1837.] 



Self-Registering Barometer, 



65 



No. 14, plate 15), from the pencil rod, because as the beam in rising 

 above the horizontal or going below it, describes an arc, this rod with 

 its weights, is moved more or less from the perpendicular, when both 

 will have a tendency to bear on one of the guide rods YY, Supposing 

 therefore the rod/5 to terminate at the scale pan T, the register appa- 

 ratus below may be moved further to the right. From this the pencil 

 rod may rise to the beam, which of course must be prolonged, in 

 order that the points on which the rod bears may be perpendicular 

 above the register roller. The disadvantage of the pencil rod deviating 

 from the perpendicular still remains in this arrangement : — but the end 

 of a beam two feet in length, moving two inches upwards or downwards, 

 deviates from the perpendicular only about quarter of an inch. Further, 

 in the tropics the ordinary range of the barometer rarely exceeds a quar- 

 ter of an inch (Daniell on Meteorological Essays), and during the storm 

 that occurred at Madras 30th October 1836, which must be considered 

 as an extreme case, the range of the barometer (as observed at the 

 Madras Observatory, and recorded Madras Journal No. 14, p. 211) was 

 only 1.135 inch ; now, as the pencil rod will be very light, it is possible 

 that the deviation under consideration, may occasion so very little dif- 

 ference between its absolute weight and its weight as bearing on the 

 beam, that the objection thence arising may be more a theoretical than 

 a practical one. It may, however, should the latter obtain, be completely 

 avoided by placing an arc at the elongated end of the beam, the radius 

 of which is the distance between it and the edge on which the balance 

 turns. The pencil rod may be attached to this arc by means of a piece 

 of platinum wire fixed to the top of the latter, which is of such fine- 

 ness, that, while sufficiently strong to support the pencil rod, it is so 

 flexible as accurately to adapt itself to the arc. It is evident that by 

 this means the pencil rod will be moved up and down perpendicularly, 

 as it will always be a tangent to the arc when this moves. If the arc 

 be a perfect segment of a circle, the weight of the pencil rod will act 

 with a uniform power on the balance, which of course it will not do if 

 the arc be npt such a perfect segment — and as there may be some diffi- 

 culty in making this, it becomes a question, which of these two modes 

 of attaching the pencil rod to the beam is the less objectionable. 

 Should both be found practically defective, the pencil rod may be laid 

 aside altogether, and the Mordan pencil furnished with its delicate 

 spring, as mentioned in the original plan, be attached direct to the end 

 of the beam at right angles to this. Thus situated, however, the pen- 

 cil will require a plane surface on which to make the register, for in 

 the case of the cylinder the pencil is a tangent to it only in a straight 

 line, while the pencil, in the position stated, will describe an arc, as the 

 beam moves up and down — consequently the hour lines instead of 

 being straight as in the cylinder, must on this plane surface be seg- 



