on the presentation of a Medal. 137 



one crossing — it does so. The points of intersection with the axis 

 ought to occur at dates in like manner calculated a \ priori ; and so they 

 do within very negligible limits of error. And, lastly, the general 

 forms, magnitudes and flexures of the curves ought to be identical with 

 those of curves similarly projected, by calculation on an assumed re- 

 sulting parallactic coefficient. This is the final and severe test : Mr. 

 Main has applied it, and the results have been placed before you : — ■ 

 oculis subjecta fidelibus. If all this does not carry conviction along 

 with it, it seems difficult to say what ought to do so. 



The only thing that can possibly be cavilled at is the shortness of 

 the period embraced by the observations: viz. from August 1837 to 

 the end of March 1840. But this interval admits of five intersecti- 

 ons of each curve with its axis ; of two maxima and two minima in 

 its excursions on either side ; and of ample room for trying its agree- 

 ment in general form with the true parallactic curves. Under such 

 circumstances, it is quite out of the question to declare the whole 

 phenomenon an accident or an illusion. Something has assuredly 

 been discovered, and if that something be not parallax, we are alto- 

 gether at fault, and know not what other cause to ascribe it to. 



The instrument with which Bessel made these most remarkable 

 observations is a heliometer of large dimensions, and with an exqui- 

 site object-glass by Fraunhofer. I well remember to have seen this 

 object-glass at Munich before it was cut, and to have been not a lit- 

 tle amazed at the boldness of the maker who would devote a glass 

 which at that time would have been considered in England almost 

 invaluable, to so hazardous an operation. Little did I then imagine 

 the noble purpose it was destined to accomplish. By the nature and 

 construction of this instrument, especially when driven by clock- 

 work, almost every conceivable error which can affect a micrometri- 

 cal measure is destroyed, when properly used ; and the precaution 

 taken by M. Bessel in its use have been such as might be expected 

 from his consummate skill. The only possible apparent opening for 

 an annually fluctuating error seems to be in the correction for tem- 

 perature of its scale. But this correction has been ascertained by 

 M. Bessel by direct observation, in hot and cold seasons, and applied. 

 Nor could this cause destroy the evidence arising from the simul- 

 taneous observation of the two companion stars, since a wrong cor- 



T 



