< PUBLICATIONS OP THE Vassar College Observatory No. 1 MARY W. WHITNEY, Director 4 * "}■. « r \\<\ Cornell University Library 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/cu31924012303529 PUBLICATIONS OF THE VASSAK COLLEGE OBSERVATORY, No. 1. Mart W. Whitney, Director. Catalogue of Stars One Degree of the North Pole OPTICAL DISTORTION OF THE HELSINGFORS ASTRO-PHOTOGRAPHIC TELESCOPE DEDUCED FROM PHOTOGRAPHIC MEASURES BY CAROLINE E. FURNESS. Assistant in the Observatory. Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Faculty of Pure Science, Columbia University. POTJGHKEEPSIE, N. Y. 1900. 61B PRESS OF THE NEW ERA PRINTING COMPANY, LANCASTER, PA INTRODUCTION. The Vassar College Observatory was established in 1865, when Vassar College was first opened to the public. It was built and equipped for purposes of instruction, the plan of founder and trustees not including scientific work for publication. The first director was Maria Mitchell, well known as a discoverer of comets, and as the recipient of a medal from the King of Denmark for the dis6overy of a comet in 1847. Professor Mitchell published a series of observations on the surface features of Jupiter and Sat- urn in Silliman's Journal and the American Journal of Science. Since 1890 an effort has been made to carry on a certain amount of observational work outside that involved in instruction, and this has been given mainly to minor planets and comets. Publi- cations have been sent with more or less regularity to the current astronomical journals. The present paper is the first regular publication of the observatory. The site of the observatory is on the grounds of Vassar College, about three miles east of the Hudson River, in Poughkeepsie, New York. The adopted latitude of the observatory is 41° 41' 18". The adopted longitude of the observatory is 4 h 55 m 33 s .6 west of Greenwich. The latitude was obtained by Professor Mitchell in 1872, by the zenith telescope method. The telescope was loaned by the U. S. Coast Survey. The longitude was also obtained by Professor Mitchell, by electric telegraph connection with Harvard College Observatory in 1877. Clock signals were exchanged for one evening only, and there was no change of observers. The observatory consists of a central dome, with wings to the south, east and north. The south wing contains a clock room and a class room, the east wing, the meridian room, the north wing the living rooms of director and assistant. In the dome, twenty-five feet in diameter, stands a twelve-inch equatorial with the usual equipment of eye-pieces and filar micrometer. The ob- ject glass of the equatorial was originally made by^ Fitz of New York, but was afterwards [recut by Alvan Clark. The telescope was remounted in 1888 by Warner and Swasey of Cleveland, Ohio. The meridian room contains a' transit circle made by Young of Philadelphia, and afterward improved in its mounting by Fauth & Co., of "Washington, D. C. The object glass of the transit telescope has an aperture of four inches. The clock room con- tains a Bond sidereal clock and a Bond chronograph ; also a machine for measuring stellar photographs, made by the Repsold firm of Hamburg. As this publication treats only of work upon photographic plates, measured with the Repsold machine, it is not necessary that a more detailed account of the observatory equip- ment should be given at this time. The Repsold measuring machine was the joint gift to the ob- servatory of Miss Catherine Bruce and Mr. Frederick Thompson. It is of the pattern described by Scheiner in Photographie der Gestirne, p. 148. The additional details of its construction are given in the following pages. The object glass of the equatorial is cut for visual purposes and we possess no correcting lens. It does not seem at present advis- able, with the means and time at our disposal, to attempt to take photographs, but the measurement and reduction of photographs taken elsewhere fit in well with our other duties, and with this purpose in view the Repsold measuring machine was secured. The first piece of reduction, based upon stellar photographs, was suggested by Professor Jacoby of Columbia University. He had in his possession twelve polar plates taken by Professor Donner of Helsingfors, Finland, partly measured and awaiting reduction. Miss Caroline E. Furness, assistant in the observatory, was placed in charge of the investigation. We express our thanks to Professor Rees and Professor Jacoby of Columbia University, who have given us much aid and sug- gestion. Our thanks are also due to our college trustee, Mr. Samuel Coykendall, through whose generosity this publication is made possible. Miss Mary E. Tarbox and Miss Louise Ware, graduates of Vassar College, have rendered much assistance in the labor of computation. We take this opportunity to give public expression to our thanks for the publications received from other observatories. Mary W. Whitney, Vassab College. Director of Observatory. TABLE OF CONTENTS. Section. Page I. Description of the Plates i II. Measurement of the Plates 3 III. Instrumental Corrections 4 IV. Method of Determining Right Ascension and Declina- tion of a Star from the Measured Co-ordinates 26 V. Method of Determining the Plate Constants 28 VI. Method of Determining the Apparent Places of the Standard Stars 1 30 VII. Determination of Plate Constants 36 VIII. Results of the Measures 42 IX. Final Results 57 X. Optical Distortion 63 Catalogue of Stars within One Degree of the North Pole, and Optical Distortion of the Helsingfors Astro-photographic Telescope, deduced from Photographic Measures. Description of the Plates. In September, 1895, at the request of Professor Harold Jacoby, of Columbia University, Professor Anders Donner, director of the Observatory at Helsingfors, Finland, made a series of photo- graphs of the region of the North Pole. Professor Jacoby had two objects in view in securing these photographs ; first, to ob- tain a catalogue of stars in "the immediate vicinity of the pole ; and second, to investigate by a new method * the optical distor- tion of the Helsingfors telescope. Twelve photographs were made in all ; four with the telescope pointed at 90 declination, and the remaining eight with it pointed at 89 °. The first four plates furnish material for a catalogue of stars within one degree of the pole, and their reduction forms the greater portion of the present paper. The discussion of optical distortion in Section X is based upon the results embodied in this catalogue. A reduc- tion of the remaining eight plates, and a catalogue of the stars found on them are reserved for a later publication. The photographs were taken in the following manner on plates provided with a reseau ; the telescope was set for the desired right ascension and 89 declination, and a plate exposed. It was next moved i° in declination, keeping the same right ascension, and a second plate exposed. The telescope was then moved i° in declination a second time, still keeping the same right ascen- sion, and a third plate exposed. For a second set of three plates the telescope was adjusted to a right ascension differing from the first by 45 , and the changes in declination just described were repeated, and similarly for a third and fourth set. The twelve plates taken together cover a circular area 2° in radius about the pole, with considerable overlapping. The four middle * Astronomical Journal, Vol, XIII, p. 190. 1 2 Stars within One Degree of the North Pole, and plates contain the region within i° of the pole. Within this limit they duplicate each other, but are taken with the telescope in different positions. Three exposures were made on each plate, 6 m , 3 m , and 20 s in duration, and between consecutive exposures the telescope was moved slightly in declination. Most of the stars were so faint that the 20* images could not be measured ; consequently they form no part of this reduction. The accom- panying figure shows the overlapping of the twelve plates : 22*21™ Fig. i. The following table taken from the Helsingfors records gives the data accompanying the exposures for the four central plates : Hereafter the plates will usually be designated by the Roman numeral in the last column. Optical Distortion deduced from Photographic Measures. 3 Table I. Plate. 1895. Obs. Barom. Tbermom. Att. Ext. R. R. Length of Expos. End of Expos. a S No. Sept. 16, No. 3 mm Dr. 757-3 +8.3 +7-7 m 8 6 3 20 h m s 2314 2 18 4 19 5 h m 22 24 1 900 I Sept. 16, No. 6 D. 756.6 +8.0 + 7.2 6 3 20 02759 3135 32 36 I 24 900 II Sept. 18, No. 2 Dr. 751.2 +8.7 + 8.2 6 3 20 2042 9 45 43 46 34 I924 900 III Sept. 21, No. 14 D. 764.5 + 2-5 +2.1 6 3 20 3 18 34 22 12 23 2 4 24 900 IV D. Donner. Dr. Dreyer, assistant. II. Measurement of the Plates. The star images were measured for the most part at the Col- umbia Observatory with the Repsold measuring machine No. 1, by Mrs. Herman S. Davis and Mrs. Annie Maclear Jacoby. A few were measured by me with the Repsold machine of the Vas- sar College Observatory. These measuring machines are of the form described in Scheiner's Photographie der Gestirne, p. 148. The reading microscope is furnished with two sets of movable micrometer threads at right angles to each other. In making the measurements at Columbia the horizontal screw was used through- out, since it was decided after trial that bisections made with ver- tical threads were more accurate than those made with horizontal threads. The plate was oriented in the machine so that one set of reseau lines was parallel to the vertical micrometer threads, and the microscope adjusted so that ten revolutions were equal to the distance between two reseau lines, or 5™"- The screw-head was then set at 47.75 and the carriage moved so as to place the parallel threads over the preceding corners of the reseau square. Micro- 4 Stars within One Degree of the North Pole, and meter settings was then made in the following order : i° upper left hand corner, 2 lower left hand corner, 3 large image, 4 middle image, 5 small image, 6° lower right hand corner, 7 upper right hand corner. The last setting was about i4 r -75- The screw was then turned forward half a revolution, to I5 r -25,and the process just described was repeated in the reverse order, ending at 5 r .25- The object of turning the screw half a revolution for- ward was to eliminate as far as possible the periodic errors, and to distribute the wear on the screw. It was deemed necessary to observe the corners of each square, because the lines for the most part were too indistinct for accu- rate bisection. In some cases even the corners were invisible and this prevented the measurement of some of the stars. Each coordinate was measured in two positions of the circle 180 apart, in order to eliminate personal errors entering into the bisection of the star image. A specimen sheet showing the form of recording the settings will be found on page 8. The first two lines of the upper and lower sections contain the actual settings on the reseau lines and star images. In addition to the original record of the micro- meter readings, each sheet contains the instrumental correc- tions, and all the steps leading to the determination of the final coordinates. Each image was reduced separately ; in fact, the means were not taken until the final right ascension and polar distance were obtained. III. Instrumental Corrections. 1°. Screw Corrections. The portion of the screw used lies between 5' and 15'. The variations in pitch of this portion of the screw were determined by comparison with an accurate quarter millimeter scale, espe- cially constructed for this purpose by Repsold, and the property of Professor Jacoby,who kindly lent it to Professor Whitney for determining the corrections of the Vassar screw. The method of comparison is described by Professor Jacoby in the American Journal of Science, 1896, Vol. 1, page 333. The following tables give the corrections to the readings of the horizontal screw in Optical Distortion deduced from Photographic Measures. 5 tenths of microns, or units of the fourth decimal place. These screw corrections, as will be seen from the specimen reduction sheet on page 8, were applied to the settings on the star-images only. Since the mean of the settings on each reseau line is very nearly $ or 15', it was assumed that the screw corrections for the four separate settings on each line would nullify one another. Table II. Micrometer. Columbia, No. 1. Vassar. 5E +0.0000 -fo.0000 6 +O.OOIO -j- 0.0028 7 4- 0.0004 +0.0038 8 — 0.0006 +0.0034 9 — 0.0024 +0.0018 10 — 0.0034 +0.0000 11 — 0.0044 — 0.0006 12 — 0.0042 — 0.0018 13 — 0.0044 — 0.0028 H — 0.0028 — 0.0030 15 +0.0000 +0.0000 2°. Reseau Corrections. The reseau used, Gautier No. 9, has been investigated by Pro- fessor Donner, and the result published.* On pages 38-40 of his paper will be found two tables of corrections to the measured co- ordinates ; one set for the ^-coordinates, and one set for the Y- cob'rdinates. They give the correction for the intersection of every vertical line with every horizontal line. The unit is a tenth micron. These corrections are applied in the following manner : Suppose I and I + 1 are the numbers of two consecutive .X-lines, and V and V + 1 the numbers of two consecutive F-lines, the four lines enclosing a square. Let K t be the correction for the line I, and K l+1 the correction for the line I + 1. Then E l is the mean of the tabular values for the intersection of the .X-line with the two F-lines V and V + 1; E l+ x is the mean of the tabular values for the intersection of the .X-line I + 1 with the same two F-lines. In other words, we find the correction for the middle of the side of the square. *Acta Societatis Seientiarum Fennicse. Tome XXI, Nr. 8. Determina- tion des Constantes necessaires pour la E^duction des Cliches pris a Helsingfors. 6 Stars within One Degree of the North Pole, and 3°. Error of Runs. The error of runs for each square can be obtained directly from the settings on the reseau lines. The method of correcting for it will be given in the derivation of the general formula for finding the star's coordinate from the measurements. 4°. Method of Obtaining the Corrected Cordinates. Let I = number of lower reseau line, l-\- i = number of upper reseau line, D = difference between the reading of the lower line and the reading of the star image, always positive. Q = error of runs expressed in units of the fourth decimal place, R = difference between the readings of the upper and lower reseau lines, always positive, = io.oooo± Q, K t = correction to position of lower reseau line taken from Donner's table, K l+l = correction for upper reseau line, X, Y= final coordinates. Then if we measure the coordinates from the line o, e. g., #-co- ordinate, X= 5 l + zi + & + (.E + i - K) | . Since B will not differ much from io r it will be sufficient if in place of -= we use — =f . The formula then becomes Ji io ioo In case we wish to measure the coordinates from the central reseau lines we should use for 5Z, $(l — 14) and 5(? — 43). This method of recording and reducing is due to Professor Jacoby. The specimen sheet appended shows in detail the method of ob- taining the final coordinate. The screw correction has been in- corporated with the general sum in order to shorten the process. Optical Distortion deduced from Photographic Measures. 1 The succeeding pages contain Table III, the table of corrected coordinates. It includes the coordinates, direct, reversed and mean, for both the big and middle images of every star on the four^plates. Stars within One Degree of the North Pole, and Polar Plate: Hels. Sept. 16, '95. No. 3. Line. December 2, 1896 : X Direct. Line. No. Micr. Big Image. 9-152 9.660 Sc. Mid. Image. Sc. No. Micr. 20 4-765 •754 5-250 .264 +15 + 6 9.142 9-639 + 15 + 6 21 I4-738 ■748 15.248 ■ 244 5.0082 9.4060 5.0082 4-3978 + 10 9-39°5 5.0082 4.3823 +10 14-9945 Low Line Read. DifE. = D. V*D Ki (Kh-i—Ki) — , D -X «?o Screw Corr'n Sum Lower Line, mm. Star's Coord. 5.OO82 9.9863 Q = —137 + 2.1989 + 6 + + 3° + 5 + 2.2030 — 3°- —32.2030 + 5 + 2.1912 + 6 + + 30 + 5 + 2.1953 — 3°- —32-1953 + 5 ■Ki+i +5 Ki= +6 Diff. = —1 Line. Noven ber 18, 3 896: X reverse d. Line. No. Mlcr. Big Image. Sc. Mid. Image. Sc. No. Micr. 21 4.776 •737 5.202 .232 10.364 10.812 — 2 — 5 10.375 IO.842 — 2 — 5 20 I4-750 •773 I5-229 .201 4.9868 10.5880 14.9882 4.4002 — 4 IO.6085 14.9882 4-3797 — 4 14.9882 Low Line Bead. DifE. = D. yi d Ki (zi+i-*i)r D Screw Corr'n Sum Lower Line. mm. Star's Coord. 4.980S 10.0014 Q = + 14 + 2.2001 + 6 + — 3 + 2 -f- 2.2006 —3°- — 32.2006 + 2 + 2.1898 + 6 + — 3 + 2 + 2.1903 —30. —32.1903 + 2 tfi+i +5 Ki= +6 DifE. = —1 Mean X — 32.2018 — 32.1928 Optical Distortion deduced from Photographic Measures. 9 Star 43. Measured by Furness. Line. No. Micr. 44 4-752 •75o 5-278 •279 5.0148 Low Line Bead. Diff. = D. y* d E, , D -#« To Screw Corr'n Sum Lower Line, mm. Star's Coord. December 3, 1896 : Y Direct. Big Image. 8.219 8.740 8.4795 5.0148 3-4647 + l-73 2 4 — 9 + 3 + 28 + 13 + 1-7359 +5- +6-7359 +31 +22 +26 + 13 Mid. Image. Sc. 8.925 9.446 9-I855 5.0148 4.1707 +2.0854 — 9 + 4 + 34 + 8 +2.0891 +5- + 7.0891 +20 +10 + 15 + 8 Line. No. 45 Micr. 14-733 .741 15-270 .250 Q 14.9985 5.0148 9-9837 -163 Ki= Diff. —9 +9 Line. No. Micr. 45 4.770 ■ .740 5-241 .260 5.0028 Low Line Bead. Diff. = D. E, «* B,-l* tPl i+l* t 0. r 2 r> 2r or In applying these formulas to the case of the polar plates, we shall assume that we know the usual constants for each plate, i. e., the coordinates of the pole, the angle of orientation and the scale value. Accordingly let X, Y be the rectangular coordinates of the star, ?, t\ be the rectangular coordinates of the point on the plate corresponding to the pole of the heavens, A toe the right ascension of the .X-axis in the sky, to be the scale value a a, - be the star's apparent right ascension and polar distance P/ f xW Y-i Fig. 2. From the figure we have the following relations, viz : p sin B = Y— v, p sin B' = — n, pcosB=X — f, p cos B' = — f, B" = B' — B. The angle B is the plane angle which we wish to convert into a spherical angle, but neither of its sides passes through the ori- gin, hence the auxiliary angle B" is used, such that B = B' — B" . The correction to be applied to B is i p 1 n = - -. sin B' cos B' 2 r ip- 2 sin B" cos B", = - -. sin B. 28 Stars within One Degree of the North Pole, and And thence a = B 4- - p 2 sin B<->* sin i" + .4, it =pu — - pp2 cos 2 .B"" 3 sin 2 i" — - p{X 2 + F 2 )u 3 sin 2 1" + ^# 3 j, A and u> if we know the right ascensions and polar distances of at least two stars. If there are more than two known stars on the plate, the resulting equations can be solved by least squares. Since we can obtain very good approximate values of these constants by a direct so- lution, it will be sufficient to find their corrections by the use of differential formulas derived from equations (1). If da and di: are changes in a and i: due to small changes in ?, ■f), A and u>, and we put a' = a -f- da, TT> = IT + dir, then we have . '. sin B cos 2? ... . a' = a-\ ; d£ r — -., dn + dA, p sin 1" p sin \" ' ■x f = it -\- pdu — u cos Bd!j — a sin Biv. Optical Distortion deduced from Photographic Measures. 29 Of these quantities a and n are computed from (i) with the approximate plate constants, a' and it' are the apparent places of the standard stars obtained by reduction from some catalogue. Each equation in right ascension should be multiplied by pa> sin i " to give it the same weight as the polar distance equations. If we make this multiplication and adopt the following substitutions Ike = (a — a') pa sin \" , n v — it — it', dA' = * On account of the symmetry of the coefficients in these last two equations the least squares solution can be simplified. Let — (J sin B = r, — a cos B = s, then the equations (2) become — rdf + sdq -\- pdA' -\- odw -\- n x = o, + sdl; + rdv + odA' + pd a + n y = 0. If a, b, c, d, n, and s, have the signification usually adopted in the Gaussian method of elimination, and the terms indicated in the square bracket are formed in the customary way, we have the following relations, [66] = [aa] , [a&] = o, [ac] = — [&d] , [ad] = [6e] , [cc] = [dd], [cd] = o. We can also make use of the ordinary check formulas, [ao] + [ac] + [ad] + [an] = [as], [aa] + [ad] — [ac] + [6m] = [6s], [ac] + [ad] + [ce] + [en] = [cs], [ad] — [ac] + [cc] -f- [dn] = [ds], * These formulas appear in a paper written by Professor Jacoby entitled "Photographic Researches near the Pole of the Heavens," and published in the Bulletin of the Imperial Academy of Sciences of St. Petersburg, V Series, Volume IX, No. 1 (June, 1898). 30 Stars within One Degree of the North Pole, and If we place F = [aa] ■ [cc] — [ae] • [ac] — [ad] • [ad], C = [ac] • [en] + [ad] ■ [an] — [cc] ■ [an], D = [ad] • [c»] — [ac] - [dn] — [ce] ■ [In], G = [ac] - [am] + [«d] • [8n] — [aa] ■ [en], J9 r = [ad] ■ [an] — [ac] ■ [bn] — [aa] • [dn], then ^ C ^ ^ .7,./ G .j ^ df=-, d, = -, d4' = -, da, = -, weight of df = weight of dn = = — , \_cc] F weight of dA' = weight of dw = -. — ; . [aaj (3) Another method of solving these equations has been published by Professor Jacoby,* the study of which suggested to me the above process. The present formulas, however, seem more suit- able for this work, particularly since they give a desirable check to the computation early in its course. Their symmetry and sim- plicity are at once noticeable. Before giving the results of the several least squares solutions., it will be necessary to explain somewhat in detail the method of computing a! and -', the apparent places of the standard stars. VI. Method of Determining the Apparent Places of the Stand- ard Stars. The plate constants f and t] are the rectangular coordinates of that point on the plate which is the projection of the pole of the heavens, but since the celestial pole is a moving point its projec- tion is not fixed, and we are at liberty within certain limits to make a choice of the epoch which fixes the exact position of the pole. £ and ij are determined through the solution of equations (2) by comparing the apparent places of the stars in the sky with their positions obtained from the measured coordinates, hence the manner of computing their apparent places must be consistent * Monthly Notices E. A. S., May, 1896. Optical Distortion deduced from Photographic Measures. 31 ■with the choice of pole. If, pursuing the usual process, we cor- rect the star places taken from some standard catalogue for pre- cession, nutation, aberration and refraction, f and jj will be the coordinates of the pole at the instant of observation, and A will be the right ascension of the .X-axis referred to the equinox of date. The places of the unknown stars obtained by the solution of equations (i) will be the apparent places at the instant of ob- servation, and must be corrected for refraction and reduced to the beginning of the year, after which a further application of precession is necessary to bring them to any desired epoch. It is possible to abbreviate quite considerably the labor of this long computation by selecting the position of the pole for some other instant than that of observation, for example, the epoch of the catalogue from which the standard stars are taken. Preces- sion and nutation do not affect the positions of the stars relative to one another, but merely change the planes of reference. Hence if we apply to the right ascensions and polar distances of the standard stars, proper motion, aberration and refraction, the con- stants " cot £ + sin ( S — A'" ) cot z] , J a, — s = — ?r w cos {6 — A'"). ) If we determine the right ascensions and polar distances of the known stars in this manner arid substitute them in equations (2) for the purpose of finding the plate constants, the values of £ and ■f) thus derived will be the coordinates of the pole of the epoch of the catalogue, considered to be refracted. A will be the right as- cension of the .X-axis reckoned from the equinox of the catalogue, and - co t*- _. hH « v -rf cO cO CO o> 6 a 8 " ■* . V *H v _(_ V CN OS O i-< ~t~ P* OS j> -5 3 K 1 O _i_ »o to 1 ioT or-*. 8* c\ . 5 co as • ^3- « OS " a CO cO IH "*1" B -cJ-"^ ^ % % * % Tt M VO CO VO ._ CM IO cO CO CN JO v v v | v W cn o m -\- M OS 3 ■o 2 ri 1 °m 4- o 10 R O HH 6 " 00 CO ff ^ o OS OS !2 ° u <0 3 ~ CO l-l 43 as as co cm 8 R 00 " sO OS -*3 ■ CD *; % 5j S; % •-« ""*■ cO CO cO a 60 M CN U*J CO CO CN M v v v , v CN tN "tf- m -h tN OS 1 -"T _i_ m io 1 o T^ o t^. ^ s S 1 ^cf> CO a . 8b io >o O "io ^b Vj vo ►H CN lO cO cO CN CN \b "d- cn -J- CO OS "3 ^5 j^o* 1 cr, 1 CN _|_ IO IO 1 o T^ o t^. N ,_; g 6 a o g ^- OS -i CN OS OS M V !>■ "* n V CO VD OS o t^ "3 E 4^ 05 io cO "? CO M O J % !j ^ rf m |>. ro r- «o M CN g c- OS ^ M H CO CO CO M CO 0) CO ** v v 1 v IO v£> co cn -h CO OS cn o J3 CO 1 \D i o r>. g 1 oS O (N, M a CO 5: % ^ 5; in r*. cn to cn HH cO iO . ■ co CO CO CN ^J- CO CN H- CO as CD to «3 ^ M CO ^. o _|_ IO io 1 OS T^ o t-^. § d rO cO lO _ "* OS • n CN l"- H O CO CN OS o f M OS +^ s3 u5 M M vo v * OS o «^ Sg ^ OS VO co o *>■ co O ro ^ -V 5; ^ ft; 5. -e M CO CO H IS rt H CO *-* M ^O a co H Tf Tt CO CO CN OS 't CN -J - CO OS "©- 60 IH CO ja o r-t ■■* _L IO IO or (L In s ts OS . TT CM OS CO ON 00 l-l *> o V V 1 ^° 1 1 *T -g cd « ^ § * 1 Mi a Optical Distortion deduced from Photographic Measures. 39 Remarks. j. The Helsingfors sidereal time corresponds to the middle of the time of exposure ; see Table I, page 3. 2. The longitude of the sun was taken from the Berliner Jahr- buch for 1895. It is there referred to the mean equinox of the beginning of the year 1895 aud was therefore decreased by the precession in longitude for seven years. 3. The obliquity of the ecliptic is the mean obliquity for 1888.0. It is taken from the Jahrbuch for 1888. 4. The effect of precession on the pole of 1888.0 was computed by. Chauvenet I. (664). The resulting right ascension and polar distance of pole of 1888 referred to the pole of date are A = i8o°2 / 43 // .34 ■k = i4o".37 The values of A'" and rJ" given in the table are obtained by the ordinary " reduction to date." using the Besselian numbers A and B taken from the Jahrbuch. They were further improved by the use of Fabritius's method. 5. The value of log h was taken from the Pulkowa refraction tables, in accordance with the data given in Table I. It was in- creased by -fa of itself to allow for the difference between visual and photographic refractions. I shall now give a rather full outline of the process of reducing one plate, selecting for this purpose plate I, middle images. An- nual aberration was computed according to equations (4); diurnal aberration according to the last term of (4) using O and 90 — z,, for and

a 1888.0 Annual aberr. Prop. Motion. Diurn. aberr. Differ, refrac. a apparent. n 1888.0 Annual aberr. Prop. Motion. Diurn. aberr. Differ, refrac. tt apparent. / a 41 26 23 + 3240 + 15 + 9 116° 833 — 924 — 5 — 9 O 1 II 193 33 33 — 16 18 — 8 + 7 2464654 — 648 — 3 + 4 O 1 II 293 55 50 + 854 — 6 + 5 — 9 4i 59 36 1297.25 + 13-97 + 0.12 — 0.42 115 58 55 3728.37 + 15-58 + 0.12 — 1. 19 193 17 14 3718.88 — 6.70 — 0.06 — i-35 37IO-77 246 40 7 2685.92 — 18. 16 — 0.15 — 0.77 294 4 34 3735-81 — 16.28 + 0-05 — 0.13 — i- 17 1310.92 3742.88 2666.84 3718.28 The plate constants are to be determined from equations (2). Approximate values were found quite easily. A, the orientation angle of the X-axis was placed equal to 66°, i. e., the right ascen- sion of the center of the plate taken from Table V, plus 90 . u> was assumed to be equal to 60" ; f and y, were obtained from the so- lution of the following simple relations : 07 = 71-'" sin {A'" — A), uf = 7r"/cos {A"' — A). A preliminary solution of (2) with equations derived from two stars gave approximate values of all the constants. This ap- proximation was so close that the further corrections did not exceed o^oi for £ and i;, and o".oi for A and u>, and the absolute terms with a few exceptions were less than 1". The following specimen equation of condition — 26dS — 54<2?? + 22dA' + o".37 = o shows that the coefficients and absolute terms are not of the same order of magnitude. In order to simplify the subsequent numer- ical work each coefficient was divided by 100, and each unknown quantity multiplied by 100, and the following substitutions made : X = ioodf , y = 100*7, 2 = loodA 1 , t = looda. We then have the following. equations : Optical Distortion deduced from Photographic Measures. 41 — .26a; — -54» +.223 +.ootf +-37 = o +.45 —.40 + .63 + .00 — .11=0 +■49 +■35 + .62 +.00 + .07 = +.01 + .60 +■45 +.00 +.01 =0 —•43 +.41 +.62 +.00 — .29 = —•54 +.26 +.00 +.22 +.19 = —.40 — 45 +.00 +.63 +.11 = +■35 — 49 +.00 +.62 +.00 = +.60 — .01 + .00 +•45 —.43 = +•41 +.43 +.00 +.62 — .84 = Solving these equations according to equations (3), we have the following values of the unknown quantities and their weights. <7f = + o mm .oo37, dy = + o mm .oo4i, dA' = — o".ooo9, dA = — 3", where s is the distance from the center, and a, b and c are con- stants to be determined from the equations. On account of the form of the terms, a cannot be separated from the scale value, hence b is the only constant obtained. The six plates in the Vic- toria region give for b the mean value, J = — o.oooooio, and the reduction of the two plates of the Pleiades furnishes for 6 the mean value, 6 = +0.0000021. * Corrections de Refraction et d'Aberration. Bull. Com. Perm., Tom. Ill, 1896. (Not yet published.) Optical Distortion deduced from Photographic Measures. 67 Calling attention to the small magnitude of these quantities and their difference in sign, Donner concludes* that the constant 6 of the term 6s 2 in the expression for the distortion does not ex- ceed a few units in the seventh decimal place, and since the cor- rections which might result are for all practical purposes con- fused with the corrections for the scale value of the plate, we can in practice neglect the distortion of the field of the objective. It will thus be seen that Donner's and Turner's methods are alike in comparing the measured coordinates with the coordinates obtained from other observations, for which purpose Donner used heliometer measurements and Turner, meridian observa- tions. Another similar investigation making use of the same stars in Gill's circular was carried out at the Observatory of San Per- nandof, one of the observatories participating in the astrophoto- graphic survey. The same formula e = ps 4- 8s 2 was employed, and its effect was found to be negligible. The de- tails of the method of research are not published in the Bulletin. It was stated in the introduction to this paper that one of Prof. Jacoby's objects in having a group of polar plates taken was to investigate the distortion of the Helsingfors objective by a new method which did not involve, except indirectly, any previous de- termination of the star places. His plan was to have the same group of stars photographed with the telescope in different posi- tions with regard to the pier, so that the object glass would be situated differently each time with regard to the stars in the group. Two different positions could be obtained by taking a group near the meridian, and reversing the telescope between the two exposures which method was indeed followed by Donner, though for another purpose. But the selection of the group about the pole made it possible to turn the telescope through any angle desired after each exposure. On reference to Table I, page 3, it will be seen that the right ascensions of the pole were b-, in* , l'-,.m ,!i~ ,m . I9»24 m , 22 n 24 m , I n 24 m , 4 n 24 n thus differing by 45". Strictly speaking, the pole has every right *Loc. cit., p. 65. t Bulletin du Comit6 Permanent, 1896; Proces-Verbaux, p. 52. 68 Stars within One Degree of the North Pole, and ascension, but those just quoted belong to the hour circles which are perpendicular to the X axes on the different plates. As before stated, the usual effect of optical distortion is to dis- place the center of density of a star image. If the displacement is only radial, it will change the distance of the star from the center of the plate, but will not change its position angle, but if, in addition to the radial displacement, there is a displacement in any other direction due to some special peculiarity of the object glass, such as would be suggested by a different scale value in two perpendicular directions, the combination of the two would alter both the position angle and distance of the star, and this ef- fect might be different over different parts of the plate. A dis- placement of this character would not be detected by Donner's method, which assumes that the distortion is the same along dif- ferent radii of the objective. Furthermore, if the plate holder always keeps the same position relative to the object glass, which is usually the case with telescopes employed only for photo- graphic purposes, and if the telescope is changed in position, then the same star appearing on different portions of the plates, will be displaced differently and a proper comparison of the right as- censions and polar distances of a large number of stars, obtained independently from the different plates would give the effect of this kind of distortion. This method of procedure was followed with the polar plates taken at Helsingfors. From the Table of Mean Results, forty -two stars (there marked with an asterisk), which appear on all four plates were selected on the basis of their distribution in right ascension and polar dis- tance. They were divided into zones, and about the same num- ber of stars was selected from each zone, as will be seen on refer- ring to the figures given below. I o"-i3oo" 2, 23, 42, 45, 46, 49, 53, 58, 60, 64, Total 10. II i3oo"-i8oo" 9, 11, 25, 26, 30, 41, 56, 62, " 8. III i8oo"-24oo" 8, 19, 2i, 28, 29, 32, 35, 39, 43, " 9. IV 24oo"-3ooo" 4, 12, 38, 44, 50, 52, 54, 65, " 8. V 3ooo"-36oo" 1, 3, 10, 14, 24, 47, 63, " 7. By means of these forty-two stars the four plates were reduced to a common standard as follows : A new set of corrections to the constants of reduction for each plate was obtained by a least squares solution according to equations (2) using the above forty- Optical Distortion deduced from Photographic Measures. 69 two stars for this purpose, and taking for the numerical terms the residuals found in the columns headed Aa sin n and Aiz of Table VII. For example the following equations are obtained from stars i and 2, plate I. V — 54." In general the effect of this new reduction was to decrease numerically the residuals to a small extent only. In fact it will be noticed in referring to the residuals in Table VII, that they are fairly well distributed both in magnitude and sign, and from this it was inferred beforehand that the corrections to ?, 1, A and <» would be small and would not materially change the characters of the residuals, as indeed turned out to be the case. A satisfactory agreement as regards the scale value appeared in the course of this reduction, and in some sense it is a test of the method. It can be quite easily perceived by an examination of the following figures which give the old and new values. Plate. (o Table VI. dot to (new). IUB + M) 59".8590 — // .0026 59"-8564 Ui(B + M) .8566 — .0010 .8556 Illi{B + M) .8528 + .0021 .8549 IVUB + M) .8561 + .0015 .8576 Mean 59".856i±o" .0017 59".856i ± o".ooo8 The following table gives the new residuals for the forty -two stars : 10 Stars within One Degree of the North Pole, and Table X. Star. B Aa Sin 7T P Air Star. s Aa sin fr p Air o // mm u mm I 294 +.07 58 +.08 12 350 +•'14 49 + •17 249 —■31 — 35 3°5 —.16 + .32 340 +■54 — .20 36 +•23 —•17 207 —•34 +■45 263 —.24 —•32 3 299 —.16 53 +.08 14 356 +.10 61 — r 3 254 — .11 +.07 311 +.15 —.07 344 +.04 —.16 4i +.14 +.09 211 +•23 + .02 268 -■36 + .11 4 317 —.04 50 —.26 19 8 —.08 40 —.12 272 —.18 —■32 322 +.06 — -°3 2 —.08 +.09 53 +.01 —•13 229 +.26 +•49 280 — .01 +•25 8 331 +.02 37 +•15 21 25 —•03 32 +.29 286 + .04 — 05 34o +.18 — .11 16 +.06 +.22 70 —■25 —.18 243 — 15 -•36 297 +.10 — .01 9 332 +•33 28 +•32 24 33 —•15 58 +.07 287 —.14 —.09 348 +.01 + •33 17 +•17 +•13 78 —.09 —.40 244 —•37 —•35 305 +.22 —•03 10 334 —.04 51 — .01 25 52 +•13 26 —.26 289 —.26 +.J5 7 +.23 +.16 19 + .00 +.10 97 +.04 +.07 246 -f.26 —.26 324 —38 +•03 ii 337 — .20 24 —•23 26 65 +•03 22 —.07 292 —.28 +•35 20 +.16 +.26 22 — .10 — 42 no — - T 5 — -37 1 250 +.56 I+.26 338 —.04 +■17 Optical Distortion deduced from Photographic Measures. 71 r rABLE X.- -( Gontir ,ued). Star. B An sin n P Air Star. B Ad Sill TT P Air 28 95 5o 140 8 +.07 +.24 —•51 +.21 mm 38 + .32 +:09 —.08 —•33 41 161 116 206 73 —•'31 +,19 +.06 +.06 mm 23 +.'i 9 " +•44 —■35 —.29 29 i°3 58 148 15 +.14 +•05 +.3° -.46 38 +•23 +.29 —.24 —.27 42 169 124 214 81 +.05 +.21 +,22 -.46 19 —.06 —.19 —.28 +•53 3° 104 59 150 16 +.41 —•17 —.09 — -13 27 — 15 —■23 + .08 +.27 43 170 125 216 82 —.14 —.19 +•31 +.04 32 —.18 +.24 +.06 —•13 32 123 78 168 35 —.32 +.18 — .02 +•17 34 +.01 —•23 +.25 —•°3 44 177 132 222 89 —•15 +.29 +.22 —•37 47 — .11 —.26 +.58 —.19 35 128 82 173 40 +.00 — .21 +.00 +.21 38 +•33 +.27 —.08 —•50 45 187 142 232 99 +•34 —.13 —•32 +.11 12 +.01 —34 — .02 +.38 38 144 98 189 56 +.07 +.21 —.14 — .12 27 —•37 +.12 +.12 +.12 46 195 150 240 107 +.19 —•33 —•35 +48 20 +.07 —■17 — .01 +.14 39 148 i°3 193 60 +.26 +.17 — .20 —.25 36 —.07 —■57 +.00 +.67 47 196 151 241 108 +.13 +.24 —•39 +.06 53 —.04 -.56 +•35 +■25 T2 Stars within One Degree of the North Pole, and Table X — (Concluded). Star. 49 5° 52 53 54 56 58 199 154 244 III 201 156 246 "3 214 169 259 126 214 169 259 126 218 173 263 130 225 1 So 270 137 230 185 275 142 +.19 —.28 —.10 +•19 +.03 —.16 — .12 +.27 —.13 +.04 —•23 +-3i +.22 —.16 —■05 —.04 +.09 — .20 +.05 +.10 — .12 +.13 +.19 —•25 —.29 +•15 — .01 +•13 mm 5 49 43 40 26 —.18 +.20 -■56 +.56 — .01 +.05 +.12 — 13 — .11 +.08 — -°3 +.06 +.15 +•25 — .20 —.18 —•15 +.01 — .01 +.18 +.00 +.21 +.25 —■47 +.19 —•05 +-°3 —•23 Star. B Aa sin it P Ajr II mm II 60 245 +.15 14 +.04 199 — 39 +•17 290 — .20 +•13 157 +•43 —•35 62 249 —.04 22 —.02 203 +•20 — .01 294 +.20 +.06 161 -•36 —.04 63 251 —.26 54 +.00 205 +•34 +.04 296 +.07 +.18 163 —.18 — .22 64 255 —.04 20 —.09 210 +.26 —■05 300 — .20 —.08 167 +.48 + 23 65 281 +.11 48 —.06 236 —.19 +•15 326 — .02 +.11 193 +.09 — .22 2 300 —.18 13 —.04 254 +.02 +.08 345 +.T2 + .24 212 +•03 —.27 23 32 —.19 14 +.09 347 —.24 +.21 77 —.19 —.07 304 +.64 —.24 Optical Distortion deduced from Photographic Measures. 73 To facilitate a comparison of the residuals they were plotted as ordinates of points whose abscissas were the position angles B; those for Aa sin jr forming the upper part of the diagram and those for An the lower. Thus the residuals from stars lying on the same section of the plates were grouped together. An ex- amination of the accompanying diagram shows that there is no decided tendency on the part of the residuals toward a grouping such as would indicate the presence of distortion. Indeed two stars close together on the plate have quite different residuals, vide 32 and 35. The plot of the points representing the residuals before the least squares reduction compared with the plot of the residuals after reduction, shows simply that the latter are more crowded together toward the axis of abscissas than the former, but there is no especial distribution. The conclusion is therefore reached that this method of com- parison does not show any indication of distortion in the field of the objective. This result is one that was to be expected from Donner's investigation, as well as from optical considerations. The tests to which the instrument maker subjects his lens before it leaves his hands as a finished product are severe and delicate. An elevation equal to the -&m two hundred thousandth part of a?t wave-length of light can be detected, and the tests for striee and variations in densit}' are equally exacting. The radial distortion is a natural consequence of the curve of the glass, but a displace- ment in position angle or a deformation such as was suggested by the discrepant scale values in the two coordinate directions, must in all probability be due to something which happens to a glass after it leaves the hands of the maker, as for instance, a strain due to the position of the lens in its cell. Hence, the general conclusion is reached, that so far as the Helsingfors telescope is concerned, there is no evidence of optical distortion, and that in this respect its performance is most satisfactory. u Stars within One Degree of the North Pole. r*7 X Pi •a is "a V "*•