474 JOHN W. G O W E N 



material for individual counting. The number of bacteria may be 

 determined by smearing a measured sample of the initial solution on 

 suitable agar media in ordinary bacteriological plates. The separate 

 bacteria then grow into small colonies, and at the end of 24 to 48 

 hours may be counted. Comparison of the numbers of organisms in 

 the untreated sample with the number in the various exposed samples 

 results in estimates of the effect of the X rays. 



Certain plant viruses, as the necrotic type of tobacco mosaic 

 (30,31), may be plated similarly on the leaves of a susceptible plant, 

 e.g., bean seedlings. The solution containing the unirradiated or 

 irradiated virus is spread on the surface of the bean leaf by moistening 

 a small pad of cheese cloth in the virus solution and lightly passing 

 this moistened cloth over the surface of the leaf with one stroke only. 

 More than one stroke is undesirable. At the end of three days to a 

 week, depending upon the kind of test, small necrotic spots will ap- 

 pear on the leaves and may be counted. The number of these spots 

 gives an estimate of the number of virus particles within the original 

 solution. On the face of it the methods seem quite crude, but when 

 a comparison between samples of the same solution is made the results 

 are in good agreement. This gives confidence in the method, although 

 some of its success must seemingly be due to the cancelling out of a 

 variety of errors that could arise in the technique. These results 

 demonstrate the desirability of a careful statistical design to eliminate 

 as many of these errors as possible or to measure their effect and take 

 them into account statistically (see Snedecor, 49). 



One of the main errors is the difference in the genetic and physio- 

 logical susceptibility of different plants of the same species. These 

 differences can be overcome by what is called the half-leaf method, 

 where on one side of the leaf one solution is placed and on the other 

 side of the leaf another solution is inoculated. As the two sides of 

 the leaves are quite similar, both environmentally and genetically, 

 the difference between the two lesion covmts on the two halves gives 

 a more accurate measure of the ratio of virus particles in the test 

 solutions. A great variety of these design techniques have been 

 worked out and should be investigated by anyone who intends to use 

 this material. The design chosen should be as simple and as free of 

 possible technical error as is adequate to do the work. 



With most animal viruses and many plant viruses no known tech- 

 nique is available by which we can plate out individual virus parti- 

 cles. To obtain results for such viruses it is necessary to depen(l 



