BEHAVIOR OF MIXED CULTURES. 73 
In the plant one organism often paves the way for others which complete the 
destruction, e. g., Bacterium campestre and Bact. solanacearum are often followed by 
soft white rots. Some of the latter, however, are able to make their way unaided, 
a fact observed and known to the writer for a white rot of the cabbage as long ago 
as 1896. 
The simplest way of studying the antagonistic action of bacteria is by means 
of crossed streaks on agar or gelatin plates. ‘These may be made either simulta- 
neously, or one after the other has begun to develop. The action of the antagonistic 
organism may also be obtained by letting its products diffuse through a collodion 
sac into bouillon inoculated with the other organism. In practice, the bottom of a 
test-tube is removed and a collodion sac is securely fastened in its place. ‘This tube 
is filled with the usual quantity of bouillon and lowered into a larger receptacle 
(tube or flask), the collodion part being surrounded by bouillon. The inner and outer 
receptacles are now plugged with absorbent cotton, and the apparatus is sterilized 
in the steamer or autoclave. ‘The two tubes are then inoculated simultaneously, or 
the outer one some hours or days after the inner one. (See an interesting paper on 
Antagonism, by Frost, in Jour. Infect. Diseases, Vol. I, 1904, pp. 599-640). Frost 
has also devised two new methods for studying this subject, viz, the divided-plate 
method and the agar-block method. ‘The first is a modification of the ordinary 
streak method. It is managed as follows: A Petri dish is divided into two equal 
parts by means of a glass rod fastened to the bottom with collodion. A tube of 
melted agar is inoculated with the antagonistic organism and poured into one half 
of this plate. Into the other half sterile agar is poured. Streaks of the other 
organism are now made crosswise of the hardened surface. If there is marked 
antagonism there will be-a decided difference in the behavior on the two sides of the 
plate, z. ¢., on the sterile agar as compared with the inoculated. To insure a uniform 
streak the inoculated loop should be swept across one half of the plate, then re- 
inoculated and swept across the other half of the plate. 
The method by agar-blocks consists in substituting agar-walls for collodion 
walls. A sterile 3-cm.-deep Petri dish is poured full of nutrient agar. When it has 
solidified it is cut into rectangular blocks, 1 by 1 by 3 centimeters, using a sterile knife 
and taking all possible precautions to avoid contamination by air-borne organisms. 
A platinum needle is now dipped into a culture of the supposed antagonistic organ- 
ism and thrust into the block lengthwise but not entirely through it. The mouth of 
the needle-track is sterilized and sealed by touching it for a moment with a red-hot 
iron. ‘The head of a small wire nail set into a suitable handle will answer the 
purpose. The block is picked up with sterile forceps and dropped into a tube of 
sterile bouillon, which then may be inoculated with the other organism. More than 
one block and tube should be inoculated, and it is best to test the sterility of the 
outer surface of the agar-block by delaying the inoculation of the bouillon for a day 
or two after the inoculated agar-block has been dropped into place. 
Still another method has been described by Frankland and Ward. ‘They use 
the walls of a Chamberland filter to keep the bacteria separate. Bouillon for the one 
