296 Mr. Pond on the Parallax of u Lyrce. 



coincides with observation ; it is only a rude agreement at 

 the points of the greatest and least variation that can be de- 

 monstrated. The disengagement of the larger equation only 

 proves therefore the existence of some regularly recurring 

 cause, acting with greatest effect at the extreme seasons. 



The reason, I conceive, why Dr. Brinkley does not find 

 parallax in y Draconis is, that with respect to the zenith point, 

 his instrument, like every one of a similar construction, is a 

 perfect instrument. No portion of the arc is employed, nor 

 can temperature here occasion any errors by its changes. As 

 the star to be examined recedes from the zenith, the instru- 

 ment becomes less and less perfect; and he finds a small 

 parallax in « Cygni, a larger in « Lyrae, and oftentimes a still 

 larger in stars more remote from the zenith. An additional 

 reason for suspecting that the discordances observed arise 

 from temperature is this : the greatest supposed }iarallax is 

 found in those stars whose maximum and minimimi of parallax 

 would fall in the extreme seasons, and it is not at all improbable 

 that irregular refraction, arising from the unequal state of the 

 temperature within and without the Observatory, may have had 

 a considerable share in occasioning the Dublin discordances, 

 combined, perhaps, with the effect of the changes of tempera- 

 ture upon the instrument itself. It is a circumstance not 

 hitherto sufficiently noticed by astronomers, that there are 

 many cases where the smallest disturbing cause will produce 

 an error quadruple of its own amount; and consequently, that 

 the greatest error to which we are liable from such a cause at 

 any one observation will be only one-fourth of the difference 

 that we can detect between the most discordant of them. Of 

 such a nature are those disturbances which, like refraction 

 for instance, introduce errors, both positive and negative, 

 into the determination of either extremity of the arc that 

 measui'es the distance between two stars. 



By a singular combination of circumstances, not probable 

 certainly when considered a prioi-i, but by no means imyiossi- 

 ble, the variation caused by change of temperature may follow 

 an annual law so little differing from that of parallax, as to 

 bring out the assumed parallax, and to leave the solar nuta- 

 tion disengaged. 



Notwithstanding the importance of these investigations to 

 the history of astronomy, and to our forming a correct notion 

 of the system of the universe, yet our decision ultimately turns 

 upon so very small a quantity, that our having reduced the 

 inquiry to these narrow limits, rather tends to show the per- 

 fection of each instrument than the defect of either. 



On former occasions I considered the question of jiarallax 



in 



