ON THE SUPPOSED NORMAL OCCURRENCE OF BACTERIA IN PLANTS. 25 
which the work was carried on. An examination of the contents of the fertile tubes also 
confirmed this view. 
Fernbach took vegetables just as they came into the market without inquiring where 
grown, assuming that any soil adapted to vegetables was rich enough in micro-organisms 
so that they would penetrate into vegetables if this was physiologically possible. He 
tested potatoes, carrots, turnips, beets, and tomatoes, judging it not worth while to try 
additional sorts, since experiments on all of the above showed the conclusions of M. Galippe 
to be erroneous. For culture-media he used neutral veal-bouillon and sugared turnip-water 
which was very slightly acid. These media were in test-tubes and Pasteur flasks. 
The surface of the vegetables was first heated to light carbonization by means of a 
thermo-cautery. The tomato juice was aspirated out by means of a pipette drawn out at 
oneend and plugged withcottonatthe other. The other vegetables were punched with brass 
cork-borers, cotton-plugged aboveand sterilized inside of cotton-plugged test-tubesat 165° C. 
I obtained with these tubes cylinders of vegetable tissues which I pushed out little by little, 
and which were sowed immediately by sectioning them with a flamed scalpel. I thus introduced 
into each tube a volume of tissue varying from 0.5 to 1 cc. 
A summary of Fernbach’s results is given in the following table: 
| Number of sow- 
- Number of vege- Number of . : 
Vegetable. tables sampled. sowings. ings ao were 
Tomatoes......... 26 | 52 2 
PRULBIOR 5 he = 5.4 Sal 36 199 19 
NEAEROUS SH os) ec es 13 101 4 
DOCS 8 ait acie sos 12 103 10 
Potatoes........... II | 100 fe) 
Totals a.5:5%: 98 555 | 35 
We see that there are a certain number of fertile sowings, 6.3 per cent. It could scarcely be 
otherwise. As a matter of fact, in experiments so delicate there are several sources of error which 
it is impossible to remove absolutely. The most important is that which arises from germs of the 
air. These vary greatly in number but are always abundant in a laboratory where there are goings 
and comings and where one is constantly exposed to currents of air. The practice of filling Pasteur 
flasks almost daily shows me that, in the laboratory where I have made my experiments (Sorbonne), 
out of 100 flasks filled there are always 4 or 5 which show growth. 
The organisms which developed in 6.3 per cent of the cultures were of various sorts— 
bacilli, micrococci, molds, etc., in general each tube being occupied by a single sort. A 
source of error also to be considered is the possible penetration of germs into vegetable 
tissues as the result of insect depredations, or of injuries due to digging or transportation. 
We conclude, therefore, that normal vegetable tissues constitute a perfect filter for microbes 
and that they can be invaded by them only as a result of causes wholly accidental (p. 570). 
In 1888 A. di Vestea repeated Galippe’s experiments. This author experimented under 
much the same conditions as Galippe, 7. ¢., with plants obtained from the market-gardens 
around Naples and grown on the lowlands where the filth of the city is dumped. He tested 
especially a variety of lettuce called Roman lettuce. In sampling he made use of a special 
glass-punch, consisting of a thin tube through which slid a thick-walled tube longer than 
the first, closed with cotton at the free end, and sliding in a plug of cotton. The whole 
apparatus was sterilized at 150° C., after the lower half had been introduced for protection 
into a test-tube plugged with cotton. The plant to be tested was then cut with a flamed 
knife. The punch was removed from the test-tube and the cutting end of the outer thin-walled 
tube was then plunged into the cut surface of the vegetable and a piece of tissue removed. 
The apparatus was then replaced in the sterile test-tube, the tissue was pushed out of the 
cutting tube by means of the inner tube, which was then used to crush it in the bottom 
of the test-tube. Finally, some of the liquid which resulted from the crushing was sucked 
