the moon is below the star. This is the angle SGM, and the comparison by which it is determined is almost independent of refraction. By similar observations at C, the angle SCM is found with equal accuracy. And the difference between these angles gives the angle CMG with great accuracy. And it is this angle upon which the distance of the moon mainly depends.
I then explained that the calculation, in point of fact, is not made by treating MGC as one triangle in a survey, but by dividing the angle GMC into two parts by the line EM, and then assuming for trial a value of the distance EM, and computing the angle EMG, and with the same assumption computing the angle EMC, and adding them together, and finding whether this sum agrees with the observed angle GMC; if it does not agree, the assumption of distance must be varied till it does agree. There is no difficulty in each of these computations; because, from the dimensions of the earth, it it easy to find the inclination of the line GE to the vertical H'GE'; and therefore from the observed angle H'GM the angle HGM is found; also the length GE is known; and the length EM is assumed for trial; and then the calculation of EMG is easy.
It is now proper to mention that astronomers very seldom refer to the actual length EM in yards or miles. I explained that the angle EMG is called the parallax of the moon at G, and the angle EMC is the parallax of the moon at C. Now, (referring for the present to the place G only,) if the line GM were perpendicular to the line EG, that is, if the moon were in the horizon as viewed from G, the parallax would be greater than in any other position of the moon, (supposing the distance EM not to be
