i^^ 
 
 
 a^ 
 
 ^■??fe, 
 
 IMAGE EVALUATION 
 TEST TARGET (MT-3) 
 
 ■/ 
 
 
 -'% 
 
 X 
 
 ?l^./4!;fiM 
 
 ■|S,fi, 
 
 1.0 
 
 
 1125 i 1.4 
 
 ■ 2.0 
 
 Mil 
 
 ■ 1.6 
 
 6" 
 
 ^Sciences 
 Corporadon 
 
 23 WIST MAM STMIT 
 
 WnSTlLN.Y. M5M 
 
 (71«)t7a-4S03 
 
 r<\^ 
 
 \ 
 
 ^v 
 
 V> 
 
 ^ 
 
 
 
 v\ 
 
CrHM/ICMH 
 
 Microfiche 
 
 Series. 
 
 CiHIVI/ICIVIH 
 Collection de 
 microfiches. 
 
 CaiMdIan Imtltuw for HInorlul Mleroraprixluetlora / ImUtut cnwdim d* nUcrorapRMluotloiM MMoriqiiM 
 
 a&s«s«iiSsi<»««fc.-9 < A#!««i«^'. ^-<-"- -**#^^. 0-.W..— 
 
 *■ 
 
Technical and Bibliographic Notes/Notes techniques et bibiiographiques 
 
 The institute has attempted to obtain the best 
 original copy available for filming. Features of this 
 copy which may be bibliographlcally unique, 
 which may alter any of the images in the 
 reproduction, or which may significantly change 
 the usual method of filming, are checiced below. 
 
 D 
 
 
 
 D 
 D 
 
 D 
 
 Coloured covers/ 
 Couverture de couieur 
 
 I I Covers damaged/ 
 
 Couverture endommagie 
 
 Covers restored and/or laminated/ 
 Couverture restaurAe et/ou pelliculAe 
 
 Cover title missing/ 
 
 Le titre de couverture manque 
 
 □ Coloured maps/ 
 Cartes g^ographiques en couieur 
 
 □ Coloured ink (i.e. other than blue or blacit)/ 
 Encre de couieur (i.e. autre que bleue ou noire) 
 
 I I Coloured plates and/or illustrations/ 
 
 D 
 
 Planches et/ou illustrations en couieur 
 
 Bound with other material/ 
 RellA avec d'autres documents 
 
 Tight binding may cause shadows or distortion 
 along interior margin/ 
 
 La reliure serr^e peut causer de I'ombre ou de la 
 distortion le long de la marge intirieure 
 
 Blank leaves added during restoration may 
 appear within the text. Whenever possible, these 
 have been omitted from filming/ 
 II se peut que certaines pages blanches ajoutAes 
 lors d'une restauration apparaissent dans le texte, 
 mais, iorsque cela Atait possible, ces pages n'ont 
 pas AtA filmAes. 
 
 Additional comments:/ 
 Commentaires supplAmentaires: 
 
 L'Institut a microf llmA le meilleur exemplaire 
 qu'il lul M AtA possible de se procurer. Les details 
 de cet exemplaire qui sont peut-Atre uniques du 
 point de vue bibliographique, qui peuvent modifier 
 une image reproduite, ou qui peuvent exiger une 
 modification dans la mAthode normale de filmage 
 sont Indiqute ci-dessous. 
 
 □ Coloured pages/ 
 Pages de couieur 
 
 □ Pages damaged/ 
 Pages endommagtes 
 
 □ Pages restored and/or laminated/ 
 Pages restaurtes et/ou pe.'MculAes 
 
 I — I Pages discoloured, stained or foxed/ 
 
 D 
 
 Pages dAcolortes, tachetAes ou piqutes 
 
 Pages detached/ 
 Pages d6tach6es 
 
 Showthroughy 
 Transparence 
 
 Quality of prir 
 
 Quality inigale de I'impressior 
 
 Includes supplementary materii 
 Comprend du materiel suppKmentaire 
 
 Only edition available/ 
 Seuie Edition disponible 
 
 r~~1 Pages detached/ 
 
 Jr~l Showthrough/ 
 
 I I Quality of print varies/ 
 
 I I Includes supplementary material/ 
 
 r~~| Only edition available/ 
 
 Pages wholly or partially obscured by errata 
 slips, tissues, 9tc., have been refiimed to 
 ensure the best possible Image/ 
 Les pages totalement ou partiellement 
 obscurcies par un feuillet d'errata, une pelure, 
 etc., ont 6t6 fllmtes A nouveau de fapon A 
 obtenir la meilleure image possible. 
 
 This item is filmed at the reduction ratio checked tMlow/ 
 
 Ce document est fiim6 au taux de reduction indiquA ci-dessous. 
 
 10X 14X 18X 22X 
 
 26X 
 
 X 
 
 12X 
 
 16X 
 
 20X 
 
 24X 
 
 Itol 
 
'e 
 
 6tails 
 m du 
 raodifiar 
 tr una 
 llmaga 
 
 B8 
 
 arrata 
 
 I to 
 
 i 
 
 I palura, 
 
 on A 
 
 n 
 
 32X 
 
 Tha copy fiimad hara has baan raproducad thanks 
 to tha ganarosity of: 
 
 _,. Library of Congrass 
 
 Photoduplication Sarvica 
 
 Tha imagas appaaring hara 9f tha bast quality 
 possibia considaring tha condition and iagibility 
 of tha original copy and in kaaping with the 
 filming contract spacification^. 
 
 Original copias in printad papar covars ara fiimad 
 baginning with tha front covar and anding on 
 tha last paga with a printad or illustratad impras* 
 sion, or tha back covar whan appropriata. All 
 othar original copias ara fiimad baginning on tha 
 first paga with a printad or illustratad impras- 
 sion, and anding on tha last paga with a printad 
 or illustratad imprassion. 
 
 Tha last racordad frama on aach microficha 
 shall contain tha symbol — »> (moaning "CON- 
 TINUED"), or tha symbol V (moaning "END"), 
 whichavar applias. 
 
 Maps, platas, charts, ate. may ba fiimad at 
 diffarant raduction ratios. Thosa too larga to ba 
 antiraly includad in ona axposura ara fiimad 
 baginning in tha uppar laft hand cornar. laft to 
 right and top to bottom, as many framaa as 
 raquirad. Tha following diagrams lllustrata tha 
 mathod: 
 
 t 
 
 ( 
 
 a 
 
 3 
 
 L'axamplaira fllmA f ut raproduit grica A la 
 gAnArositA da: 
 
 Library of Congrass 
 Ph^tpduplication Sarvica 
 
 •*»*<?•/_ 
 
 -;i# 
 
 Las imagas suivantas ont *t4 raproduitas avac la 
 plus grand soin, compta tanu da la condition at 
 da la nattat* da l'axamplaira filmA, at an 
 conformity avac las conditions du contrat da 
 fiimaga. 
 
 Laa axamplalras originaux dont la couvartura an 
 papiar ast ImprimAa sont filmte 9n commandant 
 par la pramiar plat at an tarminant soit par la 
 darnlAra paga qui comporta una amprainta 
 d'imprassion ou d'illustration. soit par la sacond 
 plat, salon la cas. Tous las autras axamplalras 
 originaux sont filmte an commanqant par la 
 pramiAra paga qui compoiita una amprainta 
 d'imprassion ou d'illustra*ion at an tarminant par 
 la darnlAra paga qui comporta una talia 
 amprainta. 
 
 Un das symbolas suhrants apparattra sur la 
 darnlAra Imaga da chaqua microficha, salon la 
 cas: la symbols — »> signifia "A 8UIVRE", la 
 symbols V signifia "FIN". 
 
 Las cartas, planchaa, tablaaux, ate, pauvant Atra 
 fllmte A das taux da rAductlon diff Arants. 
 Lorsqua la documant ast trop grand pour Atra 
 raproduit mn un saul cllchA, 11 ast filmA A partir 
 da I'angla supArlaur gaucha, da gaucha A drolta. 
 at da haut an bas. an pranant la nombra 
 d'imagas nAcassalra. Las diagrammas suivants 
 lllustrant la mAthoda. 
 
 t 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 — -«fr'" 
 
^^w 
 
 t ?■ 
 
 i I 
 
 •^ i 
 
 
 iJ 
 
 ^ 
 
 ■■ I III III 
 
 *i^ 
 
mpa 
 
 A.PPE3Sri>IX I. 
 
 ■ «0l^' 
 
 DESCRIPTION 
 
 or THR 
 
 TRANSIT CIRCLE 
 
 or THB 
 
 UNITED STATES' NAVAL OBSERVATORY, 
 
 ■w 
 
 i 
 
 WITH AM 
 
 / 
 
 INVESTIGATION OF ITS CONSTANTS, 
 
 PBnpARn> BT emtmB, or 
 Rear Admiral CHARLES HENRY DAVIS, U. 8. N. 
 
 ■ VPRItlHTBiriDXVT, 
 
 BY SIMON>^NKWOOMB. 
 
 noriMOBiwiun^uiiet, «. «. h. 
 
 /'WASHINGTON: 
 OOVEBNMENT PRINTING OFFICE. 
 1867. 
 
 ^1 
 
■I 
 
 
 ' 
 
 ■^ t 
 
mmiimmmi''^mmmmmmmmm'''mmmmmm^ii'' 
 
 ^'^mmm 
 
 •% 
 
 TABLE OF CONTENTS. 
 
 ^••«^»«hm of lfc« iaitrauMBt and Ht a^aaste 
 
 OtMty ncOod of iannal%irtl«g Ot «nran or » 1V«Mtt Canh .... .. 
 
 &m"%MiM «r tiM eoiMlMti of tin IVHub <%«d« ud ita t^fridiHy 
 IWtiftrMhwpNrfUawran , 
 
 PeifonauMorthtiaitranaat 
 
 A^MbitiMltedM. ...,. ' 
 
 1 
 
 9 
 
 ai 
 
 41 
 •46 
 40 
 
 H 
 
 
, 'f 
 
 ii 
 
 ""ww i i i lii n iiiii i i iiii i iiti M 
 
 -mm iwWiafiiiii^iiia-ii, ..:^,^. 
 
 '""T*?"ffTi1 i rT i iiHfrVY''- i '^ f-f' "■ i r'- nrnt i ftf- r U i b fa 
 
■Plp^nnRiH 
 
 INTBODDCTORY NOTE. 
 
 The imtrament denribed in the following pages wm proonred for the ObMrvatory by the 
 late Oaptaio CKUIm. On hie aeoewon to the Saperintendency, the want of a snttable Meridian 
 Circle was atrongly felt. The military operations of the government temporarily delayed the 
 supply of this want In October, 1868, however, the necessary authority was grant#.d by the 
 Hon. Secretary of the Navy, and the instrament was ordered from Messrs. Pistor and Martins, 
 of Berlin. Correspondence respecting the sine and pbu of the instrument occupied the remain' 
 der of the yew, and terminated by concluding on an object glass of at least eight Paris inches 
 dear aperture, and by leaving the plan of the instrument and its mountings altogether in the 
 bands of the artiste. 
 
 The pa^s of the instrument arrived in October, 1866. The work of mounting was com- 
 menced on the 16th of that month, and on the 28th the instrument was in position. The next 
 two months were occupied in determining the errors of flexure and division, and in making the 
 necessary preparations for active work. Begnfor astronomical observations were commenced 
 on January 8, 1866. 
 
 1 
 
 m 
 
\ 
 
 ,.:i 
 
 % 
 
 i 
 1 I 
 
 ■\ 
 
 ii 
 
 ■ \ 
 
 iMaMiMMIIMNiiliMI 
 
APPENDIX I. 
 
 DESCRIPTION 
 
 or TBI 
 
 TRANSIT CIRCLE 
 
 OrTBC 
 
 UNITED STATES NAVAL OBSERVATORY. 
 
 1 
 
 i 
 
 PAST I. 
 
 DB80BIPTI0N OP THE INSTRUMENT AND ITS ADJUNCTS. 
 OENBBAL DESCBIPTION. (PLATE m.) 
 
 (1) The instrnmeDt is moanted in the west wing of the Observatory, the room formerly 
 occapied by the Transit Instrument The interior dimensions of the room are 24.3 feet from 
 east to west, and 18.4 feet from north to south. On the north and south sides of the room are 
 built two recesses, each six feet in length, and six in depth, to make room for the collimators. 
 The slits in the wall and roof for observing are thii'ty inches in breadth, and closed by four 
 shutters on the roof, and a door on each side. 
 
 (2) The piers are solid monoliths of marble. The forur of each pier is that of a frustrnm of 
 a pyramid, surmounted by a prism. The base at the floor me.>'iures 44 inches from north to south, 
 and 38 inches from east to west. The top of the pyramidal portion is 80 inches above the floor, 
 and measures 18.6 inches from north to south, by 24.6 from east to west. The dimensions of the 
 ■lal iiBg horizontal sections of the prism are the same, while the heighv of the prisms is 28 inches; 
 the whole height of the piers 9 feet above the floor. The inside faces of the piers form a ver* 
 tical continuous plane 19 inches in breadth, and 9 feet in height. On each side of this plane 
 the stone is cut away in the form of a section of a hollow cylinder. The distance of the inside 
 faces is 64 inches. A cylindrical hole 6 inches in diameter is cut from east to west through 
 each prism for the illumination. 
 
 Into these openings the inside ends of hollow brass cylinders are flrmly set with plaster to 
 the depth of 6 inches. The outside end of each cylinder expands into a disk 8| inches in diam* 
 eter and 0.7 of an inch from the face of the pier. The arms which cai:ry the four reading micro* 
 scopes of each circle are attached to these disks, radiating from the central axis at an angle of 46° 
 from the horizontal and vertical directions. Within each disk, near its centre, is a system of 
 prismatic reflectors for illuminating the divisions of the circle. The illuminating lamps are 
 each at the large end of a conical tube, the small end of which extends through the opening of 
 the pier and fits into the interior of the cylinder carrying the microscope disks. The large end, 
 carrying the lamp, extends three feet from the outer face of the pier. One lamp illuminates 
 the field of the telescope, the other the wires. 
 
 The Ys are fastened into semi-^cyliwlrioai pieces of brara which extend inward from the 
 head of the disk carrying the microscope arms, the axis of the cylinder being a continuation 
 of that of the openings in the pier, as shown in Plate lY. 
 
2 DUBORTPnON OF THK TBAi.lIT OIBOUB OP TBI 
 
 (3) The telescope is of 12 feet focal length, and 8.6 inches dear apertare. The eye-piece 
 is furnished with a system of twenty*three fixed vertical wires, (eight of which it is intended to 
 remove,) and two borizonta' ones, distant 8". There is also a horizontal and a vertical microm* 
 eter screw, the former carrying one vertical, and the latter four horisontal wires — a central 
 pair, distant ^".f>, and two single ones, 2^' each side of this pair. 
 
 (4) The circles are each 42 inches in diameter, and divided on silvc to every S'. The 
 cylinder on the clamp-end of tne axis also has a coarser division to every 10' for setting. The 
 general character of the arrangement of circles, clamp, counterpoises, Ac, may be seen by ref* 
 erence to Plate IV. 
 
 Notwithstanding its dimensions, the instrument is reversible, and the operation of revers' 
 ing can be performed by a single person with great facility. The entire weight of the mov* • 
 able part of the instrument is only about 900 pounds. 
 
 (5) The sides of the central tube of the telescope are pierced by openings 2) inches in 
 diameter, through which the collimators may be set on each other when the instrument is ver- 
 tical. These are not shown in Plate III. 
 
 (6) The instrument is completely spanned from north to south by an arched flight of steps 
 for reflection observations of stars. They are so flgured that when the telescope is at any point- 
 ing between 120° and 240^ of zenith distance, the eye-piece will be in a convenient position to 
 look into. Above the fifth step the arch is bifurcated, so as not to interfere with the line of 
 sight. The highest step is a platform three feet in length, suspended from the roof by iron 
 bars and braces. Hand-rails, net shown in the plate, extend from the bars nearly to the floor. 
 
 (7) In the spring of 1867 another mechanical improvement, for convenience and certainty 
 in observing the nadir point, was introduced. On each side of the platform, over the axis of 
 the instrument, a seat is erected. The observer can sit astride of either seat and look into the 
 eye-piece when the telescope points to the nadir. On the inside of each seat, between the 
 observer and the telescope, a board, eight to nine inches wide, rises from the platform nearly to 
 the eye-piece. Bach of these boards is furnished with a pair of shutters of the same size, which 
 the observer can turn so that the tube of the telescope shall be completely enclosed in a wooden 
 hexagonal prism, or, more exactly, a frustmm of a pyramid, and thus protected from the heat 
 of the observer's body. 
 
 (8) The steps for reading the microscopes need no explanation except that a hand-rail 
 runs along the platform, by which the observer passes from one side of the pier to the other, 
 without descending to the floor. 
 
 DETAILED DESCRIFTIOM, WITH EXPLANATIONS OF THE PLATES. 
 
 (9) Plate I is a plan of the observing-room. 
 
 Plate II is a section of the walls and masonry below the floor in the plane of the meridian 
 of the instrument, with a view of the room as seen from the west 
 
 B is the entrance from the main building. It is dosed by two doors. The floor of the 
 room being thirty inches lower than that of the main building, a platform and flight of steps is 
 erected inside the door for convenience in entering the room. 
 
 (10) Below the floor all the masonry is of rough stone set in lime and sand mortar. The 
 base of the masonry rests upon the ground about six feet below the flooring joists. 
 
 L L are the collimator piers, the bases being of masonry, three feet square, and the upper 
 parts octagonal monoliths of marble. 
 
 S S are piers which support the turn-table, T, and the floor of the room under the in- 
 strument. ' ' 
 
 B, plate n, shows the masonry which supports the circle itself. B is a prism of the ma- 
 sonry already described, eleven feet from east to west, four and a half from north to south, and 
 
 ' 4MIHik4w«MMA<>*Wita 
 
 ^m 
 
milTID ITATia HAYAL OBSUTATOtT. 
 
 1 
 
 five io height. lU boricontal diroensioos are therefore only tuffioient to support the great piers 
 of the inatrnment. It is covered by a solid oap Q, Plate II, of hard bluok stone, of the same 
 horiaontal dimensioos, and about one foot m tbiokness. 
 
 On this cap rest the marble piers P, P'. Their bases are Ijollowed oat ho that they each 
 reet on three points, and thus remain secure without cement or other fastening. Their great 
 mass insures perfect steadiness without such aids. 0, 0, show how the inside corners of the 
 piers are cut away. The concaye faces 0, 0, 0,'0, are parts of the surfaces of four vertical cyl< 
 inders, the axes of which are in the plane of the inside face? of the piers, 23^ inches north and 
 south of the middle, and therefore 47 inches apart. 
 
 The perforation through each pier, at the end of the axis of the instrument, is shown at 
 0, Plate II. 
 
 (11) B, B, R, B, B, R, show the railroad on which the reversing carriage runs. Under 
 the instrument the rails rest upon two strong joists, supported by the piers S S. The turn- 
 table T revolves on six cannon balls. By it the reversing carriage, with the circle on it, may 
 be run into either the northeast or the northwest corner of the room. The turn •table is not 
 necessary in reversing, as the T's uf the reversing carriage themselves revolve on a pivot. 
 The whole floor is on the same levbl, except around the collimator piers, on three sides of which 
 a platform is built. 
 
 (12) Plate IV exhibits » i nd dlevation of the instrument from the south, with the hanging 
 level and so much of the reve-tiing carriage as is not concealed by the telescope. The ladders 
 are omitted. 
 
 Plate y is an isoa. ;'ic side elev> ion from the eaitt; the east pier with its appurtenances, 
 the step-ladders, and the hanf o^ level, all being removed. 
 
 (13) a is a movable rt' ji, 'or supporting a lump in observing the nadir poiut. b, b are the 
 counterpoises which lighten the weight of the instrument upon the pivots. The levers o, o 
 which support them have a siigbt movement around the pivots &, d. The hooks ,/*, / nave 
 friction rollers at the bottom for supportiut^ most of the weight of the instrument. At the top 
 they expand into a strong rectangular frame, throiu^h which the ond of the lever c paitsen, as 
 shown at o, Plate Y. Through the top of th«i frame passes the screw d^ the lower end of 
 which being rounded off, rests in a socket in the top of the lever, and thus supports the weight. 
 Small pieces of rubber have beeu placed in these sockets and under the supporting screws, 
 for reasons to be explained hereafter. When the instrument is raised from its pivots the 
 counterpoises are supported by the screws e, e, which are adjusted so as to allow a small play 
 to the levers. 
 
 (14) Jfoitn/tngr <tf the Microtoopea. — Enlarged views of the hollow cylinder D with its attach- 
 ments are shown in Plate YI. The portion P iaset into the pier with plaster; grooves are cat 
 through the flanges p, p, p to, let the plaster run through. I is the disk to whi<^h the microscope 
 arms are fastened. It expands in thickness toward its circumference. The portions J and k 
 of the microscope arms hold the disk firmly between them, being held together by means of 
 clamping screws, the heads of which are shown in Fig. 1. These screws pass through^' to k, just 
 outside the edge of the disk. To keep j in position when these screws are loosened, it is fitted 
 with pins which fit corresponding holes in k. By loosening the screws the microscope arms 
 can be moved independently around the disk, but will not come nearer to each other than 
 about 48°. 
 
 Outside of k, the microscope arms are loosely encased in mahogany shields I, to protect 
 them against rapid changes of temperature. 
 
 The outer ends of the arms are hollow, holeb about f-inch in diameter running into them, 
 to receive the holders which carry the microscope. They are encircled by the circular clamps 
 t, which are tightened by means of the screws h, partially shown in Fig. 2, but hidden by the 
 
 i 
 
 1 
 
 Kk Ji 
 
^^m 
 
 DESCRIPTION OF THE TRANSIT CIRCLE OF THE 
 
 Tjr 
 
 
 t4M( 
 
 clatnpH in Fig, 1. When /♦ is tightened, the microscope is firmly held by the arm. By loosen- 
 ing it the microscope may be drawn out, and three views, as thus removed, are given in Plate 
 VII, Figs. 1, 2, 3, each of which is from a position at right angles to the other two. 
 
 (15) a (Plate VII) in the solid metallic support which passes into and is held by the micro* 
 scope arm, as already described. The holders b and the circular clamps c, enveloping the body 
 of the microscope, form a single piece with a. When the screws d are loosened, the microscope 
 can be moved longitudinally, and the end e carrying the object glass can be slid into or out of 
 the other portion, for adjusting the focus, and the angular value of a revolution of the microme- 
 ter. For the accurate adjustment, the independent clamp y is used. Loosening d and tighten- 
 ing y by means of the screw d, a fine motion is given the microscope by turning the screw c. 
 The head of this screw ic kept pressed against b by the spiral spring ^ pressing y and b apart. 
 
 (16) is a perforated reflector for throwing light on the divisions. Its reflecting surface 
 is not polished, but is of a bright white. It can be turned in any direction by the milled head/, 
 and can be moved longitudinally by loosening the clamp g. 
 
 k \b & cylinder of white felt, extending very nearly to the face of the circle, at the same 
 time keeping out stray light and assisting the illumination from o. 
 
 (17) The micrometer screws of the microscopes revolve with the head. Instead of being 
 attached to the moving frame of the diaphragm, they screw into the latter, and thus give it a 
 slow motion to or from the bead. Each revolution of the microscope micrometers is 30" in 
 angular measure on the circle, and about one hundredth of an inch in linear measure on the 
 screw ; 30" on the circle measuring about -jaT ''^^^i ^^^ image of the divisions is magnified 
 about 3.3 times in the focus of the microscopes. The face of the circle is about 1.9 inches 
 from the object glass and the micrometer wires about 8.3 inches. 
 
 Each microscope micrometer is set upon the image of the division by a pair of parallel 
 spider lines, distant from 10" to 11". Microscope YII is also supplied with two extra pair of 
 wires, about 2' on each side of the original pair, for the convenient measurement of consecutive 
 divisions. 
 
 The micrometer head is divided into thirty second -spaces, each of which is again sub- 
 divided by a shorter line to half seconds. The entire revolutions are read from the interior by 
 a serrated scale. The divided drum may be turned on the axis of the screw by simple pressure, 
 without turning either the screw or its bearing. 
 
 (18) Supports of the Pivots. — Referring again to Plate VI, the cylinder D is hollow to the 
 depth of about an inch, the external part in fact consisting of a rim about an inch thick. The 
 cylindrical space forming the interior of this rim contains the metallic plate e, Fig. 1, into which 
 screw the antagonistic screws 8, a, the beads of which press against the inside of the rim. The 
 top and bottom of this plate are bevelled toward the pivot, and the bevelled edges are held by 
 the plates b, b'. When the screws c, c are loosened the plate e can be moved from right to left 
 by the screws s, «, but when they («re tightened, e is firmly held between the plates 6, b'. 
 
 The view actually given is that of the west plate, by which the axis is adjusted in azimuth. 
 The east one is similar, except that the plates and screws are turned 90° for adjusting the level. 
 
 n is a thin perforated plate of metal which presses against the end of the pivot, to prevent 
 longitudinal motion of the axis. As now arranged for observation, the eastern one is firmly 
 screwed against the V plate, while the western one is pressed against the pivot b}' a spring 
 behind it. 
 
 The V's are carried in the rounded plate y, which forms one piece with the plate c. They 
 consist simply of small pieces of soft metal v v, (Fig. 4,) slightly convex on their upper surfaces, 
 so that only a single point of each pivot at first rests upon them. But a slight hollow is soon 
 worn into each of them by the pivot. 
 
 "h"^^ 
 
UNTTKO STATES RATAL OBSBBYATOBT. 
 
 J 
 
 (19) The Tdeacope and ita External Appendage*. — Betnrning to Plates lY and Y, and para* 
 ing from the Y'b, we first have the pivots, which are 2.1 inches in diameter, and 1.8 inch in 
 length, and are apparently of the same diameter throughout their length. 
 
 (20) Next to the pivots the axis expands into the frustrum of a coue, the diameters of the bases 
 of which are somewhere about three and a half and four and a half inches, respectively, and the 
 height about three inches. Back of this frustrum the axis again expands perpendicularly to ita 
 length. Over these frustrums the circles C, fit with great nicety; the central perforations in 
 the latter being tightly filled by the former when the circles are pressed against the expansion 
 at their bases. The thickness of the circle is slightly greater than the depth of the frustum, 
 so that the latter does not pass quite through the perforation. The circle is held in its place 
 by the friction of the plates p, p which are pressed against it by the screws a, a, a, a, «, (Plate Y) 
 passing through them and into the top of the frustrum, without touching the circle. When these 
 screws are loos^ened, the circle can be turned round so as to take any desired position relatively 
 to the telescope. 
 
 (21) The circles each appear to be cast in a single piece. The manner in which they are 
 stiffened can be seen by a comparison of Plates lY and Y. The middle of the plane face of 
 each circumference is inlaid with a band of silver on which the divisions are cut to every two 
 minutes. The breadth of each division as seen under the microscope seems to be between two 
 and three seconds, corresponding to a thickness of about 7^ of an inch. 
 
 The circle next to the clamp has also a coarse division to every ten minutes, for setting 
 the telescope. 
 
 (22) Next to the circles come, on one end of the axis, the clamp n, and on the other end 
 a ring to counterbalance it, into which screw four handles A, h for turning the telescope on its 
 axis. The clamp is tightened by means of the screw q, q, (Plate Y,) and a slow motion may 
 then be given the telescope by another screw, not shown in the figure. 
 
 (23) The friction rollers are received by grooves cut around the axis. 
 
 On the clamp end of the axis four curved handles A', V (Plate lY) are screwed into the cone 
 for turning the telescope. 
 
 Near the base of the cone flat bands are seen, by which the instrument is supported when 
 on the reversing carriage. 
 
 (24) The central cube is 16 inches square, and forms a single piece with the axis. Its sides 
 are perforated in a direction perpendicular to the telest-ope and the axis, by openings 2^ inches 
 in diameter, through which light can pass from one collimator to the other. They are closed 
 by covers which screw into them. 
 
 The tubes of the telesi ope are each fastened to the cube by sixteen screws, as shown in 
 Plate lY. 
 
 (25) The Eye-piece. — Three views of the eye-piece and its appendages are given in Plate 
 YIII. From Figures 2 and 3 it will be seen that there are six 8<)ries or strata of plates in the 
 eye-piece. The first or outside stratum, of which a face view can be seen in Fig. I, carries 
 the ocular a and the slide for giving it a horizontal movement by means of the rack work and 
 milled head b. 
 
 The second stratum consists of the slide and fixed guides for giving a vertical aiiovement 
 to the whole of the first plate, with its guides and slide, by the milled head c. 
 
 (26) The third stratnpa consists of the declination micrometer plate with its guides. A 
 portion of its outside surface is seen at of, d, d, d, Fig. 1 . e is a fixed index for reading the 
 approximate position of the plate in micrometer revolutions. / is an opening in the plate to 
 admit of the free passage of e. An opening about two inches square is cut in the middle of 
 
 
 ■« 
 
f^m 
 
 e 
 
 DB80BIPTI0N OF THE TBAN8IT OIBOLB OT THK 
 
 61 
 
 m 
 
 the plate, and across this opening on the posterior sarfaoe of the plate are stretched the hori* 
 zontal spider Hues for zenith distance observations with the micrometer. From Fig. 3 it will 
 be seen that the micrometer plate d carries the falcram of the milled head o by which the 
 ocular is moved vertically. Consequently, when the micrometer is moved, the ocalar is carried 
 with it, and the movable spider lines are thus kept in the middle of the field without touching c. 
 
 (27) The fourth plate is moved in right adcension by the micrometer head B. It has a 
 rectangular opening like Plate III, and across this opening, on the anterior surface of the plate, 
 is stretched a single vertical spider line. 
 
 (28) The fifth plate carries the fixed spider lines. These are stretched across the anterior 
 surface of a thin annnlns about 1^ inch in diameter, which projects from the surface of the fifth 
 plate by an amount equi ! to the thickness of the fourth plate, passing through the rectangular 
 perforation in the latter. Thus, the three sets of spider lines are sensibly in the same plane. 
 The entire plate can be moved horizontally by the three antagonistic screws g hg, to adjust 
 the line of collimation. The plate is fastened to the piece through which the screws pass by 
 the projection t. 
 
 The sixth plate forms the basis for all the rest, and is pierced for the holes in which the 
 micrometer screws turn. It is fastened to the eye-tube E by a cylinder which slides tightly 
 into E, and is fastened by the six screws k. 
 
 ,. (29) The Uterometer Movements. — In all the micrometers of this instrument the screws 
 revolve with the heads. In the microscope micrometers the bearings of the heads are fixed, 
 and the screws passing into the movable plates draw them toward the micrometer heads as the 
 screws turn forward. But in the eye-piece the female screws are bored into the sixth plate, and 
 the bearings of the heads move with the micrometer plate. The projection d of the senith dis- 
 tance micrometer plate is attached to a cross n, n, n,(Fig. 3) through a circularr^be opening in the 
 centre of which the micrometer screw passes freely, and on which it presses. Thus, as the screw 
 turns forward, the screw, the micrometer head, n, and d are all slowly moved downward. Tho 
 middle n is the index from which the fractions of a revolution are read, but they can equally 
 be read from the front index, a is one of two spiral springs by which the cross is kept pressed 
 against the micrometer head, and made to follow it as it is withdrawn. 
 
 (30) Rogers^ Sdf-regiMlering Micrometer Head. — ^The zenith distance micrometer head is 
 famished with the self-register invented by Mr. Joseph A. Bogers, and described in the Astro- 
 nomische Nachrichten, No. 1493. The four arms o, o', o", o"' (Fig. 1) are movable around the 
 same centre with the micrometer, which passes through openings in the middle of them. They 
 are held to the micrometer by the friction of elastic bent plates which surround the centre. 
 Tims, when free, they move with the micrometer head; but when held, the head moves past 
 them. One end of each terminates exactly with the radius of the micrometer head, and serves 
 R8 an index, as may be seen in Fig. 3. The other end projects a little beyond the cylindrical 
 surface of the head, and may be held by notches in the levers 1 1, (Fig. 2.) When thus held, 
 the index ends are in the same vertical line with the principal fixed index, as shown in Fig. 3. 
 
 The mode of operation is as follows : The micrometer being in any position which it is 
 desired to register, one of the levers (Figs. 1 and 2) is moved back with the finger to the 
 position l\ and the movable index is thus set free. It now moves round with the micrometer, 
 and thus points continually at the reading indicated when it was set free. Four successive 
 readings may thus be registered in succession without taking the eye from the telescope. 
 
 The object of the two extra divided cylinders on the micrometer head is simply to facilitate 
 the reading of the four movable indexes. 
 
UinrBO STATES HATAL 0B8BBVAT0BT. 
 
 (31) The Oculara. — The instnimeDt n ftimished with five oonlan, of which the in«gnifyiDg 
 powers are, approximately — 
 
 No. 1. 
 
 ISO; 
 
 8, 
 
 160; 
 
 3. 
 
 180; 
 
 4. 
 
 260; 
 
 6, 
 
 360. 
 
 No. 3 18 that hitherto most used in aslronomioal observations. 
 
 (32) The Retiotde. — ^The fixed vertical spider lines are twenty-three in number, arranged 
 as in Fig. 4. Seven of these are at eqnal intervals of 12^ seconds of time, and are designated 
 by the Roman nnmerals I-VII. For convenience, a second notation is adopted, as shown at 
 the bottom of the figare, the wires being divided into five groups A to E, and the wires of 
 each group designated by subscript Arabic numerals. Thus the wires II, III, IV, T, and YI 
 are designated indiflferently by these numbers or by the symbols A„ B|, G,, D„ Ep The wires 
 of set G are 24. apart ; those of B and D It, 5 and 2«.6 respectively. Sets A and E are rarely 
 used. Stars are usually observed in right ascension ^ver B D and the three middle wires of 0. 
 In observations of planets one limb is observed ove. B and D, the other over C. 
 
 The middle of the field is shown by a single pair of fixed horizontal wires about 8" apart. 
 A single vertical wire is moved by the horizontal micrometer. 
 
 (33) The horizontal movable wires consist of a single pair 4".6 apart, between which objects 
 are usually observed, and a couple of single wires at a distance of ten micrometer revolutions 
 on each side of the central pair. The former were inserted by the machinist of the observatory 
 after the mounting of the instrument, in lien of a single wire inserted by the makers. 
 
 (34) The IttumincUion. — The flames of the two lamps L, L, Plate lY, are nearly surrounded, 
 by parabolic reflectors. The light passes through the supporting tube and the pier into the cyl- 
 inder D; here it meets a combination of lenses and prisms, shown in Plate YII, Figs. 4 and 6. 
 Fig. 4 gives an end view of the combination, as seen from the lamp. Fig. 5 gives a side view. 
 The nearly cylindrical frustum of a cone G is pierced in the direction of its axis by five holes, 
 a, b, b, b, b. The light which enters a passes through two lenses, and through the perforations 
 in the pivots of the instrument to the centre of the axis. That which passes through the b'a 
 meets four prisms, p, p, from the interior surfaces of which it is refiected at an angle of about 
 42^°, BO as to be deflected in all about 85°. G and D fit snugly in a perforation in the centre 
 of the cylinder D, Plate lY; a hole receives the pin e, and thus fixes the position of the appa*. 
 ratus. The latter can be removed by unscrewing the lamp-tubes T, and reaching the arm 
 through the pier. When in place, its position is such that the light reflected from the prisms p 
 passes down the pipes k, (Plate lY,) and into the reflectors of the microscopes, by which it is 
 thrown upon the divided limb. 
 
 (35) The central ray, passing {ihrough a, if it enters the clamp end of the axis, is conducted 
 to four prisms in the central tube of the axis, from which it is reflected down the side of the 
 tube to four more prisms in the eye-piece, and thence to the wires, in the usual way. Thus 
 the wires may be iUuminate<lf. or darkened by turning a milled head on the clamp side of the 
 eye- piece. 
 
 (36) The light which enters the other end of the axis is reflected from a prism about 4^ 
 inches from the centre of the axis, directly to the eye-piece. It shows the dark wires on the 
 bright field. It may be changed in color from yellow to blue, or cut off entirely by turning a 
 milled head on the side of the tye-piece opposite the clamp. 
 
 (37) The CoRimaiora. — Fig. 6, Plate YII, gives a general view of each collimator. The 
 object glasses have each 2| inches clear aperture, and 35 inches focal length. The eye-piece 
 
 
8 
 
 DBSORIPnON OF THB TBAN8XT OIBOLB OF THX D. 8. NAVAL 0B8BBVAT0BT. 
 
 of one, which is designated collimator A, is fornished with a pur of parallel wires, ditit»nt 
 about 9", with a single wire crossing them at right angles. The wires of the other, called col- 
 limator 3, form a simple cross. When it is required to .set the collimators opposite each other, 
 the coincidence is that of a single wire of B with the mean of the parallel wires of A. 
 
 The steel plate which carries the wires is held between four antagonistic screws, but has 
 no binding screws. 
 
 The supporting collars of the collimators are only 23 inches apart, so that each end of the 
 collimators projects about six inches from its support. 
 
 Each collimator is furnished with a delicate spirit level, which sets on the supporting collars. 
 
 (38) The Ts of the collimators, with the way in which they are fastened to the piers, is 
 shown in Figs. 7 und 8. The hemispherical bottoms of the screws a, a, a, rest in three cavities 
 in a Y-shaped block of metal, which is set into the pier with plaster. The outline of the block 
 is shown by the dotted lines. The three screws 6, b, b, pass centrally through a, a, and thus 
 bind the Y-plate firmly to the block. The level of either pivot is adjusted by loosening 6, and 
 turning a to the required extent. The azimuth is adjusted by the antagonistic screws c, e. 
 When the V is in position it is bound by the screws d, d, e, e. 
 
 m 
 
PAKT II. 
 
 GENERAL METHOD OP INVESTIQATINQ THE ERRORS OP A TRANSIT CIRCLE. 
 
 (89) The most important and •'iffioult part of this investigation is the determination of the 
 effect of gravity in changing the relative positions of the various parts of the instrnment, and 
 the correction of observations for this effect. It is therefore proposed to develop a general 
 method of determining, either rigorously, or with a near approach to rigor, the effect of gravity 
 in changing results of the various observations by which the position of a star is determined. 
 
 The necessity for a flexure correction may be regarded as an imperfection in an instru- 
 ment, since, if it were well balanced, and equally elastic in all its parts, no such correction 
 would ever be necessary. Artists generally attempt to avoid this imperfection by distributing 
 the weight of the instrument so that it shall be perfectly symmetrical with respect to the centre. 
 Thus, in the Washington Transit Circle, the two circles are of equal weight; the clamp is coun- 
 terpoised by a cylinder at an equal distance on the other end of the axis; and the tubes of the 
 telescope had their outer and inner surfaces turned simultaneously, in order to secure perfect 
 equality of thickness in every rectangular section of the tube. Yet, the effect of irregular 
 flexure cannot, by any means, he neglected, and that of the two circles exhibits a very sensible 
 difference. A similar remark will probably apply to every large instrument ever made. It 
 will not, therefore, do to assume that such an instrument is equally elastic in any of its corre- 
 sponding parte, however unexceptionable may be its construction. 
 
 (40) The only arbitrary hypothesis we shall assume respecting any law of elatitioity in the 
 instrument is, that the relative flexure of the various parts is the same as if the instrument, 
 during a revolution on its axis, were always supported by the same particles of its mass. 
 The nature and the results of this hypothesis will be rendered more clear if we suppose the 
 instrument to remain at rest, and gravity, with the forces which oppose it, to act in various 
 directions. The parts of the instrument which sustain these opposing forces are mainly those 
 which press on the friction rollers of the counterpoises. Reflecting, now, that the direction of 
 these forces always passes within two or three inches of the centre of the axis, and that our 
 hypothesis can be incorrect only by a combination of the two following circumstances, viz: 
 l' An unequal elasticity in the parts of the conical axis on which the friction rollers succes- 
 sively press; 2. Such a relation of elasticities in other parts of the instrument that the unequal 
 elasticity of this narro<v band on the conical axis causes the instrument to change its form ac- 
 cording to a different law from that which would hold true if the elasticity of the band were 
 uniform — ^it would seam that the hypothesis must be a safe one. 
 
 Experiment shows that the effect on elastic bodies of forces which are very small in pro- 
 portion to the rupturing force is strictly subject to the law of superposition of small motions; 
 that is, that the combined effect of several such forces is the sum of the effects that each would 
 produce acting separately. 
 
 ^■-mmmmmmm 
 
 
 
10 
 
 DESOHIPnON OF THE TBTMSIT OIBOLB OF THK 
 
 (41) Still snpposing the inatrament fixed, and the direction of gravity variable, let b rep- 
 resent the amonnt by which a point of the inBtrument is moved from its normal position in the 
 direction of an arbitrary fixed axis when gravity acts in the direction of that axis, and let a rep- 
 resent the displacement in the same direction when gravity acts at an angle of 90° with the 
 axis. If gravity acts at an angle W with the axis, it may be resolved into the forces g cos W 
 and g sin W, acting along and perpendicularly to the axis. The first component will, by the law 
 just referred to, produce the displacement 6 cos W, and the second the displacement a sin W, 
 and the whole effect of gravity will be a sin W-|-6 cos W. 
 
 Suppose, then, any system of rectangular co-ordinates fixed in the instrument, and revolv- 
 ing with it; then, as the instrument revolves, the law of displacement of any point in the di- 
 rection of the movable co-ordinates will be given by the equations 
 
 dx=a sinW+b cosW, 
 iy=a' Bin W-\-f/ cob W, 
 *««sa" Bin W+i" COB W. 
 
 W being the angle of position of the instrument, and a, a', a", b, 6,' 6," constants, to be 
 determined by observation. 
 
 (4*2) The method of determining such of these constants as are necessary in the correction 
 of afltronomical observations will next be shown. To admit of the application of this method it 
 is necessary that the instrument ahculd be supplied with a pair of collimators, each of which 
 admits of being accurately levelled, in order that it may be used to determine the horizontal 
 point, and an artificial horizon for observing the nadir point by reflection from mercury. A 
 vertical collimator, to be fixed over the centre of the instrument, and set vertical ither by re- 
 flection of its wires from the mercury bath, or by levelling its axis, would also be valuable as 
 an independent check on some of thd results. 
 
 (43) Let us now consider the causes which will affect the reading of any microscope. If 
 the divisions are truly cut in direction — that is, if each stroke is sensibly in the same plane with 
 the axis of the circle, the error of each point in any one division may be regarded as the same. 
 We shall therefore regard the diviftions as a series of points. When a microscope is read, the 
 position of the point under the microscope is determined by bringing a micrometer wire into 
 coincidence with the image of the point in. the microscope. This micrometer wire in the Ger- 
 man instruments is an imaginary one bisecting the space between two real ones. The division 
 point is then in the plane which passes through this micrometer wire and the optical centre of 
 the object glass of the microscope. Let us consider, as the zero of the micrometer its position 
 when the plane passing through the object glass and the micrometer wire is at right angles to 
 the face of the circle. The micrometer reading of the division will then be proportional to the 
 ttingent of the angle which the plane containing the division makes with this last plane. No 
 appreciable error will result from supposing this adjustment to be perfect. 
 
 The subsequent investigation will be precisely the same if we suppose the divisions to 
 radiate from the centre of the circle, and the micrometer wires in the microscope to be a point, 
 which is brought into coincidence with the image of the division. The micrometer reading 
 will then represent the tangent of the angle which the plane, passing through the centre of the 
 object glass and the division, makes with that passing through the centre of the object glass 
 and the axis of the circle. 
 
 If the eight microscopes were correctly adjusted in position, and the circles perfect in form 
 and position, and not acted on by gravity, the instrument might be so set that all the micro- 
 scopes should read zero at the same time. The following include all possible deviations of the 
 circles and the microscopes from this state. 
 
URITET) K'AnS NATAL OBURTATORT. 
 
 11 
 
 a. Errora in the Angtihr PoMiont of tht MRoro9Cope$. — When all the microscope micrometers 
 are set at zero, the planes passing through the micrometer wires and the centre of the object 
 glasses onght to cut the line of divisions at points making angles with the centre of the axis 
 equal to the nominal distance of the microscopes. 
 
 p. Error$ m the Oeneral PoaUixmt^ the Cirtlte.—The»9 will be six in number: small motions 
 of translation in the direotion of three co-ordinate axes, and small motions of rotation around 
 those axes. * 
 
 f. Enron in the Position </ the Divieion on the Oirde. — ^These may be three in number. I. 
 Errors in the angular position of the division. 2. Deviation of the surface on which the divis* 
 ions are cut from a plane cutting the axis of revolution of the instrument at ri|{ht angles. 3. 
 Change in the position of the division produced by the effect of gravity on the circle. 
 
 The position of a division may be in error by any of the nine causes, ]9 and y. To deter- 
 mine their effect, let us refer the position of the division to three co-ordinate axes, those of X 
 and T being in the plane of the circle, and that of Z being the axis of revolution. Let A be 
 the angle which the plane of the micrometer wire and object glass makes with the axis of the 
 circle, B the angle of position of the microscope, counted from the axis of X. 
 
 B, the radius of the circle. 
 
 c, the distance of the centre of the object glass from the plane of the circle. 
 
 r, the reading of a microscope from its own zero. 
 
 p, the mean reading of a pair of opposite microscopes. 
 
 m, the error of position of a microscope. 
 
 M, the mean error of a pair of opposite microscopes. 
 
 e, the error of division. 
 
 t, the mean error of a pair of opposite divisions, or an error in iue position of a diameter. 
 
 a, b, the flexure coefficients of any point of the circle. 
 
 a, /9, half the difference of a or b for two opposite points of the circle. 
 
 £, 17, the possible motions of translation of the centre of the circle relatively to X and Y. 
 
 at, the error of pointing of the telescope. 
 
 The differential coefficient of tan A, (which is proportional to r) with respect to the co- 
 ordinates of the division, will then be: 
 
 
 <2.tanA B.ir 
 dm °*c<te 
 
 SinB 
 
 <2.UnA Bir 
 
 cosB 
 
 c ' 
 
 <itanA Edr 
 dz cdy 
 
 tanA 
 e 
 
 The effect of small changes dx iy and dz on the reading of a microscope will therefore be 
 
 dr=:--8inB^-fco«B^+tanA^. (1) 
 
 Changes in the co-ordinate z can arise only from a motion of translation of the instrument 
 in the direction of its axis; from motions of rotation around the axis of X or of Y, in consequence 
 of irregularity of pivots; from deviation of the circle from a plane cutting the axis at right 
 angles ; or from lateral flexure of the circle. If the effect of any of these causes is appreciable, 
 9z must be determined by direct measurement, independently of the readings of the micro- 
 
 mmsmsssmsmsm 
 
12 
 
 DEBORIPnOM or THB TBAMBIT OIBOLI OF THB 
 
 % 
 
 «; 
 
 Bcopes, and the latter mast be oorreotfid accordingly. To jadge when d» can be senBible, we 
 remark that in general 
 
 tanAi 
 
 Bf->-+eonitaiiti 
 
 the constant being the valne of tan A when the micrometer reads aero, and therefore depend- 
 ing on the parallelism of the microscope to the axis of the circle. If the changes in r do not 
 
 R 
 exceed 10" and - does not exceed 12, A will not differ more than 1' from ito mean valne in a 
 
 aeries of readings. Since, by proper adjustment, this mean valne may be made ver> small, 
 there is no necessity that A should ever exceed 1' in a aeries of measures. In this case, s will 
 have to change by -y^ the radius of the circle to produce a change of 0". 1 in r. This change 
 being ten times as great as any that a well made instrument need be liable to, the last term in 
 ir may be regarded as insensible, if the microscopes are carefully adjusted. This will dispose 
 of four of the nine causes which may affect the relative position of the divifipp , and object 
 glass. The remaining five will be included in the equations of condition. > 
 
 Every micrometer reading will then, by equation (1), give an equation of the form 
 
 r«M-g tin B+l COS B+a cos B+ft sin B+e+i*. 
 
 («) 
 
 (44) The problem now is to make such mechanical arrangements and such readings, that 
 from a series of equations of this form we may be able to determine the values of the coefficients 
 in thejiecond member, so far as it may be necessary for the correction of observed senith dis- 
 tances. To obtain a complete correction for the readings of a single microscope world require 
 the solution of equations of extreme complexity ; we shall, therefore, hereafter consider the 
 corrections of a pair of oppos te microscopes. If, now, another microscope be placed at the 
 angle 180°-|-B, and its reading, taken simultaneously with (2,) be indicated by an accent, we 
 shall have 
 
 r'«m'+- sin B-l ooB B-a' eos B-»' sin B-(-e'+». 
 r It 
 
 Adding this equation to (2), and observing that by the rotation already given 
 
 r -I- r'as8/», 
 
 a — a'aeSa, 
 6 - VnmSfi, 
 e + tfasit, 
 
 we shall have, by taking the mean reading of a pair of microscopes, 
 
 />=M+aeo8B-|-j^sinB-|-c-f«. (3) 
 
 an equation more simple than (2) and equally complete. 
 
 (45) The mechanical arrangements necessary for the proposed method are these : The 
 instrument must be furnished with two finely divided circles, each read by four microscopes. 
 At least one circle must be capable of being turned on the axis of the telescopp, aqd set in 
 any required position relatively to the telescope. In what follows, we shall sUppese both 
 circles thus movable. The microscopes must also be capable of being set at different angular 
 positions. The details of the arrangement, and the mode of numbering the mioroaoopes and 
 
 ■- '¥^f-mm^'f-'jfim.fS^r^: 3?^CTa3Rtsa«»KW»««~^- w»M?»»!t!<www^5""< 
 
UNITID flTATIS KATAL OMIBTATOBT. 
 
 li 
 
 S$ulh 
 
 counting the degrMs, will be rappoied to correspond to thoee of the Wwhington transit 
 
 circle. The Accompanying diagram shows the arrangements referred to, as seen from the 
 
 west. The oater circle represents the westernmost 
 
 one, and shows the positions of microscopes I to 
 
 lY, which are on the west pier. The inner circle 
 
 represents the eastern one, and shows the positions 
 
 of the four microscopes V to VIII, which are on the 
 
 east pier. The arrows point in the direction in 
 
 which the degrees, as nnmbered on each circle, 
 
 increase. It will be seen that the two circles are ^orth 
 
 divided in opposite directions, so that when the 
 
 instmment is reversed the circle now east will read 
 
 in the same direction as the former one, and so for 
 
 the west. Also, as the instmment is turned in any 
 
 direction, the snm of the readings of any microscope 
 
 on the west pier and any microscope on the east pier ^ 
 
 will be a constant depending on the relative posi* sj 
 
 tions of the circles. 
 
 We shall suppose the axis of X to pass in the direction of microscopes I-ni, and that of 
 Y in the direction of II-IV. Thns, we have for microscopes 
 
 I and VIII. B=180O} 
 
 II sad VII, B« 90O; 
 
 III sod VI, B= OO; 
 
 IV and V, B«270O. 
 
 (46) The flexnre effect which it is required to determine is the change in the relative di« 
 rection of two lines, namely, the line joining the object glass and eye-piece micrometer of the 
 telescope, and that joining any pair of opposite divisions which chance to be under a pair of 
 microscopes. This will be determined by finding the change of direction of each line relatively 
 to that part of the central axis to which the circle is fastened. The total flexure will therefore 
 be divided into two parts: 1. The flexure of the central axis relatively to the ends of the tele- 
 scope. 2. The flexnre of the diameters of the circle relatively to the centre. The coefficients 
 of the first flexure we shall represent by / and g for the east circle, and by / and g* for the 
 west one. Suppose, then, that the line joining the object glass and eye-piece revolves uni- 
 formly; that part of the axis which passes through the circle and holds it by pressure may 
 be affected with an inequality of the form 
 
 /rinZ+^eosZ. 
 
 Again, supposing this part of the axis to revolve uniformly, it is possible that a diameter 
 of the circle may be affected with an inequality of the form 
 
 a8tnC4>i9eos(, 
 
 a and ^ being coefficients to be determined separately for each pair of divisions, and ( being 
 the angle <^ position of the diameter counted from an arbitrary axis. On the east circle we shall 
 for convenience, count C from microscope VI through V, VIII, and VII, and on the west circle 
 from microscope II through I, IV, and IIL Then, when any division D comes under microscope 
 V, the mean reading of V and VII will, by the effect of gravity, be increased by the quantity 
 
 «J>. 
 When the same division comes under VI, the mean reading of microscopes VI and VIII will 
 be increased by the quantity 
 
 uiuHii i Mmiiii i iiiiiiiw 
 
u 
 
 DKSCBIPTIOR or THB TBAMSIT OIBOUi OF THE 
 
 ,.T>. 
 
 Tba Mme stotement will apply to the we«t drole by dimioishing the nnmber of the mioro- 
 •cope by lY, to that ti.H and ^.D will be the flexare when the diviiion D comes under mioro- 
 icopea I and II. 
 
 Since changing ^ by 180° la the same as changing the algebraic sign of a and ^, we have 
 
 a.(D-fl80O)»-a.D, 
 ;9.(D+l80O)«.-/J.D. 
 In equation (3) we hereafter substitute for a cos B-f-^ sin B the two floxnrea 
 
 /■in Z-\-g OM O+a sin C+Z* eos C 
 
 Let us now take four divisions on each circle, such that they may all be under the micro- 
 scopes at once. Represent them by a, ft, o, d on the east circle, and a', V, cf, d' on the west 
 circle, the lettering being in the same direction with the numbering of the degrees, and of the 
 microscopes. We then have 
 
 cascca error of position of line joining a—e, 
 
 a.a^ error of flexure of a—c when under mierosoc^ V. 
 
 p.a= error of flexure of a— e when nnder mieroseope VI. 
 
 Suppose, also, that when a is under niicroscope V, of is under microscope I. Also put 
 
 /sin Z+j- eos Z=/ *.— /. («+180O), 
 
 -/ sin Z+^ eos Zax/je.—/. («+ ISQO), 
 and let 
 
 Zo. ^+900, Z«4-180O, Z,+270O 
 
 be the four zenith distances of the telescope toward the south when the divisions are brought 
 successively under the four microeoopes. Bepresent by fin.» the mean readings of a pair of 
 opposite microscopes. Then we have the following equations of the form (3): 
 
 First position of telescope Z^ 
 ^5.0=:M|+(.a+/jr»-|-a.a+«»,n. />1.0aiMi+('ui-4- /'jst+a'ui— •.0, 
 
 Second position, 2^+90*. 
 />fi.90=M,-|-c.i+/.(ire+90O)+a.6+«.90, />1.90»M,-f<'.<{+/'.(«k+90O)+a'.<I— «.90. 
 />6.90=M«+<.e +/(«^+90O)+i'-e+»-90, />8.90asMt+«'.a4-/'.(««+00O)+/»'.a-«.90. 
 
 Third position, Z.+ISO^'. . 
 
 /.6.180s=M,+«.c +/(«,+ 180O)+a.c-|-«».180, /jl.lSO-Mj+t'x +/'.(««+ 180O)+a'.c --.ISO, 
 />6.180sM,-f(.<2+/(««+i80O)+/}.i{+«.180, /»8.180xsM«+t'.<{4-/'.(«k+^80O)+^.<2-<».180. 
 
 Fourth position, S^+270°. 
 
 /.6J870«M,+«.<l+/(«o+2700)4.a^+»^70. ^lJ70=|li+«'.*+/'.(»o+8700)+«'A-»^70, 
 
 /.6JB70asM,+t.c+/.(««+870O)+/9.a+*.270, />2.870»M,+«'x +/'.(«»+870O)+/!'jj -i»J870. 
 
 (47) For perspicuity, let na reoapitalate the adopted notation, by ezphiinii^ the meaning 
 of the i^ve equations. 
 
 If l^ere were no errors of division or adjostoMnt, and no elasUdty of the parte of the in< 
 stmment, the readings of all the microscopes wonid be sero. The readings not being aero, and 
 />6.0 being the mean reading of microscopes Y and YII ^hen the telescope poiattf approzi* 
 mately at zenith distance Z^ the first equation indicatri that this value of />6.0 is due to the 
 five following causes: 
 
 1. Error of position of the microscopes Y and YII, (M,.) 
 
 ihmm 
 
 nvimmmm 
 
tmrrwD arAnt mawml oBsnTAioiT. U 
 
 5. Brrora of the divitioos a and o which we ooder thete miorotoopee, (t.o.) 
 
 8. RoUtion of the sxis of the telesoope reletively to iU Hne of sight, in consequence of 
 the irregolar effect of gravity, (Av) 
 
 4. Angular change of Hne joining divuiona a and o relatively to axis of rotation from the 
 same cause, (a.a.) 
 
 6. Incorrect setting of the telescope, («.0.) 
 
 In any one of the above fonr positions take the snm of />, or p„ and /i^or />,. Take alito the 
 corre»|K)nding sum in a position of the circle 180** .different, then subtrHCt the two sums, remern* 
 bering that 
 
 /»— /(«+180O), 
 •41 or/94iM— a.(a-f 180O) or /».(a+180O), 
 '. f.a«Bt.(«+180O), 
 
 and call the difference K, using the notation 
 
 K.ie.i— A>i.»-|-/»i.i— />i («+180O) — /»,.(i4- 180°). 
 
 Put also g.K=/.n-^/.% and we have the e^ations 
 
 K.15.0mi2aM+W a+2g z, K lS.90xs 9a.b'-'2a'.b+^.{z+9(P), 
 
 K.16.0»2/9.*+2o'.a+2^«, K 10.90— 2a./9-2«'.A+2f («+90O), 
 
 K.26.0»B2a.o+2/9'.4+%.«, K.26.90- 2a 5+2/J'.a+2f .(«+90«>), 
 
 K.26.0a,2liA+ail'*b+agje, ** K.2e.90aB-2i9.a+2/9'4>-|-2^.(«+g0O). 
 
 Of these eight equations it will be seen that only six are independent, any one of each four 
 being identically derivable from the other three. 
 
 Now, let the poeition of the east circle on the axis of the telescope be changed by 180", 
 and let the same divisions be again read. Treating the equations derived from the readings in 
 the same manner, and distinguishing the new K's by an accent, we shall have 
 
 KM6.0aB-8a.a+2o'4i+^.«, KM6.90ss-2a.A-2«'.i+;^.(*+90O), 
 
 K'M.0ss—2fii+2a'.a-^2g.», K' 16.90= 2fi.a-2a'.b+2g.{z+9(i°), 
 
 K'.25.0ss:-2« 0+2/9'.*+^^. K'J85.90»-2a.*+2/S'a+^.(«+90O), 
 
 K'M.0=—2fiA+2^i+2gjK, K'.S6.90s 2fiM+2fi'.a+2g.{z+9<P). 
 
 Now, turn the west circle 180° on its centre and again repeat, distinguishing the K's by 
 two accents, and we have 
 
 K".16.0s»-2«.a— 2«'.a+ai^.«» K'Mfi.90=r-2o.i-|-2a'.A+S!^.(«-|.90<'), • 
 
 K".16.0rB—2fi.b—2a'M-{-2gje, K".16.90s: 2fi.a+2a'.b+2g.{zf9(P), 
 
 K"MJ0sB^2a.a'-2fi'A+2g.z, K".26.90sai—2a.b-'2^M+2g.{z+90O), 
 
 K"J»6.0sB—2fib-2fi'.b+2g.z, K".26.90=s 2/J.a-2/J'.a-»-^.(*-f-90«>). 
 
 These equations suffice to give two values of each of the required quantities. But, to 
 have as many independent determinations as possible, a fourth series of readings are taken 
 with the east circle restored to its original position. We then have 
 
 K"'.li.0a»2mM'-2afM+2ga, K"M5.S!0» 2a.b+2afJb+2g.{+z9V>), 
 
 K"'M.0ma»fiA^2»fM+2g.z, K'".l6.90sM^2fiM+2afA+2g^z+0<P), 
 
 K"'26 0^2aM-'2fi'b+2ga, K'".86.90as 2*^—2^M+2g.{z+9(P). 
 
 K"'MJ0ss2(t.b-2fi'.b+2g.z, K'"M.90sm^2fi.a'-2/i'M+2g{z+W>). 
 
 Adding together the corresponding equations of the first and third series, and also those 
 of the second and fourth, we shall have 
 
 ^«sK. 10.O+KM5.OaBK. 16.U+K'M6.0»K. 2A.0+K".2fii0»K. 86.0+ K". 26.0 
 !=KM«.0+K''M5.0»K.'16.0+K"'.l6.0=sK'.26.0+K."25.0s=K'.26.0+K'".86.0. 
 
 i^ms^tmrxm^'sm 
 
If 
 
 DMOUPTKM or TBI nUMUT OnOLB OF TBI 
 
 *.» 
 
 Of these eight raliiec only nx un really independent, being rabjeet to the condition thnt 
 the Bam of the extreme ynlaee ie eqaal to the ram of the meuit, which fhmiihea » check on 
 the aconrecjr of the compatntion. 
 
 Again, by eimple lobtraction of the correeponding eqaatione in iracceaaive aeries, we find 
 the following four distinct valnea of each of the eight qoantitiea o^, /9.a, tuh, fiJb, af.a, fi^jo^ 
 af.b, fif.b, which gire the circle flexure. 
 
 K .lfi.O-K' .15.0. 
 K JB0.O>K' SA-O, 
 K'".lfi.O-K" .16.0, 
 K'".««.0-K" .85.0. 
 
 K .Ifi.O-K'" Ifi.O, 
 K .16.0-K"'.16.0, 
 K' .16.0-K'M6.0, 
 K' .16.0-K" .16.0. 
 
 KM6.90-K .16.90. 
 K' 26.90-K .S6.90. 
 K'M6.90-K"M6.90, 
 K".S6.90-K'".86.90. 
 
 4/9'.«» 
 K J0 9O-K"'JA.9O. 
 
 K J6.90-K"'.86.90, 
 
 K' .86.90-K" J8S.90, 
 
 K' .86.90 -K''J6.90. 
 
 K .16.90— KM6.90. 
 K .86.90-K' .86.90, 
 K'" 1A.90-K".16.90, 
 K'".8«.90-K".86.90. 
 
 K"M6.90-K .lAJO, 
 K'" 16.90-K .16.90, 
 K" .16.90-K' .16.90, 
 K" .16.90-K' .16.90. 
 
 K .16.0-K' .160, 
 K JM.O-K' M.0, 
 K'".16.0-K" .16.0. 
 K"'.86.0-K» .86.0. 
 
 K J6.0-K"'.S6.0, 
 K .86.0-K'".86.0, 
 K'.86.0-K"J6.0. 
 K' .86.0-K" .86.0. 
 
 (48) Thus, the flexure of the circles, in so far as it affiects the mean reading of any pair of 
 opposite microscopes, and the relative flexure of the two ends of the axis, are completely wnd 
 rigorously determined. It remains to determine the flexore of the line joining the eye and 
 object end of the telescope relatively to the axis itself. We have supposed, for brevity, 
 
 and have shown how to find g.u, and therefore /— ^, and g-{-f^. The coefficients / g, f g', may 
 DOW be found separately and independently by observations of the nadir point and collimators. 
 An observation of a levelled collimator, corrected for inequality of collimator pivots, diffisreuce 
 of latitude of circle and collimator, aerial refraction, and oollimation error of the collimator, 
 shows the circle reading when the line of sight of the telescope is truly horiiontal. The coin* 
 cidence of the direct and refiected images of the zenith-distance wires shows the circle reading 
 when the telescope is truly verticaL When these readings are corrected for circle flexure, the 
 horizontal and vertical circle readings will be as follows: 
 
 Ewtdrala. Wostdnl*. 
 
 Bonth horiiontal reading, 0«4-/ O't—f 
 
 North horisontsl resdiog, 0«— / O't+f 
 
 Nadir reading, 0«-jr C'^—g'. 
 
 Cg being the true reading, independent of flexure. From these equations the values of 0^ /, 
 and g, tf^f, and g', may all be determined. 
 
 The values of/ and f may also be determined independently by viewing the wires of one 
 collimator through the other, and making the horiiontal wires of the latter coincide with the 
 images of those of the formeir. If, then, the micrometer wires of the telescope be set sucoes* 
 sively upon the images of the two collimator wires, the line joining its object glass and microm- 
 eter wires will have moved accurately through the space of 180°, while the circle will indicate 
 a motion of I80°d=2/. The agreement of the two values of/ will afford a check upon the ao- 
 curacy of the collimator determinations. 
 
 (49) One precaution is, however, indispensable to an accurate result The different strata 
 of air in the room must be as nearly as possible of the same temperature. For, if there be any 
 admixture of warm and cool air, the former will ascend, and we shall thus have a temperature 
 increasing with the height If, now, this increase amount to one degree Fahrenheit in nine feet, 
 the curvature of the ray will be equal to that of the water level, so that it would never leave 
 the earth. 
 
 "■"•'*''***■*""■""■'■'' "" ' — -I TT- K ii .jj i mf in«iuii . i i i i.i]iiininn-|inr— 
 
UJfinO ITATM HATAL OBaBBYATOBY. 
 
 It 
 
 The oonrae of the rey being carved, the yalae of/ obtaiDed from the opposing oollioDitori 
 will be vitiated. On the other hand, if we adopt the other method, there is always a possibility 
 of error arising from irregalarities or nnoertaiu differences of diameter of the collimator pivots, 
 or poaeible flexure of some part of the collimator itself. However / is determined, the obser- 
 vations shoold be made with the shntters open, and under atmospheric conditions, as necrly as 
 praotioable, like those nnder which the astronomical observations are made. Considering the 
 difficulty of rigorously fulfilling this condition, it would, perhaps, be better to determine / by 
 reflection observations of stars, g also admits of being determined by a comparison of North 
 Polar Distances of stars near the zenith in reversed positions of the instrument. 
 
 EBR0B8 OF DiyiSION. 
 
 (60) The method of determining the errors of division is, in principle, that usually adopted, 
 the operation being so conducted as to eliminate all irregular flexure of the circles. 
 
 liOt us suppose the two pairs of opposite microscopes set at the distance B, the angle posi- 
 tion of the flrst pair being 0°, and that of the second pair, B. Take a series of pairs of divis- 
 ions at the distance B, and let these divisions be successively brought under two pairs of 
 microscopes. Then, using the same notation as in § (43), 
 
 Mio. I. ri»wi+^-J-ai-»-«i-f-«»i, 
 
 II. niaiMt-|--^cosB stoB+OteosB-faiflinB-fck-t-wi. 
 
 ni. fi>-M— 3^--as4-«»-f Ml, 
 
 rV. r4BM4— 4eosB-h-gsiaB— OtCOBB— i4B{nB-^«4-fi»i. 
 Whence, adding the odd and even equations 
 
 />^2^sMt4-at COS B-H/S* sin B-f-«i-f- wf 
 Turning the circles through the distance B, we shall have 
 
 (1) 
 
 ^<y»Mi.J-a,+.,+«,, 
 
 Turning again 
 
 ,<»>. 
 '<?■ 
 
 :M|-(-a3 cos B-f-^3 sin B-f ts+coi. 
 
 («) 
 
 Mi-J-a3-|-«j-f-«3. 
 p^^sMr^a^ tM B+fit sin B-(-C4+«»4. 
 
 Let 2» be the number of motions necessary to bring the circle back to its first position. 
 Then, the (t+l)th reading will have brought it round 180°, so that the same divisions as at 
 first will be under the microscopes. The readings after the tth will then give 
 
 /.<j+^)=:Mt-ai+.i+«(i+l). 
 
 /.<*+^)asM,-.a,cosB-A"inB+«,+«'(«+l). 
 ^<j+*)=Mi-«,4..,+«(i+8), ^^5 
 
 /•+2)^M,-«, eoB B-A •»» B+t,+«(H 8). 
 
 10 ■ 
 
 ice., 
 
 ace. 
 
 i ;i *ii:jii«i«» i iiiiiiiiiiift i iii! 8 
 
18 
 
 UE8CBIPTI0N OF TQE TRANSIT CIRCLE OF THE 
 
 'M 
 
 Taking the suma of readings in opposite positions of the circle, we have 
 
 Using the notation 
 
 we have 
 
 Mt— Mi + Ci— Ci=rI.1.2, 
 
 Ma— Mi+ca— ct=I*2.3, 
 
 Mj— Mi+e4--(3=I.3.4, 
 
 &c.« &c.t tee. 
 
 (4) 
 
 (fi) 
 
 Taking the sum of the entire series of values of I, and dividing by their number, the 
 c's will destroy each other, and we shall have 
 
 Putting 
 
 we shall have 
 
 Mj^Mi:=-r-. 
 
 J.1.2=:I.1.2-(tf,— Ml), 
 
 (7) 
 
 «4-C3=^.3.4, (8) 
 
 &»., Ice. 
 
 Thus, where one e is known, all the others may be found by successive addition of the 
 differences A. If, however, the number t is considerable, the accidental errors of the various 
 As may in the additions become considerable; it is therefore necessary that at least every fifth 
 c in each series be corrected by the results of another series in which the value of t shall be 
 less than in the series in question. The correction will be really applied to M',— M„ which 
 will thus be determined independently for each sub-series. Suppose, for example, that in a 
 first series the microscopes are at the distance 45°, so that t==4. The preceding formulte will 
 then give 
 
 t.ifi —e.O =4 .0.45, 
 
 C.90 — C.46 ssJ.45.90, (9) 
 
 c.l35->c.90 r=:J.90.135, 
 
 c.O — c.lSffssJ.lSff.O. 
 
 In a second series, let the microscopes be placed at the distance 16°. 
 then give 
 
 e.l5— c.O ssJ.O.ld, C.60— c.45s=J.45.60, &c.; 
 
 e.30— c.l5=J.15.30, (.75— c.60=J.60.75, &;c.; 
 
 t.45— c.30=J.30.45, c.90— (.75=^.75.90, &;e. 
 
 Adding each of (10) and comparing with (9), we have 
 
 ■A ,0.15+J.16.30+J.30.45=J .0.45, 
 J.46.60+J.60.76+J.75.90=J.46.90, 
 
 &c., &c., &c. 
 
 The formulae will 
 
 (10) 
 
 ».*lip>.vn'-;9t^-'.-WS«;i'«,»riMP^>aiSr«'KS*^^ 
 
the 
 
 UNITED STATES NAVAL OBSBBVATOBT. 
 
 19 
 
 If tbe readings were perfect, these equations ought to be exactly satisBed. Not being satis- 
 fied, the difference is due to tbe accidental errors of the readings. Let i be the probable error 
 of each of the left-hand ^^s, « that of the right-hand ones, and e the amount of the discrepancy. 
 The most probable distribution of the discrepancy will then be, 
 to each left-hand A, . ^ 
 
 and to each right hand A, 
 
 
 The corrected values of A being substituted in (10), we shall have a series from which the 
 most probable values of f„ fu. &c,, may be found by successive addition from any arbitrarily 
 assumed value of «,. 
 
 DETEBMINATION OP THE EBEOES OF SPECIAL DIVI8IONS.-PBOBABLE ACCIDENTAL EBBOB OP 
 
 ISOLATED DIVISIONS. 
 
 (51) The determination of the error of each separate division being impracticable, this 
 question arises: What is the probable error of any isolated division relatively to any consider- 
 able number of the divisions near it? Measuring a series of contiguous spaces on any part of 
 the circle we have the data for answering this question. These measures can also be made 
 aviilable for the termination of the special divisions used in observing the nadir and horizontal 
 points, or any special star the position of which is required with a high degree of accuracy. 
 The method of treating these measures is as follows: - . • 
 
 The space between two adjacent divisions being very nearly four revolutions of the microm- 
 eter, let the measure of the first, second, third, Ac, spaces be 
 
 4 rev. 4-di, 4 rev. +«i», • 4 rev. -i-«is. &«• 
 Represent also by x the error of runs of the microscope in four revolutions, and by e,, e,, Ac, 
 the errors of division. Then the measures give the series of equations 
 
 
 (1) 
 
 Here we have n equations for the determination of n-|-2 unknown quantities. The remain- 
 ing two equations are given by the condition that the sum of the squares of the errors of divi- 
 sion must be a minimum. We thus have 
 
 A* . <fcl , - <*«•• A 
 
 (8) 
 
 (3) 
 
 e in the last equation representing any error of division taken at pleasure. Regarding x and e 
 as independent variables, any change in e (« being given) will cause an equal change in all the 
 other errors of division. We have, therefore. 
 
 and therefore. 
 
 ^ <**» *.« 
 
 deM , 
 de 
 
 «^-f ei-|- Ac. • . . -f-ejiasO. 
 
 (*) 
 
 wmmtm 
 
so 
 
 DE8CKIPTI0N OF TBI TRANSIT OIBCLB, ETC. 
 
 ml 
 
 This eqnation, united with the equations (1,) will enable us to express each error of division 
 in terms of x. Multiplying the first of equations (1) by », the second by n— 1, the third by n— 2, 
 Ac, and adding, remembering that 
 
 JSemiO, 
 
 we have an equation which gives e^ in terms of x. By using the proper coefficients of elimi- 
 nation, we have each of the other errors in terms of x. These coefficients are indicated in the 
 following table, in which each column is to be read downward, in connection with the words 
 Mi the left, 
 
 To express e^ «i c^ en—1 m 
 
 the eoeffieient of ai is « —1 —1 . . . —1 ->1 
 
 * " " at is »— 1 »— I — 2 . . . —8 —2 
 
 « •• « oa is «— 2 »— 2 «— 2 ... —8 —8 
 
 
 m 
 
 tip 
 
 M 
 U 
 
 a.(>i-l)is 2 2 2 . . . -(w-l) 
 
 « " " aji is 1 1 1 . . . 1 
 
 The equations thus obtained are as follows: 
 
 -(•-1) 
 
 — « 
 
 •jn+l) 
 
 2 
 
 x+Hai+(n~-l)at+ .... +a4i+(fi+l)A,a=:0, 
 
 {n-2){n+iy 
 
 2 
 
 «— ai+(fi— 1>8,+ .... +a.»+(ii+l)«is=0, 
 
 %S;^J («-4)(ll+l) 
 
 2 
 
 X— ai— 2a,+()i— 2)a3+ . . . +an+{n+l)e,ssO, 
 
 (6) 
 
 2 
 
 X— fli— 2<i|— Sas— .... — iiaji+(fi+l)ejiassO. 
 
 We have now only to suppose the last members of these equations to be accidental errors, 
 and solve by least squares. This solution gives for the eqnation in 0, 
 
 »{n+l)(n+ 2) 
 6 
 
 x+nai+2{n—l)at+2(n'-2)a3+ .... +na.nssO. 
 
 Having thus obtained the value of a;, the quantities e^—% ^t'^* ^^'•i ^^^ immediately formed 
 from equations (1.) Thus, when we have e^ the other e's are formed in succession by the adding 
 on of successive differences. The value of e^ may be obtained from the equation 
 
 formed by multiplying the first member of (1) li>y n, the second by n — 1, Ac, and adding with 
 reference to (4.) But the most ready way of obtaining all the e's is perhaps by forming them 
 from e;,:=0, and then subtracting from each ci them the mean value of the e's thus obtained. 
 If one of the e's is known by the methods already explained, the values of the differences 
 eg— e,, Ac, will not be changed, and the vulues of aU the other errors may be found by addition 
 of the differences to this known value. 
 
 m 
 
 ««« 
 
 MMMW 
 
 ^■1 
 
PART III. 
 
 
 DETERMINATION OP THE CONSTANTS OP THE TRANSIT CIRCLE AND ITS 
 
 SUBSIDIARY APPARATUS. 
 
 « 
 
 (62) Valve <f me divitum <^ ike Spirit Lends. — It will be remembered that there are three 
 levels — one striding level for each coUimatov, and a hanging level for the axis of the in trament. 
 The former are distingmshed by the letters A and B. 
 
 1865, October 18. The old moral circle was turned until the telescope was horizontal, 
 and the collimator levels were set astride of the telescope. Microscopes A and B of the circle, 
 and the two ends of the babble of the level, were then read in different positions of the tele* 
 scope, as follows: 
 
 IficA. 
 
 lO&B. 
 
 LOTdA. 
 
 L«r«lB. 
 
 
 
 
 
 
 
 S.«d. 
 
 N.end. 
 
 S.eiid. 
 
 N.end. 
 
 1 1$ 
 
 It 
 
 i. 
 
 i. 
 
 d. 
 
 4. 
 
 w«r.» 
 
 70.8 
 
 1&4 
 
 «i.3 
 
 18.6 
 
 38.3 
 
 64.8 
 
 66.4 
 
 1.1 
 
 85.8 
 
 1.6 
 
 81.5 
 
 7S.8 
 
 77.9 
 
 87. 6 
 
 51.3 
 
 97.6 
 
 47.3 
 
 M.4 
 
 87.1 
 
 8.3 
 
 86.8 
 
 3.0 
 
 98.8 
 
 75.1 
 
 77.6 
 
 86.6 
 
 50.5 
 
 96.4 
 
 46.8 
 
 6K.S 
 
 67.8 
 
 15.0 
 
 38.1 
 
 16.4 
 
 36.8 
 
 t r^^'i 
 
 66.8 
 
 44.6 
 
 96.5 
 
 41.0 
 
 81.8 
 
 ( 
 
 flO.9 
 
 63.4 
 
 36.5 
 
 18.6 
 
 34.1 
 
 14.8 
 
 • 
 
 49.7 
 
 68.8 
 
 60.0 
 
 86.1 
 
 46.0 
 
 86.8 
 
 \. 
 
 71.0 
 
 7a8 
 
 83.8 
 
 — 0.6 
 
 81.4 
 
 1.6 
 
 a 
 
 50.9 
 
 68.8 
 
 49.6 
 
 96.5 
 
 46.8 
 
 86.0 
 
 1 
 
 68.8 
 
 71.8 
 
 86.7 
 
 8.6 
 
 84.3 
 
 4.6 
 
 5 
 
 L59.9 
 
 68.7 
 
 37.8 
 
 13.4 
 
 34.0 
 
 14.8 
 
 Temperatnre 57°. 
 
 Prom tiie observations is concluded — 
 
 One division of levd Ab0".826. 
 One division of levd B8sO".860. 
 
 The level error of the collimators being kept quite small, rarely no great as 3", the valae of one 
 division, 
 
 0".84, 
 has been adopted for both levels. 
 
 *lWil|l!M'>IW>im'*l*WU#!gi 
 
w 
 
 2S 
 
 DESCRIPTION OF THE TRANSIT OIROLB OF THE 
 
 1866, November 21. The hanging level was nnspended from the telescope of the Moral 
 Circle, and the following readings taken: 
 
 Mie.A. 
 
 IficB. 
 
 Level. 
 
 
 
 
 
 N.end. 
 
 a end. 
 
 // 
 
 // 
 
 d. 
 
 d. 
 
 16.8 
 
 81.1 
 
 JSt.l 
 
 84.5 
 
 86.9 
 
 31.8 
 
 80.6 
 
 78.3 
 
 40.3 
 
 46.1 
 
 6.8 
 
 67.8 
 
 14.9 
 
 19.4 
 
 34.4 
 
 86.3 
 
 41.5 
 
 46.1 
 
 4.0 
 
 66.9 
 
 35.7 
 
 40.6 
 
 10.8 
 
 61.1 68.1(7) 
 
 90.8 
 
 36.8 
 
 16.4 
 
 68.3 
 
 84.8 
 
 89.0 
 
 S8.8 
 
 74.9 
 
 18.6 
 
 83.8 
 
 89.8 
 
 81.8 
 
 13.8 
 
 17.8 
 
 36.8 
 
 88.8 
 
 48.9 
 
 47.7 
 
 8.0 
 
 64.1 
 
 Temperature 53*^. From which is concluded 
 
 One divi8ion=0".872a0«.068. * 
 
 The three readings preceding the last seem to indicate a diminution of the value as we 
 r^pproach the end of the scale, but the diminution is no greater than what may be due to errors 
 of the microscopes. 
 
 (53) Difference of ccUara of ooRimaiora, — It is requisite that we know the difference be< 
 tween the true level of the axis of the collimator and the readings of the spirit level set upon 
 the collars on which the collimators turn. This is effected by reading the level when the col- 
 limator is in its regular position, and when reversed, end for end, around a vertical axis, the 
 operation being repeated a number of times in succession to eliminate the possible changes of 
 level during the operation. 
 
 The following are the level indications of the two collimators in different angles of position. 
 The angle of position of the collimator indicates its position as it is turned upon its Ts, with- 
 out being raised from them. They are used only in those four positions in which one of the 
 wires is horizontal; and, in practice, the positions are almost entirely confined to two, which 
 are designated as 0° and 180°. The angle is measured by the position of the clamp which 
 binds the eye-piece, being called 0° when this clamp points horizontally to the right of an ob- 
 server looking into the eye end of the collimator, and 90° when it points upward. 
 
 The collimator is said to be direct when the collimator points toward the Transit Circle; 
 reverse when it points from it. The level readings are positive when the pivot farthest from 
 the Transit Circle is too high. 
 
 COLLIMATOR A. 
 
 Ftr«< acrtM.— Position changed from 0° to 180°, and the collimator reversed alternately 
 one pair level readings between each change. 
 
 
 PoeiUon 0°. 
 
 
 
 FMltton 180°. 
 
 Direct. 
 
 Reverae. 
 
 Diraet 
 
 Reverse. 
 
 Level A . . 
 Level B . . 
 
 d. 
 f 3.84) 
 + 8.91 J 
 — 0.46 
 -1- 0.08 
 
 d. 
 + 1.35 
 
 — 0.78 
 
 — 0.96 
 
 d. 
 
 + ^fiX 
 
 -- 8.60$ 
 ■ - l.OS 
 -• 0.56 
 
 d. 
 + 1.86 
 
 — 0.90 
 
 — 1.08 
 
 Meen. D-R. 
 
 + 1.32 
 
 d. 
 + 1.80. 
 
 tk 
 
UNITBD STATES NAVAL OBSERVATOBT. 
 
 Second »erie». — PoBition 0°. 
 
 Third amw.— Position 180°. 
 
 Diraet. 
 d. 
 +0.81 
 
 +0.32 
 
 ~0J84 
 
 0.00 
 
 Mean. D~B. 
 
 d. 
 -0.18 
 +0.36 
 -0.39 
 
 BareiM. 
 d. 
 
 -1.S4 
 
 -0.94 
 
 -1.42 
 
 d. 
 +1.42. 
 
 d. 
 -0.96 
 -1.65 
 -0.89 
 
 d. 
 + 1.09 
 
 '-m '•' 
 
 Mean, D— K. 
 
 The ooncloded correction for diflFerence of collars, when in the position 0° or 180®, is one- 
 fonrth D— E. or 0".29, the eye-pivot being too large. 
 
 Two similar, but more accordant, series gave for the correction to position 90°— 270° 
 0".09, so that the collars are not perfectly cylindrical. 
 
 OOLLIHATOB B. 
 
 The results are 
 
 The determinations have been made only for the positions 0° and 180°. 
 
 d. 
 
 D-B for 0°, +1.06 i ;r^ . 
 
 for ISOO, +0.66 ; . 
 
 Concluded correction, 0". 17, the eye collar being too large, as in the other collimator. 
 
 The collars are decidedly conical, diminishing toward the ends of the telescope, so that 
 entire dependance cannot be placed on the absolute horizontal point obtained from a single 
 collimator. But, by interchanging the collimators, this error is completely eliminated from the 
 zenith point. 
 
 (54) Periodic InequaliHeso^ i^he Micrometer Screws.— In the case of the eight microscope 
 micrometers the inequalitie" *« --6 determined by measuring the intervals between the parallel 
 wires of each pair with different portions of the screw. The circle being clamped and properly 
 set, one of the micrometer wires was brought to a distance from the edge of a division approx- 
 imately equal to the thickness of the wire. The observer retained a quite accurate idea of the 
 intervening space, though the idea could not be defined in language. The micrometer was 
 then read. The other wire was then brought into the same position, and the micrometer was 
 again read. The operation was twice repeated, making three measures in all. The circle was 
 then moved forward 6" by the tangent screw, and a similar series of measures again taken. 
 The operation was continued through the two revolutions of the screw most used. 
 
 Repeated trials showed that the wires could be set more accurately in position by this 
 method than by making them coincide with the circle division, the probable error of a single 
 setting being about 0".10, scarcely greater than that in putting the division midway between 
 the wires. 
 
 The results of a determination made in November, 1865, were: 
 
 For microacopes I, II, III, inequality InsenBible ; 
 
 IV. Ineq. = — 0".63 cos ti+0".67 sin t»+0".13 cos 2i»— 0".l 1 ain 2« ; 
 
 V. —0 -0900811+0 .028ini(; 
 
 VI. +0 .24co8«+0 .aSsiDw; 
 
 VII. — .06cos«+0 .06Bini»; 
 
 VIII. +0 .06coa«+0 JilBinu; 
 
 u being the angle of the reading of the head. 
 
 n 
 
 w^<i.„V>aw*|-jM— MWiWWWilll 
 
 "mm 
 
w 
 
 i; 
 
 24 
 
 OESOSIPTION OF THE TRANSIT OUtOLH OF THE 
 
 WW 
 
 ;,i 
 
 Microscopes V-YIII being alone employed in astronomical operations, their inequalities 
 were redetermined in March, 1866, with the following result: 
 
 V. — 0".05 COB «+0".05 sin « ; 
 
 VI. +0 •84COBU+0 -Sfisinit; 
 
 VII. —0 .07cos«+0 .OSsiiiK; 
 
 yill. +0 .llcos«(+0 JS6tin«. 
 
 The mean of the two results is adopted as the correction for the year 1866. 
 
 (55) A rough check upon the general accuracy of the screws is given by the mean dis* 
 tance of the wires as measured in different revolutions of the screw. The following are the 
 distances given by the same measures which determine the inequalities : 
 
 
 ' 1866. 
 
 1666. 
 
 Difference. 
 
 29 nv. 
 
 30 rev. 
 
 29 rev. 
 
 30 rev. 
 
 1866. 
 
 1866. 
 
 Mic. V . 
 Mic. VI . 
 Mio. VII . 
 Mic. VIII . 
 
 10.98 
 10.88 
 10.63 
 10.88 
 
 // 
 
 11.06 
 
 10.82 
 10.47 
 10.87 
 
 II 
 
 10.80 
 10.80 
 11.86 
 10.74 
 
 // 
 
 10.91 
 10.95 
 11.76 
 10.92 
 
 + 0.07 
 
 — 0.06 
 
 — 0.06 
 
 — 0.01 
 
 II 
 
 + 0.11 
 + 0.16 
 - 0.10 
 + 0.18 
 
 The method of observing is such that the microscope micrometers are seldom moved through 
 more than a fraction of a revolution . No very accurate investigations have therefore been 
 made to find whether the value of the revolution of any one of them changes progressively; but 
 the measures occasionally made for runs show that the change, if it exists, is entirely inappre- 
 ciable. 
 
 There are, however, outstanding discrepancies, amounting sometimes to three or four 
 tenths of n second, which I have not been able to refer to any law. 
 
 (56) The Dedination Micrometer of Telescope. — Neither the collimators nor the telescope 
 micrometer were originally furnished with double wires; the usual method could not, therefore, 
 be used to determine the inequality of the lalter. The plan was, therefore, adopted of meas- 
 uring successive half revolutions of the telescope micrometer with the microscope micrometers, 
 by setting the wire of the former upon the collimator. The following are the value of eight 
 successive half revolutions, as given by two microscopes of each circle, in one series, and the 
 four microscopes V-VIII in the other : 
 
 
 Fint Mriei. 
 
 Mid. 
 
 Mio. III. 
 
 Mic VI. 
 
 MicVm. 
 
 Meui. 
 
 r. r. 
 
 // 
 
 // 
 
 // 
 
 // 
 
 // 
 
 24.0-24.6 
 
 7.70 
 
 7.61 
 
 7.91 
 
 8.39 
 
 7.90 
 
 24.5-85.0 
 
 7.77 
 
 7.76 
 
 7.86 
 
 7.49 
 
 7.78 
 
 85.0-85.6 
 
 7.66 
 
 7.37 
 
 7.66 
 
 7.66 
 
 7.66 
 
 85.5-86,0 
 
 7.81 
 
 7.41 
 
 7.46 
 
 7.41 
 
 7.68 
 
 86.0-86.5 
 
 7.75 
 
 7.86 
 
 8.00 
 
 &86 
 
 7.96 
 
 86.5 — 87.0 
 
 7.86 
 
 7.61 
 
 7.63 
 
 7.38 
 
 7.60 
 
 87.0 — 87.6 
 
 7.65 
 
 7.65 
 
 8.03 
 
 7.86 
 
 7.77 
 
 87.5 — 88.0 
 
 7.83 
 
 7.63 
 
 7.61 
 
 7.76 
 
 7,71 
 
 The mean of the odd measures is 7 ".80, and of the even ones, 7". 64. 
 
 •3 ■•V«r»WW««rfj*tTjn'^f«.:-*,'s^5»:tBBi 
 
UNITED STATES NAVAL OBSEBVATOBY. 
 
 
 Second lerioi. 
 
 MicL 
 
 Mio. III. 
 
 Mlc. VI. 
 
 Mi«.VIII. 
 
 Mean. 
 
 r, r. 
 
 II 
 
 // 
 
 H 
 
 // 
 
 // 
 
 31.25 — 31.75 
 
 7.64 
 
 7.fl9 
 
 7.51 
 
 7.36 
 
 7.58 
 
 31.75-38.85 
 
 7.54 
 
 7.9fi 
 
 7.73 
 
 8.02 
 
 7.82 
 
 38.25 — 32.75 
 
 7.75 
 
 7.66 
 
 7.S6 
 
 7.86 
 
 7.78 
 
 32.75-33.25 
 
 7.51 
 
 7.49 
 
 7.24 
 
 7.22 
 
 7.36 
 
 33. 2C- 33. 75 
 
 7.37 
 
 7.50 
 
 7.64 
 
 7.71 
 
 7.56 
 
 33.75 — 34.25 
 
 7.57 
 
 7.76 
 
 7.6« 
 
 7.61 
 
 7.66 
 
 34.25 — 34.75 
 
 7.94 
 
 7.64 
 
 7.75 
 
 7.85 
 
 7.79 
 
 34.75 — 35.25 
 
 7.49 
 
 7.57 
 
 7.62 
 
 7.60 
 
 7.57 
 
 J} 
 
 \% iM'-'%h' '- 
 
 The meau of the odd lueasare is 7". 66, and of the even ones, 7". 60. 
 
 From the differences 0".16 and 0".06 of the half revolutions the screw would seem to be 
 
 affected with the error . 
 
 0".04co8«+0".02 8in«; 
 
 an error so Hmall that it has been neglected, especially as the screw is used about equally 
 
 through several revolutions for every class of observations, by which the periodic errors will 
 
 disappear from the final result of the observations. After double wires were placed in the 
 
 collimators, measurements of successive half revolutions of the micrometer were made, which 
 
 gave for the periodic correction, . , 
 
 0".07 COB «— 0".05 sin «. 
 
 (57) Value of. a Revolution of the Micrometer /Screws.— 1866, June 12. — The telescope was 
 set on collimator B. The record does not state whether it was north or south. The microscopes, 
 however, indicate that it was south. 
 
 The following readings of the microscopes and zenith distance micrometer were taken 
 successively in different positions of the telescope. Each micrometer reading is the mean of 
 nine; three for coincidence of collimator image with each wire, and three for a position midway 
 between wires: 
 
 I 
 
 •■ 
 
 ' 
 
 JS 
 
 
 
 
 .a 
 
 1 
 
 1^ 
 
 Readings of microecopes— 
 
 
 Corr. for— 
 
 "^* 
 
 
 •s-^ 
 
 
 Mean. 
 
 
 
 !^ 
 
 *\ 
 
 1^ 
 
 
 
 
 
 ^i 
 
 1 
 
 
 
 
 
 
 
 1 
 
 V. 
 
 VI. 
 
 vn. 
 
 vra. 
 
 
 lueq. 
 
 Div. 
 
 c 
 
 5 
 
 ( 
 
 r. " 
 
 // 
 
 // 
 
 41 
 
 II 
 
 
 
 / It 
 
 r. 
 
 56 
 
 9 19.5 
 
 SO. 3 
 
 21.2 
 
 23.7 
 
 21.18 
 
 - .16 
 
 - .04 
 
 56 50.99 
 
 28.408 
 
 .54 
 
 16.5 
 
 18.0 
 
 19.2 
 
 21.8 
 
 18.89 
 
 - .14 
 
 — .12 
 
 53 48.63 
 
 20.392 
 
 58 
 
 21.4 
 
 22.8 
 
 23.2 
 
 26.0 
 
 23.35 
 
 ~ .12 
 
 + .16 
 
 57 53.39 
 
 .%.377 
 
 64 
 
 15.8 
 
 17.9 
 
 18.9 
 
 21.2 
 
 18.45 
 
 - .14 
 
 — .12 
 
 53 48. 19 
 
 20.373 
 
 58 
 
 21.6 
 
 22.8 
 
 23.4 
 
 25.9 
 
 23.43 
 
 - .12 
 
 + .16 
 
 57 53.47 
 
 36.393 
 
 66 
 
 19.0 
 
 20.6 
 
 21.4 
 
 23.7 
 
 21.18 
 
 - .15 
 
 -.04 
 
 65 60.99 
 
 28.395 
 
 These readings give from 20 r. to 28 r., 1 rev. =15".286, 
 
 28 r. to 36 r., 1 rev. =15 .337. 
 An increase in the value of a revolution as the turns increase seems to be indicated. 
 
 In the great majority of observations the micrometer is used between 26 and 34 revo- 
 lution?; a determination of the value between those limits was therefore made which gave 
 I5".296. The value actually adopted was the mean of the two determinations, or 15".303. 
 
 A third determination, made on September 3, 1866, gave 
 
 From 26 r. to 34 r., 1 rev. =il5".311 ; 
 34 r. to 4fi r., 1 rev. sl5 .375 ; 
 which seems to confirm the suspicion of an increase in the value of a revolution as the screw 
 
 is advanced. 
 
 4 
 
 '^^*h~^--m&^St 
 
 M Pt WM 
 
 mmKmieMms &satmmmm s i/^mwtmim ^xm^ 
 
hi 
 li 
 
 26 
 
 DESCBirnOH OP THS TRAKSrr OIBOLS OP THE 
 
 ^: 
 
 (58) Irregularity of the Screw. — It is desirable to know whether this change in the valae 
 of revolution is regularly progressive, or whether it is subject tr sudden changes. To learn 
 this, advantage was taken of the fact that the distance of the pair of wires in collimator A, 
 and of the close pair of wires moved by the micrometer, are together nearly equal to a revo- 
 lution of the latter. Consequently, turning the telescope on that collimator when its double 
 wires are horizontal, we have a measure of a constant space by setting first the upper microme- 
 ter wire on the lower image of the collimator wire, and then the lower micrometer wire on the 
 upper image. The measures were commenced, as nearly as practicable, at an even revolution, 
 and therefore ended nearly at the beginning of the next revolution. The measures were made 
 on three different dates, and three measures of each revolution made on each date. But the 
 measures were not always continued through the entire range of the screw. About 30 revolu- 
 tions they were prevented by the interference of the fixed horizontal wires of the reticule. 
 
 a 1 
 
 
 I 
 
 
 IfeaniN. 
 
 
 1 
 
 i 
 
 ^ 
 
 
 
 
 
 h 
 
 1 
 
 
 Her. 
 
 Arc. 
 
 Cor. arc 
 
 1 
 
 r. 
 
 r. 
 
 r. 
 
 // 
 
 II 
 
 // 
 
 
 17.7 
 
 18.7 
 
 0.964 
 
 15.06 
 
 14.97 
 
 + 0.13 
 -- 0.18 
 
 1 
 
 18.7 
 
 19.7 
 
 .964 
 
 15.06 
 
 14.97 
 
 1 
 
 19.6 
 
 20.6 
 
 .978 
 
 14.88 
 
 14.80 
 
 — 0.05 
 
 1 
 
 30.0 
 
 21.0 
 
 .986 
 
 15.09 
 
 15.03 
 
 -f 0.17 
 + 0.08 
 
 3 
 
 80.5 
 
 21.5 
 
 .980 
 
 15.00 
 
 14.93 
 
 4 
 
 81. 
 
 88.0 
 
 .971 
 
 14.86 
 
 14.79 
 
 — 0.06 
 
 4 
 
 28.0 
 
 83.0 
 
 .976 
 
 14.94 
 
 14.88 
 
 -1- 0.03 
 
 4 
 
 33.0 
 
 84.0 
 
 .978 
 
 14.88 
 
 14.63 
 
 — 0.02 
 
 3 
 
 84.0 
 
 85.0 
 
 .973 
 
 14.89 
 
 14.85 
 
 0.00 
 
 3 
 
 85.0 
 
 86.0 
 
 .966 
 
 14.77 
 
 14.73 
 
 — 0.13 
 
 3 
 
 86.0 
 
 27.0 
 
 .973 
 
 14.88 
 
 14.85 
 
 0.00 
 
 3 
 
 27.0 
 
 88.0 
 
 .965 
 
 14.77 
 
 14.75 
 
 — 0.10 
 
 3 
 
 88.0 
 
 80.0 
 
 .964 
 
 14.76 
 
 14.75 
 
 — 0.10 
 
 3 
 
 31.0 
 
 33.0 
 
 .970 
 
 14.85 
 
 14.86 
 
 -1- 0.01 
 
 3 
 
 38.0 
 
 33.0 
 
 .968 
 
 14.88 
 
 14.83 
 
 — 0.08 
 
 3 
 
 33.0 
 
 34.0 
 
 .976 
 
 14.93 
 
 14.94 
 
 -)- 0.09 
 
 1 
 
 34.0 
 
 35.0 
 
 .961 
 
 14.71 
 
 14.74 
 
 - 0.11 
 
 1 
 
 35.0 
 
 36.0 
 
 .965 
 
 14 77 
 
 14.81 
 
 — 0.04 
 
 I 
 
 36.0 
 
 37.0 
 
 .966 
 
 14.79 
 
 14.83 
 
 - 0.08 
 
 1 
 
 Of the three columns headed "Measure," the first gives the measures in revolutions of a 
 micrometer; in the second these revolutions are turned into arc, using 1 rev.ssl6".303 ; in 
 the third they are corrected for progressive change in the value of the revolution. The resid- 
 uals show the excesses of the individual corrected measures over the mean value 14". 85. I 
 conceive that they proceed mainly from the accidental errors of reading and temporary derange- 
 ments of the motion of the screw by dust, displacement of the oil, and other causes, and that 
 the screw itself may be regarded as sensibly regular. 
 
 ' (59) J7. A. Micrometer. — ^The value of a revolution of this micrometer seems to be exactly 
 the same as that of the other. Wide measures give 15". 300. Owing to its limited use, no 
 special investigation of its movement has been entered upon. 
 
 (60) Flexure of the Cirdes.—The work of determining separately the flexure of the different 
 parts of each circle was commenced in 1866, January 19. But after taking one series of read- 
 ings, it was found that the axes of several of the microscopes deviated quite sensibly from the 
 perpendicular to the face of the circle; some deviating as much as 40'. To adjust them with 
 entire accuracy appeared to be a difficult and troublesome operation ; but a kind of ^ T-square 
 was made by which they could be set without an error exceeding 5', and they w^re adjusted 
 by it on April 7. The set of readings previously made were not used. 
 
 The following are the details of the operations by which the definitive values of the flexure 
 coefficients were obtained: 
 
UNITBD STATU VAVAL OBSKBYATOSY. 
 
 %1 
 
 In the first series of readings the position of the circles wm snch that when the telescope 
 pointed to the senith, the divisions of circle A, which were nnder microscopes V-VIII, were: 
 
 MIc. V, ISfiO; VI, 2SfiO; VII, 31 5° j VIII, IfiO 
 The divisions of circle B, under microscopes I-IV, were: 
 
 Mie.1, 3160. II, 4AO; III, ISfiO. IV, SSfiO. 
 
 When the telescope pointed at senith distance s, the above readings of microscopes V-VIII 
 wonld be increased, and those of I-IV diminished by i. 
 
 The operation was began on divisions 0°, 90<^, 180°, 270°. The division 0° of circle B was 
 brought under microscope I, and the eight microscopes were read , one observer reading each 
 circle. The telescope was then turned 90°, and the microscopes were again read. The tele- 
 scope was again turned 90°, and the operation of turning and reading continued until the 
 microscopes had been read three times in each 6( the four positions of the telescope. The 
 mean of the three readings was taken as that corresponding to each position. 
 
 The telescope was then set 15° backward from its first position, and the same operation 
 was performed on divisions 75°, 165°, Ac, of circle A, and 15°, 105°, Ac, of circle B. The 
 operation was repeated through every 15° of the quadrant. The first series was now complete. 
 
 In the second series circle A was loosened and turned 180° on its axis, so that when the 
 telescope pointed to the senith, the division 315° was under microscope V. A series of read- 
 ings exactly like the first was then made on the two circles. 
 
 Circle B was then turned on its axis 180°, and a third series of readings were made. 
 
 Finally, circle A was returned to its original position, and a fourtb series of readings were 
 made. 
 
 The mean of the three readings for each position of the instrument is given in the foUoW' 
 ing table: 
 
 8EBIES I. 
 
 I 
 
 I 
 
 I 
 
 46 
 
 las 
 
 986 
 315 
 
 30 
 190 
 310 
 300 
 
 15 
 106 
 196 
 986 
 
 
 
 90 
 
 160 
 
 £70 
 
 346 
 
 75 
 
 165 
 
 966 
 
 330 
 
 60 
 
 160 
 
 940 
 
 9.0 
 
 180 
 
 00 
 
 
 
 986 
 
 106 
 
 106 
 
 16 
 
 300 
 
 910 
 
 190 
 
 30 
 
 316 
 
 996 
 
 136 
 
 45 
 
 330 
 
 940 
 
 160 
 
 60 
 
 345 
 
 965 
 
 165 
 
 78 
 
 180 
 
 970 
 
 
 
 00 
 
 166 
 
 966 
 
 346 
 
 76 
 
 160 
 
 940 
 
 330 
 
 60 
 
 136 
 
 996 
 
 315 
 
 45 
 
 190 
 
 910 
 
 300 
 
 30 
 
 106 
 
 105 
 
 985 
 
 15 
 
 Beading* of micnwcopM. 
 
 97.00 
 97.40 
 93.43 
 94.37 
 
 96.57 
 96.43 
 93.47 
 99.97 
 
 96.87 
 97.3:1 
 94.03 
 93.17 
 
 96.57 
 97.73 
 96.47 
 93.30 
 
 96.03 
 97.13 
 96.37 
 99.86 
 
 94.77 
 97.43 
 96.30 
 99.70 
 
 n. 
 
 95.63 
 95.60 
 93.70 
 91.00 
 
 33.67 
 95.10 
 93.70 
 90.13 
 
 84.77 
 95.03 
 93.63 
 91.50 
 
 93.00 
 96.00 
 94.60 
 99.35 
 
 99.43 
 
 96.07 
 94.17 
 99.60 
 
 91.97 
 96.87 
 95.57 
 99.37 
 
 m. 
 
 4.97 
 4.77 
 3.73 
 3.07 
 
 3.40 
 3.83 
 4.40 
 1.80 
 
 4.67 
 4.77 
 3.30 
 9.47 
 
 4.53 
 5.33 
 4.70 
 9.60 
 
 3.30 
 4.70 
 4.03 
 9.60 
 
 3.10 
 4.60 
 4.93 
 3.13 
 
 lY. 
 
 0.67 
 99.13 
 97.83 
 98.50 
 
 98.03 
 98.87 
 97.00 
 87.17 
 
 99.70 
 90.40 
 97.07 
 97.03 
 
 90.97 
 
 0.40 
 
 97.77 
 
 97.75 
 
 98.13 
 
 0.93 
 
 87.47 
 
 97.36 
 
 98.57 
 
 0.03 
 
 88.67 
 
 97.57 
 
 V. 
 
 4.07 
 5.33 
 5.43 
 4.50 
 
 6.93 
 6.30 
 5.80 
 6.10 
 
 4.83 
 5.97 
 6.00 
 6.17 
 
 4.67 
 4.67 
 4.93 
 5.00 
 
 5.83 
 6.17 
 5.33 
 5.45 
 
 5.80 
 5.97 
 5.40 
 6.33 
 
 VI. 
 
 7.07 
 6.87 
 4.63 
 6.73 
 
 8.60 
 8.00 
 4.47 
 8.60 
 
 6.83 
 7.43 
 5.60 
 6.77 
 
 7.30 
 6.67 
 6.07 
 6.90 
 
 8.13 
 7.00 
 6.43 
 5.35 
 
 7.70 
 7.77 
 6.30 
 6.77 
 
 YU. 
 
 8.87 
 0.67 
 8.17 
 4.57 
 
 4.77 
 1.83 
 1.63 
 5.57 
 
 3.67 
 1.67 
 1.43 
 4.63 
 
 a03 
 1.50 
 0.63 
 6.80 
 
 4.17 
 8.87 
 1.03 
 8.96 
 
 5.03 
 3.30 
 0.93 
 9.83 
 
 Yni. 
 
 3.43 
 a 10 
 3.80 
 5.63 
 
 4.13 
 9.37 
 3.97 
 5.16 
 
 6.60 
 9.93 
 3.73 
 4.80 
 
 ■MwiwilMlM 
 
 m u mm 
 
M 
 
 
 '#' 
 
 315 
 
 45 
 
 136 
 
 886 
 
 300 
 
 30 
 
 190 
 
 SIO 
 
 985 
 
 15 
 
 106 
 
 196 
 
 870 
 
 
 
 90 
 
 180 
 
 856 
 
 346 
 
 75 
 
 166 
 
 840 
 
 330 
 
 <M) 
 
 150 
 
 DESCRIPTIOir or THE TKANSIT OIBOLK OW THK 
 8EKIEB II. 
 
 o 
 
 870 
 IHO 
 
 90 
 
 16 
 886 
 196 
 106 
 
 30 
 300 
 310 
 180 
 
 45 
 315 
 885 
 136 
 
 60 
 330 
 840 
 150 
 
 75 
 345 
 856 
 165 
 
 195 
 
 885 
 
 15 
 
 105 
 
 Raitdingf of microMsopM. 
 
 89.00 
 
 1.70 
 
 86.90 
 
 85.36 
 
 86.70 
 
 0.80 
 
 87.40 
 
 84.73 
 
 84.80 
 98.97 
 96.70 
 89.80 
 
 83.67 
 87.77 
 86.47 
 83.30 
 
 89.73 
 86.90 
 87.03 
 83.80 
 
 89.50 
 86.80 
 87.47 
 84.87 
 
 II. 
 
 88.30 
 
 0.90 
 
 96.65 
 
 96.35 
 
 96.00 
 99.93 
 97.63 
 96.57 
 
 94.93 
 87.90 
 96.17 
 83.37 
 
 83.77 
 85.90 
 96.73 
 83.33 
 
 83.60 
 95.13 
 97.10 
 99.97 
 
 83.10 
 96.37 
 97.13 
 94.10 
 
 III. 
 
 6.70 
 7.30 
 9.96 
 3.45 
 
 4.60 
 6.87 
 3.00 
 a90 
 
 9.33 
 4.67 
 8.17 
 0.67 
 
 0.66 
 .^57 
 9.07 
 0.63 
 
 0.83 
 3.13 
 9.97 
 0.10 
 
 0.80 
 3.99 
 3.47 
 0.77 
 
 IV. 
 
 0.45 
 
 1.50 
 
 96.56 
 
 96.86 
 
 98.73 
 
 0.80 
 
 86.30 
 
 84.77 
 
 86.63 
 88.53 
 95.63 
 89.83 
 
 84.57 
 87.03 
 86.97 
 88.73 
 
 83.83 
 96.77 
 87.30 
 89.40 
 
 83.93 
 87.80 
 97.33 
 93.93 
 
 4.70 
 5.10 
 6.06 
 6.80 
 
 6.67 
 5.73 
 6.37 
 6.40 
 
 6.17 
 6.33 
 6.57 
 6.57 
 
 7.17 
 7.33 
 7.90 
 7.13 
 
 7.00 
 7.10 
 6.87 
 7.00 
 
 7. S3 
 
 7.07 
 6.97 
 7.97 
 
 VI. 
 
 VII. 
 
 8.40 
 5.80 
 8.36 
 9.66 
 
 9.90 
 6.80 
 7.60 
 9.97 
 
 11.10 
 7.63 
 7.00 
 9.77 
 
 11.93 
 9.33 
 8.97 
 
 10.53 
 
 10.80 
 
 10.47 
 
 7.90 
 
 10.93 
 
 11.70 
 
 10.43 
 
 7.87 
 
 9.83 
 
 0.40 
 
 98.76 
 
 3.80 
 
 4.60 
 
 1.83 
 
 99.67 
 
 9.77 
 
 4.53 
 
 a60 
 0.10 
 9.60 
 6.67 
 
 4.70 
 1.47 
 9.77 
 6.80 
 
 5.60 
 1.83 
 1.60 
 6.37 
 
 6.03 
 9.37 
 1.33 
 5.90 
 
 SERIES m. 
 
 VIU. 
 
 90.70 
 
 30.30 
 
 4.66 
 
 9.56 
 
 1.50 
 0.87 
 4.93 
 9.37 
 
 9.13 
 0.77 
 4.43 
 4.90 
 
 9.67 
 9.07 
 4.07 
 6.07 
 
 3.00 
 9.03 
 9.90 
 6.70 
 
 3.98 
 1.93 
 3.07 
 5.83 
 
 135 
 
 
 
 90 
 
 96.30 
 
 94.80 
 
 6.40 
 
 97.97 
 
 6.57 
 
 9.40 
 
 4.47 
 
 4.93 
 
 885 
 
 870 
 
 180 
 
 89.57 
 
 97.80 
 
 6.77 
 
 98.93 
 
 8.77 
 
 10.37 
 
 6.17 
 
 3.50 
 
 315 
 
 180 
 
 870 
 
 88,77 
 
 98.63 
 
 6.87 
 
 96.50 
 
 6.90 
 
 10.87 
 
 3.37 
 
 9.90 
 
 45 
 
 90 
 
 
 
 86.17 
 
 97.80 
 
 7.17 
 
 96.63 
 
 7.80 
 
 ao3 
 
 1.97 
 
 a90 
 
 ISO 
 
 15 
 
 76 
 
 83.70 
 
 93.17 
 
 4.97 
 
 96.40 
 
 7.37 
 
 0.07 
 
 4.87 
 
 6.77 
 
 810 
 
 985 
 
 166 
 
 98.40 
 
 85.87 
 
 4.30 
 
 97.70 
 
 7.17 
 
 10.60 
 
 6.73 
 
 3.93 
 
 300 
 
 196 
 
 956 
 
 98.07 
 
 93.03 
 
 4.40 
 
 96.90 
 
 7.10 
 
 11.47 
 
 3.70 
 
 9.97 
 
 30 
 
 106 
 
 345 
 
 96.53 
 
 97.87 
 
 6.97 
 
 94.93 
 
 7.00 
 
 7.47 
 
 1.93 
 
 9.33 
 
 106 
 
 30 
 
 60 
 
 94.93 
 
 84.40 
 
 6.47 
 
 96.10 
 
 6.70 
 
 8.37 
 
 3.10 
 
 4.63 
 
 196 
 
 300 
 
 150 
 
 87.60 
 
 94.97 
 
 4.47 
 
 97.97 
 
 6.83 
 
 0.97 
 
 6.67 
 
 4.97 
 
 986 
 
 910 
 
 940 
 
 98.40 
 
 97.30 
 
 4.40 
 
 96.70 
 
 e.'v 
 
 11.67 
 
 4.17 
 
 9.87 
 
 16 
 
 180 
 
 330 
 
 96.17 
 
 97.03 
 
 6.17 
 
 94.80 
 
 6.93 
 
 8.97 
 
 OOS 
 
 1.60 
 
 90 
 
 46 
 
 45 
 
 93.63 
 
 84.80 
 
 6.97 
 
 96.00 
 
 7.63 
 
 8.07 
 
 8.93 
 
 4.60 
 
 180 
 
 315 
 
 135 
 
 96.67 
 
 83.07 
 
 4.63 
 
 96.47 
 
 7.37 
 
 10.87 
 
 6.77 
 
 6.13 
 
 970 
 
 886 
 
 985 
 
 87.63 
 
 96.70 
 
 3.60 
 
 96.87 
 
 7.67 
 
 19.40 
 
 6.90 
 
 3.47 
 
 
 
 136 
 
 315 
 
 96.80 
 
 96.87 
 
 6.00 
 
 94.17 
 
 7.70 
 
 9.00 
 
 1.80 
 
 9.63 
 
 78 
 
 60 
 
 30 
 
 93.90 
 
 86.83 
 
 6.00 
 
 95.80 
 
 7.37 
 
 7.40 
 
 1.90 
 
 8.43 
 
 166 
 
 330 
 
 180 
 
 96.80 
 
 93.97 
 
 6.77 
 
 96.93 
 
 7.80 
 
 10.40 
 
 6.30 
 
 6.70 
 
 966 
 
 940 
 
 810 
 
 87.87 
 
 96.97 
 
 3.00 
 
 96.97 
 
 7.30 
 
 11.30 
 
 5.60 
 
 a30 
 
 345 
 
 160 
 
 300 
 
 96.77 
 
 96.M 
 
 6.70 
 
 84.90 
 
 7.93 
 
 10.07 
 
 9.10 
 
 9.40 
 
 60 
 
 76 
 
 15 
 
 94.37 
 
 96.70 
 
 6.90 
 
 96.00 
 
 7.93 
 
 7.n 
 
 1.47 
 
 3.«7 
 
 160 
 
 346 
 
 105 
 
 96.77 
 
 93.80 
 
 6.63 
 
 97.30 
 
 7.98 
 
 9.67 
 
 5.13 
 
 6.87 
 
 840 
 
 966 
 
 195 
 
 98.17 
 
 96.60 
 
 4.63 
 
 97.30 
 
 7.13 
 
 11.40 
 
 6.77 
 
 4.03 
 
 330 
 
 166 
 
 886 
 
 97.50 
 
 98.00 
 
 6.90 
 
 86.97 
 
 7.93 
 
 10.63 
 
 9.63 
 
 9.13 
 
 BM 
 
uirmD iTATn watal OBtnnrAToxr. 
 
 8EHIE8 IT. 
 
 1.' 
 
 li 
 
 if 
 
 
 
 Baadiofft of tlM 
 
 niiofOfloopM. 
 
 
 
 I. 
 
 II. 
 
 m. 
 
 IV. 
 
 V. 
 
 VI. 
 
 VII. 
 
 vm. 
 
 o 
 
 
 
 o 
 
 // 
 
 II 
 
 
 II 
 
 II 
 
 // 
 
 II 
 
 // 
 
 «» 
 
 316 
 46 
 
 
 
 870 
 
 ISO 
 
 00 
 
 870 
 
 
 
 00 
 
 180 
 
 88.87 
 96.33 
 93.90 
 19.40 
 
 90.97 
 93.63 
 99.33 
 91.67 
 
 3.60 
 
 8.47 
 
 90.87 
 
 1.73 
 
 95.60 
 96.17 
 90.77 
 90.17 
 
 5.90 
 5.87 
 5.67 
 6.37 
 
 8.80 
 6.43 
 7.97 
 8.80 
 
 1.10 
 
 99.93 
 
 3.33 
 
 3.63 
 
 0.33 
 1.87 
 4.73 
 a 17 
 
 190 
 
 no 
 
 300 
 30 
 
 16 
 966 
 106 
 106 
 
 956 
 
 346 
 
 75 
 
 166 
 
 91.07 
 97.10 
 83.80 
 90.33 
 
 90.90 
 93.53 
 93.13 
 98.37 
 
 3.47 
 3.17 
 0.17 
 1.93 
 
 96.03 
 96.77 
 91.93 
 19.70 
 
 6.13 
 5.13 
 6.07 
 5.90 
 
 8.80 
 5.33 
 7.00 
 8.00 
 
 1.90 
 
 90.90 
 
 9.17 
 
 3.10 
 
 1.70 
 0.73 
 a 03 
 8.07 
 
 106 
 
 106 
 
 9H6 
 
 16 
 
 30 
 
 30O 
 910 
 190 
 
 940 
 
 330 
 
 60 
 
 150 
 
 90.70 
 96.80 
 94.47 
 90.47 
 
 80.33 
 83.10 
 88.93 
 88.10 
 
 9.40 
 9.83 
 0.93 
 1.30 
 
 93.07 
 96.80 
 91.93 
 19.93 
 
 6.13 
 6.17 
 5.90 
 5.93 
 
 0.37 
 6.13 
 6.70 
 7.37 
 
 9.93 
 
 90.33 
 
 1.13 
 
 a 10 
 
 •8.00 
 0.80 
 
 a 80 
 
 8.80 
 
 00 
 
 180 
 
 870 
 
 
 
 46 
 316 
 985 
 136 
 
 986 
 
 315 
 
 45 
 
 136 
 
 90.43 
 86.40 
 84.70 
 90.63 
 
 81.47 
 88.50 
 83.33 
 81.80 
 
 1.97 
 3.60 
 0.60 
 0,37 
 
 99.83 
 96.80 
 93.93 
 18.60 
 
 6.17 
 6.30 
 6.07 
 5.33 
 
 8.83 
 6.57 
 6.60 
 8.90 
 
 9.17 
 
 99.47 
 
 0.83 
 
 a33 
 
 1.53 
 0.67 
 9.57 
 a 43 
 
 76 
 166 
 986 
 346 
 
 60 
 330 
 940 
 160 
 
 910 
 
 300 
 
 30 
 
 190 
 
 19.07 
 93.97 
 95.13 
 91.47 
 
 81.30 
 81.00 
 8a 63 
 81.37 
 
 1.70 
 9.80 
 1.40 
 0.10 
 
 91.63 
 94.97 
 94.73 
 18.47 
 
 5.90 
 5.30 
 6.97 
 5.13 
 
 8.93 
 8.03 
 5. .17 
 
 7.83 
 
 9.73 
 
 0.07 
 
 89.90 
 
 8.63 
 
 1.80 
 1.03 
 1.47 
 a 90 
 
 60 
 150 
 940 
 330 
 
 76 
 346 
 956 
 
 166 
 
 195 
 
 386 
 
 16 
 
 106 
 
 10.30 
 83.60 
 85.77 
 89.90 
 
 81.10 
 81.43 
 83.63 
 88.17 
 
 1.07 
 9.03 
 1.67 
 0.17 
 
 90.00 
 96.60 
 94.03 
 19.77 
 
 6.63 
 6.70 
 5.60 
 5.47 
 
 9.17 
 8.53 
 6.83 
 7.80 
 
 a 67 
 0.83 
 0.10 
 a 87 
 
 9.73 
 
 0.97 
 9.07 
 4.90 
 
 One of the most obvioua couclasioiis, from the above table, is that the two ciroleo do not 
 give the same result for the distance the telescope has moved. If they did, the sum of the 
 eight microscope readings would be constant for each series and each set of circle divisions. 
 In the first set of the first series, for instance, there is a difierence of 1".45 after the circle has 
 been turned 180°. One or both circles are therefore affected with a quite sensible flexure. 
 
 The above readings were corrected for inequality of screw, and the flexare coefiicients 
 were then computed from the formula) already given. We give an example of the form of 
 computation adopted, by presenting in full so much of the computation as relates to the four 
 cardinal divisions. 
 
 8EBIES I. 
 
 ■» 
 
 
 
 
 Zo=3160; a=90o o'=0° 
 
 
 
 
 
 
 
 
 
 z 
 
 
 
 
 
 
 9K.15 
 
 9K.I6 
 
 9K.96 
 
 8K96 
 
 8a.« 
 
 8/?.« 
 
 e^.zo 
 
 
 
 
 
 
 
 
 
 
 ap.i 
 
 ^9 
 
 9p.b 
 
 9pA 
 
 Ml+S) 
 
 s^d-HJ) 
 
 »>>.(9+5) 
 
 S^-(9+«) 
 
 
 
 
 
 
 
 
 o 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 // 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 // 
 
 // 
 
 316 
 
 98.0 
 
 19.9 
 
 9.1 
 
 19.7 
 
 7.1 
 
 10.7 
 
 99.0 
 
 9.6 
 
 -1.0 
 
 —1.7 
 
 -1.9 
 
 -1.0 
 
 -8.7 
 
 ^8.3 
 
 + 0.8 
 
 46 
 
 39.9 
 
 96.0 
 
 7.8 
 
 10.5 
 
 10.7 
 
 ia4 
 
 33.8 
 
 6.5 
 
 +5.9 
 
 +5.0 
 
 +5.9 
 
 +5.9 
 
 —8.3 
 
 —1.7 
 
 - 0.5 
 
 136 
 
 39.9 
 
 94.3 
 
 5.9 
 
 10.9 
 
 8.1 
 
 13.4 
 
 30.3 
 
 4.5 
 
 ^ ^ 
 
 . 
 
 . 
 
 ^ ^ 
 
 -1.9 
 
 —1.4 
 
 + 0.4 
 
 895 
 
 87.9 
 
 90.3 
 
 7.6 
 
 10.3 
 
 4.8 
 
 7.5 
 
 97.9 
 
 0.6 
 
 • • 
 
 • • 
 
 
 • • 
 
 —9.1 
 
 -1.9 
 
 - 0.3 
 
 SERIES n. 
 
 o 
 
 
 
 
 
 
 
 
 
 9K'.16,«te. 
 
 
 
 %a.h 
 
 -S-5 
 
 // 
 
 315 
 
 36.3 
 
 98.4 
 
 5.0 
 
 8.4 
 
 11.3 
 
 14.7 
 
 a4 
 
 6.8 
 
 ti;; 
 
 +1.5 
 
 +a7 
 
 +1.1 
 
 KS 
 
 --1.3 
 
 0.0 
 
 46 
 
 39.0 
 
 39.1 
 
 4.8 
 
 6.0 
 
 ia8 
 
 16.0 
 
 6.9 
 
 8.1 
 
 +6.1 
 
 -^.8 
 
 —8.8 
 
 -- 0.9 
 
 135 
 
 99.8 
 
 88.5 
 
 9.8 
 
 ia4 
 
 9.6 
 
 ta8 
 
 9.3 
 
 6.9 
 
 
 . 
 
 
 , ^ 
 
 -1.7 
 
 -0.8 
 
 .. 0.6 
 + 1.5 
 
 986 
 
 98.8 
 
 91.4 
 
 10.4 
 
 18.5 
 
 9.9 
 
 11.3 
 
 1.8 
 
 a9 
 
 - • 
 
 • • 
 
 - • 
 
 - - 
 
 +8.9 
 
 —1.0 
 
 '-.%^- 
 
 :.y^gifM^ppm-pt-,-!S-a6=;»,-;it-«Jf-n ■ . WMi«W!«.u»ilWMi««I»l«»iB* 
 
ti 
 
 SO 
 
 
 DiaOBIPTIOIf OP THK TRAMfllT OBOLK OF THE 
 SERIEfl ni. 
 
 z 
 
 Z,r=316°; «»:90O M'saOO 
 
 9K".16.«to. 
 
 Sa'ui 
 
 8^.« 
 
 
 
 
 
 
 
 
 
 
 
 vg\gtvrw)\ 
 
 
 %>.! 
 
 8^.8 
 
 8p.5 
 
 %>.6 
 
 a>,.(l+6)a>..(l+6)9^.(8+6)|9^.(8+6) 
 
 
 
 
 
 
 o 
 
 II 
 
 // 
 
 // 
 
 t/ 
 
 // 
 
 II 
 
 // 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 
 // 
 
 315 
 
 4.6 
 
 84.5 
 
 10.3 
 
 14.0 
 
 14.9 
 
 18.6 
 
 4.8 
 
 8.5 
 
 +1.8 
 
 +1.8 
 
 +1.6 
 
 +1.6 
 
 -0.3 
 
 +10.8 
 +10.1 
 
 
 0.0 
 
 45 
 
 8.3 
 
 99.7 
 
 9.1 
 
 11.7 
 
 11.4 
 
 14.0 
 
 1.6 
 
 4.4 
 
 -6.9 
 
 —5.6 
 
 —6.6 
 
 -6.1 
 
 —8.1 
 
 + 
 
 0.4 
 
 135 
 
 8.7 
 
 98.9 
 
 11.0 
 
 14.7 
 
 13.7 
 
 17.4 
 
 3.8 
 
 6.9 
 
 
 • • 
 
 ^ , 
 
 ^ , 
 
 +0.5 
 +0.3 
 
 +11.6 
 +10.3 
 
 
 0.6 
 
 896 
 
 6.3 
 
 96.3 
 
 19.0 
 
 14.9 
 
 17.3 
 
 19.6 
 
 6.3 
 
 10.5 
 
 • • 
 
 • • 
 
 • • 
 
 • • 
 
 — 
 
 0.8 
 
 
 
 
 
 HERIE8 IV. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 8a'.» 
 
 9fif.h 
 
 
 
 315 
 
 83.1 
 
 19.6 
 
 0.0 
 
 13.8 
 
 38.1 
 
 6.3 
 
 81.6 
 
 96.8 
 
 -0.7 
 
 +0.4 
 
 —0.6 
 
 +0.6 
 
 -10.1 
 
 -0.7 
 
 — 
 
 0.6 
 
 45 
 
 9t.l 
 
 11. a 
 
 * 9.9 
 
 18.4 
 
 30.3 
 
 3.5 
 
 80.5 
 
 83.7 
 
 -4.8 
 
 -4.1 
 
 —4.3 
 
 -4.8 
 
 —10.0 
 
 -9.6 
 
 — 
 
 1.4 
 
 135 
 
 86.5 
 
 15.8 
 
 6.3 
 
 0.4 
 
 39.8 
 
 6.9 
 
 89.1 
 
 85.8 
 
 • • 
 
 • • 
 
 ^ ^ 
 
 ^ _ 
 
 —10.5 
 
 —0.5 
 
 + 
 
 0.8 
 
 885 
 
 88.8 
 
 19.1 
 
 6.7 
 
 6.8 
 
 34.5 
 
 7.6 
 
 94.8 
 
 87.9 
 
 • • 
 
 • • 
 
 • • 
 
 • • 
 
 — 9.8-0.7 
 
 
 0.0 
 
 The oompntation for the other divisions was performed iu the same way. The complete 
 results for every 15° are given in the following table, which is so arranged that the two coeffi* 
 cients which correspond to the same pqsition of the circle are under each other. For this 
 purpose some of the angles are changed by 180°, and the signs of the coefficients are changed 
 to correspond. 
 
 CIBCLE A. 
 
 «=s 
 
 OO 
 
 16° 
 
 30° 
 
 45° 
 
 600 
 
 76° 
 
 90O 
 
 106O 
 
 190" 
 
 1360 
 
 160° 
 
 1660 
 
 VkluMofSa^i 
 
 II 
 
 —1.3 
 -0.8 
 —1.7 
 -41.8 
 
 -0.19 
 
 II 
 —4.0 
 —3.9 
 —3.1 
 -A8 
 
 —0.41 
 
 -3.8 
 —3.3 
 -8.4 
 -8.6 
 
 -0.38 
 
 -0.9 
 —0.8 
 -9.9 
 —3.1 
 
 -0.84 
 
 // 
 
 -9.4 
 —8.5 
 —3.8 
 —3.3 
 
 —0.38 
 
 II 
 
 -9.7 
 -«.8 
 -9.6 
 -9.7 
 
 —0.34 
 
 II 
 
 —9.7 
 -8.3 
 —1.9 
 -8.1 
 
 — o.8e 
 
 II 
 —0.7 
 —9.1 
 —1.6 
 —9.1 
 
 —0.90 
 
 II 
 
 -1.8 
 
 0.0 
 
 —0.6 
 
 —1.4 
 
 -0.10 
 
 1/ 
 
 -1.9 
 —0.9 
 
 -0.01 
 
 II 
 
 +0.9 
 +0.6 
 +1.3 
 +0.8 
 
 +0.11 
 
 II 
 
 +1.6 
 +0.8 
 -.0.9 
 +1.7 
 
 +0.16 
 
 — 
 
 90O 
 
 106O 
 
 180° 
 
 136° 
 
 160O 
 
 1660 
 
 180O 
 
 1950 
 
 910O 
 
 9800 
 
 940O 
 
 966° 
 
 f 
 
 YalnnofS^.* 
 
 // 
 -9.9 
 
 —1.9 
 —0.89 
 
 II 
 
 -3.1 
 ^2.3 
 —3.0 
 —9.7 
 
 -0.35 
 
 II 
 
 -3.1 
 —8.6 
 —8.9 
 -3.1 
 
 -0.37 
 
 // 
 
 -8.6 
 —8.6 
 
 -8.8 
 -8.9 
 
 — 0.3« 
 
 II 
 
 -3.3 
 —8.4 
 
 — 4. i 
 —8.6 
 
 — 0.«li 
 
 // 
 ».'3.0 
 
 — ».e 
 
 ™3.f' 
 — •./.38 
 
 It 
 
 -3.9 
 -9.8 
 -0.8 
 —1.0 
 
 -0.84 
 
 II 
 
 —0.8 
 —1.9 
 —1.1 
 —1.7 
 
 —0.16 
 
 —0.7 
 
 +0.5 
 
 0.0 
 
 -0.9 
 
 -0.03 
 
 II II 
 
 +0.8 +0.6 
 +1.8 +0.9 
 -0.6 +1-3 
 —0.3 -^.8 
 
 +0.03 +0.09 
 
 II 
 +9.0 
 +1.3 
 +0.6 
 +1.4 
 
 +0.17 
 
 CIBCLE B. 
 
 •^ 
 
 0° 
 
 le'' 
 
 30° 
 
 4^ 
 
 60O 
 
 760 
 
 90° 
 
 106O 
 
 190" 
 
 1360 
 
 160O 
 
 1860 
 
 
 II 
 
 II 
 
 II 
 
 // 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 II 
 
 Valne«of8«'.« J 
 
 -0.3 
 —8.1 
 +0.6 
 +0.3 
 
 -9.8 
 -9.8 
 —9.1 
 —9.1 
 
 -4.1 
 
 -4.8 
 —5.8 
 -6.6 
 
 —4.7 
 -6.7 
 -6.7 
 -0.9 
 
 —8.4 
 -8.0 
 —7.0 
 —7.8 
 
 -0.1 
 
 -8.8 
 -8.9 
 -8.7 
 
 -10.1 
 —10.0 
 -10.5 
 — 9.9 
 
 —9.7 
 -0.9 
 -9.1 
 
 —8.6 
 
 —8.9 
 —7.9 
 -8.6 
 -8.6 
 
 -8.9 
 —7.7 
 -6.8 
 -6.3 
 
 -4.8 
 -.4.8 
 -f6.& 
 -6.5 
 
 -8.6 
 —1.8 
 —1.7 
 -1.9 
 
 a'.«=r 
 
 —0.06 
 
 —0.97 
 
 -^.63 
 
 -0.77 
 
 -0 98 
 
 -1.09 
 
 -1.84 
 
 —1.16 
 
 —1.04 
 
 -0.89 
 
 -0.64 
 
 -0.99 
 
UmraD ITATBS MATAL 0B8BBTAT0BT. 
 
 n 
 
 OIBOLE B.— OonMniMd. 
 
 ■ SB 
 
 90° 
 
 106O 
 
 180° 
 
 135° 
 
 1600 
 
 166° 
 
 160O 
 
 I960 
 
 810° 
 
 886° 
 
 840» 
 
 885'> 
 
 ViivmolS^.* 
 
 II II 
 
 -0.7 -1.9 
 -8.6 -1.9 
 -0.6 —1.8 
 —0.7 —1.8 
 
 -0.14 ~«.83 
 
 ./ II 
 
 -4.8 -4.1 
 —4.9 -6.1 
 -6.0 -6.4 
 -&.1 -6.6 
 
 -0.60 -0.79 
 
 II 
 
 -8.8 
 —8.4 
 
 -ai 
 
 -8.9 
 —1.07 
 
 // 
 
 -9.3 
 -9.0 
 -S.9 
 -9.4 
 
 —1.14 
 
 II II 
 
 -10.9 - 9.6 
 -10.1 - 9.7 
 —11.6 —10.4 
 -10.3 — 9.8 
 
 - 1.38— 1.83 
 
 II 
 
 -8.4 
 -8.1 
 -8.7 
 -«.7 
 
 —1.06 
 
 II II 
 
 -«.0 —5.1 
 —7.5 -5.1 
 -6.0 —3.8 
 -6.0 —3.7 
 
 —0.86 —0.66 
 
 -8.7 
 -8.0 
 -8.7 
 -9.8 
 
 -0.30 
 
 (61) It will be remembered that a represents the excess of reading of microsoopea Y-YII, 
 when a division of circle A is brought under microscope Y, and ^ the excess of YI-YIII, when 
 a division is brought under microscope YI. Also, when the division a is under mioroncope Y, 
 the division a-f90° is under microscope YI. The same remarks hold true for circle B, by 
 diminishing the number of the microscope hy lY. Now, comparing the values of a with those 
 of p immediately under them, it will be seen that there is generally a quite dose agreement, 
 the difference amounting to one*tenth of a second in only one case out of the tw«nty-four, and 
 the mean difference being less than 0".05. If we suppose, as seems probable, that these differ* 
 ences are no greater than the unavoidable errors of the determinations, we arrive at tbe con- 
 clusion : 
 
 Hie geometriced form of the oirde rdalive to any ayatem of Jhced aae» rem^ii'\n invariable a» 
 U revolve$. 
 
 The large values of a and fi show that if the central part of the cirde revoi untforndy, the 
 drcutnftrmoe don not revolve tm^ormly, ha is qfected with a periodic inequality 
 
 I am disposed to attribute this singular phenomena to a slight deviation of the centre of 
 gravity of the circle from its centre of figure. The circles weigh about 80 pounds each, and 
 a weight of a few ounces on their circumference is sufficient to produce a flexure of 1". But 
 to whatever cause we attribute it, the circumstance of invariability of form of the circle involves 
 the law that the flexure shall be of the form of a sin x+b cos s. If, then, we suppose 
 
 a a A sin (D +460)+B cos (D + 4fiO). 
 /9 sA sin (D -460)+B COS (D -450),' 
 o'sbA' sin (D'-f 460)+B' cos (D'+460), 
 /ysA' sin (iy-460)+B' cos (D'-4fiO). 
 
 D being the circle division, we flnd by equating the preceding values of a and ^, and solving 
 
 by least squares, 
 
 A=-0".37; B«i+0".01} 
 
 A'—0".84; B'=!+0".86. 
 
 (62) The outstanding apparent errors are seen in the following table, which includes the 
 combined errors of the two hypothesis; first, that the geometrical form of the circle remains 
 invariable, in other words, that 
 
 «ua=/J.(a+90O). 
 
 Second, that a and are each of the form, 
 
 AsinD+BeosD. 
 
 The first column of the table gives the reading of the finding microscope, which, for circle 
 A, is midway between Y and YI, and for circle B, midway between I and II. 
 
 The second gives the flexure for that position of the circle as computed by the formnle, 
 which is assumed to be the same for each pair of microscopes. 
 
 ^jiEjsS«i»saM».;i5s%5!S: 
 
I 
 
 h 
 
 t fl- 
 
 am- DESCBIPTION OF THE TRANSIT CIBCLE OF THE 
 
 The third gives the observed flexnre of the mean of microscopes VI- VII, or I-III, for that 
 position of the circle, in other words, the value of a.(R— 45°.) 
 
 The fourth gives the observed flexure of the mean of microscopes VI-VIII, or II-IV, for 
 the same position of the circle, or the value of ^.(R+45°.) 
 
 The lifth and sixth give the outstanding errors. 
 
 R 
 
 Flexure 
 formula. 
 
 ObBorved flexure. 
 
 Errors, 
 
 R' 
 
 Flexure 
 
 by 
 formula. 
 
 Obserred flexure. 
 
 £rron. 
 
 V-VII. 
 
 VI-VIII. 
 
 V-VII, 
 
 VI-VIII. 
 
 i-m. 
 
 II-IV, 
 
 i-in. 
 
 II-IV. 
 
 o 
 
 46 
 
 60 
 
 75 
 
 90 
 
 105 
 
 120 
 
 135 
 
 150 
 
 165 
 
 180 
 
 195 
 
 210 
 
 225 
 
 // 
 
 —0.25 
 —0.31 
 -0.36 
 -0.37 
 
 — o.:}6 
 
 —0,32 
 -0,27 
 —0,19 
 -0,11 
 —0.01 
 4-0. 09 
 +0,18 
 +0.25 
 
 // 
 
 —0.19 
 —0.41 
 —0,38 
 —0,24 
 
 — o.:« 
 
 —0,34 
 —0.28 
 —0.20 
 —0.10 
 —0,01 
 +0.11 
 +0. 15 
 +0.16 
 
 // 
 
 —0.22 
 —0.35 
 —0.37 
 —0.34 
 —0.45 
 —0.38 
 —0.24 
 —0.16 
 —0.03 
 +0.03 
 +0.1J9 
 +C. 17 
 +0,19 
 
 // 
 
 +0,06 
 —0,10 
 —0,02 
 +0. 13 
 —0,02 
 -0,02 
 —0,01 
 —0.01 
 +0.01 
 0.00 
 +0.02 
 —0.03 
 —0.09 
 
 // 
 
 +0.03 
 —0.04 
 —0,01 
 +0,03 
 —0,09 
 —0.06 
 +0,03 
 +0.03 
 +0.08 
 +0.04 
 0.00 
 —0,01 
 —0.06 
 
 o 
 
 45 
 
 60 
 
 75 
 
 90 
 
 1(» 
 
 120 
 
 135 
 
 150 
 
 165 
 
 180 
 
 195 
 
 210 
 
 225 
 
 II 
 
 +0,02 
 —0,30 
 —0.69 
 —0.84 
 -1.03 
 —1.16 
 —1.20 
 —1.16 
 —1.05 
 —0.86 
 -0.61 
 -0,32 
 —0.02 
 
 —0.05 
 —0.27 
 —0.63 
 —0,77 
 -0.98 
 —1.09 
 —1,24 
 —1.16 
 —1.04 
 -0,89 
 —0,64 
 -0,22 
 —0,01 
 
 —0.14 
 —0,23 
 —0.60 
 —0.72 
 -0.07 
 —1.14 
 —1.32 
 —1.23 
 —1.06 
 -0,86 
 —0.55 
 —0.30 
 +0.14 
 
 // 
 
 —0.07 
 +0.03 
 —0.04 
 +0.07 
 +0.05 
 +0.07 
 —0.04 
 0.00 
 +0.01 
 —0,03 
 —0.03 
 +0.10 
 +0.01 
 
 -0,16 
 +0.07 
 —0,01 
 +0.12 
 +0.04 
 +0.02 
 —0.12 
 —0.07 
 —0,01 
 0.00 
 +0,06 
 +0,02 
 +0.15 
 
 The lower line of the table is, it will be seen, only a repetition of the upper with the 
 sign changed. We always have 
 
 V w . , . /R+/(R+180O)=0 
 
 by the fundamental hypothesis of the investigation. 
 
 (63) Values of g z. — Taking the differences of the K's, we have eight distinct values of 
 8 g.z, the sum of which gives the value of 64 gji, derived from the observation. The mean is 
 as follows: 
 
 
 
 +.01 
 
 16 
 
 +0.14 
 
 30 
 
 +0.07 
 
 46 
 
 -0,02 
 
 60 
 
 +0.08 
 
 75 
 
 +0.10 
 
 90 
 
 +0.C8 
 
 105 
 
 +0,06 
 
 120 
 
 +0.02 
 
 136 
 
 -0,04 
 
 150 
 
 + 0.02 
 
 165 
 
 —0.02 
 
 Though these values are quite well marked, indicating a twisting flexure coefficient of 
 0".06, I am not at all satisfied of their reality, and have therefore preferred to dispense with 
 their use, and derive the telescope flexure for each end of the axis directly from the observations. 
 
 (64) Flexure of the Telescope. — The preceding investigation gives the flexure of the circle 
 divisions relatively to the central nucleus of the circle. We next wish to know the flexure of 
 the line joining the micrometer wire, and the optical centre of the object glass relatively to 
 the same nucleus. During the early part of the year 1866, I was greatly trooble4 by finding 
 a constant difference of a large fraction of a second between the horizontal flexnre determined 
 from the opposing and that from the levelled collimators. It seemed to follow from this, that 
 if the axes of the collimators were so*^^ optically in the same line, tbe difference of their level 
 
 H 
 
 mm 
 
UNITEU STATES NAVAL OB8EBVATORY. 
 
 88 
 
 9 mean is 
 
 errors would not be equal to their d fference nf latitude, m it should be. On April 17, 1860. 
 this was tested directly, in the following way: The error both of level and coUitnation of each 
 collimator was reduced as much as possible. The telescope was set vertically, the cube opened, 
 and collimator A turned till its double wires were horizontal. The shutters were closed, as 
 usual, in observing the collimators. Three observers were employed; one to read the level 
 of each collimator, and one to make the images of the horizontal wires coincide. The latter 
 looking into collimator B, set the wires opposite in each of the four combinations of position of 
 the collimators, A 90°, B 180°; A 90°, B 0°; A 270°, B 0°; B 180° and in each combinatitm a 
 set of level readings of each collimator was taken. To eliminate any possible personal error 
 in setting, the observer then went to collimator A, and the operation was repeated. The result 
 
 was as follows: 
 
 Mean level of A, (North,) p".87, (8. end high.) . 
 
 " " B, (SoQth.) .29, " " • 
 
 Dififorence - - • 
 Oorr. for diff. of collars, 
 Gorr. for diff. of latitude, 
 
 +0" 58 8. 
 +0 .26 
 +0 .25 
 
 Sum 1".09 ,..,:>..-.„■:.,-' ^ _;■ 
 
 This sum ought to be zero, so that there is a seeming discrepancy of 1".09. That this is 
 due to refraction, I entertain no doubt, for the following reasons: (I.) A p-iori; an increase 
 of temperature amounting to 1** Fahrenheit in two feet, will entirely account for it; and the 
 actual increase frona the floor to the roof is found to exceed this on a sunny day. (2.) A 
 few days afterward the observations were partly repeated with the shutters open, and a cold 
 wind blowing through the room. The discrepancy was 0".61 in the opposite direction. The 
 images were quite unsteady, and the wind troublesome. 
 
 (66) The flexure by the opposing collimators was determined by setting the telescope on 
 one collimator, and reading the telescope and microscope micrometers. The telescope was then 
 pointed upward, and the horizontal wires of the other collimator set on those of the first. The 
 circle reading was then determined for the other collimator, and the telescope again pointed to 
 the zenith. The first collimator was then set independently on the second, and the two colli, 
 mators were thus alternately set and read as often as was deemed advisable. The following 
 are the separate results obtained on different dates: 
 
 1865. Dec. 16, /=4-0".15; 
 
 1866. Mar. 29, 
 
 Apiil 16, 
 
 26, 
 
 May 31, 
 
 Jane 9, 
 
 +0 .83, 
 
 + 1 
 +0 
 -♦-0 
 +0 
 
 .42, 
 
 /'=0".77; 
 1 .30; 
 
 .71, wt. =1 ; 
 .89. wt. =2; 
 .49, wt. =:1. 
 
 After the first determination the screws of the object end of the telescope tube were tight- 
 ened. The two next v,ere made without suspecting that the results might be vitiated by re- 
 fraction, and therefore without attention to the equality of temperature in the different strata 
 of air. They are therafore rejected. The last three were made with the shutters open, at 
 times when the int'jraal and external temperatures were nearly equal. That of May 31 was 
 particularly satisfactory, and depends on four readings of one collimator, and three oi the other, 
 the separate readings being 
 
 N. 
 48".39 
 
 47 .81 
 
 48 .12 
 47 .87 
 
 B. 
 50".77 
 
 50 .87 
 
 60 .09 
 
 mmnm i m mm 
 
 fcVti*a!»^,Uiit'WPrt^'ailK«»«S- It'^'V^'ftW 
 

 'M DESCBIPTION OF THE TRANSI1' CIRCLE OK THE 
 
 The valae of /, concluded from obsttrvHtions of the opposing collimatora. is 
 
 0".76. 
 
 In the beginning of 1867 the object glass was taken out and cleaned. Gonceiviug a 
 change in the elasticity of its bearings possible, a careful determination of/' was made on 
 September 9, 1867. The circumstance taken advantage of to secure equality of temperature 
 was a cold ruin. Two thermometers were fastened to the stairway below the line from the 
 object glass of the telescope to each collimator, and two more were suspended just under the 
 roof. The upper pair indicated a higher temperature of 1°.2 before tho observations, and 2°.2 
 afterward. The separate readings uncorrected for circle flexure were: 
 
 SonthcoU. North coll. 
 
 fi'=90O B'b270O 
 • 14".69 lfi".13 
 
 15. 10 15. 18 
 14. 97 16. 26 
 
 16, 43 16. 67 
 .,^^ 16.61 
 
 The readings were commenced on the south collimator. On looking into it to set it on the 
 north one, preparatory to its second reading, the mean of wires were seen to differ quite sensibly 
 from coincidence with th«i wire of the other collimator. The first reading is therefore regarded 
 as doubtful. The result of these readings is 
 
 /'=+0".78. 
 
 During the autumn of 1866 the levelled collimators were regularly observed at night with 
 the shutters open, so that the mean result ought to be free from refraction. The result, from 
 observations made by Messrs. Hall and Rogers and myself, was: 
 
 Mean excess of reading for S. collimator, 1".86 ; 
 , > t Uncorrected flexure coefficient - - 
 
 , u^ - Correction for difference of latitude 
 
 1, *i «rv difference of pivots - 
 
 circle of flexure • - 
 Resulting coefficient . . - . . 
 
 o there is a discrepancy of more than half a second between the flexure coefficients found by 
 the two methods. The error is probably in that determined from the levelled collimators, the 
 conical character of their shoulders rendering their results uncertain. 
 
 The discrepancy is so great that I think it best to try alio the method of comparison of 
 direct and reflection observations. 
 
 (66) Fertiocd Flexure. — Thus fiir, the coefficient of cos Z has been found only by the method 
 already set forth, namely, by comparison of the nadir reading obtained from observations of the 
 collimators, and that obtained directly by coinoidenoe of the wires with their images reflected 
 from mercury . The observations were so conducted as to completely eliminate every constant 
 error of the collimator itself, the following b^ing the usual order: 
 
 (1) Nadir; 
 
 (2) OoUimator B (north); 
 
 (3) OoltimatorA(sooth); ' ' ' 
 
 (4) OoUimator B (soadi); 
 (6) OoUimator A (north); 
 (6) Nadir. 
 
 
 
 .93; 
 
 -0 
 
 .12; 
 
 — 
 
 .23; 
 
 -0 
 
 .37; 
 
 - +C ,21. 
 
 •Jl 
 
UNITED STATES MAVAL OBSSRVATORT. 99 
 
 If. now. eitlier collimatoi' be nifected with any conntHnt cnnse of error wlion on one »ido of 
 tlie inatrament, lliat canoe will act in the opposite direction on the circle reading when t' o col- 
 limator is carried to the other side. By interchanging the collimators we therefore eliminate 
 all constant errors pecniiar to them. 
 
 The following are the separate results obttined in this way: 
 
 
 Co 
 
 C-g 
 
 t 
 
 Wt. 
 
 Co 
 
 C\-f 
 
 
 m. 
 
 1865. 
 
 II 
 
 II 
 
 II 
 
 
 t/ 
 
 tl 
 
 
 Dee. 39 
 
 13. S9 
 
 13.76 
 
 —a 54 
 
 8 
 
 16.35 
 
 15.96 
 
 +0.39 
 
 3 
 
 30 
 
 88.34 
 
 88.44 
 
 -0.10 
 
 a 
 
 33.61 
 
 33.61 
 
 0.00 
 
 3 
 
 1886. 
 
 
 
 
 
 
 
 
 
 Jan. 30 
 
 88.34 
 
 88.48 
 
 —0.14 
 
 3 
 
 68.88 
 
 68.78 
 
 --0.04 
 
 3 
 
 Apr. 7 
 
 84.87 
 
 84.96 
 
 —0.09 
 
 3 
 
 18! 43 
 
 14.88 
 
 •■0.04 
 - .0. 11 
 
 3 
 
 16 
 
 19.80 
 
 80.00 
 
 —0.80 
 
 3 
 
 18.38 
 
 .3 
 
 17 
 
 17. « 
 
 17.40 
 
 +0.88 
 
 3 
 
 16.50 
 
 16.66 
 
 —0.16 
 
 3 
 
 . J8 
 
 1 
 
 11.68 
 
 11.88 
 
 -0.88 
 
 3 
 
 16.44 
 
 16.17 
 
 +0.87 
 
 3 
 
 From which results 
 
 g =:-0".U, 
 
 ^=+0".09. 
 
 There must always be a possibility of the nadir determinations being affected with undis- 
 coverable sources of error, depending either upon the habits of the observer, or the disturbing 
 conditions to which the instrument may be subjected, as, for example, the heat of the observer's 
 body. I think it best, therefore, to depend for the final value of gr upon the comparison of ob- 
 servations made in reversed positions of the instrument, the effect of the cosine flexure being 
 reversed with the instrument. For the present, therefore, the quantity — 0".14 is regarded 
 simply as the reduction of an observed nadir reading of circle A to the mean of the horizontal 
 readings. 
 
 ' (67) In the flexure of the telescope is included the effect of gravity in changing the posi- 
 tion of the declination micrometer slide relatively to the fixed plates of the eye-piece. As the 
 telescope turns, the reading of the- micrometer for coincidence of the fixed and movable wires 
 is affected with the inequality 
 
 — 0r.0876 sfa Z— 0r.0197 cos Z, 
 
 Z being the zenith distance of the telescope counted in such a direction that sin Z is positive 
 when the micrometer head is above the screw, and negative when below it. 
 
 The flexures alretidy found being corrected for this inequality, the value of the sine coef- 
 ficient would be quite small, while that of the cosine coefficient would be increased to 0".44. 
 
 (68) In observing tSe sun, the aperture of the telescope is diminished to about three 
 inches by means of a cap weighing 5.3 ounces. It is found, by experiment, that this weight 
 causes a flexure of — O'MO fet'n Z. A further flexure correction of 
 
 +0M08inZ 
 
 is therefore required in reducing observations of the sun. 
 
 »»»<««f'ffi-?'*!W,'«B!(S*SW*-M«'^«fl'^»«**w^»'»" 
 
 ■<^^U<-v4 JJiil! ■ immisiihW'" 
 
S6 
 
 L*R8CRIPriON OF THi'. TBAN8IT CIRCLE OF THE 
 
 , ERRORS OF DIVISION. 
 
 (G9) The readings Cor errors of diviaion of every 5° were made during November and 
 December, 18()5, before the cummenceraent of astronomical observations with the instrument. 
 The observers, beside mys«!lf, were Professors Hall and Eastman, and aides Rogers and Thirion. 
 Each observer rend two microscopes. Any personal error in reading will appear only in the 
 distance of the microsfopes. 
 
 A few readings were tirst taken on the 0° and 90° divisions, with the microscopes 90° 
 apart, to determine their angle. This, however, is of little importance, since any error, in its 
 value, is eliminated in the mean of four microscopes, the number always read in astronomical 
 observation. 
 
 To determine the error of every 45°, the microscopes were set 45° apart, and both circles 
 were read thirteen times in each of the eight positions. Assuming the error of 0° to he zero, 
 the following are the resulting values of 4e, or the negative of four times the correction for 
 error of division for every 45°: 
 
 * , V''; Circle A. Circle B. 
 
 
 
 0.00 
 
 II 
 ... 0.00 
 
 46 
 
 -1 19 
 
 —2 87 
 
 90 
 
 -1.12 
 
 —0.80 
 
 130 
 
 -0.67 
 
 -0.69 
 
 To find the errors of every 15°, two determinations were made: the first being made with 
 the microscopes 60° apart, the second with the microscopes 75° apart. The circle was read 
 five times in each position in each series. For circle A another and more exact series was 
 made with microscopes 75° apart. 
 
 The tirst series could not give an independent determination of the 45° spaces, but the 
 latter did, and small corrections were applied to them accordingly; not, however, with exact 
 reference to the forniulsB for weights already given. The following are the results: 
 
 
 
 Circle ▲. 
 
 
 
 Circle B. 
 
 Mic.60. 
 
 76° (l«t.) 
 
 75° (8d.) 
 
 Concluded. 
 
 60°. 
 
 75°. 
 
 Concluded. 
 
 o 
 
 II 
 
 It 
 
 tl 
 
 
 // 
 
 // 
 
 II 
 
 II 
 
 
 
 0.00 
 
 0.00 
 
 0.00 
 
 
 0.00 
 
 0.00 
 
 0.00 
 
 0.00 
 
 15 
 
 — 0.02 
 
 - 0.81 
 
 — 0.18 
 
 
 
 0.18 
 
 - 1.73 
 
 — 8.58 
 
 — 8.13 
 
 30 
 
 + 0.14 
 
 + 0.38 
 
 + 0.36 
 
 + 
 
 0.31 
 
 — 1.07 
 
 — 1.48 
 
 — 1.86 
 
 46 
 
 — 1.19 
 
 — 1.37 
 
 — 1.11 
 
 
 1.19 
 
 — 8.93 
 
 L- 8.93 
 
 — 8.93 
 
 60 
 
 — 8.89 
 
 — 8.81 
 
 — 8.88 
 
 
 
 8.86 
 
 - 3.91 
 
 — 3.88 
 
 — 3.90 
 
 76 
 
 — 1.86 
 
 — 1.79 
 
 — 8.05 
 
 
 
 1.79 
 
 - 1.51 
 
 — 0.99 
 
 - 1.86 
 
 90 
 
 — 1.18 
 
 — 1.13 
 
 - 1.18 
 
 .^ 
 
 1.18 
 
 — 0.80 
 
 — 0.80 
 
 — 0.80 
 
 106 
 
 — 0.68 
 
 + 0.03 
 
 — 0.08 
 
 _ 
 
 0.13 
 
 - 1.11 
 
 — 1.07 
 
 — 1.09 
 
 180 
 
 — 1.14 
 
 r- 8.19 
 
 — 1.11 
 
 
 
 1.39 
 
 - 1.75 
 
 h- 8.3S 
 
 — 8.03 
 
 136 
 
 — 0.67 
 
 — 0.76 
 
 — 0.68 
 
 _ 
 
 0.07 
 
 - 0.43 
 
 — 0.43 
 
 — 0.43 
 
 180 
 
 - 0.60 
 
 - 1.43 
 
 — 8.88 
 
 ,m^ 
 
 1.68 
 
 — 0.94 
 
 — 1.43 
 
 — 1.18 
 
 166 
 
 + 1.16 
 
 + 1.16 
 
 ^ 0.90 
 
 + 
 
 1.08 
 
 — 0.13 
 
 — 0.86 
 
 — 0.19 
 
 For the errors of every 5° two series of readings were made; one with a distance of 50°, 
 the other with one of 55°. Three readings wore made in each po.»ition of the circle in each 
 series, except the second series of circle B, when only two readings were made. The following 
 are the separate results. The lust column in each table gives one-fourth the negative of the 
 mean by weights of the two preceding columns, and is the correction to be applied to the mean 
 of opposite microscopes on account of errors of division: 
 
UNITED STATES NAVAL 0B8BRVAT0BT. 
 
 m 
 
 1 
 
 Circle A. 
 
 
 
 Circle B 
 
 
 
 60°. 
 
 66°. 
 
 Concluded 
 — e. 
 
 60". 
 
 65°. 
 
 Conclnded 
 
 o 
 
 
 // 
 
 
 // 
 
 •/ 
 
 
 // 
 
 // 
 
 
 ti 
 
 
 
 
 0.00 
 
 
 0.00 
 
 0.00 
 
 
 0.00 
 
 0.00 
 
 
 0.00 
 
 5 
 
 + 
 
 1.44 
 
 + 
 
 0.89 
 
 — 0.88 
 
 — 
 
 0.86 
 
 — 1.06 
 
 + 
 
 0.34 
 
 10 
 
 + 
 
 1.55 
 
 + 
 
 1.39 
 
 + 0.37 
 
 — 
 
 0.99 
 
 — 0.76 
 
 + 
 
 0.88 
 
 15 
 
 
 0.17 
 
 
 0.18 
 
 + 0.04 
 + 0.86 
 
 _ 
 
 3.08 
 
 — 8.13 
 
 
 0.58 
 
 80 
 
 — 
 
 0.93 
 
 — . 
 
 1.10 
 
 — 
 
 1.96 
 
 — 1.91 
 
 0.48 
 
 86 
 
 — 
 
 1.56 
 
 — 
 
 8.05 
 
 + 0.45 
 
 — 
 
 1.88 
 
 — 8.03 
 
 + 
 
 0.48 
 
 »0 
 
 + 
 
 0.34 
 
 + 
 
 0.3i 
 
 - 0.08 
 
 
 
 1.88 
 
 — 1.33 
 
 + 
 
 0.33 
 
 35 
 
 
 0.64 
 
 
 0.96 
 
 + 0.80 
 
 — 
 
 3.77 
 
 — 3.07 
 
 + 
 
 0.78 
 
 40 
 
 — 
 
 1.41 
 
 _ 
 
 0.83 
 
 + 0.88 
 
 — 
 
 3.19 
 
 - 8.63 
 
 
 0.74 
 
 45 
 
 _ 
 
 1.19 
 
 - 
 
 1.19 
 
 + 0.30 
 
 — 
 
 3.93 
 
 — 8.93 
 
 
 0.73 
 
 50 
 
 
 
 1.09 
 
 
 
 0.08 
 
 + 0.14 
 
 ,_ 
 
 3.80 
 
 — 3.33 
 
 ■1- 
 
 0.90 
 
 55 
 
 — . 
 
 1.08 
 
 — 
 
 8.89 
 
 4- 0.43 
 
 — 
 
 3.35 
 
 — 3.44 
 
 -\. 
 
 0.85 
 
 60 
 
 _. 
 
 8.88 
 
 — 
 
 8.18 
 
 + 0.55 
 
 
 
 3.98 
 
 — 3.84 
 
 + 
 
 0.97 
 
 66 
 
 _ 
 
 8.80 
 
 _ 
 
 8.38 
 
 + 0.56 
 
 — 
 
 3.47 
 
 — 4.06 
 
 + 
 
 0.9H 
 
 70 
 
 — 
 
 0.88 
 
 
 
 0.36 
 
 + 0.16 
 
 — 
 
 3.83 
 
 — 3.93 
 
 t 
 
 0.88 
 
 78 
 
 _ 
 
 1.79 
 
 
 
 1.70 
 
 + 0.44 
 
 
 
 1.87 
 
 — 1.31 
 
 0.38 
 
 80 
 
 + 
 
 1.79 
 
 + 
 
 8.85 
 
 — 0.51 
 
 _. 
 
 O.I"8 
 
 — 0.96 
 
 + 
 
 0.13 
 
 85 
 
 
 0.66 
 
 
 0.38 
 
 ■+■ 0.13 
 ■f 0.88 
 
 + 
 
 0.84 
 
 — 0.91 
 
 + 
 
 0.06 
 
 90 
 
 — 
 
 1.18 
 
 _ 
 
 1.18 
 
 
 0.80 
 
 — 0.80 
 
 + 
 
 0.30 
 
 95 
 
 
 
 8.05 
 
 
 
 1.78 
 
 + 0.48 
 
 — 
 
 1.34 
 
 — 1.46 
 
 + 
 
 0.35 
 
 100 
 
 — 
 
 8.84 
 
 — 
 
 1.36 
 
 + 0.45 
 
 — 
 
 1.81 
 
 — 0.88 
 
 + 
 
 0.36 
 
 105 
 
 — 
 
 0.88 
 
 — . 
 
 0.13 
 
 + 0.04 
 
 — 
 
 1.06 
 
 — 1.04 
 
 + 
 
 0.86 
 
 no 
 
 + 
 
 0.71 
 
 + 
 
 0.09 
 
 — 0.10 
 
 + 
 
 0.88 
 
 — 0.89 
 
 
 0.01 
 
 115 
 
 
 1.09 
 
 
 1.04 
 
 + 0.87 
 
 
 1.33 
 
 -1.30 
 
 
 0.3:1 
 
 180 
 
 — 
 
 1.51 
 
 — 
 
 1.43 
 
 + 0.36 
 
 ' — 
 
 8.07 
 
 — 1.96 
 
 0.51 
 
 185 
 
 _ 
 
 8.86 
 
 — 
 
 0.90 
 
 4- 0.40 
 -j- 0.48 
 
 
 
 1.10 
 
 — 1.81 
 
 
 0.34 
 
 130 
 
 — 
 
 1.79 
 
 _ 
 
 8.07 
 
 — 
 
 0.84 
 
 — 0.78 
 
 A. 
 
 0.80 
 
 135 
 
 — 
 
 0.63 
 
 — 
 
 0.67 
 
 + 0.16 
 
 — 
 
 0.43 
 
 — 0.43 
 
 + 
 
 0.11 
 
 140 
 
 + 
 
 1.48 
 
 + 
 
 1.51 
 
 — 0.37 
 
 + 
 
 0.56 
 
 — 0.03 
 
 
 0.08 
 
 145 
 
 
 0.87 
 
 
 0.08 
 
 + 0.04 
 
 
 0.16 
 
 + 0.10 
 
 M 
 
 0.08 
 
 150 
 
 — 
 
 1.50 
 
 — . 
 
 1.53 
 
 + 0.38 
 
 — 
 
 1.30 
 
 - 1.81 
 
 ^ 
 
 0.38 
 
 155 
 
 
 
 0.40 
 
 
 
 0.44 
 
 + 0.10 
 
 __ 
 
 1.19 
 
 — 1.41 
 
 + 
 
 0.33 
 
 (60 
 
 + 
 
 1.61 
 
 u. 
 
 0.97 
 
 — 0.38 
 
 
 
 0.83 
 
 — 0.44 
 
 + 
 
 0.08 
 
 165 
 
 
 0.97 
 
 
 1.09 
 
 — 0.86 
 
 
 
 0.18 
 
 — 0.80 
 
 + 
 
 0.06 
 
 170 
 
 0.89 
 
 
 0.66 
 
 — 0.19 
 
 — 
 
 0.38 
 
 — 0.11 
 
 
 0.07 
 
 175 
 
 T" 
 
 0.87 
 
 T" 
 
 0.51 
 
 — 0.17 
 
 "■• 
 
 0.88 
 
 — 0.78 
 
 "T" 
 
 0.18 
 
 
 We have here two entirel)- independent determinations of each division error, except 
 those which are multiples of 16°. The mean difference between the two values of 4« is 0". 52 for 
 circle A, and 0".4I for circle B. The difference of accuracy between tlie circles arines from 
 differences in the eye^sight of the observers. It appears then that the probable value of a 
 concluded e is about 0".065 for circle A, and 0".05l for circle B. The probable error of the mean 
 reading of four microscopes, when these divisions are under them, will therefore be about 
 0".046 for circle A, and 0".036 for circle B, a quantity smaller than the probable accidental 
 error of the isolated divisions. Little advantage would therefore be gained by making the, 
 determination more exact. 
 
 (70) When the mean of divisions, 90° apart, is taken, the progression of the errors in the case 
 of circle B are so regular that a determination of the intermediate divisions was supposed to be 
 hardly necessary. But, as circle A was used for all the observations in 1866, it was thought 
 desirable to determine at least some intermediate points on that circle. Accordingly, the micro- 
 scopes of that circle were placed at the distance 43°20', and one complete series of readings were 
 made, which gave the error of every 10°40'. The result ghoived that the errors of the intermediate 
 divisions were systematicaUy greater than those of the 5° divisions, the mean difference being -f-0". 39. 
 
 This was supposed to indicate a cyclical inequality in the error, the period of which was 
 5°. In order to determine it accurately, it was necessary to determine the error of every de- 
 gree of both circles. This was done on September 17-19, 1866. The microscopes of each 
 circle were placed at the distance 48°, and the circles were both read once in each position 
 Again the result was anomalous, and the cyclic hypothesis bad to be modi6ed or abandoned. 
 The systematic mean difference between the 5° divisions and the intermediate even degrees 
 was only 0".14 for circle A, and 0".08 for circle B. 
 
 
w 
 
 f^^ 
 
 38 
 
 DBHCItlPTION OK THE TRANSIT CIRCLK OP THE 
 
 mJi 
 
 A larit Htteiiipt lo discover the f^enernl liiw of the inequality without the laborioiii) operation 
 of deterniiiiiiifj; the error of every 10' wus made in January, 1867. The 6° interval wa» di 
 v'uUn\ into six parts by setting the microscopes of circle 8 44" 10' apart, and this circlw was read 
 on<!o in each position. The cycle now seemed to be reduced to 30'. 
 
 To show the nature of the law the correction for the mean of four microscopes, as it re- 
 sulted from the above determination, is shown for each circle in the following tt»ble: 
 
 _ CIRCLE A. 
 
 
 
 Correction for 
 
 error of diviBion of— 
 
 A 
 
 
 A+ 
 
 A+ 
 
 A+ 
 
 A+ 
 
 A+ 
 
 A+ 
 
 
 A 
 
 o 
 
 O ; : 
 
 o 
 
 o 
 
 ( 
 
 o 
 
 
 
 1 
 
 1 40 
 
 8 
 
 3 
 
 3 20 
 
 4 
 
 
 
 II 
 
 // 
 
 // 
 
 // 
 
 // 
 
 ft 
 
 // 
 
 
 
 --0. 14 
 
 +0.08 
 
 —0.34 —0.10 
 
 -0.05 
 
 -0.84 
 
 +0.07 
 
 5 
 
 - -0. 13 
 
 +0.10 
 
 —0.20 —0.33 
 
 —0.40 
 
 —0.18 
 
 -0.88 
 
 10 
 
 --0.04 
 
 -0.24 
 
 —0.56 +0.25 
 
 —0.22 
 
 —0.23 
 
 -0.41 
 
 15 
 
 --0.04 
 
 +0.15 
 
 -1-0.02 +0.06 
 
 -0.18 
 
 —0.18 
 
 +0.24 
 
 W 
 
 --0.08 
 
 —0.05 
 
 —0.33 +0.14 
 
 +0.17 
 
 +0.15 
 
 -■0.15 
 
 •25 
 
 +0.35 
 - -0. 14 
 
 -0.17 
 
 -0.64 - 
 
 -0.08 
 
 -0.16 
 
 -0.18 
 
 ■-0.08 
 
 :w) 
 
 +0.18 
 
 —0.14 - 
 
 h0.07 
 
 —0.25 
 
 -0.08 
 
 +0.10 
 - -0. 10 
 
 35 
 
 ■■0,30 
 
 - -0. 18 
 --0.26 
 
 -0.14 --0.24 
 
 +0.18 
 -f.0.35 
 
 —0.08 
 
 40 
 
 --0.38 
 
 -0.08 --O.IB 
 
 +0.02 
 
 +0.22 
 
 45 
 
 --0.23 i-»0,06 
 
 -0.44 ;- 
 
 -0.10 
 
 -0.04 
 
 -0.24 
 
 -0.25 
 
 50 
 
 -fl. 12 i 0. 13 
 
 -0.45 --0.02 
 
 +0.08 
 
 —0.30 
 
 +0.18 
 
 56 
 
 -i . 23 04 
 
 -0.34 - 
 
 -0.15 
 
 +0.18 
 -}-0.20 
 
 —0.18 
 
 -H).30 
 
 GO 
 
 -.•0.4S r'..58 
 
 -O.07 - 
 
 -0.18 
 
 --0.10 
 ■■0.04 
 
 -H).08 
 
 65 
 
 U0.33 
 
 +0.15 
 
 —0.12 - 
 
 -0.26 
 
 0.00 
 
 -0.05 
 
 70 
 
 -0.08 
 
 -0.08 
 
 —0.76 ;— 0.18 
 
 -0.16 
 
 —0.78 
 
 -0.16 
 
 75 
 
 i-'.Of 
 
 -0.2(' 
 
 _-fl.48 —0.29 
 
 -0.44 
 
 -0.46 
 
 —0.18 
 
 80 
 
 -fl.3i ''.S* 
 
 •..E5 —0.40 
 
 -0.43 
 
 —0.48 
 
 —0.29 
 
 85 
 
 —0.02 
 
 --0. !2 
 
 -0.39 —0.88 
 
 -0.07 
 
 -0.85 
 
 -H).35 
 
 Mean . . 
 
 +0.13 
 
 +0.02 
 
 —0.33 ; 
 
 0.00 
 
 -0.08 
 
 —0.19 
 
 +0.08 
 
 M.— <)."13 
 
 0.00 
 
 —0.16 
 
 -0.46 —0.13 
 
 -0.81 
 
 —0.32 
 
 -O.U 
 
 CIRCLE B. 
 
 A 
 
 . — 
 
 
 
 Correction for error of division of— 
 
 
 
 
 A+ 
 
 A+ 
 
 A+ 
 
 A+ 
 
 A+ 
 
 A+ 
 
 A+ 
 
 A+ 
 
 A+ 
 
 
 A 
 
 O / 
 
 o 
 
 / 
 
 
 
 O / 
 
 o 
 
 O 1 
 
 o 
 
 / 
 
 
 
 50 
 
 I ♦ 
 
 1 40 
 
 8 
 
 8 30 
 
 3 
 
 3 20 
 
 4 
 
 4 10 
 
 
 
 II 
 
 It 
 
 // 
 
 // 
 
 // 
 
 It 
 
 // 
 
 II 
 
 // 
 
 II 
 
 
 
 -fO.lO 
 
 -0.39 
 
 0.10 
 
 —0.05 
 
 0.18 
 
 -0.09 
 
 0.11 
 
 —0.06 
 
 0.06 
 
 -0.88 
 
 5 
 
 .30 
 
 .04 
 
 — .01 
 
 .31 
 
 — .06 
 
 .08 
 
 .85 
 
 — .06 
 
 .18 
 
 - .01 
 
 10 
 
 .89 
 
 - .09 
 
 .14 
 
 .00 
 
 .16 
 
 .86 
 
 .37 
 
 .24 
 
 .04 
 
 - .05 
 
 15 
 
 .39 
 
 — .01 
 
 .32 
 
 .34 
 
 .31 
 
 .40 
 
 .10 
 
 .14 
 
 .39 
 
 .14 
 
 20 
 
 .24 
 
 .88 
 
 .88 
 
 .10 
 
 .30 
 
 .31 
 
 .38 
 
 .34 
 
 — .09 
 
 .01 
 
 25 
 
 .40 
 
 .80 
 
 .84 
 
 .06 
 
 .88 
 
 .86 
 
 .58 
 
 .34 
 
 .16 
 
 .04 
 
 30 
 
 .42 
 
 .16 
 
 .59 
 
 .36 
 
 .33 
 
 .41 
 
 .36 
 
 .54 
 
 .58 
 
 .31 
 
 35 
 
 .53 
 
 .24 
 
 .37 
 
 .46 
 
 .46 
 
 .31 
 
 .38 
 
 .39 
 
 .04 
 
 .16 
 
 40 
 
 .47 
 
 .20 
 
 .47 
 
 .30 
 
 .40 
 
 .61 
 
 .45 
 
 .58 
 
 .43 
 
 .11 
 
 45 
 
 .42 
 
 .08 
 
 .88 
 
 .01 
 
 .03 
 
 .19 
 
 .10 
 
 .80 
 
 .10 
 
 .30 
 
 50 
 
 .41 
 
 .30 
 
 .85 
 
 .16 
 
 .88 
 
 .85 
 
 .81 
 
 .19 
 
 .88 
 
 .16 
 
 55 
 
 .44 
 
 .16 
 
 .39 
 
 — .04 
 
 .35 
 
 .84 
 
 .15 
 
 .18 
 
 .46 
 
 .89 
 
 60 
 
 .64 
 
 .66 
 
 .82 
 
 .51 
 
 .56 
 
 .64 
 
 .48 
 
 .54 
 
 .30 
 
 .86 
 
 65 
 
 .ei 
 
 .40 
 
 .44 
 
 .48 
 
 .48 
 
 .60 
 
 .58 
 
 .61 
 
 .64 
 
 .89 
 
 70 
 
 .48 
 
 .85 
 
 .53 
 
 .36 
 
 .88 
 
 .39 
 
 .68 
 
 .86 
 
 .37 
 
 .44 
 
 75 
 
 .18 
 
 .80 
 
 .48 
 
 .06 
 
 .88 
 
 .00 
 
 .86 
 
 .11 
 
 .86 
 
 -.04 
 
 80 
 
 .10 
 
 — .01 
 
 — .04 
 
 — .06 
 
 .40 
 
 .80 
 
 .15 
 
 .88 
 
 .88 
 
 — .04 
 
 85 
 
 0.09 
 
 —0.48 
 
 0.08 
 
 -0.88 
 
 -0.08 
 
 -0.89 
 
 -O.06 
 
 0.01 
 
 -0.16 
 
 -0.88 
 
 
 // 
 
 tl 
 
 // 
 
 II 
 
 // 
 
 // 
 
 // 
 
 // 
 
 If 
 
 'ii 
 
 MOMI . . 
 
 0.36 
 
 0.13 
 
 0.31 
 
 0.16 
 
 0.86 
 
 0.87 
 
 0.38 
 
 0.86 
 
 0.23 
 
 0.09 
 
 M.-0'.'36 
 
 0.00 
 
 —0.83 
 
 —0.05 
 
 -0.80 
 
 ~C 10 
 
 -0.00 
 
 —0.08 
 
 —0.10 
 
 —0.13 
 
 -0.8T 
 
UNITED STATES NAVAL OB8BBTATORY. >< flf 
 
 The 8yAtematic irregnlarity is well marked in this table, and HeiMnx to follow the Rnme law 
 in the two circles, except that it is nearly twice hh great for A a» for B. Two important quoH> 
 tions now present themselves respecting the nature of the law. 
 
 1. Do the intermediate errors really depend upon those of the 5° spaces on each side of 
 them more than on any other part of the circle? If not, the attempt to determine the errors 
 with precision might as well be abandoned. Innpection of the above tab^es, however, shows 
 that for circle B, at least, the question is to be answered in the affirmative; the corrections of 
 the intermediate divisions increase and diminish with those of the 5<^ ones, and to about the 
 same extent with the latter. 
 
 2. Do the systematic errors of the 5° spaces peculiar to them affect them alone, or is the 
 law of error continuous? For example, are the systematic errors of the divisions 4° 68' and 
 5° 2' the same as those of 6°, or are they the same as those of 4° 10' and 5° 60'? If the former, ' 
 the spaces on each side of the 6° divisions will be equal; if the latter, they will differ by half a 
 second. Twenty pain, of i itervals of circle B — those adjacent to every 16° from 0° to 136°, 
 and from 180° to 316° — were measured and compared, and the mean difference found to ' e 
 0".07. The law of periodic error, whatever it maybe, is therefore continuous. It seems 
 result, in great part at least, from two cycles in the errors of division; the one having a peric? 
 of 6°, the other a period of 30'. 
 
 (71) The course which it seems best to adopt is this: instead of attempting to determine 
 the errors of division with the last degree of precision, we shall seek to eliminate them by 
 changing the position of the circles from year to year, so that the poxition of any one star will 
 depend on different divisions in different years. If the periodic errors be entirely ne;;lected, 
 the effect of their probable amount, at least in the case of circle B, will be lass than C". I. and 
 the total probable effect of the difference between the actual and the adopted error uf any iso- 
 lated error of division will not much exceed that amount. The uncertainty of the moan of 
 four divisions will, therefore, scarcely exceed the probable error of the declinations of funda- 
 mental stars derived from all the observations hitherto made. 
 
 (72) To obtain a general table of the corrections of the divisions, we have first corrected 
 the determinations given in the preceding tables for periodic error, so as to take the mean of 
 the entire degree divisions, those which are a multiple of 6 excepted, as the standard. This 
 has been effected by applying the following correction;) to the different vertical columns: 
 
 Colomn. Circle A. Circle B. 
 
 o t It It 
 
 A —0.14 —0.19 
 
 A+0 60 + .14 
 
 A4-1 .00 .00 
 
 A+1 40 + .32 -f- .11 
 
 A+2 .00 .00 
 
 3 +2 30 .00 
 
 A+3 .00 .00 
 
 A.+3 20 + .16 .00 
 
 A+4 .00 . .00 
 
 A+4 10 + .18 
 
 The corrections are thus reduced to what they would have been had there been no peri- 
 odic error, and arranged consecutively in a table. The mean difference between consecutive 
 numbers was now found to be 0".160 for circle A, and 0".136 for circle B, indicating a prob- 
 able error of each individual determination, combined with the accidental error of division, of 
 leas than 0".l. 
 
 ; 
 
 ; 
 
 I 
 
 1 
 
40 
 
 Tahit ^ rerreetumt to 
 
 OESCRIFTIOM OP THE TEAMSIT OIROLK OP THE 
 
 (ffjimr mieroieope* fur errort vf divuiom. 
 microwope. 
 
 ArgumefU, mdimg ^ hori*<nUal 
 
 Arg. 
 
 Circle A. 
 
 Circle B. 
 
 Aiv. 
 
 Circle A. 
 
 Circle B. 
 
 o 
 
 II 
 
 (/ 
 
 o 
 
 // 
 
 // 
 
 
 
 +0.18 
 
 +0.30 
 
 45 
 
 +0.08 
 
 -0.01 
 
 1 
 
 + .06 
 
 .83 
 
 46 
 
 .00 
 
 .00 
 
 s 
 
 .00 
 
 .18 
 
 47 
 
 — .08 
 
 .00 
 
 3 
 
 — .06 
 
 .18 
 
 46 
 
 — .04 
 
 .00 
 
 4 
 
 — .11 
 
 .88 
 
 49 
 
 — .04 
 
 + .06 
 
 6 
 
 — .17 
 
 .87 
 
 50 
 
 — .04 
 
 .18 
 
 6 
 
 - .10 
 
 .87 
 
 51 
 
 — .08 
 
 .18 
 
 7 
 
 — .08 
 
 .86 
 
 68 
 
 — .88 
 
 .18 
 
 8 
 
 .00 
 
 .86 
 
 53 
 
 — .80 
 
 .13 
 
 9 
 
 + .04 
 
 .87 
 
 54 
 
 — .16 
 
 .14 
 
 10 
 
 .07 
 
 .89 
 
 55 
 
 — .18 
 
 .15 
 
 11 
 
 .08 
 
 .89 
 
 56 
 
 — .12 
 
 .16 
 
 IS 
 
 .10 
 
 .89 
 
 67 
 
 — .18 
 
 .17 
 
 13 
 
 .13 
 
 .36 
 
 68 
 
 — .18 
 
 .19 
 
 14 
 
 .80 
 
 •<< 
 
 59 
 
 — .18 
 
 .80 
 
 15 
 
 .88 
 
 .68 
 
 60 
 
 — .05 
 
 .88 
 
 16 
 
 .87 
 
 .66 
 
 61 
 
 + .05 
 
 .86 
 
 17 
 
 .84 
 
 .57 
 
 68 
 
 .06 
 
 .86 
 
 18 
 
 .83 
 
 .57 
 
 63 
 
 .05 
 
 .87 
 
 10 
 
 .80 
 
 .58 
 
 64 
 
 .08 
 
 .85 
 
 80 
 
 .16 
 
 .48 
 
 65 
 
 .00 
 
 .88 
 
 ai 
 
 .17 
 
 .49 
 
 66 
 
 .08 
 
 .28 
 
 82 
 
 .16 
 
 .68 
 
 67 
 
 .09 
 
 .88 
 
 8:» 
 
 + .10 
 
 .53 
 
 68 
 
 .15 
 
 .82 
 
 84 
 
 -.08 
 
 .50 
 
 69 
 
 .14 
 
 .83 
 
 85 
 
 — .14 
 
 .46 
 
 70 
 
 + .08 
 
 .84 
 
 86 
 
 — .18 
 
 .44 
 
 71 
 
 -.03 
 
 .85 
 
 87 
 
 — .88 
 
 .48 
 
 78 
 
 — .18 
 
 .87 
 
 88 
 
 — .84 
 
 .40 
 
 73 
 
 - .18 
 
 .88 
 
 89 
 
 — .80 
 
 .34 
 
 74 
 
 — .06 
 
 .31 
 
 30 
 
 — .16 
 
 .87 
 
 75 
 
 + .04 
 
 .35 
 
 31 
 
 — .83 
 
 .84 
 
 76 
 
 .04 
 
 .40 
 
 38 
 
 — .89 
 
 .88 
 
 77 
 
 .02 
 
 .45 
 
 33 
 
 — .38 
 
 .19 
 
 78 
 
 .00 
 
 .46 
 
 34 
 
 — .36 
 
 .15 
 
 79 
 
 .01 
 
 .45 
 
 35 
 
 — .37 
 
 .11 
 
 80 
 
 .18 
 
 .44 
 
 36 
 
 — .36 
 
 .18 
 
 81 
 
 .16 
 
 .41 
 
 37 
 
 — .36 
 
 .15 
 
 88 
 
 .17 
 
 .38 
 
 38 
 
 — .36 
 
 .14 
 
 83 
 
 .18 
 
 .36 
 
 39 
 
 — .89 
 
 .08 
 
 84 
 
 .19 
 
 .36 
 
 40 
 
 — .88 
 
 + ,08 
 
 e& 
 
 .20 
 
 .36 
 
 41 
 
 — .17 
 
 -.07 
 
 86 
 
 .23 
 
 .40 
 
 48 
 
 — .18 
 
 — .12 
 
 87 
 
 .25 
 
 .44 
 
 43 
 
 — .06 
 
 - .13 
 
 68 
 
 .25 
 
 .44 
 
 44 
 
 - .08 
 
 — .07 
 
 89 
 
 .19 
 
 .37 
 
 45 
 
 +0.03 
 
 -0.01 
 
 90 
 
 +0.18 
 
 +0.30 
 
 The correctioua thuB obtained wore now made continaous, and the above table was formed 
 in the following way. Represent the correction for y° by (Y). Then, for circle A was taken 
 
 [1^1- >^ (i)+(l|)+(2) J.. 
 
 [3H=i<{ (3)+(3t)+(4) ^ 
 &e.t Ace. 
 
 [Oj=i'{ [88i]+(0)+[l|] J., 
 
 l«]=H [3i]+(«)+[6|U. 
 [2J]«H (2)+(3) ^ 
 
 &. 
 
 dee. 
 
 &c fcc. 
 
 ■''^^^^S^SS^f^-y? 
 
UMITBD STATn HAYAL OBSBBYATOBT. m. 
 
 The concluded corrections were then interpolated between [0], [[2^]], [6], [[7|]], Ao. 
 For circle B was taken 
 
 [i*]«H (t)+(i)+(ii)+(8) y, 
 
 [Sil-H (3)+(34)+(4)+(4j) }, 
 • tcct cbc. 
 
 (0]-t^[8^)+(0)+fli]}.. 
 
 [7i]^u*m]Hn)+m]>. 
 
 oCC.t «cc> 
 
 The condaded corrections were then interpolated between [0], [2)], [5], [11^], Ac. 
 In the table the argument is changed 45°, so as to correspond to the reading of the finding 
 microscope. 
 
 ERRORS OF CERTAIN ISOLATED DIVISIONS. 
 
 (73) During the year 1866 the circle was so set that when the telescope pointed toward 
 the nadir, the reading of the finding microscope was 359° 56' . It therefore becomes necessary 
 to determine the error of the particular divisions then under the microscopes, relatively to 
 the others. For this purpose microscope YII was furnished by the machinist of the Observ* 
 atory with two extra pair of spider lines at a distance of, as nearly as possible, 2^ on each side of 
 the central pair. Each division, from 44° 42' to 45° 8', was then brought in succession under 
 the middle pair of wires, and at each setting the three pairs were placed in succession over 
 their corresponding divisions. Thus, two measures of each space were obtained. These 
 measures, being treated in the way already set forth, gave the following corrections for each 
 division relatively to the mean of the fourteen divisions from 44° 42' to 45° 8', 134° 42' to 
 136° 8', Ac: 
 
 DlY. 
 
 Cor. 
 
 Div. 
 
 Cor. 
 
 DIv. 
 
 Cor. 
 
 Dlv. 
 
 Cor. 
 
 Mora. 
 
 O ' 
 
 (/ 
 
 / 
 
 // 
 
 o / 
 
 II 
 
 O ' 
 
 it 
 
 II 
 
 44 48 
 
 +0.86 
 
 134 48 
 
 +0.87 
 
 384 48 
 
 -0.01 
 
 314 48 
 
 +0.18 
 
 +0.16 
 
 44 
 
 + .19 
 
 44 
 
 + .07 
 
 44 
 
 — .09 
 
 44 
 
 — .09 
 
 +('.08 
 
 46 
 
 -.09 
 
 46 
 
 -.31 
 
 46 
 
 — .83 
 
 46 
 
 — .11 
 
 —0.18 
 
 48 
 
 — .06 
 
 48 
 
 - .01 
 
 48 
 
 + .03 
 
 48 
 
 + .01 
 
 0.00 
 
 60 
 
 — .06 
 
 60 
 
 + .03 
 
 60 
 
 — .07 
 
 60 
 
 -.83 
 
 -0.08 
 
 68 
 
 + -*ih 
 
 68 
 
 — .04 
 
 68 
 
 + .10 
 
 68 
 
 + .« 
 
 +0.08 
 
 64 
 
 — .18 
 
 64 
 
 — .09 
 
 54 
 
 .00 
 
 64 
 
 -.86 
 
 —0.18 
 
 66 
 
 - .06 
 
 66 
 
 — .13 
 
 — .81 
 + .81 
 
 56 
 
 -- .09 
 
 56 
 
 — .08 
 
 —0.04 
 
 68 
 
 + .36 
 
 68 
 
 68 
 
 -- .40 
 
 68 
 
 + .09 
 
 +0.16 
 
 46 
 
 — .03 
 
 136 
 
 885 
 
 -• .04 
 
 315 
 
 + .10 
 
 +U.08 
 
 8 
 
 — .08 
 
 » 
 
 + .43 
 
 8 
 
 — .01 
 
 8 
 
 -.08 
 
 --O.08 
 
 4 
 
 — .86 
 
 4 
 
 — .03 
 
 4 
 
 — .11 
 
 4 
 
 — .14 
 
 —0.13 
 
 6 
 
 — .14 
 
 6 
 
 — .19 
 
 6 
 
 + .04 
 
 6 
 
 + .a 
 
 —0.08 
 
 8 
 
 + .19 
 
 8 
 
 — .01 
 
 a 
 
 — .80 
 
 H 
 
 +0.16 
 
 +0.04 
 
 TESTS FOR OTHER POSSIBLE ERRORS. 
 
 J - 
 
 (74) The errors of an instrument may be divided into two clatwes; those which we expect 
 to find, determine, and allow for in the reduction of observations, and those we expect the 
 6 
 
42 
 
 DBSosipnoif or ths tbariit oisolb op ths 
 
 1(* 
 
 ,. i 
 
 111 
 
 
 I* § 
 
 artist to avoid entirely, or at least render insenaible. The olaaaifioation is somewhat arbitrary, 
 depending, as it does, upon the degree of precision sought by the astronomer, and the degree 
 of excellence attained by the artist ; yet the custom of astronomers has rendered it quite 
 definite. The errors we have investigated are generally recognised as of the first class; we 
 shall now consider those of the second. 
 
 (76) IrregularUy of Pivot*. — No apparatus fur determining directly the influence of possible 
 irregularity of pivots upon the axis of rotation was furnished with the instrument, but the 
 artists did furnish an extremely delicate instrument for determining any difference nf diameters 
 of the same pivot. It consists of a pair of calipers, which are screwed upon one of the Y 
 bearers, and grasp the horissontal diameter of the pivot. The telescope being turned, any 
 change in the distance of the calipers amounting to the two hundred thousandth of an inch 
 will be rendered sensible by a pair of multiplying levers, the end of the last of which moves 
 over a divided scale. The telescope being turned through an entire revolution with the cali> 
 pers on one pivot, no difference of diameters so great as this was detected. Only one pivot 
 was thus tested. 
 
 As an additional test, the hanging level was placed upon the pivots, and read at every 20° 
 of zenith distance of telescope, from 20° to 160°; the telescope being moved by one of the 
 west handles. It was then returned by the same handle, and the readings repeated. The effect 
 of the pressure of the handle in changing the level of the pivot was quite sensible, amounting 
 to 0". 25 ; but this effect was reversed by the backward motion, and the extreme range of the 
 mean level reading for different positions of the telescope was 0".05. It was conc'uded that 
 the form of the pivots might be regarded as perfect. 
 
 (76) Another possible source of irregularities in the motion of the optical axis of the tele- 
 scope at first caused me considerable solicitude. It has been seen in the description that the 
 fulcrum of the levers of the great counterpoises are not, as I conceive they should be in so large 
 an instrument, knife edges, but pivots. Owing to the unavoidable friction of these pivots, and 
 also to the friction of the end springs, the instrument, when balanced by the counterpoises, 
 will allow changes of weight of perhaps eight or ten pounds without causing motion of the 
 lever. Consequently, the division of the weight between the friction rollers and the pivots 
 will be uncertain to this amount. If, now, there bo uny irregularity in the grooves of the axis 
 by which the friction rollers act on the instrument, this division may vary in different positions 
 of the instrument, the friction roller acting more powerfully on points more distant from the 
 centre of rotation. And such a change of pressure will produce a vertical flexure of the axis, 
 which will change the direction of the optical axis of the telescope. To avoid any possible 
 error from this source, small pieces of rubber cloth were inserted in the sustaining sockets on 
 the levers, the elasticity of which would take up any minute irregularities of the kind referred 
 to, and make the pressure nearly constant. 
 
 To test the effectiveness of this contrivance, a piece of thin paper, probably 7^ of an 
 inch thick, was drawn under the friction rollers, and the effect upon the level of the axis and 
 the verticality of the teloscope was noted. The former waa not changed at all. The latter, 
 which was determined by comparing the position of the vertical middle wire, and its image 
 reflected from mercury, did not admit of exact measurement, as the image was nearly hidden 
 by the wire. Certainly, however, there was no change as great as 0".S. The paper being 
 many times as thick as any probable irregularity in turning the axis, there is little danger of 
 error from the source in question. 
 
 (77) Part of the same general investigation was the determination of the effect en'the level 
 of the axis when the action of the counterpoise was changed from its maximum to its minimum 
 amount. One pivot being raised from its Y, by pressing on the counterpoise, was gently let 
 down ngain, and the level carefully noted. It was then pressed downward by pressing the 
 
UMITKD 8TATM NATAL OBSmVATOST 
 
 48 
 
 oonnlerpoiM upward, and the level again noted. The changes of reading varied from 0".3 to 
 1".0. They are, I conceive, almnat entirely due to flexure of the Y'a under the weightH of the 
 "ivots. 
 
 (78) Ooincidenre of the Divided Facet <^ the Cirdet with Ptama perpendtcvlar to the Axis of 
 Botation — There is no error in tbia respect which affects the definition of the divisionH in the 
 Belds of the microscope, and I have made such measures as to satisfy myself that no appreciable 
 error arises from the product tan A it, p. 11-12. 
 
 (79) Oenercd Bemark. — Beside the above systematic examinations, the instrument is from 
 time to time in an irregular way examined for every cause which I can think of, as liable to 
 vitiate the results of observations. Nothing serious has yet been detected. 
 
 #.1 
 
*«* 
 
PART IV. 
 
 REMARKS ON THB PBRFORMAi^CB AND USB OF THE TRANSIT OIROI.E. 
 
 (80) The general design of the inttrament ia entirely that of the makem. Specific direc- 
 tions were sent them only on a few minor points, saoh as the arrangement of the micrometers, 
 the self- registering micrometer head, and the wires of the retioale. It was the opinion of 
 Captain Oilliss, that, considering the repataion and expent.aoe of the artists, he would be 
 more likely to secure a good instrument by allowing tiiem U carry out their own vIbwh, and 
 holding them responsible for the performance of the instrument &han by designing it himself. 
 
 As a general remark, it may be said that the mechanical execution of every part of the 
 instrument is of the first order of excellence. I cannot speak with certainty of the object glass, 
 as it has not been severely tested. Oertainly, however, it has no defect which interferes with 
 the performance of the instrument. After being transported by land and water a fourth of the 
 way round the globe, all tli delicate and complicated parts of the instrument were put together 
 without impediment or dumy, and immediately went into successful operation. 
 
 As a knowledge of the defects in design and performance, which we have thus far suc- 
 ceeded in discovering, may be valuable to astronomers, I shall set them forth. 
 
 The only defects of design which can yet be pronounced upon with certainty have already 
 been alluded It). They are : 
 
 1. Making the zenith distance micrometer carry the slides of the ocular, thus causing the 
 screw to carry too much weight when the head points downwards. 
 
 2. The form of the supporting fulcmms of the counterpoise levers. 
 
 3. The instability of the collimators and their levels, owing to the small distance (22 inches) 
 between the supporting shoulders. The only inconvenience which results from this construc- 
 tion is the increased labor of levelling the collimator. 
 
 (81) StabUUy qfthe Instrument. — ^Invariableness of instrumental constants is generally con. 
 sidered one of the most desirable qualities in a meridian instrument. A deficiency in this 
 respect, however, need not vitiate the results of observations, if only the astronomer can deter* 
 mine and apply the constants with a frequency proportioned to the instability of his instrument. 
 
 The most precise way to measure and indicate the variableness or uncertainty of the instru- 
 mental constants is to take the mean difference between consecutive determinations of the con- 
 stants. This I have done for the latter part of the year 1866, and the results are given in the 
 following table. During the summer of 1867 the stability of the zenith point has decidedly 
 improved, the mean difference being reduced to 0".60. 
 
 Collimation, October, 1866, to July, 1867 - - • -interval 1 week ■ • - - (K'.23 
 
 Level, Jnly to November, 1866 interval 1 to 4 days- - 
 
 Level on consecative days interval 1 day • • • 
 
 Mean difference of consecative transit of Polaris - - • interval 1 day • - - 
 
 Resoltiog mean change of axfmath interval 1 day • - • 
 
 Zenith point interval 3 to 12 hours - 
 
 Inclination of £. pier ........... interval 1 day . . - 
 
 0"51 
 0".89 
 3t. 2 
 1". 6 
 0". 1 
 0".65 
 
 i 
 
 |B |Hi lS 'M'^W l ii i W ' ''M WBtt>^^ 
 
46 
 
 DK8CR1PTION OP THE TRANSIT CIRCLE OP THE 
 
 * .> 
 
 If e be tlio meun error of the detormination itself, we raay expect a mean difference of 
 ■\/2e owing (o tiiat error alone, supposing the instrument to remain invariable. The stability 
 of the line of collimiition may therefore be regarded as perfect, and that of the level error 
 practically so, if determined for each day of observation. 
 
 It is far otherwise with the zenith point and the azimuth. Not only are they variable to 
 an annoying degree, but the causes of the variations are not definitely determined. Home light 
 may, however, be thrown upon them. 
 
 (82) Zenith Point. — Previous to any trial of the instrument, its most objectionablo feature 
 seemed to bo the mode of mounting the microscopes, and many astronomers would have pre- 
 dicted instability from this cause. But the relative positions of the microscopes have proved 
 unexpectedly steady. The following table shows the amount by which the line through the 
 zero of V and VII was in excess of 90° from that through VI and VIII, at various dates between 
 June 29 and September 17, 1866, the longest interval as yet in which the microscopes have 
 been subjected to no disturbance. They are formed by subtracting the mean reading of V and 
 VII for the two collimators, from that of VI and VIII, and are therefore the difference between 
 the nadir points of the circle as given by the two pairs. The dates are taken at random, except 
 with reference to the observer: 
 
 Date. 
 
 A. 
 
 1866. 
 
 •M'^ 
 
 y&'ik'^'ii-- 
 
 ;'.&}^y:. 
 
 June 29, 
 
 // 
 + 0.05; 
 
 29 9. 
 
 + 0.02; 
 
 July 2, 
 
 -0.12 ; 
 
 
 —0.08 ; 
 
 9, 
 
 -0.05 : 
 
 12. 
 
 -0.32 ; 
 
 26, 
 
 -0.10; 
 
 30. 
 
 + 0.30; 
 
 Aug. 6, 
 
 + 0.32; 
 
 15, 
 
 —0.20 ; 
 
 20, 
 
 —0 28; 
 
 29, 
 
 + 0.15; 
 
 Sept. 3. 
 
 -0.16; 
 
 10, 
 
 +0.12; 
 
 14. 
 
 -0.16. 
 
 I am persuaded that this degree of steadiness has never been exceeded, so that if the cen- 
 tral core has been sot in the pier in such a way as m secure immobility, the positions of the 
 microscopes relative to the pier will be as invariable in this mounting as in any other. 
 
 Passing from the microscopes to the circle — the nicety of fit and firmness of connection of 
 every part, from the divisions of the circle to the ends of the tube of the telescope, is beyond 
 reasonable doubt. 
 
 The constancy of position of the optical axis of the tube is rendered highly probable by the 
 steadiness of the error of coiiimation. Moreover, the object end of the telescope: has been sub- 
 jected to shocks several times greater than it ever receives in ordinary use, without any effect 
 upon the nadir point. 
 
 The constancy of the reading of the zenith-distance micrometer head for a given position 
 of the wires is all that could be desired. 
 
 (83) Supposing, from these considerations, that the changes observed must be due to move- 
 ments of the pier itself, a horisontal cylinder was fastened to it in July, admitting of being 
 levelled by one of the collimator levels and the changes in the inclination of the pier thus 
 determined. But the changes of nadir point were still only partially accounted for, and the 
 correct nadir reading sometimes exhibits a progressive change, continuing through a period of 
 several days. 
 
T7NITBD STATES NAVAL OBSERVATORy. 
 
 47 
 
 (84) Apparently, the only untested link in the chain ia the stability of the setting of the 
 microscope holder (D, plate lY, D-P, plate YI, Fig. 2) into the pier. To insure the solidity 
 of this setting, a hole was cut from the top of the pier to the perforation which received the 
 core of the holder, and the plaster poured in until the hol«) was full, when it oozed out on all 
 aides of the core. The setting is, therefore, as solid as it can be with plaster. It will, indeed, 
 yield, and allow the microscopes to turn by a strong pressure of the hand ; but under pressures 
 several times as great as they are ever subject to in observing, the microscopes are not dis- 
 turbed at all. Still, consi'jering the known hygrometric qualities of plaster; considering, also, 
 that the relative readings of the microscopes on the two piers are subject to changes of the 
 same general character and magnitude with the zenith point, I decidedly think that ''^e greater 
 part of the instability of the zenith point is due to the want of firmness of the plaster setting. 
 
 (85) The Azimuthal Error. — 0^ the cause of the variation of azimuth I. entertain little 
 doubt. The extreme breadth of the masonry on which the piers are supported is only four 
 and a half feet from north to south. With so narrow a base injurious changes of inclination of 
 the piers, from motion of the ground and consequent tipping of the masonry, seem to me un- 
 avoidable. And that such clfanges do take place is shown conclusively by the levelling appa* 
 ratus attached to the piers. If every part of the masonry tipped equally, the nadir point 
 alone, and not the azimuth, would be affected. But since the masonry is not perfectly rigid, 
 this condition would be fulfilled only by the ground giving way equally at each end of the 
 pier, which we have no reason to suppose the case. The piers tipping unequally, we may look 
 for changes of azimuth us well as of nadir point. The height of the axis above the centre oK 
 the pier being three times the distance of pivots, the change of azimuth from the cause in 
 question will be three times the change of relative inclination of the piers. 
 
 (86) In the spring of 1867 a levelling cylinder was attached to the west pier also, and the 
 difference of tipping of the piers compared with the changes of azimuth. The latter were not 
 accounted fur, a fact which may be attributable to the imperfections of the apparatus itself. 
 
 (87) Dependence of the CoUimation Error on Temperature. — Toward the end of the year 1866 
 the amount of this error was found to be dependent on the temperature, varying 0".05 for ewery 
 degree of Fahrenheit. The cause was discovered when the object glass was taken out at the 
 end of the year. It was then found that the glass was held in its cell by the pressure of three 
 chucks, 120° apart, two of them being fixed, Aud the third pressed in by a strong spring. The 
 direction of action of the spring was horizontal. Hence, owing to the different expansibilities 
 of the brass and glass, the centre of the latter would take different positions relative to the 
 centre of the former at different temperatures. The calculated change of collimation on the 
 hypothesis of perfect rigidity of both object glass and cell, is 0".07. The difference between 
 this and the observed change is probably due to the fact that the spring is not perfectly flex- 
 ible, nor the glass and brass perfectly rigid. 
 
 (88) It is probable that such changes will ultimately be made in the mounting of the object 
 glass, the microscope holders, and the great piers as may seem sufficient and necessary to 
 secure greater in>:iiobility of the collimation, azimuth, and zeiiitii point, the instability of the 
 two latter arising, as has been seen from defects of mounting rather than of construction. 
 
 ■i, 
 
 :::i 
 
 
 -J 
 
 i 
 
 1 
 
 '*■; 
 
wrr 
 
 t 
 
 ■ f 
 
 Wi.- 
 
 " , * 
 
 .£;; 
 
 «».ai* ) » « tl - i ) i4i— ^Wii r ut w 
 
 ■"'■■*•""■■■■ 
 
 ^HSUBtitasuimmtiii-miiiiimlmtii 
 
ALPHABETICAL INDEX. 
 
 (THE NU1IBEB8 HEFEB TO TH|^ PABAGBAPH&) 
 
 Azb, dataila of 
 
 twkt of, daring roTolaUon 
 
 teste of ite motion 
 
 lorel of, afftcted hj oonntorpoiM. . . 
 
 end tprinff of 5. 
 
 AiimnUi, amonnt of ite ehuigM.. 
 
 33 
 «3 
 75,76 
 77 
 18 
 80 
 85 
 4.21 
 SI 
 
 Cireles, diMnetor anddivisfcwia 
 
 owHagof 
 
 Law faatenedtoub ao 
 
 perpendienlvitjr of, to azie • 78 
 
 lawf of floxnre of. 61 
 
 table of flasnin.. .4 02 
 
 Cirale dirifiona. (flte DiTiakna.) 
 
 Clamp, deaoriptionof 89 
 
 CSolUmation error, ohanfesof. 81 
 
 depead on tamporatare , • 87 
 
 OolUmatoia, deseriptionof 37 
 
 Y'« of. 38 
 
 maj be aet on each otber ^ 5,!i4 
 
 difforenoea and form of eoUara of 63 
 
 diacrepaneiea o^ due to refica^ion 64 
 
 reoaaaaafor pladng 1 
 
 Connterpdaea, action of. „ 13 
 
 aflact level 77 
 
 posaible bad ellitet of, bowteatad 76 
 
 Cube, oantial, of axia S4 
 
 Deaignof tba inatmmant, Uatory and defcota of „ 80 
 
 Diviaiona of eireia, number and thtckneas 91 
 
 formula for wrofa '60,51 
 
 errora of; iaoiated 51-73 
 
 table for evei7 6C< 69 
 
 10 78 
 
 cjoUeal, inequality of. 70 
 
 bowteatad 71 
 
 Bnma, general pncaotion againat 79 
 
 inatnuHintal, table of inrtabUity of 81 
 
 Eya-fieoe, dewsription «rf 86-88 
 
 Flwnre . 39 
 
 f hnd a mant al hypotberia in Inireatigattng 40 
 
 geaatal tbaomnof 4i 
 
 requititH for determining 48-45 
 
 of drda, apedal invaatigation of 60 
 
 e b awrved lawaof 61 
 
 tabloof 68 
 
 of t a l ae e ope, general metbod ^ 48 
 
 preeautiona in daterminiog 40,64 
 
 produoad b7 ann oqt 08 
 
 vertteal, or ooaine 48,66 
 
 "M 
 
 3 
 
 i 
 
■W^ 
 
 50 
 
 ALPHABETICAL INDEX. 
 
 ^f 3 
 
 Flazureof dedinAtion micrometer snppoiti., 07 
 
 twisting, of (uii * 63 
 
 Floor, level of , 9,11 
 
 IlluminatioD, lamps for 34 
 
 of the field 38 
 
 ofthewirea on dark field 36 
 
 of mieroeoopes » • 10,34 
 
 Lerela, tpirit, valne of divisioiu ..i 6S 
 
 Level error, cbuigei of 81 
 
 Micrometer, general deacription of ■erawi 39 
 
 declination, how moved 89 
 
 Bogeri'B Mlf-registering head df* 
 
 periodic ineqoalitiee of..... « 66 
 
 - value of ravoiation of 67 
 
 irregularity of motion of 66 
 
 right aacension, deacription of 96 
 
 revolution, &o., of 69 
 
 wireaof. (SMWirea.) 
 
 of nticroacopea. (8m MicroscqMa.) 
 
 Mioroacopea, at^a for obaerving ....« <...... 8 
 
 mounting of » • 14 
 
 a^jnatment of focua * ••• 16 
 
 illnmiMtiou of 16,34 
 
 dimenalona and power 17 
 
 relative ateadineas of , ... 88 
 
 probable uuateadineaa of holder 84 
 
 micrometera of • 64 
 
 deacription of Uwir pwiodic iaeqoalitiea 66 
 
 of their irregnlaritiea 17 
 
 Nadir point, platform and aeata for obaarving ..«. 6,7 
 
 Object glaaa, aperture of •. 3 
 
 mode of mounting 87 
 
 Piera, form, material, and dimenaioua. ^ 8 
 
 aupporting maaonry of. 10 
 
 changea of their inclination 81,83,86 
 
 Pivota, teat of their form • 78 
 
 dimenaiona of. --.. 19 
 
 Powera, magnifying, of oculars 31 
 
 B^lroad, for reversing carriage 11 
 
 Ejection observations, faoilitiea fi>r making 6 
 
 Befraotion, atmospheric, inimical to flexure inveatigatiou 49,64 
 
 Room, dimensions of. •- J 
 
 details of - » 
 
 Stability of instmment, tables of. , 81 
 
 Telescope, dimenaiona of ^ - 3 
 
 Wirer, of declination micrometer • 80,33 
 
 of right asceaaion micrometer 87,38 
 
 of microscope micrometers >•. '• 17 
 
 fixed, of diaphragm, how carried • 88 
 
 ufiangemant uid notation • 38 
 
 how used in observation 39 
 
 Zenith point, changes of 81,88,84 
 
 ••>« ^ ' ''iti'trm^nn u timim iimMMmiUlimm 
 
 •••MWMMMMMiMtMil 
 
 Hi.^g i l i 
 
-.■» .-'iv 
 
 
 er 
 
 63 
 9,11 
 M 
 M 
 » 
 16,34 
 68 
 81 
 » 
 
 » 
 66 
 67 
 68 
 96 
 60 
 
 
 
 L-- * 
 
 8 
 
 14 
 
 16 
 
 16,34 
 
 17 
 
 8i 
 
 84 
 54 
 66 
 
 17 
 
 6,7 
 
 3 
 
 87 
 
 S 
 
 10 
 
 »,83,86 
 
 75 
 
 1» 
 
 it 
 
 11 
 
 6 
 
 49,64 
 
 ) 
 
 9 
 
 81 
 
 3 
 
 «U,33 
 
 87, 3S 
 
 17 
 
 88 
 
 88 
 
 39 
 
 U,88,84 
 
 iif'.' " -'^«^*"3i; ,, 
 
 ' rJ. 
 
 '-'*^f.-;-\'/, '■ 
 
 ■■■ .-■^- 
 
•^. 
 
 *mt<utiMtitiA> 
 
 U.4^4. 
 
 W. Baumann deJ. 
 
■ .^- 
 
 ^1 
 
 ^Ip^^^^f^p^ 
 
 3r 
 
 M^^^^^^M 
 
 £ 
 
 PLATE I 
 
 vy,- / -. - 
 
 m^ 
 
 iiiui^iMiiiiui^UJdm^ 
 
 ■A 
 
 T 
 
 "1 ; i. Jtiii 
 
 =*==»= 
 
 ■^ ! - 
 
 4*= 
 
 ^ 
 
 
 m. 
 
 ^ 
 
 Scale of fi>el . 
 
 ■»<-f -:''^M« » '• S; ■ -t-rj*-;.. 
 
 •v-r'-;- 
 
 Si- 
 
 <^6Kondrui) £<tt|^. 
 
i»^ 
 
 fr^wW'' ♦**%«'•«•» -i.i'-iCf-f 
 
 \V RrturnanTi del. 
 
Scale of Vert. 
 
 HATK II 
 
 ■-■^»T4f ilhUuBPi I I ilMMw.nniii,|^ 
 
 i-^^^-Jw., . 
 
 
 |3 
 
 -:1' ■ 
 
 J.CJCbndntp Biigr* 
 
 '■■rtV^''^*S^-. ■^■^^ 
 
^i^ 
 
 WBaumaim del 
 
IM.ATK 111 
 
 J.C.Kaiidrup Eng^ 
 
"WBRumaiui del. 
 
 m¥ 
 
 f ' fMll in i.i Bumiw Mi W B nj M L i i i 
 
IM^ATK IV: 
 
 J.C.Kondriip Engl" 
 
 > ^ 'Ww »WB i jlnil l MiWli n i0. t lM 
 
^ 
 
 WiSnunuum del. 
 
i - i 
 
 J 
 
 J.r Kdiidi'i-t) Kiigf' 
 
 yf-^i^ mnaimmmim iwiiiWiiiimiwiiaimi[iif(iiiiiniiji.B)i ^i.m 
 
■•■kw,-»„L'i-i*>-:5a,,^,j.y.L3i*>>;,^iiii. K^-.i~-- 
 
«•«!••*■*«•■ !ir 
 
 S 
 
 Bamsi*r°tiO> Philaila 
 
/__»[) ^__..^j1J\ 1 
 
 ^ 
 
 W baumjim <i»-l 
 
.IV Kniidriij' K1151'' 
 
 :' . ,^\ 
 
 -Stew 
 
'.V ,\«iimmn\ ili'l. 
 
1 IV 
 
 iL,;;-^{{» » 
 
 f«iito* • •■*>'««»»• 
 

 :i-ism,^>'----"'-»--~'~^''''--^''-'-f^'itMX.A.^-,.--A^»t,m.fm^ i.,-4i--™.:'- '