®0meU Hwei^^itg Jibi^arg THE GIFT OF •<>^HboL ^'^jT^^ vO^Jf^ Y 692- PHOTOGRAPHS OF THE COROiXA TAKEN DURING THE TOTAL ECLIPSE OF THE SUN, JANUARY I, 1889. STRUCTURE OF THE CORONA. BY DAVID P. TODD, Ph.D., Director Amherst College Observatory. CITY OF WASHINGTON: PUBLISHED BY THE SMITHSONIAN INSTITUTION. 1889. The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924032176590 692- PHOTOGRAPHS OF THE CORONA TAKEN DURING THE rn TOTAL ECLIPSE OF THE SUN JANUARY 1, 1889. STRUCTURE OF THE CORONA. DAVID P. TODD, Ph.D. Director Amherst College Observatory. CITY OF WASHINGTON: PUBLISHED BY THE SMITHSONIAN INSTITUTION. 1889. V INTRODUCTORY NOTE. The accompanying plates have been prepared from positive copies on glass of photo- graphs of the total eclipse of the sun of January 1, 1889, kindly presented to the Smith- sonian Institution by Professors Pickering, Holdcn, and Payne, Captain Floyd, General Irish, and Mr. Burckhalter. The copies have, for tlie f^ake of comparison, been reduced to a uniform diameter by Mr. Smillie, the photographer of the Institution, and a descriptive note with remarks on the structure of the corona has been added at my request by Professor Todd. It is not intended to include these plates in the Contributions to Knowledge, but a limited number of carefully prepared prints will be distributed to astronomers and others specially interested in solar phy.sics. S. P. LANGLEY, Secretary. Smithsonian Institution, October 1, 1889. (3) ON THE STRUCTURE OF THE CORONA AS INDICATED BY THE PHOTOGRAPHS TAKEN 1889, JANUARY 1. By Professor David P. Todd. On occasion of the eclipse of this date clear skies were everywhere prevalent. A great variety of photographic apparatus was in the field and a rich harvest of pictures was gathered. At the request of Professor Langley, glass positives of all the better photo- graphs were forwarded to the Smithsonian Institution for comparison and preser- vation. These positives being contact prints, thei'e was, of course, great diversity of scale, and the first step was to enlarge or reduce the photographs to conformity with an arbitrary unit. For the unit of the moon's diameter Professor Langley chose 25 millimeters, and the necessary negatives of this size were prepared by Mr. Smillie, the photographer of the Institution. Prints from these secondary negatives form Plate I. Of course the results would have been better had all the original negatives been available. Captain Floyd alone transmitted an original negative. Also a series of lantern positives was prepared, and the photographs were exhibited to the National Academy of Sciences on the 17th April, 1889. Certain conclusions were drawn from the collation of the photographs, and suggestions made for the observation of future eclipses. In the main these follow. In the subjoined table are presented all the important circumstances and conditions pertaining to these nine photographs of the corona. The top and bottom of each print are north and south, and the right and left are west and east, respectively. The sun's axis is inclined at an angle of between one and two degrees with the north and south line (position angle = -I- 1°.4), and the vertex or highest point of the sun and the trace of the ecliptic are shown with sufficient accuracy for each figure by the diagram in Plate II. (5) 6 STRUCTURE OF THE CORONA. The excellence of Nos. 1 and 2 indicates the desirability of photographing the corona with reflectors in the future. In No. 3 the poor definition appears to be due partly to the lack of exact focal adjustment and partly to the deficient clock-work, a temporary apparatus having been devised for turning the polar axis by hand. That a glass of less than two inches aperture and unadapted to photography should have produced a photograph (No. 4) comparable with that obtained (No. 5) by a 13-inch objective specially corrected for the photographic rays is a matter requiring minute investigation, and is suggestive as to eclipse outfits in the future. There appear to be other effects than those due to difference of aperture merely. Both these pictures are shown in Plate II of the size of the original negatives. In No. 8 the effect of the dry-plate granules is conspicuously brought out by excessive enlargement. To a slight extent this is apparent in No. 9 also. In commenting on this collection of photographs before the Academy, I ven- tured the following observations : (I) The axis of symmetry of the corona does not coincide with the axis of revolution of the sun as determined from the solar spots. The corona aj^pears to be at least a triple phenomenon* made up of — (a) The polar rays, seen most prominently about the poles, but probably extending into the equatorial regions, and not there seen because projected upon the filaments which have their proper origin there. {b) The inner equatorial corona, the lower regions of which bear some re- semblance to an outer solar atmosphere, and have perhaps a closer connection with truly solar phenomena than any other part. (c) The outer equatorial corona, consisting of the long streamers, for the most part visible only to the naked eye, and having perhaps no necessary con- nection with the sun. (II) The polar corona consists of rays, straight or nearly so, and radial from neither the sun's centre nor the sun's poles. Rather they seem to radiate from areas the centres of which are adjacent to the sun's poles ; and the law of their inclination to the polar axis of the corona appears to be susceptible of precise empirical determination. A few of these rays appear to be double nearly their whole lenffth. For the most part, if not entirely, the rays or beams have parallel sides. They can * Note in this connection Young and Huggins on the compound spectrum of the cor.iiia, consisting of three superposed spectra. Silliman's .lournal, vol. ll)2, p. 'iS, and Proc. lloyal Society, liilO, 18S'i, pa"-e 121. STRUCTURE OF THE CORONA. i be subdivided with increase of magnifying power, and are sharply defined like the lines of the sun's spectrum. Between them the dark rifts sometimes extend quite to the disk of the moon. Occasionally a ray appears to have a slight curvature, in general from, but often toward, the solar axis, and this may be due to optical or to photographic illusion, or both. (III) The inner equatorial corona emits a large percentage of the total light of the corona. Photographs taken with clock-work show great detail in this region, though the streamers are not generally so sharply defined as about the poles. ^Many of these streamers appear to hare a real curvature. Four large prominences are visible, two on each side of the sun and at about 35° of solar latitude, as if to suggest some connection between the protuberances and the corona. (IV) The equatorial streamers of the corona are for the most part lost by the operation of enlargement. These streamers are very slightly curved, while they are convergent on the east side of the sun and divergent on the west. On the latter side and about 1° from the sun's centre there is a suggestion of wider divergence, as if there were electric rei)ulsion between adjacent streamers, as Huggins' theory would imply. The photographic evidence as to the existence of a meteoritic ring or equatorial envelope surrounding the sun is inconclusive. The fact of chief importance established appears to be the periodicity of the outer corona in a cycle probably of equal duration with that of the solar spots. A comparison of the corona of 1889 \vith those of 1867 (Cxrosch), 186S (Bullock , and 1878 (Langley, New^comb, and others) is sufficient to establish this periodicity beyond reasonable doubt. The epoch of greatest extension of the equatorial corona appears to coincide very nearly with the epoch of minimum sun spots. (V) No rapid change in the structure of the corona can safely be inferred from a comparison of photographs taken at the diflferent stations. The difi'er- ences are slight, if any, and may well be due to diiferences of objectives, plates, and development. The time-difference between the photographs at Bartlett Springs and at Wil- lows is about one minute. The only safe inference appears to be that, while the corona mav change from hour to hour, there is no present indication of change from minute to minute. In order to investigate this question fully in the future, it will be necessary to make two like series of exposures at widely separate stations, and combine each series into an accurate representation of the entire corona, if possible, by means of composite photography. O STRUCTURE OF THE CORONA. (VI) A few suggestions for the coming eclipse bearing on coronal structure are pertinent. They are the more important as the decade of the nineties con- tains only two eclipses likely to be well observed — 1893, April 16, in Brazil and West Africa, and 1898, January 22, in East Africa and the west of India : (1) For the minute study of the detailed structure of the corona it is abso- lutely essential that the photographic telescope be equatorially mounted and driven by perfect clock-work. (2) These photographs should be taken on a scale as large as convenient in order to avoid effects of the granulation of the dry-plate films. (3) Means must be provided for the most accurate orientation of the polar streamers. If a plumb-line or the parallel cannot be photographed on the plate, exact orientation should be accomplished by optical measurement of the position angle of a suitable prominence. (4) Attention should be directed to photographing the outer coronal streamers by the most delicate apparatus available. (5) In view of differences in the photographic correction of objectives, pairs of reflecting telescopes should be used, widely distant on the earth's surface, and with identical plates. In \ie\\ of the greater absorption of the H K rays by silver on glass, the reflector would preferably be made of speculum metal. If dioptric instruments have to be used, the actinic focus for the coronal rays should be experimentally determined with the greatest precison. (6) Comparisons of different photographs and generalizations upon the coronal structure are not worth attempting unless the precise I'elation of the solar to the lunar centre is known. It is therefore essential that the exact time of the middle of each exposure be known, together with the longitude and latitude of the station within 15". Oi GO 00 in < I— I o p^ o < H O H W O w < p^ o H O W p-t ^uiod qijou uiojj sjxB s,uins JO a[§UB-uo!i!so^ •5[aoio-saiAi.iQ f^ •qiSuaq •iCii[E}0} JO puooag UB§8a , jeqranu ja^srao^isueg o a. . t^ 03 o 00 T-H T-H < •qjSuai [Boo^ .g S » S ^ •X}I[B)0^ JO UOIJB.ind ■asd;ina ibj}ii3o jo 8ui] uiojj 90UB^sip a}i3uiixo.iddy > m Q o -* 00 c-1 CO CM '^ c^ lO CD 00 o r- o o o o ■9.inV8dl3 JB810 g O O O CO CO CC ^ !M CO ^ 1^ -f— «is ^_ 5ic -M- s '-' ^ rH ^ ■-' "-^ ^ ■-' r-H ^^ -rj tn oj :^ :zi ^ 1^ •ii ^2^ i CI Cl (M CO as Oi cc CO Ti ■ L~i o 1 1 o o CO CO o •[8A8I-«3S OAOqi? UOl'Jl? A8ia: ^ CO CO ' ' (M 1— 1 o V O o , 1 , , o - t^ r.-^ -^ -* "* I-H CO CO TfH ^ ■Jj CO CO c^ Gi cr- — 05 C".' o c3 CO CO CO CO CO CO CO CO "^ h^ >i ^ ^ ^ ^ fz; !z; ;z; 1^5 o o Tj< 1 o o _ 1 1— 1 CO -X o -U93.i^ JO :^saAV opn^jiS UCJ - '-' ■-I ^ < CO 00 00 CO CO CO CO CO CO 1' bn [ c 1J OJ Cl. ^ 03 V C3 7J ^ > ^ c3 = ,£3 c o- c3 u bfl fcn a; c T( rl oj ?. tij s ?3 Ja Cl. rr,' -C u P. X o P. o o 'O > O -S o ,r3 o O' ^ ^ 1 O O CD in « «J K O t) o M •s-' * -1- ++ ss&o t= — Eh p O) Nl e i- M h H o e< M (J o m K a H 13 p4 ^ a rf ■< OT « H ^ -o lO O (M CO lO t- I-H o G^ ■-' i-- I-H T-H T-H r-H -^ C^ CO t— (N -* Oi (M l-H 'o lO r- 00 1>- CD CO 1— t (M co iO t^ 05 i—t r-i 1—1 r^ r- - ^ I-H CM CO CD CO O 1-H to CD Oi OS 00 I— ( CO 05 t^ CO 00 ^ o o (N CO lO t- o CO rH ^ I-H I-H CM CO Ol tH t7~ CO b- o lO t^ TtH o :^ 00 t- o CD o o -* lO 00 (M T-H I— 1 ^ T-l -< -' 1-i cq CO V t- 02 C• CM -ntl r-H •^ lO 00 CM CO ^ ^ I-H I-H -- -^ '- CM CO o o ■* CO lO w ■^ ^ CD CD CM Ol Oi 00 (M I-H T— 1 CM ■^ LO CO CM ° ■-H i-H "■ '-' l-H I-H ^ '-^ CM CO ^ . zn > > 2 '^ o M 8 O rH l-H CN CO •-< O t^ 1-1 rH CO o o o o o o d S O i-i w THE SOLAR CORONA. U The lines of force are constructed by the equation : N = 2//(+ I //CR')^'f, or, calling the constant n R'* C unity, sin' tf . 8 // N = o — r 3 The successive integral numbers may be given to N, . 1 . 2 3, etc., and the corresiDonding values of 6 computed. They are : N=0 0=0 N=O0=O 1 20° 12'.7 5 00° 35.0 2 29 15.0 6 57 48 .(i 3 36 4.5 7 66 4.7 4 43 42.5 8 77 44.7 when we assume in the formula that r = 1. These give us the points at which the lines of force of integral orders depart from the surface of the sphere. But more conveniently for our purposes we may assign values to the angle 6, such that the cosine of the successive angles differ by one-tenth radius, and compute the values of N under this case : U e = 25° 51' N = 1.593 If = 72 33 N = 7.624 36 52 3.016 78 28 8.043 45 34 4.272 84 15 8.294 53 8 5.362 87 8 8.357 60 6.283 90 8.378 66 25 7.037 To trace out the path of a line of force of any order N, take sin^ 6 ■ 3r N 8 ri assume the required N, assign successive values to r at convenient distances, and compute 6 ; e. g. : IfN = 1.593 andr 1 ^ = : 25.51 IfN = 1.593 andr =4 = 60.42 2 38. 4 5 77.10 3 49. 3 6 or assign values to Q and compute r. (A table of Lines of Force is given on the following page.) To find where the lines of any order N cut the. equator axis, take 3r N _ 8 // - assign the values to N and compute r N= 1.539 r= 5.248 " 3.016 2.779 4.272 1.961 5.362 1.562 6.283 1.333 7.037 1.191 ] or / 8 // 3N N = 7.624 r = 1.099 8.043 1.042 8.294 1.010 8.357 1.002 8.378 1.000 12 THE SOLAR CORONA. Q O O CO V ^ lO 10 00 \o j_, CM CO , OS CO 00 00 CM 00 CM 00 CO 00 en CJ 10 Jr- tM (M 1 >c (M .—1 t- iC> ■* CO CM (M ,-H T-H i-H ,_, i-H ] t- -* (M >— 1 (M CO »o c^ UO uO -* 00 00 CI- t- !M CM H >— ' r-> r~- CM ^ ■^ CO Gi CD CO ■M T— 1 t- iQ -* CO !M r-{ r-" 7-A f— ' r— 1 "X) CM Ol r^l V -TtH '^ Oi Oi tH 1—1 CM UO -h 1^ Tt: 00 1-- L- Q H CD 10 "^ c» a> Ir- t-~ -* rH Jr^ T-H CD 03 I— 1 V t- 01 I— t-^ (-, ■* >~> CO CO t~ CD 8 t— i >— 1 1— 1 t' I— 1 CO CO t~~ Tjl (M CO 'M CO LO ^ ^ t-M CO CO Oi »o 0^ ir- CM 00 t- uO 1 1 1 C^ CO Oi 00 00 CO to H 00 ^ CO CO CD 00 ^^ *o C^l CD -^ CO (M r-i T-i T-i ^ -* ^ V CO CM t^ Oi ir^ C1 tr^ CO (-J cq -# r- ^ t^ I— 1 H C7S LO t- 00 CO lO CM I— 1 .-4 1—1 lO c^ fM ,_l r_i '^ r-( I^ -M 00 01 c^ CO 00 C5 « CO cc (M 00 05 -* CO CM -—1 T-H ,_l CO ■"■ 1 1 CO 00 8 ct CO 00 -^ i 1 ' CO I-- CO CM ,— I ,—1 r-i ""' t- t- CO CO ir- 8 CO Ci CM CO CO uO t- Til --D (M 1 :n 05 '^ CI ^ I-S ^ 1 lO 8 ^ 00 1-^ 10 X 1 1 t^ 00 a^ CO 1 1 CN I- CO ^ ^ ^ I ] Jr- r^ I 8 GO m ^ 1 1 LO oi '-' " J 1 -;-- CD ] ! ] 1 1 1 8 c:i OS 1 J 1 CD CO 1—1 ] 1 1 1 00 1 j (M 8 00 01 1 1 1 I— 1 ,_, ,_, [ 1 1 V 8 1 1 1 1 GO ^ 1 [ V, I 1 1 1 1 1 1 I— ' CO -t1 CM Ol tz; ^ l~~ (M 1^ 10 00 10 CO CO l- GO CO (M rjn S l^ ■"1 CO CO CD CJ CO CM to CM Ki CO 02 o 00 r- 00 -'• 1 C-l c/j 00 -H CM CD CM CO (M CO CM t^ ■-^ CI CO rtl m 1 01 01 CD CD CO JSi^o i 01 crJ d d d d d CD d d d CM --, -H ■X -M ;d >o >o CXJ x> THE SOLAR CORONA. 13 To find the order of line of force at the earth's mean distance from the sun, talce the mean semi-diameter of the sun, 962", and the mean parallax of the earth, 8".84S, and the earth is 108.7 radii of the sun distant from it. For r = 109, N = 0.07686. The angular distance from the pole at which this line of force leaves the sun is 5° 29' 47". Graphically the lines of force cut the equipotential lines orthogonally, and may be so drawn, starting at the points of the surface heretofore marked by the equipotentials. These lines are ovals cutting the equator perpendicularly and becoming tangent to the polar axis at the centre of the sphere. A test of the accuracy of the drawing is found by taking the sides of any of the quadrilateral figures, wherein the ratio of the mean distance between consecutive equipotential surfaces is to the mean distance between consecutive lines of force as the half the distance of the centre of the figure from the polar axis is to the unit of measure. APPLICATION TO THE CORONA. An analysis of these lines of force appears to be a description of the visible solar corona, and this analogy first suggested the explanation of the phenomena now given. The concentration of potential at each pole throws lines vertical at the polar region, bending gradually each side, and at a distance of 26° losing one-tenth of the force, — the angle of the line of force to the polar axis bring nearly 45° ; this curve closes on the equator at o.2o radii from the centre. The next decimal line leaves the sphere at an angle of 67° to the vei'tical axis, and having a potential of eight-tenths closes on the equator at 2.8 radii. The tliird line of force is inclined at 76° to the axis, and having potential seven-tenths closes on equator at 1.96 radii. The fourth line starts perpendicular to the vertical axis, leaves the sun at jaolar distance 53°, closing on equator at 1.56 radii. The other lines rapidly become more nearly parallel with the surface and close in as they lose iDotential. The solar corona can now be analyzed. The straight polar rays of high tension carry the lightest substances, as hydrogen, meteoric matter, debris of comets, and other coronal material, away from the sun, and they become soon invisible by dispersion. Next we come to the strong quadrilateral rays of poten- tial .9 .8 .7 .6, which united form the aj)pendages conspicuously seen at periods of great solar activity. They rapidly diminish in intensity, and at the distance of one radius have generally a potential of one to two tenths. The explanation of the long equatorial wings, with absence of well-marked quadrilaterals, seen at periods of minimum activity, is that they are due to the closing of the lines of force about the equator. The re-entrance of these lines forms along the equator, 14 THE SOLAR CORONA. the place of zero potential, a sort of pocket or receptacle wherein the coronal matter is gradually carried by the forces, accumulated and retained as a solar accompaniment. During periods of inactivity or low maximum potential the streams along the region 40° — 60° polar distance diminish in intensity, so that huge volumes are not carried away from the sui'face, but none the less what does leave the sun is persistently transported to the equatorial plane of the corona. In fact, the zodiacal light may be the accumulation at great distances from the sun along this equator of such like material, being carried by forces, all of which approach the equator perpendicularly, but there become zero. Here the zodiacal coronal material has no way of escape, being once deposited. We have a test of the accuracy of our theory which may be applied to any portion of the coronal rays, using the caution that we deal with true rays undisturbed by perspective and diffraction, and notably the polar and the outer boundaries of the quadrilaterals are best available. From the centre of the sun, with a radius vector r, draw a circle at any chosen point of such ray where the curvature is well marked, and a tangent to the circle, prolonged to intercept the polar axis with which it makes an angle. (See Fig. 2.) Let/ equal the angle at which the line of force crosses this tangent ; draw another tangent to the line of force and prolong it to the polar axis, then tan / = 2 cot = 3 tan X. The inter- cept cut off from the centre of the sun by the force-tangent is one-third the inter- cept cut of by the circle-tangent. I believe that this criterion holds good on the photographs taken during recent eclipses, as the following readings show : (A table of Readings is given on the opposite page.) These readings were taken from Professor Holden's diagram (Monthly Notices R. A. S., April, 1889) by centering one side of a right triangle on the sun with radius 2 rotating to the several angles 6, previously selected to mark the prominent rays, and reading the other side of the triangle on the axis ex- tended and marked on a scale with the radius as unit. Finally an edge was laid on the local lines at (rO), and the reading on the axis again taken. It must be clearly kept in mind that it is not the direction of the whole ray from its base on the sun to the point (r 0), but the direction tangent to the ray at the point (r 0). Besides the local inaccuracies there may be a slight error in placing the direc- tion of the axes, and the readings in the S. E. quadrant suggest this pi'esump- tion. Still the approximation of the ratio U> 3.0 is so evident as to show the application of this theory to the solar corona and also to witness the fidelity of Professor Holden's drawing. I have just had the pleasure of seeing one of the photographs of the inner corona and one of the outer taken by the Harvard College party Januarv 1, 1889, and the details are shown so clearly that our theory is at once able to be THE SOLAR CORONA. READINGS ON HOLDEN'S DIAGRAM. 15 Radius Angle Intercepts by- "'• Circle-Tangent. Force- Tangent. N. W. Quadrant- N. E. Quadrant. S. W. Quadrant- S. E. Quadrant- s'" 16 22 29 35 47 3° 10 16 18 23 28 36 53 64 72 2.01 2.07 2.15 2.26 2.42 2.83 2.00 2.03 2.07 2.09 2.13 2.26 2.4.3 3.32 4.62 6.70 3° 2.00 6 2.02 11 2.04 19 2.11 41 2.62 50 3.07 56 3.60 6° 2.02 12 2.04 19 2.10 41 2.59 51 3.08 56 3.60 .65 .66 .71 .50 .59 .85 .65 .72 . ( o .72 .75 .74 . 77 1.07 1..35 1.20 .50 .55 .•58 .70 .84 1.02 1.21 .61 .57 .60 .72 .90 1.20 Ratio. 3.09 3.14 3.03 4.51 I 4.10 3.32 3.20 3,08 2.82 2.83 2.96 2.83 3.01 3.18 3.10 3.45 15^58 1 2.97 4.00 3.67 ^53 8.01 3.12 ■3.01 2.97 3.03 3.32 3.-52 3.50 3.59 3.42 3.00 3.39 Local curvature too straight. Mean. Local curvature faulty. Mean. Local lines in error. Mean. Whole Quadrant may need adjustment as to its axis. Mean. The necessity is obvious of rejecting freely such lines of force as are not natural, and the difficulty of obtaining true lines is at present great. 16 THE SOLAE CORONA. tested. I give a table of the measures whea our rule of polar intercepts is ap- plied to tlie ray structure. They can be verified by any one possessing a Pickering photograph on celluloid. THE PICKERING PHOTOGRAPHS. THE INNER CORONA. ^i" ^' bfs Kadius ?. Angle 9- O