June 19, 1885.] 



SCIENCE. 



503 



instrument needed is a round burette, tapering at 

 both ends, and graduated so that the mark 100 is 

 exactly at the lower orifice, and the mark a few 

 centimetres below the upper (fig. 7). On each ex- 

 tremity is placed a rubber tube closed by a spring. 

 Before using, I disinfect the apparatus by passing 

 sulphurous acid through it, and then attaching it to 

 a Papin's pot filled with water at 110° C. for an hour. 

 In fifteen minutes all trace of the sulphurous acid 

 has disappeared, both rubber tubes are closed, their 

 ends plugged with sterilized 

 flax, and the burette left to 

 cool. In cooling, a perfect 

 vacuum is produced in the 

 tube. I affix to the lower 

 rubber tube a pointed canula, 

 sterilized at the time by a cur- 

 rent of steam or the flame, 

 introduce it into a flask of 

 bouillon kept for three or four 

 weeks at 30° C. (to prove its 

 complete sterilization), open 

 the lower spring, and the bu- 

 rette is filled immediately: the 

 fluid is allowed to rest at the 

 mark 0. 

 The apparatus being thus prepared, a very dilute 

 portion of the fluid, or a small piece of the substance 

 containing the organisms to be separated, is intro- 

 duced at the top of the burette. The dilution must 

 be great; for the contents of the burette can only b<; 



Fig. 8. — Glass cul- 

 tcre-flask. 



FlG. 9. — "WlESNEGG STOVE. 



distributed among twenty-five tubes, and more than 

 two-thirds of these tubes must become inoculated in 

 order to the success of the experiment. If it be 

 desired to determine the number of germs in a given 

 specimen of water, put a very minute quantity into 



the burette filled with sterilized bouillon ; mix the 

 two thoroughly, thus obtaining an equal distribution 

 of the germs; introduce the canula of the burette, 

 immediately after heating, through the plug of a 

 sterilized tube; allow four cubic centimetres of the 

 fluid to flow from the burette ; and so on for twenty- 

 five tubes. 



If all the tubes become cloudy, it is because the 

 amount of water used was too great; and this amount 

 must be reduced until only a portion of the tubes 

 show any sign of life. With water full of bacteria, 



Fig. 10. — Chamberland's filter. 



the quantity to be used is too small for exact measure- 

 ment; and then two burettes are used, the one full 

 of water, and the other of bouillon. A drop of the 

 suspected water is placed in the first, and then this 

 dilution is used with the bouillon, as before. A simple 

 arithmetical calculation then gives the approximate 

 number of germs in a given quantity of water. 



The first experiments of Pasteur and Tyndall were 

 imperfect in method ; Miguel used a flock of cotton 

 in a glass tube, through which he filtered the air, and 

 then washed the cotton in sterilized bouillon : my 

 improvement is to substitute a powder, soluble in the 

 nutrient fluid, for the cotton; and for this purpose I 

 use common salt, w r ell sterilized. This salt may be 

 turned into the burette, and the calculation made as 

 before. 



The results obtained by fractional culture are at the 



