STOKES. — PENDULUM OBSERVATIONS. .9 



ever apparently firm, may not be subject to small motions in 

 azimuth, wliicli would vitiate the transits. 



Clear weather should be chosen for the observations, that 

 transits may be observed. 



The observers are assumed to be already acquainted with the 

 mode of making pendulum observations, and therefore it will only 

 be necessary to mention some precautions. 



It is recommended that great care be taken as to the mode of 

 illuminating the bright patch on the clock-pendulum. Sir George 

 Airy found a gold-leaf surface of an oblique section of a cylinder 

 projecting from the bob towards the observer to be best. The 

 light is then to be lateral, and may be distant. 



As even an astronomical clock cannot be trusted to go for 

 short intervals of time with a rate equal to its mean rate for 

 24 hours, it is desirable to take a series of consecutive swings 

 extending over 24 hours, which would have the further advantage 

 that the mean temperature of the pendulum would more accurately 

 correspond to the mean indication of the thermometers. The 

 time chosen for commencement should be about the middle of the 

 time most favourable for transits. As a swing may be expected 

 to last about four hours, and it is sufficient to observe two or three 

 coincidences at the beginning and end of each swing, the observer 

 would have time enough to take transits and to rest in the intervals 

 between observing coincidences. The observer must remember, 

 however, that he is responsible for the number of coincidences 

 that have taken place, and therefore he would do well to take at 

 first, or in preliminary trials, one or two intermediate coincidences, 

 merely as counters not intended for reduction, and not leave off 

 this practice till he has convinced himself that it may be safely 

 dispensed with. 



In observing coincidences the observer must, of course, register 

 both the disappearance and the reappearance of the mark. But 

 as it is somewhat perplexing to observe and register four events 

 which succeed one another at intervals of a few seconds, namely, 

 the two disappearances (those of the right and left edges of the 

 mark) and the two reaj)pearances, the observer (unless he can 

 thoroughly depend upon himself to record the four events without 

 confusion) is advised to be careful in the adjustment of the mark 

 and diaphragm, so as to secure the two disappearances or the 

 two reappearances taking place on consecutive seconds even when 

 the pendulum is swinging in the smallest arc that will be observed 

 with, in which case it will, of course, suffice to observe one 

 disappearance and one reappearance for each coincidence. 



The barometer and the thermometers hung near the pendulum 

 should be read at the beginning and end of each swing. Should 

 there be much variation of temperature, the thermometers should 

 also be read at noted times once or oftener during the swing. It 

 is to be remembered that what we want to know is, not the exact 

 temperature at the moment of coincidence, but the mean tempera- 

 ture during the swing. 



In one of the swings, or, if more convenient, in a preliminary or 

 subsequent special swing taken for this sole object, and in which 

 36122. h 



