194 CARNEGIE INSTITUTION OF WASHINGTON. 



It is not to be supposed that this will produce perfect results. The variation 

 of differential refraction with observed meteorology is, though fair, not abso- 

 lute. Difficulties of this sort inhere in any series of observations. Moreover, 

 there will be anomalies due to wholly different causes. Such will be hopelessly 

 disguised by any ironing-out process; but by systematically removing one 

 anomaly the others will stand out more clearly. 



This in itself is cause sufficient for removing the differential refraction effects 

 from observations, but there are further developments of great importance. 



DEVELOPMENTS FROM DIFFERENTIAL REFRACTION. 



Reference is made to the various astronomical constants, to variation of 

 latitude, and more especially to the law of refraction. The last has been 

 studied many times, with results which do not carry us far beyond the point 

 reached by Bessel in his Fundamenta Astronomica about a century ago. The 

 reason is that, already in his day, the study of refraction had gone about as 

 far as it could by the method of averaging results and hoping that anomaly 

 would balance anomaly. Further progress can result only from a study of 

 these anomalies themselves. Their removal will clear the way, we may 

 reasonably hope, to further progress with the law of refraction itself. 



The study of most of the astronomical constants, so called, and of the varia- 

 tion of latitudes, has been hitherto made upon the basis of differential obser- 

 vations especially designed for the purpose. The possibility of something 

 having been overlooked in such studies, which might make their results not 

 strictly applicable to observations made under routine meridian-circle con- 

 ditions, has never been sufficiently considered. 



Through a revision of our reductions of the observations, on the basis of 

 inclusion of the differential refraction term, we hope to reap the benefits 

 delineated above. The combination of observations taken north and south of 

 the equator with the same instrument, and largely with the same observers, 

 offers a special opportunity for the study of the fundamental laws of refraction 

 and for the establishment of a more perfect system of star positions stretch- 

 ing from pole to pole. This in turn will help materially in improving the 

 proper-motions of the stars. 



It is well to stress a few points which should be carefully considered in all 

 fundamental meridian programs of the future. 



1. Circumpolar stars must be observed at successive culminations. For 

 the sake of continuity, and to avoid personal equation, the same observer 

 should continue his observations over that period of time. (According to our 

 practice at Albany, the same observer is on duty for the period of a week.) 



2. Time stars should be observed 12 hours apart, preferably 6 hours apart, 

 and successive transits of the same time star should be taken. Again all these 

 observations should be taken by the same observer. 



3. Daylight as well as night observations should be taken, some in the 

 morning and some in the afternoon. Zenith-distances as well as transits of 

 these stars should be observed. 



4. Sun-temperature and humidity records should be included hereafter, 

 as the effects of differential refraction seem to respond to these two elements. 



