a stationary object .
During the last eclipse the regularity of the clock motion was , as we have seen , not sufficient to give more perfect images ; the same cause has very likely stood in the way of more perfect during previous eclipses .
We have at present no reason to suppose that the corona possesses any structural detail that could not be seen with an aperture of two inches , and it would be very desirable if eye observations could be taken at some future eclipse which would give us some idea of the detail of the structure visible with larger instruments .
But it is very remarkable that the definition obtained is just what we should have to expect , if we take the error of clock motion into account .
If the motion of the clock is regular , we should by shorter exposures get in great part over the difficulty of adjustment , but in the instruments which have been , at our disposal the mechanism of transmission from the clock to the telescope is not as perfect as it might be made , and the irregular vibrating motion produced by this cause would be sufficient to damage the sharpness of the image .
There does not seem to be any advantage at present in using larger apertures in future eclipses , at least if these larger apertures are accompanied by increased focal length .
Most of the adjustments have to be more accurate for them , in order bo get the advantage of the increased aperture , and the difficulties of mounting and adjusting are * Owing to an oversight , the aperture was in the Preliminary Account stated to be 2 inches .
The focal lengbh of the leus was only half that used during the later eclipses , but the aperture was the same .
ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29 , 1886 .
greater .
The photographs would be larger and more convenient to copy , which would be an advantage , but it seems more rational to improve the definition obtained with -inch lens , before larger apertures are used .
If we could increase the perture without increasing the focal length , we could reduce the time of exposure and gain an advantage , provided that the clock motion is sufficiently uniform .
All our investigations , then , seem to point to the desirability of an improvement not only in the average rate of the clock motion , for that could easily be effected , but in the diness and regularity of the motion .
I give , in conclusion , collected together , some formulae which may be useful to future eclipse observers .
In the equations is the displacement in seconds of arc which is allowable , consistent with full definition as above defined .
is the true aperture of the lens in centimeters .
is the effective aperture of the lens .
is the time of exposure in seconds .
is the longest time of exposure compatible with full definition .
is the greatest allowable angle between the true pole and the pole of the instrument in seconds of arc .
is the change of Solar declination measured in seconds of arc per second of time .
is the error of the clock rate , measured in per cent. In equations ( 2 ) , ( 4 ) , and ( 6 ) it is assumed that in adjusting the pole of the equatorial no account is taken of the change of the Solar declination during the eclipse . . . . . .
( 1 ) , . . . . .
( 2 ) , . . . . . . . . . .
( 3 ) ; or , if the longest exposure is to be used , . . . . .
, . .
, ( 4 ) .
Finally , to determine , we easily find .
( 5 ) .
This equation will allow , to determine the time of exposure allowable for a given effective aperture if the clock rate is known .
If the longest exposure is to be used , we get . . . .
( 6 ) . .
THE PHOTOGRAPHTC CAMERA .
By ARTHUR SCHUSTER .
1 .
The lens used for obbaining photographs of the corona had a clear aperture of 4 inches ( l01 cm and a focal length of about 5 in .
( 160 cm Ib therefore gave images of the Sun of about 3 in .
in diameter .
The accurate adjustment of the camera is a matter of some difficulty , as distant objects are not sufficiently sharp to admit of delicate focusing .
It was found better to focus on sharp object at a moderate distance , and to find by calculation the correction for parallel rays .
The correction to infinity Sf for a distance of the object focussed is given by If , for instance , were 1 kilometer , and , as mentioned above , 160 cm .
, the correction would be mm. , and , as it proved impossible to adjust the focus more nearly than to about ] a millimeter , error of 20 per cent. in the estimate of the distance is allowable .
It is a matter of importance to be able to determine accurately the position of the corona in reference to the coordinates fixed in space .
In Egypt and at Carolme Island this was done by means of a platinum wire stretched across the camera directly in front of the ] ; the shadow of this wire appears on all photographs , and its position was determined by a succession of instantaneous photographs of the Sun 's crescent taken directly before and after totalit , with a stationary telescope and at short intervals of time .
This plan has some disadvantages .
The shadow of the wire hides certain parts of the corona , and , for instance , in the photograph taken at Caroline Island the shadow is very awkwardly placed .
Then , , it is difficult to stretch the wire sufficiently tight .
The camera has to be turned directly on the Sun the crescent is being photographed , and the heat thus concentrated on the wire lengthens it and may change its position .
A different plan was therefore adopted .
Two needles were placed one on each side of the camera , the line joining the points passing approximately through the centre of the plate ; one of the ) eedles was intentionally inserted somewhat obliquely , so that its shadow might always be recognised on the ) raphs .
The image of needles would thus completely determine the position of plate in the camera .
The direction of the line joining the needle-points can be accurately determined with reference to the declination circle .
We need only bring the image of some , either terrestrial or celestial , just in coincidence with one needle-point , then with the other , and note the change of reading of the hour and declination circles .
For safety , and as a check , photographs of the Sun 's crescent might be taken after totality .
The needles , in the present instance , did not reach sufficiently fa into the centre of the camera .
In previous eclipses the general illumination of the sky in the ON TOTAL SOLAR ECLIPSE AUGUST 29 , 1886 .
of the corona was sufficient to affect the whole plate ; but it did nob on this occasion , the images of the needle-points only appear on one plate , owing to a cause to be mentioned .
This one is , however , quite sufJicient to fix satisfactorily the orientation of the corona .
I believe that the method could be rendered perfectly safe in future eclipses , and it is certainly more accurate than that hitherto used .
2 . .
P. F. , the resident officer of health in Grenada , and Mr. MURRAY , who acco1npanied the expedition as a ) offered their assistance during tot lity .
This was atefully accepted , and my best thanks are due to them . .
MURRAY undertook to even the camera while the slides were or pushed in , so as to prevent a confusion of .
due to king when the ] ides are touched .
Dr. MACLEOD took charge of slides before and after exposure , handing them , they lyere wanted , and placing them in a bag aftel they had been taken out of the carnel.a .
I had intended to take eight raphs , but had to change all ugenlents suddenly when I saw the corona obscured by a cloud at the beginning of totality .
As the brighter parts of the corona were faintly visibIe , a.od as I did not know whether the cloud would clear away in time , two photographs were taken , referred to in Mr. WESLEY'S report as Plates 1 and 2 .
) exposure of these must have been about 15 seccnds each .
It is only in a very indirect way that I could , after the eclipse , obtain an idea of the various exposures given , and not much value is to be placed , therefore , on the numbers .
I knew the time I intended to expose , but I generally closed the camera sooner , as I was vays afraid of a fresh cloud coming on .
The corona came out clearly while the third exposure took place .
Instead of a plate , a piece of sensitized paper had then been put into the slide , but the result was not good .
Three further plates were taken with the corona clear , the last about 15 seconds before the end of totality .
After the spectroscopic camerits had also been shut , intended to take slides .
Unfortunately I loosened the wrong hook , and drew , shutter instead of the slide , but , noticing my mistake at once , pushed it back .
That moment the Sun came out , and the last plate was partly spoilt , but parts of the corona show well , and it has been used in making out the details of the corona .
The accident was fortunate , sso as the image of the needles necessary for orientation came out clearly owing to the illumination by the reappearing Sun .
3 .
Mr. made drawing of most of the recent eclipses , the corona it is shown on the on the present occasion , need llob add anything to his remarks .
But I should like to ttention to one or :304 CAPTAIN L. DARWIN , DR. A. SCHUSTER , AND MR. E. W. MAUNDER two points which have a bearing on the speculations respecting the origin of the corona .
It is generally accepted now , I believe , that the Sun cannot have an atmosphere of anything like the extent of this luminous appendage .
The reasoning ] which seems conclusive , is that the pressure on the Sun 's surface must , to jlldge spectroscopic evidence , not exceed that on the surface of the Earth , and that an atmosphere comparable in size with the of the Sun would necessarily produce an enormous pressure .
I do not know , however , whether attention has been drawn to the fact that , even if such a gaseous atmosphere existed , it could not be luminous to any great height unless it contained a great number of or liquid The same laws of conyectiye thermal equilibrium which regulate the decrease of temperature in the atmosphere act still more strongly on the surface of the Sun , and even baking the highest estimate which can reasonably be made of the Sun 's temperature , the gases in the Solar atmosphere would quickly fall in temperature below the point at which they can be luminous .
The ] way in which they can be kept luminous is by containing sufficient matter to absorb the radiation from the body of the Sun , or by some independent cause , such as friction between solid gaseous matter , or electric discharges .
If electric discharges are the cause of the luminosity , the matter through which the discharge takes place must either be supplied by the Sun or by something outside , such as streams of meteoric matter .
The la.tter hypothesis has the great adyantage that may possibly account for the periodicity of Sun-spot phenomena , and it deserves , therefore , the attention of scientific men .
Now , it occurred to me several that if meteoric streams fixed in space were in any way connected with the corona we should find some eyidence of them in the general shape as seen from the Earth .
During eclipses which take place about the same periods of the year the Earth and the orbits of these hypothetical meteor streams would occupy the same relative positions .
We might then expect some periodicity in the shape of the corona depending on the time of the year .
To my surprise , I found that , as far as the evidence went , it seemed indeed to point in that direction .
In the report of the Eclipse of 1875 a paragraph was inserted , which , without laying any stress on the poiut , )tion to similarity between that eclipse and the one which had taken place the previous year during the same month .
The eclipses which have taken place since have added to the evidence , and it seems worth while , therefore , to draw attention to i-t , without however in the least wishing to imply that I consider the fact as proved or even as probable .
The Solar corona often shows a rough symmetry about its axis of rotation , but deviates from ] symmetry owing to one of the halves being broader than the other .
Hence , it often appears in the form of a trapezium .
Fig. 1 may serve to illustrate this .
AB is the axis ; the points of the longest rays of the corona , the distance DE being longer than the distance CF .
Now , during the ON THE TOTAL QOLA R ECLIPSE OF AUGUST 29 , 1886 .
eclipses which have taken place ] in April the eastern half of the corona is one which is broadest , while in the eclipses observed in July and August the opposite held good and the western half was the broader .
eclipses of December and May have hitherto shown no difference between the two halves .
It will no doubt be considered that the number of well ascertained coincidences is too slight to prove anything , and with this opinion I quite agree , especially as the changes in the corona which seem to depend on ) Sun-spot cycle have to be taken into account .
eless , it seems worth while to give the evidence here .
To begin with the eclipses of which we possess , the only ones , perhaps , we can safely take into account , the following Table will illustrate my meaning:\mdash ; The Sun-spot numbers represent frequency of spots as determined by Mr. R. WOLF .
As as 1878 these numbers are given in Mr. RANYARD'S clipse volume ( ' Roy .
Astron. Soc. Memoirs , ' vol. 41 ) for the actual date of the eclipse .
After that date the nulnbers are those given from time to time for the average of the whole ear 'oy Mr. R. WOLF in the ' Comptes Rendus .
' 306 CAP liAIN L. DARWIN , DR. A. SCHUSTER , AND MR. E. W. MAUNDER As regards the drawings , we have the following:\mdash ; rawings RIGHT aEGERMANN s We have , therefore , not one case in which the eastern side was the broadest in the autumn or the western in the spring .
If we take the time at which the greater eastern width changes to the greater western width to be about the middle of June , and middle of December , we should expect that in the eclipse which hasjust taken place the two sides would be about equally developed , but if is a slight asymmetry the eastern side should be the broadest .
* CORONAGRAPH .
By DARWIN .
1 .
D.escription of the Instrument its ustments .
The coronagraph was designed by Dr. HUGGINS as the inst1ument which would give the best chance of rendering it possible to obtain photographs of the corona in Sun light .
For this purpose a reflector is to be preferred to a refractor , and special precautions are taken to avoid internally reflected light .
The telescope is of the Newtonian form .
The instrument which I look out is by , of Dublin , and was mounted on an ordinary equatorial stand .
The light enters a tube 4 feet long , fitted with numerous diaphragms .
It then passes into a tube 5 feet 6 inches ( a little less than the focal length of the mirror ) , of about double the diameter of the first tube , the mirror * [ Note addel July 14 , 1889.\mdash ; This prediction has not been verified .
In the Eclipse of January 1 , 1889 , although there is not much difference between the sides , the western half is the bl.oader .
Nevertbeless , the considerations in the text are of some use , as they show the importance of taking account of the relative position of and Earth in discussing the shape of the corona in different eclipses .
] ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29 , 1886 .
at further end .
The beam enters chis tube at one side of the centre , and is reflected back at a small angle from the mirl.or at the further end to the aphic plate , which is alongside of the place where the light enters .
A partition in this tube separates the incoming and reflected rays for as long a distance as Qible .
Below this partition there are ova ] diaphragms , through which both beams pass .
I may here remark that last is , I rather think , a mistake ; for the backs of most of these di phragms are necessarily visible from the late , and , to whatever extent they are nseful in preventing stray light from getting the bottom of the tube , are proportionately harmful in reflecting this light directly back on to the photographic plate , which is very to them .
I was so much afraid of the direct Sun light striking the , of these oval phragms that I placed a temporary diaphragm of only 2 inches in diameter at the bottom of the outer , thus ensuring that the beam of Sun light should pass clear through the diaphragms the inner tube .
I should gladly have dispensed with this if I had dared .
The two tubes are held in position ]-n a long wooden frame .
There is a half-open space between the tubes , which has to be covered up by a cloth or other temporary contrivance .
This opening should be provided with a permanent cover .
I arranged instantaneous shutter close in front of the plate , that is to say , in the position which Dr. HUGGINS sidered most advantageous .
It consisted of a long wooden slide , with a rectangular opening 6 inches wide , which was drawn across an opening of inches diameter by two pieces of elastic .
It was released by cutting a thread .
Both the elastics and the thread were fastened to the body of the telescope , and not to the catnera or slide which takes the plate .
This latter part is not too firmly fixed to the , and I was anxious to avoid vibrations .
As Dr. SCHUSTER has fully described the adjustment of his innstrument for equatorial movement , focus , \amp ; c. , I will remark that the same hods were in the case of the coronagraph in so far as they were applicable .
2 .
Obse,.vations in Sun Light and the Partial The most important observations to -be made with the coronagraph were with the view of testing the practicability of photographs of the corona during light by this instrument .
This could be done in two ways : ( 1 .
) By obtaining } ) shortly before or after the eclipse , and comparing any irregularity that might appear in the halo round the Sun with photographs of the corona taken during totality\mdash ; a similarity of form indicating that the corona had been photographed .
( 2 .
) By taking photographs during the partial eclipse .
Then , if the of the corona produced any effect on the plate , the dark limb of Moon would be seen advisable in trying the first of the above experiments to a series of CAPTAIN L. DARWTN , DR. A. , AND MR. E. W. MAUNDER in each of the reversed positions of the instrument .
Then any irregnlarity in the illuminafion round the Sun which was reversed on turning the instrument over would clearly be proved to be due to instrumental causes ; whereas any larities which were visible in photographs taken in both positions , and which were not thus reversed , must be due to some outside causes , either atmospheric or coronal .
In my instrument reversal was only ] ) ) ssible near noon ; and on this account , as well as on account of the Sun being at its greatest elevation , the best results would be obtained by taking one series of photographs shortly before noon in one position , and ulOtb e. series in the other position soon afterwards .
With .
to the length of exposure , I thought that there was no advantage to be gained for either test by a longer exposure was necessary for the air glare to produce a the plate .
I estimate the exposure that I gave with the autonlixtic shutter at between one-fifth and one-tenth of a second .
The plates which I used for these tests consisted of gelatino-chloride dry plates on and on glass .
These were for me by two makers : by Mr. A. COWAN ( for Messrs. MARION and Co to whom I am indebted for making for me a specially batch of plates , and by Mr. WARNERKE .
Chloride plates choecn in preference to bromide plates , because it was considered that their relatively greater to ultra-violet rays would be advantageous for distinguishing between the corona and the air glare .
The plate to be chosen is no doubt the which is most sensitive in that part of the spectrum where there is the maximum difference of intensity of light in the of the corona and of the sky .
I may here remark that Dr. SCHUSTER informs me that , according to his photograph , the maximum difference would appear to be in less refrangible part of the spectrum , rather than in the violet ; and , if that is the case , bromo-iodide plates woul d I think , have been preferable .
was considered better than glass , as tending to educe the halation to a minimum ; as a fact , I relied chiefly on the paper negatives .
I used various developers for these ) , without observing any marked erence in the results:\mdash ; Hydrokinone , 1 grain to the ounce .
rTo 2 oz. of the above add dr .
of a saturated solution potassic carbonate , with common salt as a restrainer .
( 2 .
) Potassic citrate , 136 .
; potsssic oxalate , 44 .
; water , 1 oz. To 3 parts of the above add 1 part of ferrous sulphate , 140 .
; sulphuric acid , 1 drop ; water , 1 oz , .
Potassic bromide was used as a restrainer .
( 3 .
) A weak ferrous sulphate developer .
Some of the plates were developed in the West Indies , and some on my return home .
The results appeared to be the same , but they were , I think , much easier to develop in England .
Ice was used freely in Grenada , but I do not think the gelatine became quite hal.dened during the time of development , and the plates appeared to be more liable to fog .
* See p. 841 .
ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29 , 3 .
Results of First Test , or Photc before the For the first test , that is , the one by means of photographs taken or after the eclipse , I had intended to have taken several series , but fortune did not favour me .
We reached Prickly Point on a Tuesday .
By the following Sunday , great deal remained to be in the way of final preparations and adjustments , I was able to take my first .
This left me six days before the eciipse in which to make the final arrangements , to obtain that experience so necessary when .
in a new climate , and , in my case , with an instrument with I was not very lniliar , and also to obtain , if possible , one or more series of for the test in question .
I may here remark that , before starting , Dr. had given me eyery possible assistance and advice , without which I could hardly have undertaken the work ; and I also spen seyeral mornings in London in ctising with the , but from want of time I had to entrust a great deal of the troublesome preparation at home to Mr. , to whose careful attention at this period I owe a great deal .
As already remarked , the weather was very unfavourable , and on several photographic work was an impossibility ; on others I could only get casual photographs in the intervals between the tropical showers , or in gaps in the cloudy sky .
For two days after the the weather was also unfav urable .
In fact it was only on the before the eclipse that I succeeded in a series of about noon in both positions ; but , the rapidity of change in form of the corona being an unknown quantity , the the interval the more valuable would the photographs be .
The sky on that day was variable , but generally clear\mdash ; as clear as it ever was during our stay in the West Indies , but not , I think , as clear as it often is in England .
Soon after coming home , Mr. WESLEY , who has had great erience with regard to of a milar nature , very kindly undertook to examine this series .
Several of the } ives exhibited col-like markings , but Mr. WESLEY could not find any detail on any one of the which was confirmed by ) others .
This was done before he had had an opportunity of seeing the photographs of the corona taken during totality .
Had he succeeded in making a drawing of the supposed corona , its comparison with the true corona would have therefore afforded a valuable test as to its genuineness ; under the tances t was , of course , impossible .
WESLEY also informs me that he thinks that " " the photographs taken on the same side are as little comparable with each other as with those taken on the opposite sides , \ldquo ; which , as far as it goes , indicates that the irregularities are due to atmospheric rather than to instrumental causes .
My own opinion on the above points coincides with that of Mr. WESLEY ; but I was especially glad to obtain his assistance , because of large experience , and also because he had not at that time seen any photograph of the corona .
4 .
liesults of Test , or raphs taken the Eclipse .
The second of the two tests could be applied by taking photographs either the ipse , before or after totality , or sufficiently near first or last contact for th ' Moon to be still f , he corona .
I succeeded in taking over twenty PJotograp during these periods .
Many of these , developed , showed what I may a false corona , that is , an increase of density near the Sun , including the between the cusps , and therefore in front of the Moon .
In none of these can the limt of the Moon be seen ainst the corona as a background .
Besides subsequent exami lation , I watched these photographs very carefully during development , without result I mention this circumstance because they appear to have gone back considerably , that , in instances in which the air glare was clearly visible before fixing , it now di scernible .
I have also searched for any trace of a remarkably ] arge prominence in the plac I kn ew it should be found , but without result .
This prominence was hardly , by the Moon after totality , that is , during the period in which nearly al my phot ographs were obtained .
l'rominences are certainly more actinic than the corona we should erefore expect them to appear on the plate if the corona is obtain .
able by this method .
However , if the air glare increases much rapidly ir intensity than the corona does as the is approaclJed , this argument is not sound as the prominences then be more oyerpowered by sky light than the outer of the corona would be .
F ) .
Pl , It will ) observed that the two experiments or tests just described were made before and after the eclipse , and during the partial phases .
during totality the trument was not idle .
The following was ) eprogramm which had laid for myself the 3 minutes 50 seconds available:\mdash ; During the first l1linute a photograph was to be taken with the prismatic camera After that , six plates were to be exposed with the \mdash ; fonr chloride with the same length of exposure as that , during Sun light , and two with exposures of 5 and 10 } ) ectively .
This mme could not be followed exactly .
Immediately after I had com. menced exposing the prismatic camera , I looked up and found that a light cloud drifting across the rona .
The sky became clear again in about 50 seconds .
I anxious not to take any with the coronagraph during the : the prismatic camera , for fear of vibration ; but , as nearly a minute had been lost something had to be sacrificed , and I decided to take some of the raphs with before the cap on the prismatic camera .
I do not think that the suffered in consequence , at all events I obtained all the plates I ha , allowed in progra1nme .
TOTAL ARE AUGUST 29 , 1886 .
rfhe instantaneous rapbs of the corona were complete blanks , proving , think , that the exposure had been far too .
I developed them with the same solutions , and for at least the same length of time , as when developing , immediately beforehand , some of the ] ales exposed during the eclipse ; the instrument in the same condition as before and after totality , when successful photographs were taken .
These circumstances are worth mentioning , to show that I did not fall into some of the commoner traps with which the photographer is surrounded .
Three of the showed of , which was probably only due to the lenotho time I allowed them to stay in the developer .
The long exposure photographs were not taken any special object beyond that of obtaining record of the corona .
The plates used were bromo-gelatine dry plates prepared by Capt in , and I used the ordinary alkaline development .
The extension of the corona shown on these plates is not very great , and they show signs of vibration ; they have , however , I hope , been of use to Mr. ] in his drawing of the inner parts of the corona .
As my object was to get instantaneous photographs , these plates had to be taken without the automatic shutter ; the shutter had , therefore , to be worked by hand , and this probably caused the vibration .
It may , however , have been caused by a puff of wind ; and on another occasion I should take fiar greater precautions ainst this danger by surrounding myself with canvas screens in all exposed directions , and as high as possible .
6 .
Conclusions .
Returning again to the consideration of my observations with regard to the special uses of the raph , it will be seen my results are adverse to the possibility of obtaining photographs of the during Sun light with this instrument .
It is , however , I consider , by no means proyed that the method is impossible , for t , here are several reasons why this trial should not be considered collclusive .
( 1 .
) The atmospheric conditions were very unfavourable .
The air was fully charged with moisture , and on the morning of the eclipse the sky was certainly of that dark blue which , no doubt , atmospheric purity .
It was slightly hazy , and not , I think , as clear as an English blue sky .
About a minute after totality I noticed a halo with prismatic colours round the Sun\mdash ; an indication , I presume , of suspended matter .
( 2 .
) The Sun was at a low elevation during the eclipse , and the station was only about a couple of hundred feet above the sea .
Both these circumstances , no doubt , increased the air glare .
( 3 .
) Professor THORPE'S observations at this eclipse show that the light from the corona was not so bright as on other occasions .
This also appears to be the general impression amongst other obseryers who had seen previous eclipses .
If this was due to the corona not being so luminous , the opportuniby was , no doubt , an unfayourable one independently of the state of the atmosphere .
But this effect is not to be CAPTAIN L. DARWIN , DR. A. SCHUSTER , AND MR. .
W. MAUNDER distinguished from the light ) .
diminished by an impure or dense atmosphere .
Experiments oaring the light of the corona with Sun light immediately before and after the eclipse would be necessary to settle this point .
( 4 .
) The exposure too short .
1 feel confident that at Grenada an alteration in this respect would not have materially altered the result ; but , even with the atmosphere in its condition , I think a larger exposure should have been .
Before leaving England I had come to the co1lcltlsion that the right length of exposure would be that sufficient , and not much more than sufficient , for the air glare to produce an effect on the plate .
I thought that the corona would be more readily visible if the background of the sky faintly developed on the negative ; and in this opinion I think Dr. HUGGINS would have agreed , as his photographs seem to have been treated in that way .
I found my instantaneous shutter , when set at its slowest rate , gave required result , and I was the time quite contented .
I have , vever , very carefully reconsidered this subject , and now come to the conclusion that this a mistake .
The question to be settled is\mdash ; At what photographic density do we get the greatest ratio of small changes of shade in the negative , as seen by the eye , compared wit , the changes in luminosity of the objects photographed ?
Captain gives some results in the form of curves in his Text-book of Science on , which , if modified to suit this , would appear to indicate that the is greatest when the photographic density is about one-third or half way between white and the deepest shade due to the full development of the image of a bright object .
If this is the right interpretation\mdash ; and I believe it to be then it is evident , in order to have the best chance of photographing the small difference of shade between the sky and the corona , I should have given an exposure which would have allowed me to bring up the sky to what I may call a third or a half full density .
In order to have done this I should have lengthened my exposure .
Acting in accordance with the advice of Dr. , I adopted slow development with weak solutions .
I could not have followed a higher authority , but I must confess that I have some doubts whether a longer exposure with a quicker but wellrestrained development might llot have produced batter results .
The best time for watching for faint outlines is during development , and with a well-exposed plate we have an opportunity of observing the image in every stage .
All these considerations appear to me tu point to longer and more varied es than I gave , together with a well-varied development .
My conclusions are , therefore:\mdash ; ( 1 .
) That my results do not prove the impossibility of photographing the corona in Sun light .
( 2 .
) But they prove that under certain circumstances the light of the corona is not * See p. .
The ordinates of his c , nrves represent the shade as measured by the proportion of black to white ; the scale of shade as measured by the eye is somewhat different from this .
The abscissae represent the length of exposnre , which I to be eqnivalen- to differ nces of luminosity .
ON OLAR E OF AUGUST 29 , 1886 .
sufficient to produce effect on dry ) lates with an which is sufficient for the air glare or false pointed out to me\mdash ; and I think \mdash ; that the sky or background to the Sun , as shown on ] , is due to thses : , the light of sky , , secondly , the reflected from the interior of ] instrument .
Dispersion of light from the its would shed a over the whole surface of the exposed plate ; whereas the sky is brightest the Sun , and ot quibe uniform or regular in its appearance .
Now , on several of that I took there is a perfectly uniform oyer the whole sky , with no apparent increase of brightness near the liJnb of the Sun .
This constitutes the only ) I seen ; because , if the bove v correct , it indicates I not rivcn a sufficiently exposure to the sky round the , and giyes hopes that the sky I did obtain might be diminished by alterations .
7 .
on the of ' of th Light .
But there are certain considerations which appear to me to that a practical method of photographing the ) rona during Sun light is not likely to be obtained .
If I am right in considering that the increased density round in these is a true picture of the sky , and not due to diation o. internal reflection , then it will be seen this sky is not uniform round the Sun , that tlJe f of the corona thus would always be difficult to distinguish from the of trnc corona , TJld vould be liable to be llistaken it .
It is quite possible that in the of the be visible when photographed than seen , and tha , these nges m be as sudden as that from blue sky to clond .
At high altitudes difficulty would no doubt be lessened , but it ] , I think , always be source of trouble .
If , however , these ' in the false corona are llot due to ularities in the , I can only say that the instrumental defect which causes them is s ) very difficult one to In considering the utility of the of the corona taken in Sun , if obtainable , effect of the beinoo .
superimposed 011 the much denser picture of the air glare should ) carefully considered .
In the first place , it is to be observed that the ir or light of the sky incl.eases rapidly as the Sun is ched , cTnd that the ligh of the corona also in a similal manner .
This will , I think , cause the corona to be uished with in ) in Sun light .
not be so much the when ccnsid e ) pt ) shade in the corona , but , in so the of the corona is guished by gradual charlge of , it certainly will be .
In the ing , which have no numerical significance whateve , and which are merely inten ded to illustrate my argument , I have assumed the corona to be only from 10 to 15 tit less than .
In fig. let the lines round the Sun represent lines of equal intensity corona ) to be non-existent .
In fig. 3 , let the Sun be the represent lines of equal intensity of corona light .
Place one on the other , lines of equal intensity of the combined of corona .
glarc will be shown in , in which it will be seen how materially the of the corona ) to be altered .
It is to be noted also that an alteration in the of the here by altering the ratio of air glare to corona light would to alter the form ] of the corona as shown on the Sun light photographs .
For the of folloning discussion of the effect of combining the sky and , ib is assumed that particular difficulty does not exist ; that is to say , that the air glare is uniform , or , what comes to nearly the same we a radial inequality of the corona .
In dealing nulnerically with this ) , very little trnst must be placed in figures used ; but , however ulty t be , they may , at ] ] events , help to make the consideration of the subject rn Captain ABNEY informs me his experiments with regard to sky light when compared with the results obtained by Professo THORPE at this eclipse ) that at 30 ' from the Sun 's the corona , as see1l at Grenada , was about 60 times less brilliant than the air the same distance from the Sun 's limb , as seen at South Kensington under the most ble circumstances , that is , with the darkest sky measured .
As the sky on this occasion at South Kensington was probably darker than the sky at this latter place was probably greater than 1 to 60 ; but , on , wibh a arer sky the corona would have been brighter and the ratio .
Hence the ratio of 1 to CO is aboul half way between the ratio under nces South Kensington and the at Grenada near the time of the cclipse .
the ) oses of argument , let it be assumed that ab the of the light of the corona to the fit of the sky at 30 ' from the Sun 's limb ils 1 to 50 , a , if the above figures are correct , no doubt considerably less than the true informs me that experiments he has made show that an abrupt of cent. in ) derlsity of a photograph is about the mimmum of shade that seen by the eye ; or , in other words , that a photograph may be in diHerent shades can be used .
Tt has ] been shown to obtain the best results a photograph at a half or ON TOTAL SOLAR LCLIPSE AUGOST 29 , 1886 .
a third full density should be obtained .
Let us assume the most vonrable circumstances , and that the development the ) of the nnder consideration up to full density .
Then , if there on lie 200 ) in any photograph , there are , , only 100 possible this which for convenience may be represented by 100 .
Let the light which caused this shade be also represented by 100 .
Now , if the shade on the varied proportionately with the light , each of the 100 parts of light would be represented by one shade on the photograph , and a change of light less than 1 per cent. would not produce a change of shade on the photograph which would be visible to the eye .
But there is no doubt that under certain coltditions he shade on these photographs varies proportionately rapidly than the which ib represents , and , from Captain ABNEY'S curves , before mentioned , it is not quite a fanciful position to assume that under the most favourable nces it ries twice as rapidly in parts of the photograph ; that is to , that a of per cenb .
in the minosity of the object is the minimum change which would be visible on the photograph .
Now , in a photograph taken in Sun of the sky at 30 ' from the Sun 's limb , the 100 parts of light will consist of 98 of sky light and 2 of corona light .
sky light being assumed to be constant , the minimum visible change of per cent. in totad light must be due to a change of 25 per cent. in the total ] of the corona ; whereas , with a photograph taken during a total eclipse , on similar suppositions , a of per cent. in the light of the corona will produce a visible of shade on the photographs , or , in other words , the totality raphs will show 50 times as much detail the Sun light photographs .
It will be observed that , for the purposes of comparison of the photographs taken in Sun light and totality , neither an error in the assumption of the shade to which the raphs were developed , nor in the assumption of the ratio of of light to chance of shade , nor in the number of possible shades in a photograph , would viti te the esult , as they would apply equally to both cases .
It therefore appears to me most that , under nces , by whatever ratio the air glare is brighter than the corona , in very nearly the same ratio will the detail of the corona be obliterated in a photograph taken compared with one taken during a total eclipse ; tha6 is , unless a cf shade in the COl.ona were considerably than 50 as abrupt as the least nhangs visible in the totality raphs at Grenada , it would be invisible in the photographs taken there in Sun light .
This rally leads to the question\mdash ; Are there abrupt of shade in the corona , or does the light diminish ) dually as the from the Sun increases As far as I ] ) seen , an increase of exposure of the botality regularly brings with it an increased extension of the corona , as } ) that is to say , that the light of the corona gradually diminishes in intensity ) the Sun outwards .
The detail of the inner parts of the corona soon gets obliterated as .31 ItWIN , DR. W. MAUNDER the increases , llows t detail is due to delicate rather than to in sity .
conclude , of } light of hc sky on ) raphs taken during ) , when the conditi o are at comparable with those at , would be to entirely alter the of the corona , in so far as that form is visible by gradual of ; and that , with regard to abrupt of luminosity , they would have to be exceedingly well marked before they would bc the Sun light photo.oraphs , whereas , as a rule , the form of the corona is indicated by delicate Thus , the possibility of pbotographing the corona in Sun light appears to depend on the to which c high altitudes , or other localities , more favourable atmospheric conditions be found .
Witb dearer air , both the Sun and the corona will be hter tlJd the glare less , but to what extent this may be hoped for I do not .
It is to be observed , however , besides glare we have to deal yith internal cflcction , which increases with ) the increased brilliancy of the Sun .
If my exposure , short as it reality sufficient for the internal reflection to produce an on the plate , with longer and a brighter Sun , but und-er otherwise lllol'G vourable conditions , this might become a serious source of trouble .
The limb of the Moon undoubtedly been seen for some time after totality ( not at no doubt the atmosphel.ic conditions were too unfayourable ) , and , if visible , it should , I think , be possible to raph it .
But we must be careful in using this as an rgument in your of the possibility of obtaining photographs of the corona in Sun light .
In the first place , this ) henomenon has generally been observed within a few minutes of .
For example , M-r .
S. P. , in 1878 , at an elevation of 14,000 , was able to observe the of the Moon for 4 minutes 12 seconds the ance of the Sun without taking any utions to shield his eyes .
* Mr. H. H. has kindly calculated for me that this time only of the Sun 's exposed ; and , as the outer rim of the Sun has only about the ] of central , it may be safely assumed the total light emitted was not more one-twentieth of full Sun .
If the air glare varies the total emitted , and if , } been assumed for the purposes of argu- meant , the ratio of air glare to corona were.50 to 1 in full light , it would only be to 1 llnder these circumstances .
This would , of course , make the limb of the Moon much more readily visible inst the corona .
This argument does not apply to the cases where the limb of the Moon has been seen long before or after totality , but as long as the cusps are fairly sharp another circumstance must be considered .
In all the photographs which I took during the partial eclipse in which there is a false corona , it will be seen that it is distributed rouud the nlass of the Sun as a centre , and does not uniformly fringe its outline .
The cusps often protrude out of it , and quite free from it .
Hence , the corona will be less oyerpowered by the air glare * Nature , ' November 18 , 1886 .
ON TOTAL SOLAR CLTPS OF ' near the cusps than at other parts of the limb of the Snn .
In the second place , in the se of the Moon ) seen the corona , we are dealing with a sudden change of shade , and not one , as in the case of the corona .
This , as I have , I think , proved , would make it more easily raphed .
It would have been .
to have ined photographs of the limb of Moon during the partial eclipse ; but , unless they were obtained when a large part of the Sun exposed , it would have done little the possibility of getting photographs of the corona in full Sun light .
It is to be observed that many of these remarks indicating the imrobability of obtaining results b.y this method do not hold good in the case of exceptionally well marked features of the colona , where the changes of shade really abrupt , and especially if they are radial in their direction .
, the first successes tlJat we should expect to obtain would be in such cases as hat of the " " remarkably formed rift\ldquo ; in the corona of 1883 , to which such a distinct likeness is reported to have been seen in photographs taken in Sun light by Dr. in England .
As for .
to obtain any photographic records Sun light of the outer limits of the corona as visible to the eye during totality , it is , I think , out of the question , as the considerations will show , although the actual quoted may be considerably in error .
Judging from my own results , 10 seconds with a bromide plate would appear to be too , -mall an exposure for the extreme limits .
Assuming that the exposure I gave ring Sun light was one-fifth of a second , and that the chloride plates are 30 times less sensitive than the bromide plates , then it would take 1500 times the I gave during Sun light to rapb the outer parts of the corona during totality .
On this point it is instructive to compare drawing from the photographs with Captain ARCHEB'S direct drawing , probably one of the most accurate ever taken .
* 8 .
Should the be Tried ?
conclusion , the questions naturally arise : Is the experiment worth trying again , and , if so , under what conditions ?
As to the latter point , experience to that :\mdash ; ( 1 .
) It should be tried at a station at great elevation above the sea .
( 2 .
) It should be tried when the Sun is nearly vertical , during an eclipse of the corona by the Moon , Mercury , or Venus .
( 3 .
) The exposure must be more varied and longer .
At present I see no reason why increase of exposure should not simply increase the density of the air glare in the same proportion as that of the corona , thus no advantage by the change .
Unless the air glare and internal eflection can be so diminished as to make an exposure of one second with an instrument like that used by me a possibility , I fear Appendix to ABNCY and Professor 's paper on the Photometric Inteusiby of he Coronal Light during this eclipse .
no good are to be expected .
As to whether the experiment should be it is to be observed that the coron graph is admirably adapted for taking of the corona uring totality , for ordinary records ; for reflectors have several advantages over refi.actors for purpose , and the special mces of this instrument would all be more or less useful under circumstances .
Hence , another can easily be combined with what may be described as the ordinary routine work of an eclipse .
Atmospheric or other conditions are now certainly unfavourable , and time only can settle the question whether this is a permanent or a temporary difficulty .
* I should therefore recommend the experiment being made in connection with other work , but not before the next occasion when the path of a totad eclipse passes over high ground ( say , over 7000 feet ) can be conyeniently reached , and where there appears to be a reasonable chance of having a clear atmosphere .
But on all occasions it would , I think , be worth while arranging that some one , not necessarily a trained astronomer , should be detailed to observe for how long before and after totality the outer limb of the Moon can be observed .
This was not done at Grenada .
During partial eclipses a look-out should also be kept for this phenomenon .
Most of what I have said has been unfavourable to Dr. HLTGGINS ' method .
But there is very material evidence in its fayour , and , for the benefit of any one reading up this subject , I have given the necessary rences in a footnote .
VII .
THE PB,1SMAT1C CAMERA .
By CAPTAIN 1 .
Description Adjustments .
The instrument was the same as that used by Dr. cHUSTER , in Egypt , in 1882 ( see 'Phil .
Trans. ' 1884 , Part I and by Mr. LAWRANCE , at Caroline Island ( ' Phil. Trans , 1889 ) .
The camera has a lens of 3 inches aperture , and of 20 inches focal length .
There is slit , and the prism , which is placed directly in front of the lens , has a refracting of A photograph taken with this instrument of any small object which has a bright line spectrum will shcw several , each one to some is a curious fact , that the limb of the Moon beyond the has seldom been reported to have been observed long before totality during the last 15 years , although before that date it is said to have seen on sevcral occasions .
See ' Astron. Soc. Mem vol. 41 , 1879 , pp. ( 1 .
) " " On a Method of Photographing the Solar Corona without an Eclipse W. HUGGINS , ' Roy .
Soc. Proc No. 223 , 1882 .
( 2 .
) " " On some Results of Photographing the Corona without an Eclipse W. HUGGINS , ' .
Assoc. Rep 1883 .
( 3 .
) " " On the Coronn of the Sun.\ldquo ; \mdash ; Bakerian Lecture for 1885 .
W. HUGGINS , ' Roy .
Soc. Proc No. 239 , 1885 .
( 4 .
) " " Photographing the Corona withour Eclipse .
L. TRONVELOT , 'Observatory , ' No. 117 line in the spectrum .
Thus , in a photograph taken during totality , there will be several distinct images of each prominence , and , in measuring the distances between them , we are , in , measuring the distances between the corresponding lines in the spectrum of the prominence .
A numbel of sInall ) rominences close together gives the appearance of a segment of a circle , while the continuous spectrum of the corona shows as a confused band across the plate .
This will explain why it is difficult to obtain a clear interpretation of the various appearances shown in these raphs .
The accurate ad.justment of the prismatic camera in its present form is difficult .
In order that the photographic plate be in focus for a large range of the spectrum , the camera is provided with a back which can be inclined to the optic axis of the instrument .
In order to find the proper angle and position of the plate , Sun light which had passed through a collimator accurately adjusted for parallel rays was thrown into the instrument .
The spectrum thus formed was focussed on the plate , and the instrument , therefore , adjusted for parallel rays elltering in the direction of the optic axis of the collimator .
The position of ] blue and red end of the spectrum was then marked on the ground glass of the camera , and the collimator removed .
If the instrument were then fixed on the equatorial , it would have been in focus , provided the camera were pointed relative to the Sun.as it was before relative to the collimatol .
This was the awkward part of the adjustment ; the camera was turned until the spectrum formed by the Sun occupied the position marked .
The spectrum , of course , .
blurred , owing to the finite size of the Sun , but the adjust1nent could be made approximately , and the result shows that definition was obtained .
The slant of the plate essential for securing fair focus for rays of different refrangibility must be detrimenta ] to the simultaneous of diflerent parts of the Sun 's circumference .
The instrument can , therefore , never be a perfect one , but it affords valuable information in every eclipse in which a nnmber of sharply defined prominences appear .
2 .
Observations and Results .
one photograph was taken during totality .
The estimated at nearly two minutes , but for about fifty seconds the corona was nearly obscured by clouds .
It has been explained how other photographs with another instrument , but on the same mounting , had to be taken at the same time as , and it , therefore , came about that the plate shows signs of a shake .
The images of five prominences can be distinguished , but of these only two were ured , which will be called prominences Nos. I. and II .
No. I. is eviderltly the large prominence on the north-west of the Sun which forms such a marked feature in some of the photographs .
There are also two segments of rings in a different part of the plate .
The instrument used for measuring the distances apart of the images of the prominences was very kindly lent to me by the Royal Observatory .
Greenwich .
It was for uring t ) of Venus raphs , and consists of two parallel micrometers , rigidly fixed together nd capable of sliding in one direction .
With one lnicrometer the image is observed , with the other the is read off on a scale .
Two of the of all the prominences are especially densL3 , from their position , by reference to previous results , these can be ecognised without much doubt as being due to the Solar and lines .
Taking these as a basis , three other lines , , and , can ] ' nised in the images of prominences I. and II .
with gree of certainty .
The accuracy of this assumption can best be tested by seeing nearly the wave-lengths of three of these five lines with the wavelengths of three known lines , the assumption that the two others are .
For this ) and were taken as known yilh wave-lengths and respectively .
Then the wave-lengths of the other lines were calculated on the assumption that distances lneastlred on the plate varied inversely as the square of the wave-length .
This last ssumpti o is , however , known to be inaccurate , and a correction had to be found and applied .
This was done in the following manner:\mdash ; A photograph of the olar spectrum was obtained with the prismatic camera , but -placing a slit and collimator in front of the prism .
The distances apart of known lines of wave-length as near as possible the same as those believed to have been found in the spectrum of the prominences were measured ; the wave-lengths of these lines were then calculated as before on the assumption that the same two lines were known .
We , therefore , get in this Solar spectrum the known and the calculated wave-lengths of lines ; the differences between the two give corrections which , if applied to the calculated wave-lengths of the lines of the prominences for the same part of the spectrum , should give their true The results thus obtained are recorded in the following Table:\mdash ; three other prominences showed on lie in the and lines , and llot , therefore , measured .
It already been explained how it came about that the instrument was shaken ON THE ] TOTAL SOLAR ECLIPSE OF AUGUST 29 , 1886 .
during exposure .
The result of the shake is that there are two distinct sets of images , corresponding to two positions of the instrument .
The second set of images is far more dense than the other , and was alone used in the measurements .
in this case are lengthened out in one direction , either by vibration or through imperfect movement of the clock , and prominence No. I. presents a curious appcarance in consequence .
The image is wide , and is also separated into two parts by a band of white , giving the appearance of another shake , but a comparison with other photographs shows that it is in reality due to the peculiar formation of this prominence .
In this instance this white band was used in measuring the tances of the images .
The first set of images is much more faint than the other , and , as far as the prominences are concerned , is only visible as the and lines of No. I. prominence .
The existence of two segments of circles has been mentioned ; these were extremely puzzling at first , until it was noticed that they coincided in position with this at first unobserved faint set of images ; that is to say , if the circles had been completed they would have run through the bases of these prominences .
No other rings corresponding to other lines can be seen .
The and lines are in all cases by far the strongest , and it is , therefore , not surprising that these should be visible when the exposure is not sufficient to produce any others .
But it is remarkable that , in the second or stronger set of images , the and lines of the same prominence should not have rings corresponding to them , for they are far more dense ; that is to say , it would appear at first sight that the same object produced an image during a short exposure , and yet failed to produce one during a considerably longer exposure .
It is probable from its position that this bright inner circle was covered by the Moon during the period when the one set of images was being formed , but was visible during the other exposure , which is thus presumed to be the first .
It is also to be observed that the part of the limb where these rings occur was that which was least obliterated by clouds at the commencement of totality .
VIIr .
THE SPECTROSCOPIC CAMERAS .
By ARTHUR SCHUSTER .
1 .
Resolving Power of the Instruments .
In order form a clear idea of the amount of detail we may expect in a photograph of a , it is necessary to enter further than is generally done into the details of the construction and adjustment of the instruments used .
Lord RAyLEIGH has given the theory of the spectroscope for the case of a sufficiently narrow slit .
It is easy to extend his results so as to apply also to the case of slits of which the width cannot be ected .
If is the horiz ntal of the beam entering the prisms ( the refracting edge being supposed vertical ) , that of the beam leaving the prisms , the equation Si/ r defines the angle through the emergent beam is turned , owing to a small rotation Si of the incidental beam .
If is CAPTAIN L. RWIN , DR. , AND MR. .
W. MAUNDER the width of the slit , the focal length of the collimator , the geometrical image of the slit will subtend an angle at the centre of the lens of the telescope or camera , and to this we must add the width of the diffraction band of an indefinitely slit in order obtain the actual lgular width of the image .
The angle between two rays having a refractive index differing by S , is on leaving the prisms , as shown by Lord RAYLEIGH , fS/ r , where is the gregate effective thickness of the prisms .
This angle must be equal to the angular width of the slit in order that a double line may just be resolved .
We then obtain , when stands for means the ] subtended by the useful horizontal aperture of the collimator at the slit .
If SX is the difference in wave-length of the two lines , we find @ In order to define the actual possible power of a oectroscope , it is conyenient to distinguish between the resolving of a spectroscope and the purity of a spectrum , ( to troduce the following definitionls ( see 'Encycl .
Britann otvcr of spcctroscope part of the spectrum is that the of two lines differing by the of their own -lengfh .
The umit of of a spectrnm is that purity allous the of two lines differing by the of their -length .
We speak , therefore , of the resolving power of a spectroscope , which is a constant for each instrument , and of the purity of a spectrum , which differs according to the width of the slit .
Both resolving power and purit .
vary inversely as S/ , and are equal to unity if that fraction is equal to .
Hence , as from RAyLEIGH'S investigation for narrow slits , @ or , from ) aboye equations , , or .
Here stands for the purity , and for the resolving power .
They are both equal if the slit is indefinitely reduced .
If the is weak , narrowing ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29 , 1886 .
the slit will not diminish the intensity until the width of the slit , as calculated from geometrical optics , is equal to the width of a ction band .
In case and the purity is half the resolving power .
Hence we , when light is of consideration , only be able to make use of half the of a ectroscope .
Whether slits which are wider than the here given show the fainter lines on a photograph more distinctly owing to their increased width is matter .
in vestigration ; but it would , perhaps , be well in a future eclipse to adopt that width of slit which makes Two spectroscopes were to be used on the present occasion , and the following numerical data will allow us to calculate their power and the purity of the spectrum which they gave : Spectroscope Focal lens of condenser to form image of the COl.ona on the slit . . . . . . . . . . .
Focal of collimator . . . . .
Aperture of , , Width of slit . .
, 003 Greatest thickness of prism Focal length of camera lens The beam was limited by the aperture of the collimator lens .
As the full aperture of the collimatol was used , we must put 6/ 106 and ; hence , for a wave-length of With the width of slit used , we have , therefore , made use of the fifth part of the resolving power .
To calculate the tter , we have , in the firsl place , to consider that the effective thickness of the prism used was only cm .
, as the aperture of the collimator lens was not sufficient to use the full width of the prism .
The dispersive power of the glass was sufficiently near to that for which Lord RAYLEIGH has calculated the value cm .
as the necessary thickness to resolve the sodium lines .
As , according to our definition , a resolving power of 98 is required to separate the sodium lines , we may take the to give a resolving power of 1 for each centimeter thickness in the neighbourhood of the sodium lines ; it will inversely as the third power of the wave-length , and will , therefore , be for each centimeter in the violet .
We have also to reduce the resolving power in the ratio of 10 : 12 , as the beam is limited by a circular aperture .
We have , therefore , finally in the violet The purity was such , therefore , that near the wave-length , two lines differing by the six-thousandth part of their own wave-length should be resolved .
Spectroscope II .
cm .
Focad lens of condenser to form image of the corona on the slit . . . . . . . . .
Focal lens of collimator for G. . . . . . . .
telescope , , . . .
, , Aggregate thickness of two prisms Width of slit . . . . .
01 The beam was ited by the prisms , the aperture of the collimator and camera being more than sufficient to fill the prisms and transmit the beam .
The calculation is carried on as before , except that we must now take the useful instead of the full aperture of the collimator in order to find .
As the prisms were placed in minimum devi tion for , this can be calculated , and we find it to be cm .
; this determines ; or , roughly speaking , the purity in this case was only the twentieth part of what it might been if the slit had been narrower .
The spectroscope was intended originally for the outer parts of the corona , and the slit was , therefore , intentionally kept ather wide .
There was at the time no means of measuring the width of the slit , and I did not ] how much of the resolving power was lost by widening the slit .
The full resolving power in the violet was 32 , therefore the purity realised 2 .
Adjustments of Instruments .
In order to be able to differentiate between the spectra of different parts of the , its must be thrown on the slit of the spectroscope ; and for that purpose the plane of the slit must be placed in the focal plane of a condensing lens .
This was done as follows : an auxiliary telescope was focussed for parallel rays , and the slit of the spectroscoI ) looked at from behind through the condensing lens , which was moved backwards and until a sharp image of the slit was seen in the telescope .
The method is the same as that usually employed for adjusting the collimator of a spectroscope ; only that , instead of observing the slit through the coll iating lens , we observe it through the condensing ] ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29 , 1886 .
The backs of the cameras employed could be tilted so as to have a larger portion of the spectrum in focus at once .
It would be worth while to determine by calculation the best shape of the camera lenses in order that the foci of the different rays shall be as much as possible in a straight line .
An achromatic lens is not as good for the purpose as an ordinary lens , but a combination might be found which would give better results thaI ) the ordinary lenses .
3 .
Results .
The general appearance of the spectrum of corona , shown on the plate exposed during the whole of totality in the spectroscope described above as No. II .
, is given in Plate 9 , fig. 1 .
It has a streaky appearance , consisting of a series of bands stretching from the ultra-violet into the green , and is crossed by the bright lines of corona and prominences .
The horizontal bands are parallel to the dust ]ines , but they must be chiefly due to the rent brightness of the portions of the corona cut by the slit .
This is shown by the fact that they are wider and more diffuse than the dust lines shown on the reference spectrum , and that we are able to recognise in the photograph of the corona itself the brighter portions cut by the slit .
A few dust lines do actually appear , but can be easily distinguished .
Fig. 5 , p. 326 , shows the section of the corona cut by the slit .
The left-hand part of the drawing is a tracing of the streamers in the corona .
The line AA ' is the line along which I had intended to place the slit .
It is parallel to the declination circle , and tallgential to the western limb of the Sun .
It was , however , impossible to place the sIit accurately in the position aimed at , and I think that , for reasons presently to be mentioned , the slit really ran along the line .
It must have crossed one of the prominences , for the prominence lines are very strongly marked , and this prominence will allow us to fix exactly the position of the image on the slit .
The right-hand part of the drawing gives the distribution of light and shade in the spectrum of the corona near the Solar line reduced to the same scale as the outline of the corona .
The images of the prominence appear along the line marked P. If we go from this line towards the southern side we find , in the first place , a strongly marked band , due no doubt to those parts of the corona which are closely adjacent to the Sun 's limb .
This band and the next are separated by an interval in which the photographic intensity is very small along the spectrum , and I believe that here the slit must to some extent have overlapped the image of the Moon .
Next comes a band , marked ( 2 ) in the drawings , in which all corona lines appear very strongly .
Here the slit must have crossed the most intense streamer of the Solar corona , as will be seen by comparison with Mr. WESLEY'S drawing .
The most southerly band ( 1 ) coincides with the section of another streamer .
It was the light interval between ( 2 ) and ( 3 ) and the portion of this band ( 1 ) which induced me to fix on the line BB ' as the most p robable- line of the slit .
ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29 , 1886 .
scattering of minute particles , only the smaller part can be owing to this cause , as otherwise the maximum of actinic intensity in the spectrum would be displaced towards the violet and not towards the red end of the spectrum .
The faintness of the Fraunhofer lines is further evidence in the same direction , and also makes it probable that there is not much matter round the Sun sufficiently large to scatter Sun light without polarising it .
If we now turn to the lines shown on the photographs , our attention is at once arrested by the lines .
They form in this , as in previous eclipses , the most intense feature of the corona .
In the Eclipse of 1882 , these lines were so bright that the atmosphere scattered enough of their light to make them appear over the disc of the Moon and at a considerable distance outside the corona .
In the Eclipse of 1883 , the same lines appeared reversed over the lunar disc ; on that occasion , therefore , the light scattered by the atmosphere in front of the must have been derived from some source showing absorption lines .
In our photographs the same lines end sharply with the corona , and we must con clude , therefore , that , in spite of the unfavourable atmospheric conditions , there was but little light scattered by our own atmosphere in the neighbourhood of the Sun .
This is confirmed by the photograph of the corona itself ; for , while on previous occasions the sky light formed a bright round , on which the shadow of the wire stretched across the camera showed distinctly throughout the whole plate , the two needles do not show at all on our plates , even on those which had the longest exposure .
The analysis of the light seen in front of the Moon , or in the neighbourhood of the corona , seems to give us important information on the general light emitted by it , for in three successive eclipses bright lines , dark lines , and no lines at all appeared on the photograph .
The inence lines on the plate present a curious winged appearance towards the violet .
At first sight it might seem that this is due to some defect in the spectroscope , such as a reflection on the inside of the slit ; , after careful consideration , I do not think that this is possible .
One edge of the wing is perfectly sharp , and almost as black as the prominence line itself The and lines in the corona seem perfectly sharp along their whole length , and do not show anywhere a trace of fuzziness such as ] be produced by a defectiye slit .
The reference spectrum shows the finest lines with perfect definition .
The wing appears in all prominence lines , and , if it is not due to an instrumental defect , must be owing to a rapid motion of the matter forming the prominence .
I find that an approaching motion of about 247 miles per second in part of the prominence would account for the displacement .
According to YOUNG , such motions have been observed , though seldom .
The limb of the Sun on which the prominence appears receding .
Another fact of some importance is this , that , while the continuous spectrum on the northern hemisphere of the Sun is weaker than on the southern , the and lines are stronger .
The hydrogen lines and do not appear at all on the southern side , while I can , under favourable conditions , trace some distance on the northern .
It will be noticed from Mr. WESLEY'S drawing that the side on which these lines appear is the side which was rich in prominences .
A similar phenomenon was shown in the photograph taken in 1882 , when the hydrogen and calcium lines appeared strong in the corona on that side on which they were strong on the edge of the Sun , presumably in a prominence .
The explanation which first suggests itself cannot be the true one .
We might think that , if many prominences exist over one part of the Solar surface , the corona above , which undoubtedly does to some extent scatter the light falling on it , will show the prominence lines more strongly than other parts which have no protuberances near them .
No bright line could , however , be produced in this way , for the light scattered by the solid and liquid particles of the corona must have the same composition as the direct Sun light which reaches the Earth .
The particles are illuminated not only by the prominence light , but by the light from the whole Solar surface beneath them .
If we imagine ourselves placed near the Sun , the spectrum there seen cannot differ on the average from that which we observe here , and therefore the scattered light must also be of the same composition , and must show dark lines on a bright background , and ) bright lines .
Only one other explanation occurs to me .
We must imagine that sufficient matter is actually thrown into the corona from the prominences to show the lines .
This does not seem surprising if we look , for instance , at the large prominence on Mr. WESLEY'S drawing .
The life of such a prominence is short ; had the eclipse taken place a few hours later , it might not have appeared as a prominence ; but would it not have left behind it a sufficient quantity of calcium and hydrogen to show their characteristic lines ?
We need only consider the remarkable effects produced in our atmosphere by the Krakatoa eruption , and reflect that matter is projected into the Solar atmosphere on a vastly larger , to see that the coronal ectrum , as we observe it , may contain injected calcium , which only very gradually sinks back again into the Sun .
According to this view , we must separate the true coronal spectrum , which will be seen evenly all round the Sun , from the spectrum due to prominence matter , which will differ much in intensity during diflerent eclipses .
The strongest of the true corona lines has , according to my measurements , a wavelength .
It is slightly less refrangible than the calcium line which forms a conspicuous feature in the Solar spectrum .
That it is not identical with it is proved not only by the measurements , but by the fact that on the corona photograph the Fraunhofer line appears faintJy by the side of the corona line .
As the well known green corona line is always present in the spectrum of the chromosphere , it is very likely that this new line , which appears so strongly on our photographs , should frequently make its appearance there .
YOUNG , in his excellent book on " " The Sun gives , indeed , a line as one of about 30lines which appear there on " " very slight provocation It therefore seems highly probable the two lines are identical , and that the comparatively weak Fraunhofer line is a corona line .
ON THE TOTAL SOLAR AUCUST 29 , A faint Fraunhofer line has just been mentioned appearing in the corona spectrum .
ppearance of dark lines has been a source of trouble in the reduction of the raphs .
After all the measurements had been , it struck me that some lines which I had put down as corona lines were really only the intervals between Fraunhofer lines , and 1 to subject the to a further careful ination .
The effects produced by the .
of a spectrum of dark lines over one of bright lines is very ] icated ; especially , apparently , some of the weaker Fraunhofer lines can be traced , while of the ones do make their appearance .
I believe that this is due partly to the of and dark lines , bub to an optical effect of contrast .
A bright line shows black on the negative and is bounded on both sides by an apparently hter b ) is a well-known contrast effect .
The and lines , for instance , to be surrounded by a lighter , which follows the contour not only of the lines , but also of the wing by the qide of the prominence .
If , now , Fraunhofer line happens to be by side of a bright line , the ltrast is bhened , and both the bright and the dark lines appear more distinctly than they otherwise would .
This is the only simple way in which I can explain some of the appearances of the potorahs .
The triplet with a wave-length , and 40368 , which is represented in fig. 1 , Plate 9 , is a case in point ; the group of Fraunhofer lines at 4031 , about , is weaker than the strong Fraunhofer line 404 nd yet is much more easily on the eclipse photographs .
This I believe to be due simply to the fact that the two lines set off by contrast the group of Fraunhofer Such effects may , in some , have materially ected my surements .
I have made these as carefully as I could , and I believe that the great majority of the lines I have put down as corona lines are really such , but some of the weaker lines may be due to the optical effects I have just described .
Possibly the true spectrum of the corona may be still further complicated owing to the ause : base of the corona giyes us eyidently a strong continuous spectrum , and ib is possible that lines of outer corona therefore opear as dark lines in a bright background .
Captain ABNEY found some Fraunhofer lines reversed over the face of the Moon , while bhe band was absent .
The cause I have suggested may count for this .
All these tions show how careful we must be in interpretation of raphs of the coronal spectrnm .
I had intended to have a .
of all I see on the raphs .
Figs. 2 and 3 , Plate 9 , are specimens of certain portions of the spectrum on a scale 40 times the I had , however , to give up the work , I found it too ying to the eyes .
The length of spectrum represented in fig. 3 is in the origi taboLlt 2 mm. * The distances between the horizontal bands ) ( 'nified on a Yhtly drcnt scale in the drawings , so as to make them , while as before explained , real distances the violet .
3.30 APTAIN L. DARWIN , DR. A. SCHUSTER , AND MR. L. W. MAUNDKR By the photograph inst a sky , and exanlining it with a lens of about half an inch focal , I can see what I have tried to repl.esent in the .
But I could only carry on the work about an hour at time , and it was always followed by strong pain and , in the eye , lasting sometimes for days .
I am sorry , therefore , not to be able to accompany this report with a complete set of drawings .
The two specimens which I can give will , however , give a fair idea of the appearance of the lines 011 the photograph .
band reaching from a to a has a curious shape .
It is broad near the centre of field .
where it is widest , and there nearly covers the bright space on the least side of the band , comes out so strongly in phototaken on a small scale .
If it was of equal width hout , I should have taken it for ) band , coming cut by contrast between the Fraunhofer line and the group of lines at .
But the manner in which the line becomes thinner towards the outer parts of the corona shows that it must be a real band .
As we go from this band towards the less refrangible side , we come to very complirkings until we reach the prominence .
These markings I believe to be due to an overlap of Fraunhofer lines and COl.ona lines ; ether the difference in the on the northern or southern side is real or not I cannot be certain .
in is the conspicuous one in this part of the spectrum .
By its side we can trace the line .
The series of lines here seem to widen at the base of each band of the corona , and they are weak on the northern side .
I Dot been able to trace with certainty a difference in lines of the spectrum of different of the corona , except that already mentioned of the calcium and hydrogen lines .
The group of lines at 4076 ( fig. 3 , Plate 9 ) at first sight looks as if there was such a difference ; , we have here possibly only one broad band , and the lighter appearance in the centre ma be a defect in the photograph , or it may be due to a unhofer absorption line , which ought to be here , or , finally , to a reversal of a corona line against a hotter background .
The results obtained with Spectroscope I. are much less satisfactory .
The plate had a very narrow escape dI1ring development .
Owing to the hot damp weather the films were often found to detach themselves from the glass in the developirg bath ; but we not found any difficulty when the bath was kept cool by ice , when the film soaked with alum .
Captain DARWIN had kindly to do the whole of the photographic part of the work for me , and this , as well as for continuous assistance in other ways , I have to offer my sincere .
The plate in question was developed during the afternoon of the ecli p. For some time nothing appeared , and when at the image showed itself plate at once to frill at the edges .
It was only by repeated treatment with ) ice alum that Captain DARWIN saved the plate , but the image is very faint .
What this is due to , I cannot say with certainty .
The plate was one of Captain ABNEy'S red end plates , and probably was less sensitive in O-X OTAL SOLAR OF AUGUST , lb'86 .
the blue and violet than those used the other spectroscope , which also were supplied by ptain A. The oscope w in good adjust meant , the reference spectrum taken with it of the is perfectly .
If the developcould have been carried ltlrtl]er without , a better have been obtained .
The plate shows a cenlent of ctinic intensity towards the led in the spectrum .
We can on the plate the absorption the band and the corona line , and thus fix the extension of the spectrum .
The reference speclrum reaches from into the :md is very between H. It , in addition , a band in the yellow red , , however , is very faint .
The corona spectrum rins about , is between , and off very rapi in about .
It be barely aced between and H. The lines shown on this will be given the annexed Table .
I must now describe the method adopted to determine position of the lines which appear on the photographs .
In 1882 a reference spectrum had been taken on the same plate previous to totality , but that reference spectrum fonnd to be of little use in measuring the lines .
It had to be near the of the plate in order not to interfere with the spectrum of the corona ; and , to the curvature of the lines , we cannot directly the position of a line coming from the centre of the slit with one from near one of the order to re tely two spectra on the same plate , they must be in contact with each other , which would be impossible unless part of the coronal spectrum were sacrificed .
A reference spectrum on rent plate is as useful as one taken on the plate , if are taken to fix well the position of the plate in the holder .
All good phog.aphs of the corona have been taken therto s ) the calcium lines coincident with and , and we can take these always as the starting point for nents .
The distance of these lines from any unknown line will be the same in the centre of the reference } as in the centre of the coronal spectrum , provided no of the has place during deyelopment .
But , if has taken place , is no reason to suppose that it is unifo ] all oyer the plate , and it may , , be just as different in the reference spectrum and the coronal spectrum , if these are on the same plate , as if they are on different , plates .
Besides the and lines we can generally ' some other ture in the coronal spectrum , consisting either of prominence line or some wellmarked Fraunhofer line ; and in that interpolation becomes easy with the ] ) of the lines on the reference spectrum .
If to these considerations the danger of exposing to Sun ] even day light , a plate which is to be used during the ipse , merely for the of the reference spectrum , I think it be conceded that I was justified in .
the re ence spectrum on a ) is , which I pointed out before the eclipse , is not imaginary , is shown by the failure of the Carriacon the present hydrogen lines , and those in the ultra-violet : in the prominence spectrum as as and , and give us sufJicient to determine the of the lines ppear in the spectrum of the corona .
The measuremen}s were taken by means of a dividing machine reading to 002 mm. following giveH in centimeters the distance of three of hydrogen lines from , as measured on the reference spectrum and on the spectrum taken , the eclipse .
and appear as prominence , and their distance , from are practically the same on the two plates .
About there is more doubt .
There is a faint mark on the eclipse spectrum which I had inally taken to be the image of the prominence which I expected to find there , but which may easily be an accidental spot on the photograph .
The distance of this spot from is or the tenth part millimetre further than F. The discrepancy seemed to me to be too great , and on the photograph I found on the more ible side of spot a distinct absorption line , which , when measured , as distance from H. Considering the of the image near , the measurement seemed now to agree , as well as could be expected , with the reference spectrum , on the supposition that shows as an absorption line in the corona .
I conclude , therefore , that the distance of a given line from on the eclipse plate is the same as that of a corresponding line on the reference plate .
In order to find the wave-lengths of the in the corona I follows .
The distances of the following lines were measured on the reference spectrum from Fe . . . . .
4004.9 Fe . . . . .
4045.1 . . . . .
Ca . . . . .
4226.4 4666.5 . . . . . .
4860.7 ON TOTAL OF of and of to be , we can , from the observed positions , calculate by some interpolat.ion the of the intervening reference lines .
The formula I used is the usual where and are the two known leno t that to bc determinod .
is the distance measured on the plate becween and between .
The jghs thus found are approximate only ; but , by comp , ring the calculated values with found A ' and given alTove , we can construct a table of corrections for the reference li nes .
A curve was drawn ted wayelengths as abscissae and corrections as This curve found to be quite regular in shape .
The interpolation form u used to calculate approximate wave-lengths for the corona lines , and the ections were read off from the curve .
I have found this combination of interpolation by calculation by a graphical method to be very convenient .
In order to show accuracy which may be ined in this way I give the I left out originally the reference line ether .
There was , thel.efore , very large gap between the lines and .
The had to be , more or less , as a straight line between the points to the two lines .
the reference line , then , first like an ordinary corona line , thab is , finding its ) proximate wave-length by calculation and the cox.rection from the , I found as its wave-length , being very near the correct value .
The curve was improved by making it still further agree with this line , and , as there is no gap of equal magnitude in other parts of the spectrum , I concluded now the .
due to tion are negligible .
It is more difficult to gi an idea of tlJe possible error due to faulty measurement .
Most of the important lines were measured several times , and , the way in the -len gths , I should that error of near and 1 near in metres will not often occur in the strc ) lines .
The position of the weaker is , of course , more uncertain .
As the strong corona line had to be used as the starting point for the measurements of the second plate , considerable trouble was taken to find its position .
Perhaps the best idea of the accuracy which can be reached in such measurements can be got from a comparison of the individual measurements .
The following then are in centimetres , the distances from of the corona and of the calcium line , as measured on the corona plate and on the reference spec.trum respectively .
The measurements were taken at different times during the last two vears .
Those marked were taken by my assistant , Mr. ARTHUR CAPTAIN L. DARWIN , .
A. SCJIUSTElt , AND MR. E. W. Dist irom H. cium 1 ; Reduced to , the rence is , and , taking the wave-length of the calcium line as , I arrive at the adopted wave-length The taken with Spectroscope I. more difficult to reduce .
When the the stroncrc , corona line had been recognised , the distances of the other lines were measured , and the corresponding wave-lengths found by help of the eference spectrum .
All measurements of this plate were taken by Mr. STANTON , as my eyes could no longer stand the strain .
The measurements are very difficult , as the band which has left a mark on the plate is exceedingly narrow , and it is sometimes impossible to distinguish accidental spots from real lines .
Mr. STANTON has found a number of apparent absorption lin es in the spectrum .
Most of these agree with strong FrauIlhofer lines , and this justifies the belief that his measureY ents are sub , stantially correct as giving the of certain on plates ; but it is quite possible that specks of dust or other accidental marks have orcasionally been taken for corona lines .
The Table gives a list of the lines which have been measured on the two plates .
For reference , the corona lines obseryed in 1882 and 1883 have been added .
A query is atto c to those wave-lengths in 1883 which have not been included in Captain ABNEY'S final list .
The lines are in thicker type are those 1 have measured repeatedly , and about can speak therefore with greater confidence ; but all lines given have been measured at least twice .
In the colulnn headed in tensity represents the intensity , ( 1 ) he lowest .
The intensities refer to the appearance of the lines in the centre of the spectrum , that is , on the prominence line .
The numbers given are , of course , very approximate only , as the sities much depends on the fatigue of the eye and other circumstances .
to intensity ( 3 ) the lines are easily visible ; ( 2 ) means more fficult doubtful ; hnes marked doubtful .
CAPTATN L. DARWIN , DR. A. SCHUSTER , AND MR. E. W. MAUNDER If compare ether the measuremen ts taken during different eclipses , we find that the agreements are good , especially those takell with Spectroscope II .
in 1886 comI)ared with those taken in 1883 .
The resolving powers in the two cases were abouf equal , and double that used in 1882 .
Nearly all lines found in 188:3 have a repre .
sentation of intensity ( 3 ) , or above in 1886 .
Our data as yet are insufficient to judge whether there is any great difference in the spectra during different eclipses .
The slit in 1883 was radial ; that in 1886 was placed tangential to the Sun 's limb , and this would account for some difference in the relative intensities of lines , for some of the lines the radial spectroscope fade undoubtedly more rapidly as we go outwards in corona than others .
A number of lines are wide in those parts of the corona which show the continuous spectrum , strongly thinning out where chat spectrum is weak .
Nevertheless , I cannot help hat , if the line had been as strongly marked in eclipses as it has been in this , it would have attracted special notice .
On the other hand , a line 4015 , which was marked strong in 1882 , and which was present in 188.3 , does not appear in our photographs .
There is , therefore , some ground to believe that the spectrum of the corona differs much on different occasions , although we want further evidence to settle the point definitely .
I have collected in the following Table the most persistent lines in the spectrum of the corona ; that is to say , those hich appear in all three eclipses .
It seems curious that the discrepancy is greatest for our strongest corona line .
The line 4227 in 1883 may have been the calcium line .
That our corona line is less refrangible is proved not only by the measurements which are given above in detail , but can be shown by simply placing the corona spectrum film to film against the reference spectrum .
If the lines are made to coincide , the corona line is shown distinctly to be by the side of and not above the Fraunhofer line belonging to calcium .
For the greater part of these lines the agreement is as good as can be expected .
ON THE TOTAL SOLAR ECLIPSE AUGUST 29 , 1886 .
Alnongst the stronger lines the following have been observed in 1886 and 1883 , but not in 1882 :\mdash ; 1883 .
40.31 40:37 Here again the agreement is good for the greater number of lines .
If we endeavour to trace coincidences between the corona lines and the lines of known elements , we meet with serious difficulties .
Owing to the multitude of lines , accidental reements will be frequent , and no certain conclusions can be drawn unless we can trace a number of coincidences , or , at any rate , discover some group epeated ith ] its characteristic features .
Nor must we forget that the green corona line , which , before the Egyptian Eclipse , was the only known corona line , ] no representative in the spectrum of terrestrial substances , and the same holds for a large number of the lines occurring most frequently in the Solar chromosphere .
It should , therefore , not surprise us if we can not recognise any of our elements in the corona .
I have been much pnzzled , however , by a series of coincidences ; which , although I have finally come to the they are accidental , yet seem to deserve being mentioned in this report .
If we look at photographs of the line spectrum of the most striking features in the yiolet are as follows:\mdash ; Now almost the three strongest corona groups occur exactly at the same places as the groups 2 , 3 , and 4 .
When we compare the individual lines within each group , however , the is not so satisfactor the defective coincidence in group 3 tles strongly against its reality .
The strongest corona line is at in the middle between the two nitrogen lines ; but the line 4228 is weak nitrogen line , and while ) other two are about equally strong and so broad , ; any rate at atmospheric pressure , that they would not be separated on our , but as a broad band with 4239 as centre , I conclude that the ongest corol ) line call have nothing to do with nitrogen , and this makes the other coincidences very doubtful , as the strongest of all violet nitrogen lines at has no representative in the corona spectrum .
Nevertheless , I think it would be worth while to look for the strong green nitrogen line in the spectrum of the corona on the next occasion .
I have also examined the spectrum of oxygen , a number of curious coincidences which are also very likely accidental only .
, there are some of the corona lines which seem to lie very near lines of nese .
As Mr. has shown that the spectrum of anese plays ortant part in cosmical spectroscopy , the coincidences deserve careful consideration .
I have , therefore , added a column for the .
lines } the above Table .
The strongest anese l in the violet are and ; the mean of these two numbers is , or exactly at the place at which the sbrongest corona line is placed .
I think that our opinion as to the presence of manganese in the spectrum of the corona must depend on the question whether the two manganese lines could possibly , with the instrunlents used , look as a single line .
My own opinion is ainst such a view .
Our photograph resolves lines which are quite as close or even closer together thau the two lines in question , which are quite sharp .
The corona line on our plate has a certain width filling a space between 4231 and 423 , , that is to say , it fills exactly the two spaces between the two maIlgaI ) lines , but does not overlap them .
It is strongest near its centre .
From the width of the slit used , I calculate that nn indefinitely thin line would , in this part of the spectrum , cover a space of about units , so that the manganese lines should be separate , and reach respectively from , and from to .
The corona line should , therefore , show very decided signs of duplicity , I cannot reconcile its actual appearance with the supposition that it is the representative of the two manganese lines .
For the present , then , our attempt to identify corona lines has only led to negative results .
4 .
of Results and Suggestions .
In conclusion , I give a summary of the principal results of the spectroscopic , together with a few sugge.stions which may prove dseful to future eclipse observers .
ON THE TOTAL SOLAR ECLIPSE OF AUGUST 29 .
of Results .
1 .
The continuous spectrnm of the corona has the maximum of actinic intensity displaced considerably towards the red , when compared with the of Sun light .
This proves that it can in small part be due to light scattered by small particles .
2 .
While on the two previous occasions on which photographs of the spectrum were obtained lines showed outside the limits of the corona , this was not the case on this occasion .
Hence there must have been less light , due to the scattering in our atmosphere .
Calcium and hydrogen do not part of the normal spectrum of the corona .
The hydrogen lines are visible only in the parts overlying strong prominences ; ) and lines of calcium , though visible everywhere , are on that side of the corona which has many prominences at its base .
4 .
The strongest corona on the present occasion was at ; this is probably the same line as ofcen observed by YouNG in the chromosphere .
5 .
Of the other strong lines , the positions of the following seem pretty well established : \mdash ; The lines printed in thicker type have been observed also at the Caroline ] and Egyptian Eclipses .
6 .
A comparison between the lines of the corona and the lines of terrestrial elements has led to negative results .
ASuggestions the Spcctroscopic gements in future Eclipses .
1 .
In order to distinguish better any difference in the spectra between diffel.ent parts of the corona , a larger image should be thrown on the slit .
A lens of 4 or even 5 feet focal length might be enlployed with advantage .
The aperture of the lens need not be than that required to fill the collimator with light .
2 .
The width of slit be equal to , where is the useful aperture of the collimator lens , the wave-length , and the focal length of the collimator .
In order to prevent difficulties , due to dark lines , \amp ; c. , should be about 4 or feet .
3 .
A resolving power of about 12 in the yellow , if fnll use is made of it , seems sufficient .
can be obtained either by one prism or two small ones .
IX .
SULTS O AT CARRIACOU By E. W. The wovk lotted me in the servation of the eclipse was purely photographic , was intended to be , in its general character , a duplication of that undertaken by Dr. HUSTER .
pbotographs which 1 was to take were to be both of bhe corona itself and of its spectrnm .
For the form mer I was ) rovided with a lens of about inches ected for the raphic rays , and having a focal length of about .
) feet .
diameter of the image of the Moon on the , which taken in primary , was therefore about six-tenths of an inch .
For the of the spectrum I had two spectroscopes , the second of which was only provided immediately before ] were packed up for shipment .
The first spectroscope had two eight and inches in base , and with cting .
angle of itnd it was used in conjunction a condensing lens of inches aperture , focal length of ] ches .
The second spectroscope had one prism , inches both in height and base , and with refracting angle of ; the condensing lens used to throw of the Sun on the slit of this spectroscope was 3 inches in aperture , and had a focal length of inches .
These three instruments were all attached to the same fralne , which was mounted equatorially and supplied with clock-work .
The polar and declination axes and the R. A. and declination cles were those of the Corbett Equatorial of the Royal Observatory , Greenwich ; the driving-clock also belonged to the same instrument , but the stand to which these were attached was made specially for the expedition .
It was a tripod stand , composed of pieces of angle iron boIted together , and found to be light and portable , and at the time strong steady .
In addition to the camera and spectroscopes , a telescope inches ur and 5 feet focal length , together with its finder , was nlounted on the same stand ; and a lens of 1 inch aperture and 4 feet focal length , with a little screen in its prin ) focus , was attached to the side of the coronal camera as a findel to a delay in the selection of the equatorial mounting to be igned to my use , to that the second spectroscope was added to the equipment as , the preparation of the entire instrument was thrown so Iate that it was completed only just in time to be packed for shipment , and I had no opportunity , for a ) , to test its performance as a whole , or of the different parts separately , except the coronal camera , until my arrival at Carriacou .
On setting np the instrument , it was at once seen the driving-clock drove in the wrong direction , the Corbett aving been last used in the Southern hemisphele .
By the help of one of the al.tihcers of the " " Bullfrog\ldquo ; this was altered , and he driving of the clock rendel.ed fairly good .
Its actual rate was not determined , as the necessary alterations not ] ) until the day before the eclipse .
But the gearing ON THE TOTAL SOLAR ECLIPSE AUGUST 29 , 1886 .
34.3 of the hour circle into the -screw of the clock remained loose and unsatisfactory , and it was not found possible to remedy it .
The spectroscopes , too , when put together , proved to require many alterations in small details , which cost the most labour , but which were all effected very satisfactorily before the day of the eclipse .
The programme which I to out with the oove instruments comprised the taking of one of the of the corona during totality with each of the spectroscopes , and of seven photographs of the corona itself with the foot coronal camera .
The slit of the two-prism spectroscope was to be usted so as to point east and west , and to lie across the centre of the Sun ; the slit of the single- prism spectroscope was to lie north and south , and to form a tangent to the east limb of the SLIn .
The expedition reached Grenada on afternoon of Thursday , , and on the ernoon of the following the huts which had been made for the .
S. J. PERRY and myself were taken on board H.M.S. " " Bullfrog which left St. George 's at daybreak on August ] 4 , in order to convey us to Carriacou , an island some twenty miles north of Grenada , and the largest member of the chain of islets known as the Grenadines .
We anchored in Grand Ance Bay , off ) orough , the principal village of the island , and made the necessary examination of the country near , and inquiries respecting other portions of the island , in order to be able to select the most suitable site for our station .
As there were no safe anchorages for the gunboat on the east of the island , and as Grand Ance was on the west of the island , in its broadest part , and was , moreover , surrounded by steep and lofo hills , we were compelled to fix upon Tyrrell Bay , a little to the south , where the island was much narrower , and where the hills were only about 200 feet in height .
A position was finally selected on the top of the ridge , whence a good eastern horizon was obtained .
The site chosen was about 300 yards to the north of a slllall house known as ' The \ldquo ; the residence of Mr. DRUMMOND , find its approximate position as given by the niralty chart was W. long , , and N. lat. A heavy rain storm on the of the tornado which wrecked the neighbouring Island of St. Vincent\mdash ; rendered the steep slopes of the hill so muddy and sodden that the work of conveyinoo .
the huts and instruments to the summit a lengthy and laborious task , and the huts were not finally completed and the uments erected until Friday , .
The eight days still remaining before the eclipse were then to the necessary adjustments of the various instruments and the alteration of the driving clock .
The work was continually interrupted by short showers and passing clouds , bnt all the cameras were brought into good focus , the several telescopes and condensin lenses all rendered truly parallel , and the elevation and azimuth of the mounting ascertained to be correct within a of arc\mdash ; the circles with which the instrument was provided being divided to half degrees , but reading to minutes by means of a vernier\mdash ; before the day of the eclipse .
The entire programme of the eclipse had also been rehearsed on two occasions , the morning.s of Thursday and Saturday , August 26 and 28 .
Friday , August 27 , was wet .
On the day of the eclipse the Sun rose behind cloud , and the sky was generally overcast , though with breaks here and there .
A smart shower fell shortly before the eclipse , which necessitated a hasty closing of the observing huts , but it passed off before totality began , and the total phuse was observed in what appeared to be an entirely clear space of sky .
An alarm clock , ringing at every tenth second , was set up in the observing hut , and the timekeeper started the clock at the moment totality commenced , and called out at every second the number of seconds yet remaining before the first re-appearance of light .
The clock face had been numbered for 205 seconds , and the total phase was over about one second and a half after the last had been counted , so that , half a second for the delay in the clock , the duration of totality must have been 3 minutes and 27 seconds .
As totality approac , head I watched the Sun in the finder of my telescope , and gave the word " " Start the clock\ldquo ; to the timekeeper as the last ray of Sun lignt disappeared .
He started the striking-clock very promptly , and called the seconds very sharply and cleally throughout the eclipse .
I drew back the slide of the camera of the two-prism spectroscope first , then that of the single-prism spectroscope , and then proceeded to expose seven plates upon the corona itself with the five-foot camera , the seyeral plates exposed for the following times:\mdash ; First plate second .
Second plate Third plate Fourth plate Fifth plate Sixth plate Seventh plate The fourth plate was exposed at the word " " 120 seconds i. e. , 85 seconds after the commencement of totality , and closed at the word " " SO seconds The seventh plate was exposed before the word " " 20 seconds It would have been quite possible to have obtained at least two more short exposure plates had I had them ready , but I had not judged it wise to attempt more than the seven of the original programme , as I had not been able to manage more than that number during the rehearsals , but I found that I was able to work more rapidly and collectedly during the eclipse itself than during the preliminary drills .
At " " 10 seconds\ldquo ; the timekeeper gave the word " " close cameras and I closed the single-prism spectroscope first , and the two-prism spectroscope afterwards .
Both THE TOTAL SOLAR ECLIPS OF AUGUST 29 , were closed some seconds before the end of totality .
I was able to look up at the corona during the exposure of my plates , and I ) tched it the finder the 40 , exposure of the fourth plate .
I saw no trace of a red or rosy tint in either chromosphere or prominences , but I remarked two exceedingly bright and beautiful prominences of the inteloeest silver whiteness .
The taller of these is very well shown on some of my photographs .
The light during totality was feeble , but was just barely bright enough to enable me to read the programme which I had written out in a bold round hand , and had pasted on the top of the coronal camera .
After totality the Sun was covered by light cloud almost immediately , but a photograph was secured to give the direction of the two needles which were fixed in the east and west sides of the coronal camera , close to the sensitive plate ; and later on the Sun was brought on the bottom of the slit of each of spectroscopes and the plates were re-exposed for a second in order to secure reference spectrum .
The were not developed until after the return of the expedition to England , when Captain ABNEY kindly consented to undertake the operation .
No ice could be at Carriacou , and many of the best trial , taken for the purpose of ascertaining the focus of the different cameras , had been spoiled or completely lost by the heat .
It was , therefore , thought unwise to run the risk of deveiopin the eclipse photographs Carriacou , and the plates were accordingly securely sealed up , and brought home undeveloped .
Of the seven photographs of the corona with the five-foot coronal camera , five proved to be good , one sho.wed some deformation of the image , and the seventh was spoiled .
The spoiled plate was the fourth in order of exposure , and was exposed fur 40 seconds ; the accident which it useless was brought aboul in the following manner : .
DRUMMOND , the owner of the estate on which we had lixed our observing station , and who had been our most self-sacrificing host , had looked through the little telescope attached to the coronal camera during the first 80 seconds of totality , but immediately on the exposure of the plate in question he epped down and I took his place .
Unfortunately , in making the transfer in the semi-darkness , the instrument received a severe jar , a jar rendered the more serious by the unsatis- factory character of the gearing of the R. A. circle alluded to above .
The clock , however , drove the telescope very both before and aftel this occurrence .
The other plates were placed , with the photographs of the other obseryel .
S , in the hands of Mr. W. H. ESLEY , Assistant-Secretary of the Royal Society , who a drawing from the collation of the entire series .
The plates exposed upon the corona were supplied by Captain , and were inches by inches .
The two spectrum plates were also supphed by ABNEY , and were inch by inches in size .
Both these unfortunately proved to be useless , for , on development , the coronal spectrum was found to be masked by an Solar spectrum .
.3 CAPTATN L. DARWIN , A. SCHUSTER , AND MR. W. It most probable that , whilst the reference spectrum , I ently to full Snn shine for a moment or so ; for it does not appear possible at the actual time of the eclipse can have been prolonged beyond tlJe duration of the total .
It is to be much retted that the attempt was to secure beyond the coron al spectrum upon the same plate with it , and that any instrument not absolutely necessary should have been mounted on the same stand as the cameras and their accessories .
But for this mistake I should probhad to report the success of all the nine photographs instead of that of only ix of them .
X. OF CLIPSE ND DRA WING OF THE CORONA .
CAPTAIN IRWIN C. MALING , SECRETARY .
The Total Eclipse of the Sun on the 29th August , 1886 , was observed by me from Prickly Point , Grenada , Wesb Indies , the station selected by Captain L. E. , and .
A. SCHUSTER , F.R.S. , of the Expedition ; they kindly reque me to take of disc , and the following are the results of my evious to the commencement of totality my eyes were covered for 10 minutes to enable the sight to be as strong as possible ; I had , however , scarcely begun my when small drift of cloud passed over the eclipse , hiding it for about 40 seconds , after which time it was perfectly clear , and I was enabled to continue my observations .
Moon was surrounded by a bright halo resembling that painted round tlJe heads of saints in old pictures , from which streams of light extended , varying much in length , form , and apparent density .
The longest ray was on the upper .
It was of a bright pale yellow , fading into white at the extreme point ; it ) eared to be about two and a half times the diameter of my disc .
next gest , and I admit most beautiful , streamer was on the left side , in about the 3 , counting from zenith to the , and immediately above a small red rominence .
This ray differed from its companion on the right , inasmuch that it was of a conical shape , dense the edges and upper curve , and gradually thinning towards the corona .
Its colour appeared to be of a whitish-yellow , and the centre seemed be hollow , as if one could see through it .
This description also applies to the smaller conical shaped rays on the and the lower part of the corona .
I further observed two small of a red salmon colour on the left of the Moon .
upper one was round , the lower irregular and angular , apparently in contact with the Moon , or immediately contiguous to it .
It will be observed that a small space between the long conical streamer and zenith is left bare .
I was unable to complete my drawing , owing to totality being over . .
MAUNDER'S Plate .
Exposure sec. Corona well defined , but not extending further than 7 ' from linlb .
Three prominences on N.E. limb , and lower part of great prominence on N.W. Plate 2 .
Exposure 2 secs .
Corona well defined , but not extending further than from limb .
Same prominences visible as on Plate 1 .
Plate 3 .
Exposure 10 secs .
Corona can be traced on N.W. to nearly a Lunar diameter from limb , but is ext .
] ill defined , showing scarcely any detail .
Negative so dense that prominences can hardly be made out .
Plate 4 .
Exposure 40 secs .
Corona of great extent , but ill defined .
There has been much shake , and two separate images , about 13 ' apart , are superposed upon the plate .
Plate 5 .
Exposure secs .
Corona can be traced on N.W. to about 27 ' , but detail very imperfect and indefinite .
Not quite so dense a negative as Plate 3 .
Plate 6 .
Exposure 4 secs .
Greatest height of corona 22 ' .
Dense negative , but detail very ill defined .
Plate 7 .
Exposure sec. Scarcely a trace of corona .
Plate fogged , but prominences on N.W. and W. limb perfectly defined and better seen than on any of the plates .
Professor Plate 1 .
Exposure 15 secs . ?
Lower portions of corona only arejust visible .
Two prominences on N.E. limb , and lower part of great prominence on N.W. Plate 2 .
' Exposure 15 secs .
) Corona slightly more shown than in Plate 1 .
One prominence visible on N.E. limb , and part of great prominence on N.W. .
Exposure 20 secs . ?
Corona of greater extent in any of the plates of this series , reaching on the N.W. to a height of nearly 26 ' .
Details of coronal structure very well shown .
Negative dense near limb ; prominences not very distinct .
Plate 5 .
Exposure 15 secs . ?
Coronal detail well shown , but extension somewhat abruptly cut off at a height of about of a Lunar diameter .
Dense near limb ; prominences not ] seen .
Plate 6 .
Exposure 5 secs . ?
Somewhat thin negative ; details of corona in N. and S. polar regions very clearly defined , but a superposed image ( a few minutes of arc from the principal image ) , with a trace of the re-appearing crescent , has blotted out the corona on the W. side .
* The was behind a film of clouds while these photographs taken .
ON THE TOTAL SOLAR OF , 1886 .
inences .
These have been drawn from Plates 1 , 2 , and 7 of bIAUNDER'S , and Plates 1 and 2 of Professor SCHUSTpR'S .
Twelve prominences are visible on the plates .
Of these , four inconspicuous ones are near together in the N.E. quadrant .
On the W. limb is a well-marked group of seven prominences extending N. to S. of the equator , and at about N. is a very remarkable branched one , rising to a height of 5 ' .
The accompanying woodcut gives an enlarged view of this prominence , which is perfectly shown on Mr. MAUNDER'S Plate 7 .
This has been mainly drawn from Plates 4 and 6 of Professor ScHUSTER'S .
somewhat greater extension is given in Mr. MAUNDER'S Plate 3 , but its is so imperfect that but little use has been made of this .
To ensure greatel accuracy , two pel.fectly independent outline drawings were made to scale ; these were then superposed , and further measul.ements made in every case of diffel.ence .
The rifts at the N. and S. poles , which have generally the corona , were well shown in 1886 .
They were very wide , the northern rift along the limb for a distance of more than , and the southern for They are almost symmetrically placed about bhe Sun 's axis .
Both rifts are filled with somewhat similar to the rays of 1878 , but not nearly so fine , numerous , regular .
The southern rift is less obyious than the northern , the it broader , longer , and less definite .
These rifts are bounded on each side by groups of more or less synclinal structure , which are , clearly separated from the masses of equatorial rays .
The synclinal groups bounding the great rifts very unsymmetrical , those on the east being mparatively low and depressed towards the , while the corresponding rifts to the west rise to a much greater height , and are in direction .
The general lnc'lSS of coronal structure on the western side is therefore much greater than that on the east , although the rifts are symmetrically placed .
This want of metry e also to the masses of equatorial rays , that on the west extending along the limb nearly twice as far as the mass on the east ; it also rises to a much greater height , and has a far more complicated structure than the comparatively low and structureless eastern equatorial group .
The specially structure is best seen in the mass bounding the northern to the west .
In the centre of this mass is the tall prominence before referred to , and ovel this prominence the coronal rays bend vards each other on either side .
The of this group is encroached upon by the broad equatorial mass , which appears to overlap it .
group bounding the northern rift on the east shows but little structure .
The synclinal group to the west of the great southern rift is nearly radial , narrow , and conical , and extends to a height of quite 24 ' from the limb .
The group to the east extends not much more than half this height , and is broad and depressed towards the equator .
On comparing the corona of 1886 with those of fJer years , it appears decidedly different from any previously photographed .
The great northern and southern rifts extending for some distance along the limb , and the character of the rays filling the rifts , recall to some extent the more extreme form shown in 1878 .
The depression the main mass towards the equator , howeve eyen on the eastern side , is not so great as in 1878 , and the western side is totally different in character .
In , the western half of the corona of 1886 shows a striking resemblance to the corresppnding side in 1871 , but the compressed eastern half in 1886 and the wide polar riits have nothing in common with the corona of 1871 .
In 1875 there was the same decided want of symmetry between the two sides of the corona , but in this case it was the western half which was more depressed towards the equator .
In lS75 , also , there was a greater tendency towards the extreme polar depressiol of 1878 .
The corona of 1886 has no resemblance to that of 1882 , which had no conspicuous polar rifts .
Almost equally dissimilar was the corona of 1883 , with a single great rift at the north pole only , opened at an angle of about , but hardly extending to the limb , and very unsymmetrically placed with regard to the Sun 's axis .
Ofl , he Eclipse of 1885 , I believe the only successful negatives were those taken by Mr. RADFORD , at Wellingon , N.Z. , now in the possession of the Royal Society , but which have not yet been published .
I am not certain about the orientation of these , but , as far as I can judge , they present no resemblance to the corona of 1886 .
In conclusion , it may be sald that this corona occupies a middle place ween the extreme forms of 1871 and 1.878 .

