Description of an improved Hygrometer. 323 



The principal objection to the hvgrometer described in 

 my former paper *, arose from the necessity of shortening 

 the beard of the Ooheena hooloo, in order to reduce the scale 

 to a convenient length ; this was to be obviated only by 

 giving the instrnment a circular form, and invcntkig some 

 mode of ascertaining, without difficulty, the number of re- 

 volutions made by the index. 



A, B, C, D, (fig. 1 . Plate IX. )t is a frame made of small 

 square bars of brass or silver; this frame is soldered to a 

 square plate B,E, the edges of which are turned up, as re- 

 presented by the dotted lines, to secure the ijidex from in- 

 jun,'. On the face of the plate is engraved a circle (see 

 fig. 4.) which is divided into one Immlrcd equal parts : three 

 holes a,i',c, (fig. i .) are made through the frame and plate 

 in the same direction; the holes a and Z' are of a conical form 

 as representfd by the dotted lines, and are highlv polished 

 to lessen friction : tb.e hole at c receives a screw, one end of 

 which is tapered, and has a notch cut in it with a fine saw, 

 which may be closed by means of the sliding ring d. 



The axis e,f, is made of silver wire, very smooth and 

 straight, and of the size of a large knitting-needle; on the 

 axis a screw is formed by twisting a smaller silver v,'ire tightly 

 round it from left to r.iglit. This screw should be fourteen 

 or fifteen threads in length. The end of the axis, y", is di- 

 vided, and is to be closed by a small sliding ring. As this 

 is the most important part of the hygrometer, fiir. 2. repre- 

 sents it on an enlarged scale, 



A loop and drop (fig. 3.) is made of fine gold wire, of, 

 Ruch a size that, v. hen suspended on the screw, it luay slide 

 along it with perfect frecdoni by means of the revolution of 

 the axis, but not escape from one interval to another by anv 

 other motion. Sliould the ^oop, on trial, be found too 

 large, (as indeed it ought to be,) it may easily be closed a 

 little by placing it on the screv/, and pulling it.gcntly by 

 the dro|) ; it will then asnnne an cl'-plical form, as in the 

 figure. This loop is iiUendpd to regii'ier the number of re- 



* The (lojcription of an llyc;iomct.?r mad? of tiie s.iinr^ materi;il, bjt of a 

 difTerciU and inCciiof r. nsirucliyn. 

 f 'Jiic plait- heie .■•efcrryd to will be (j'.frcn in oit j)exl Number. 



X Li volutions 



