EXAMINATION OF WATER. 435 



is derived from a single cell this figure then represents the 

 number of. bacteria present on the plate examined. If this 

 plate is made from the tube to which 1 c.c. of water was 

 added then the results are expressed, at once, as so many 

 bacteria per c.c. 



In case the plate is made with the gelatin that re- 

 ceived c.c. of water the result is multiplied by 2. In 

 order to express the results obtained with the plate con- 

 taining 1 drop of water, it is necessary to know how many 

 drops the particular pipette employed will discharge from 

 1 c.c. of water. Drops vary a great deal in size but, on an 

 average, 1 c.c. will yield 20 drops. As indicated above, the 

 results should always be expressed as so many bacteria 

 per c.c. 



Wolffhiigel's apparatus can also be employed to count the number 

 of colonies in a Petri dish. The latt.er, if possible, should be inverted 

 and the ruled plate then brought into contact with the bottom of 

 the dish. The average number of colonies per square are determined 

 as above. Since each square is actually 1 cm. square, it is necessary 

 now to determine the number of square centimeters covered by the 

 gelatin in the dish. The area of a circle is TT R ? . Hence, on multi- 

 plying the square of the radius by 3.1416 we obtain the area of the 

 circle expressed in square cm. This, multiplied by the average number 

 of colonies per square, will give the number of colonies on the plate. 



Several modifications of the above apparatus have been devised 

 especially for use in connection with Petri dishes. That of Laf ar 

 (Fig. 58) is widely used. It consists of 5 concentric rings which are 

 divided by 18 radii. Each of the several spaces, thus resulting, has 

 an area of 1 sq. cm. Three of the sectors are still further subdivided 

 in order to facilitate the counting of colonies when very numerous. 

 The Petri dish, unlike the ordinary gelatin plate, is not strictly flat. 

 The center is almost invariably slightly raised, and, as a result, the 

 medium over the center will be thinner than that near the edge. 

 Obviously, the number of colonies in a square cm. over the center 

 will be considerably less than in one farther removed from this point. 

 It is, therefore, not advisable to count at random the number of 

 colonies in a square cm., as in the case of the WoliThiigel apparatus. 

 The counting should be done by sectors. The lines are etched on a 

 glass plate which is fixed in a brass collar. The bottom of the Petri 



