1838.] 



a7id Contraction of Wood. 



113 



9. — The whole of these woods were straight grained, and free from 

 curls or knots; they were planed up in slips or rods one inch wide and 

 two-tenths of an inch thick : and were cut to the length of three feet one 

 inch, having a piece of brass for an index rivetted on them, distant 

 three feet from one end. 



10. —The following account will explain the method I adopted to 

 ascertain the variation of length in these rods occasioned by the hygro- 

 metric changes of the atmosphere. 



11. — I drove into the interior wall of a dry, well sheltered bungalow 

 two stout iron spikes about three feet apart ; the bottom spike had its 

 upper surface slightly rounded, and the top spike had a slit cut in 

 it, and a piece of silver inserted, as shewn at B, figure 2, and 

 across the slip of silver an index line, or mark, was made at ex- 

 actly three feet distance from the surface of the lower spike. The 

 brass plates on the rods, as shewn at A, fig. 3, had index lines made 

 on them at the outset of the experiment, exactly corresponding with the 

 distance from the surface of the bottom spike to the line on the silver 

 in the top spike ; that is, each rod was placed in position, and a fine 

 line was then made on the brass, and the length of the rod was adjust- 

 ed till the line exactly coincided with the index line on the silver. A 

 stop and spring, a, b, fig. 3, were fixed near the lower spike, in order 

 to ensure the rods always fitting exactly in the same place, and another 

 spring c pressed upon the face of the rod to keep it flat against the 

 wall, which was a perfectly plane surface : for in reading off microme- 

 trical measurements to a minute fraction of an inch, some care and 

 nicety are necessarj^ 



12. — it is only requisite to add in further describing the apparatus 

 employed, that a micrometer was fixed over the upper spike so as to 

 read off any variation that might occur between the lines made respec- 

 tively on the brass plates fixed on thai rods, and the immoveable index 

 line upon the silver. One division on the head of the micrometer screw 

 read off to g-g-Vo- P^^'^ of an inch, and the error of the screw in twenty- 

 five full revolutions of the micrometer head did not exceed the To,io o" 

 part of an inch. The readings therefore may be depended on with 

 safety to the ^-^^-^ part of an inch, which is greater accuracy than I re- 

 quired for the purpose I had in view. The wall of the bungalow being 

 quite dry and well sheltered, I should apprehend that the quantity of its 

 expansion or contraction from the humidity or the dryness of the at- 

 mosphere, would not be sufficient to interfere with the result of these 

 experiments j however if it be thought that some allowance should be 



