288 J. ALFRED GRIFFITHS. 



even, commencing at eighteen and ending at fifteen revolutions 

 per minute. 



In the diagram b the thread from the paper roller to shaft T S 

 was led to a larger pulley than in diagram a, hence the difference 

 in the vertical scales. In both diagrams the envelope of the 

 vertical strokes is a fairly even curve of a somewhat parabolic 

 outline, with minor irregularities due to variations in the train 

 and spring friction. 



Fig. 2 is a portion of record extending over three days (4-6 

 October, 1894) taken at Parramatta with the special recorder for 

 wind, and showing the daily wave in a marked manner, as well 

 as some of the more prominent irregularities inside of each hourly 

 average. During this experiment the wind gauge was coupled to 

 both recorders, and Fig. 3 is a duplicate of the same record from 

 4 p.m. to midnight on the 5th October, taken with a four minute 

 period instead of one hour, and giving the minute changes in 

 greater detail. 



Fig. 4 is a short example of one hour taken 19 October, with 

 an eighty second period, and showing that with wind there is no 

 such thing as uniform velocity, even for a few seconds, but that 

 however small the period be made yet succeeding intervals always 

 show great and irregular changes of velocity. This has been 

 tested, both in calms and gales, down to intervals so short as 

 four seconds. 



The instruments may be used for recording any class of velocity, 

 such as the motion of a railway train or steamship, the fluctuations 

 of steam engine speed on various duties, or the variations of 

 current or potential in electrical transmission of power or light. 



