Sept. 13, 1883] 



NATURE 



475 



freedom from accidental errors for individual stars, the final 

 catalogue was obtained by combining with the Pulko.va series 

 the Greenwich observations from 1836 to 1S76, the Harvard 

 College observations for 1S71-72, the Leipzig observations in 

 declination only, between i865 and 1870, and the Leyden ob- 

 servations in declination between 1864 and 1S70. Befjre this 

 combination was made, however, these observation) were all 

 reduced to the Pulkowa system. 



The following observatories have taken part in the zone 

 observations : — 



Limits of Limits of 



The zone between - 2° and + 1° was originally undertaken at 

 Palermo, that between +i°and +4° at Neuchatel, that between 

 + 4° and +10° at Mannheim, and that between +35° and +40° 

 at Chicago. 



In the latter case the great fire at Chicago crippled the 

 resources of the observatory to such an extent that Safford was 

 compelled to relinquish the work, which was at tint time quite 

 far advanced. 



Attention was called at an early date to the importance of 

 continuing the survey of the northern heavens beyond the 

 southern limit fixed by Argelander. The preparation necessary 

 for the execution of this work consisted in the extension of the 

 Durchmustaitng to the tropic of Capricorn. This was under- 

 taken by Schonfeld at Leipzig. 



In the report to the Gesdlschaft at the meeting held at Stock- 

 holm in 1S77, he has given an account of this work, in which 

 he stated that it was sufficiently near completion to invite the 

 consideration of the question of the meridian circle determina- 

 tions of the places of stars to the ninth magnitude. The lack 

 ofsouthern fundamental stars whose positions were well deter- 

 mined was still a hindrance to the immediate commencement of 

 the work. Relatively more stars of this class are require .1 than 

 in the northern observations, in order to eliminate the inequalities 

 due to refraction. Schonfeld stated that, while the burden of 

 the determination of the places of these southern fundamental 

 stars must rest mainly upon southern observations, it seemed 

 necessary to connect them with the Pulkowa system by a con- 

 necting link (Mittelnlied), through observations at some observa- 

 tory well situated for this purpose. At this meeting Sande 

 Bakhuysen, of Leyden, gave notice of intention to take part in 

 this work. Gylden urged the importance of securing the co- 

 operation of Melbourne, and Peters suggested the advantage of 

 securing Washington as an additional "mean term" (V.J.S. 

 1877, p. 265). 



The next reference to this work is contained in the Gesellscha/t 

 for 1881 (V.J.S. xv. p. 270). A list of 303 southern stars is 

 here given whose exact places were at that time being deter- 

 mined at Leyden and at the Cape of Good Hope. This list 

 was selected by Schonfeld and Sande Bakhuysen, in a way to 

 meet the requirements referred to in previous discus-ions. 



A final catalogue of eighty-three southern fundamental stars 

 by Auwers appears in this number of the Gisellscha/t. The 

 places depend upon the same authorities as for the northern 

 stars, with the addition of the Cape of Good Hope catalogue for 

 i860, Williamstown, Melbourne for 1870, and Harvard College 

 (Safford) for 1864. For stars not observed at Pulkowa, the 

 general catalogue of Yarnall (185S-1861), and the Washington 

 observations, with the new meridian circle between 1872 and 

 1875, were employed. As in the case of the northern stars, 

 these observations are all reduced to the Pulkowa syste n for 1865. 

 It is understood that the coordinates of the list of 303 stars are 

 to depend upon this extension of the general syste a of Publica- 

 tion xiv. to the limits required by the southern Durchmuslcrung 

 of Schonfeld. 



It would be surprising if all the conditions of success were 

 fulfilled in the first execution of a work having the magnitude 

 and involving the difficulties of the scheme of observations under- 

 taken under the auspices of the Gesellscha/t. The extent of the 

 discordances which are to be expected between the results 



obtained by different observers can only be ascertained when 

 the observations by which the different zones are to be con- 

 nected have been reduced. Each observer extended the work- 

 ing list of his own zone 10' north and south ; and it is expected 

 that a sufficient number of observations of this kind have been 

 made to determine the systematic relations existing between the 

 coordinates of each zone with those of its neighbour. 



It is probable, however, that the experience of Gill will be 

 repeated on a larger scale. In 1S71 he solicited the cooperation 

 of a-tronomers in the determination of the coordi iates of twenty- 

 eight stars, which he desired to employ in the reduction of his 

 heliometer observations of the planet Mars for the purpose of 

 obtaining the solar parallax. The results obtained at twelve 

 observatories of the first class are published in vol. xtxis. p. 99, 

 of the Monthly Notices of the Royal Astronomical Society. Not- 

 withstanding the fact that the final values obtained at eacl 

 servatory depend upon several observations, the average differ- 

 ence between the lea-t and the grcate-t results, obtained by 

 different observers for each star, is 024s. in right ascension, and 

 2 '3" in declination. In four cases the difference in right ascension 

 exceeds 30s., and in four cases the difference in declination 

 exceeds 3'o". 



Even after the results are reduced to a homogeneous system, 

 the following outstanding deviations from a mean system are 

 found : — 



Autho 



Authority. 



The observations of a second list of twelve stars, one-half of 

 the number being comparatively bright, and the remaining half 

 faint, showed no marked improvement, either with respect to 

 the magnitude of errors which could be classed as accidental, or 

 in regard to the systematic deviations from a mean system. 



This discussion revealed one source of discordance which will 

 doubtless affect the z me observations : viz. the difference between 

 the right ascensions determine 1 by the eye-and-eir method and 

 tho-e determined with the aid of the chronograph. 



The programme of the Gesellschaft makes no provision for the 

 elimination of errors which depend upon the magnitude of the 

 stars observed ; but special observations have been undertaken 

 at several observatories for the purpose of defining the relation 

 between the results for stars of different magnitude-. At 

 Harvard College Observatory, the direct effect of a reduction of 

 the magnitude has been ascertained by reducing the aperture of 

 the telescope by means of diaphragms. Be-ides this, the obser- 

 vations have been arranged in such a manner that an error 

 depending upon the magnitude can be derived from an investi- 

 gation of the observations upon two successive nights. 



At Leyden, at Albany, and perhaps at other observatories, the 

 effect of magnitude has been determined by observations through 

 wire gauze. But notwithstanding all the precautions which 

 have been taken in the observations, and which may be taken in 

 the reductions, it will undoubtedly be found that the final results 

 obtained will involve errors which cannot be entirely eliminated. 



In the experience of the writer two other sources of error have 

 been detected. It has been found that there is a well defined 

 equation between the observations, which is a function of the 

 amount, and the character of the illumination of the field of the 

 telescope. It has also been f.mnd that observations made under 

 very unfavourable atmospheric conditions differ systematically 

 from those made under favourable conditions. When the seeing 

 was noted as very bad, it is found that the observed right ascen- 

 sions are about "oSs. too great, and that the observed declinations 

 are about o'S " too great. 



There are doubtle-s other sources of error which the discussion 

 of the observations will bring to light. The effect of the dis- 

 covery of these and other errors will probably be to hasten the 

 repetition of the zone observati ns under a more perfect scheme, 

 framed in such a manner as to cover all the deficiencies u hich 

 experience has revealed or may yet reveal. One would not pro- 

 bably go far astray in naming the year 1900 as the mean epoch 

 of the new survey. If the observations are again repeated in 

 1950, sufficient data will then have been accumulated for at 

 least an approximate determination of the laws of sidereal 

 motion. 



