THERMAL RELATIONS. 79 



They cost 9 marks each when ordered direct from Berlin, and can be had without 

 delay. Good American thermometers are made by Henry Green, New York. 



With this open bath it is easy to keep the range of temperature down to 

 o.i to 0.2 of a degree, and the writer has frequently exposed tubes for ten minutes 

 without appreciable change in temperature. Temperatures may be read easily to 

 o. i degree by means of a Zeiss aplanat lens magnifying six times (fig. 25), 

 and should be recorded for each half minute during the exposure. Under 

 no circumstances should exposures be made iu water which is not agitated. 

 Of course, for accurate reading the eye and the center of the lens must be 

 level with the top of the column of mercury. The lens may be supported 

 at the proper level on a grooved piece of cork. If possible the thermom- 

 eter used should be compared with some standard instrument. If not, it 

 should at least be compared with several other good thermometers in the 

 same laboratory. The test-tubes are supported by perforated corks thrust 

 into holes bored through a rectangular piece of hard, heavy wood. 



The writer formerly made use only of the first four tests. It seemed 

 hardly worth while to recommend that all bacteria be tested for the killing 

 |i i effect of cold, so long as we had nothing but the inconvenient and more 



or less inexact methods of salt and pounded ice or of ether and frozen 

 CO* ; but now that liquid air may be obtained at a small price in many 

 of the larger cities, can be shipped long distances, and can be used with 

 so little inconvenience, there is no good reason why the effect of freezing 

 should not be determined in all cases, since in some instances it is likely 

 to prove a valuable means of differentiation. The bacteria may be exposed 

 iu 5 cc. portions of distilled water or bouillon in block-tin test-tubes, or 

 preferably iu tubes of resistant glass, for standard periods, e. g., one-half 

 hour, i hour, 6 hours, 1 2 hours, 24 hours, 48 hours, etc. They may also 

 be exposed to alternate freezing and thawing every fifteen minutes or 

 thirty minutes until all are dead. To avoid endospores, the depressing 

 effect of by-products, etc., young cultures should be used, and, of course, 

 all should be of the same age and grown in the same medium, /. e., 

 bouillon cultures 24 hours or 48 hours old. The tests should be quanti- 

 tative rather than qualitative. They may be made as follows : Into 5 cc. 

 of sterile water or standard bouillon a carefully-measured quantity, i. e., 

 one loop, 5 drops, *^ cc., etc., of the culture is placed, stirred very thor- 

 oughly, and allowed some time for diffusion. To avoid zoogloeae, which 

 form early in some species, and to reach more uniform measurements, 

 it is recommended to take the loop from a bouillon culture rather than 

 from agar or other solid media. After sufficient time has elapsed for 

 uniform diffusion, six Petri-dish poured plates are made from each of the 

 inoculated tubes. The plates should be of the same diameter (area of 60 sq. cm.). 

 The amount of agar used for each plate should be 10 cc., and the amount of infec- 



*Fic. 63. Anschiitz normal thermometer with degrees divided into fifths (Centigrade scale). 

 For use in thermal death-point tests. About three-fourths actual size. 



