492

HISTORY OF PHYSICAL ASTRONOMY.

the position of the latter being ascertained by observations of circumpolarstars. The instrument, according to this practice, served to determine thepolar distance of an object, and therefore the results were totally inde-pendent of the place of observation. The error of the instrument wasascertained by comparing the observed places of certain standard starswith their places as deduced from a fundamental catalogue. The differ-ence between each observed and calculated result gave a correspondingvalue of the error of the instrument for the day of observation, and themean of all such differences gave the mean error of the instrument, or,as Pond termed it, the index error. This quantity, being then applied toall the observations, gave the polar distances of the various objects, freedfrom the effects of instrumental error. In the case of the standard stars,the application of the index error supplied the means of obtaining a cor-rection of their places as assigned by the catalogue, and it is manifest thatby frequently repeating the same process the catalogue would be renderedmore and more perfect*. This method of observation obviated the neces-sity of employing the plumb-line, the use of which was not, indeed, con-sidered by Trougliton in his original design of the instrument; while atthe same time it dispensed with an exact knowledge of the latitude of theObservatory f.

In 1816 the Pioyal Observatory was enriched by the accession of a newtransit instrument ten feet long, constructed by Troughton. Hence-forward the Greenwich determinations, both in polar distance and rightascension, acquired a degree of precision which has hardly been surpassedby the most accurate observations of the present day.

In the course of his famous controversy with Brinkley on the parallaxof the fixed stars, Pond was led to employ occasionally a new method ofobservation which he subsequently practised with great success. Tt is essen-tially founded upon the principle of optics that when a ray of light is re-flected from any surface, the angles of incidence and reflexion are equal toeach other. In virtue of this principle, it is clear that when a celestial ob-ject is observed with the mural circle by direct vision, and also by reflexionfrom the surface of a fluid, the half of the angle contained between thetwo readings of the circle will assign the altitude of the object, and thereading which corresponds to the point of bisection will indicate thehorizontal point. Hence, knowing the latitude of the place of observa-tion, the polar distance of the object may at once be determined. Pondfirst applied this principle by observing the object at one of its transitsby direct vision, and then at the next transit by reflexion from the sur-face of a trough of mercury. This method was attended, however, withinconvenience, inasmuch as, in consequence of unfavourable weather,several days might intervene before a complete observation could beobtained. Moreover the result was liable to some degree of uncertainty,arising from the different effects of refraction at the beginning and endof the interval, and from causes tending to produce a slight derange-ment in the adjustment of the different parts of the instrument. Im-pressed with a sense of these defects, Pond contrived a method of observ-

* The process thus indicated would not free the polar distances of the catalogue of acommon error, but Pond shewed how this object might be effected by observations ofcircumpolar stars..—( Phil. Trans., 1818, p. 410.)

+ A knowledge of the latitude was necessary for transforming the polar distances intoaltitudes, with a view to compute the effects of parallax and retraction, but it is manifestthat, for such a purpose, extreme precision was not required.