86 HISTORICAL INTRODUCTION TO CHEMISTRY CHAP. 



barrel, with a lead syphon attached to the end. The gas 

 was carried by the syphon through a vessel of water xx 

 into an inverted flask ab, where it could be stored for any 

 length of time without risk of leakage. 



The great importance of this new design will be realised 

 on comparing it with the apparatus used by Cavendish 

 (Fig. 1 8), by Lavoisier (Figs. 16, 21, 25), and by Priestley 

 (Fig. 20). 



Priestley's apparatus. The apparatus which Priestley 

 used for manipulating gases is described at the beginning 

 of the first volume of his Experiments on Air and illustrated 

 by two plates (numbered and lettered consecutively) which 

 are reproduced in Fig. 20. 1 



The most important feature was the trough, often called 

 a PNEUMATIC TROUGH, shown with all its accessories at i in 

 the first plate. The original trough a was of earthenware, 

 about eight inches deep, with thin flat stones, bb, just below 

 the surface ; afterwards a large wooden trough was used 

 with a shelf fixed an inch from the top. The various gases 

 were stored in cylindrical jars, cc, which could be immersed 

 in the trough, or stood on the shelf, or lifted out in dishes 

 as at 2 2 2. Mice were stored as at 3 in a receiver standing 

 on a perforated tin plate and provided with a perforated 

 cover held down by weights ; in order to test the goodness 

 of a sample of air, a mouse was held by the back of the 

 neck and passed through the water into a tall beer-glass, d, of 

 two or three ounces capacity, in which a mouse could 

 usually live from twenty to thirty minutes. The tapering 

 cork, 4, was used to close or open a small bottle inside a jar 

 of air, whilst the wire stand, 5, served to support small 

 dishes as at /. The funnel, 6, was used to pour gases from 

 a vessel with a wide neck into a vessel with a narrow neck, 



1 In the second plate, the blackboard has been moved and the table 

 altered a little in order to reproduce on as large a scale as possible. 



