224 THE I'LOATING-MATTEE OF THE AIE. 



differentiating itself by thick turbidity from a number 

 of perfectly pellucid neighbours, the yielding of the flask 

 being immediately traced to the fracture of its sealed 

 end. 



With the view of showing how readily, unless ex- 

 treme care is taken, contamination may enter her- 

 metically-sealed vessels, the following experiment was 

 made on the 6th of December. Four retort-flasks 

 were charged with cucumber-infusion, boiled for the 

 usual time and sealed, not during the outrush of the 

 steam, but a moment after ebullition had ceased. On 

 the 9th of December three of these four flasks were 

 faintly but distinctly turbid. The reason is obvious. 

 On the cessation of the ebullition, a momentary conden- 

 sation of the steam above the infusion caused an 

 indraught — slight, no doubt, but still sufficient to 

 contaminate or vivify the infusion. 



The source of the contagium was also indicated by 

 the following experiments. A large number of retort- 

 flasks, embracing infusions of snipe, wild duck, partridge, 

 hare, rabbit, mutton, turbot, salmon, whiting, mullet, 

 turnip, and hay, had remained over from my stock of 

 1875-. After a year's exposure to the temperature of 

 our warm room not one of these flasks showed the 

 slightest trace of turbidity or life. On the 7th of De- 

 cember the sealed ends of forty of them were snipped 

 off in the laboratory. Five days afterwards twenty- 

 seven of them were found swarming with organisms — a 

 considerably higher percentage than that obtained by 

 the same process in the same laboratory a year pre- 

 viously. 



It is needless to dwell with any emphasis on the 

 obvious inference from all this, namely, that the conta- 

 gium is external to the infusions, that it is something 

 in the air, and that at different times we have different 



