286 J. ALFRED GRIFFITHS. 



ascertain the absolute scale by rotating T by hand for 10 — 20 — 30 

 etc. revolutions, and moving the drum a little by hand at the 

 end of each ten turns. This eliminates the possibility of error in 

 determining the radius of the sector and friction roller, and also 

 the effect of obliquity of the pen clip. As the length of pause 

 depends solely on the notch at I, it is the same whether the drops 

 take place once per hour or at any of the faster intervals, and 

 may be only one to two seconds when the ten-second cam does the 

 lifting, but the scale used in measuring the diagram must allow 

 for this pause, when its magnitude is a sensible fraction of the 

 whole interval. The special scale of speeds to be used for any 

 particular experiment depends on the time interval and pause. 

 In the illustration the proportions of roller and sector are such 

 that on the absolute scale about one hundred revolutions of T equal 

 one inch. If, for example, the recorder is coupled to the prime 

 motion under test, so that T makes say two revolutions for each 

 one of the prime motion, and the tripping takes place once per 

 minute with two seconds pause, then fifty-eight on the absolute 

 scale will correspond to thirty revolutions per minute of the prime 

 mover, and the vertical width of the paper (five inches) will allow 

 of velocities up to about three hundred revolutions per minute 

 being recorded. 



Two experimental recorders have been constructed by the writer, 

 the first (made in Brisbane in June 1893), had great range of 

 adaptability, and was correspondingly troublesome in adjustment. 

 Its periods varied from four seconds to eight minutes, and there 

 were also two sets of friction rollers. The lines on the diagram 

 could be spaced either at one hundred and forty lines per inch, 

 giving a shaded diagram, or at fifty per inch separating the lines. 

 The second instrument (made in Parramatta in August 1884), was 

 specially modified for anemometry. The differenting period is one 

 hour with a pause of four and a half seconds, which is negligable 

 in comparison with one hour. The drum is driven continuously, 

 and revolves in a week, the paper moving one inch and a quarter 

 per day, the verticals being spaced about twenty per inch. At 



