44 



outlet, and carries .some, water aloiij;- with it. The accuracy of this instrument is very consitlei'al)lc 

 and easily tested ; its management is simple, its construction easy. Previously two ways had 

 been tried. In one case a similar vessel about 15 cm. diameter had been used, having a graduated 

 glass tube on the outside communicating with the vessel. The readings, made with a magnitier, 

 were facilitated by a tioat. A rise of 1 mm. requiring about 17 cub. cm. would have to distinguish 

 one-seventeenth of a millimeter. Without further discussing the ditticulties of this nu'thod, it may 

 be stated that it proved by no means satisfactory. The second way was to allow the water to flow 

 over a broad, short and steep spout with a clean, sharp edge, and then to weigh the water, as is 

 done at present. This gave much better results than the method just described, but it is a slow 

 method ; the water recjuires too long to arrange its level, and any slight sh< tck or disturbance readily 

 vitiates the experiment. 



\pparatus for ili'tiTiiiiiiiiiK speiilic gravity. 



DISCUSSION OF VARIOITS METHODS TO ASCERTAIN VOLUME. 



Having described the mode of doing, the application of the method of measurement by im- 

 mersion for jiorous bodies like wood requires some consideration. The fact that it has been, and 

 is to-day, extensively emidoyed by the best of experimenters is certainly a recommendation, but 

 can not serve as an argument for its accuracy, and it appeared necessary, therefore, to resort 

 to spcM'.ial exi)eriments. Before describing these experiments and their results, a consideration of 

 the objections and their justiflcation from a general standpoint may not prove amiss. The objec- 

 tion of inaccuracy of the apparatus ceases at once with an instrument whose accuracy can be 

 demonstrated any monu^nt, and expressed in grams or i)arts of such, ^ot so with the more com- 

 mon objection that " wood soaks up water while the measuring is going on." This statement 

 generally appears to express two more or less clear notions, one implying a swelling of the wood 

 oil imbibing water, and thus a change of dimensions during the process of the expcnnient; tlu^ 

 other a. soaking of water into the cell cavities anah)gous to the rushing of water into the capillary 

 tubes of plaster of Paris, and thus a decrease of w-ater in the vessel and a consequent inaccura(\\' 

 of the experiment. The correctness of the tirst of these two notions, experiment alone coidd es- 

 tablish ; that of the second necessitates a consideration of general wood structure. A i)iece of pine 

 wood, for instance, resists the passage of water in a radial direction even when under considerable 

 pressure, less so in a tangential direction, and, of course, very much less longitudinally. In a 

 radial direction it is a process of soaking from cell to cell, often, in split wood commonly through 

 the tliick cell walls of the summer Avood, anil even the tnedullary rays, owing to the small radial 



