182 Report of the Department of Bacteriology of the 



Two field plats, in Dunkirk clay loam, 8 about twenty feet apart, 

 and both cropped to millet, were sampled nearly thirty times during 

 the two years. The counts are listed in Table I, taken from the 

 second of the two articles already mentioned. These results are 

 also plotted in Graph V. In this graph the relations between the 

 bacterial count, the moisture content of the samples, and the average 

 weekly temperature are shown. 



The following points were brought out by these analyses : (a) The 

 highest counts were all made while the soil was frozen. Out of 

 seventeen counts made from frozen soil or from soil recently thawed, 

 all were over 10,000,000 and only four under 15,000,000 per gram; 

 while of the forty other samples only fourteen were over 10,000,000 

 and but five over 15,000,000. The highest count in unfrozen soil 

 (22,000,000) was exceeded by seven of the winter counts. (6) During 

 the winter the numbers of bacteria increased while the soil was well 

 frozen, but tended to decrease when it thawed, (c) In general, 

 increases and decreases in the numbers of bacteria accompanied 

 rises and falls, respectively, in the moisture content. In January 

 and February, 1911, however, a series of fluctuations occurred which 

 seemed to be closely associated with the freezing and thawing of 

 the soil, but which were plainly independent of changes in moisture 

 content. 



This relationship between the number of bacteria and the soil- 

 moisture suggested that the increase in germ content during the 

 winter might be due to the greater moisture content rather than 

 to the difference in temperature. There are two ways in which 

 this may be possible: the added water, even though frozen, may 

 furnish better conditions for bacterial growth; or, provided this 

 increase in moisture is due to a capillary rise of water from lower 

 depths during freezing, it may carry bacteria up with it, thus increas- 

 ing their numbers in the surface soil without actual multiplication. 

 To test out this point a pot of soil was kept frozen with a constant 

 moisture content of 40 per ct. The experiment was not carried on 

 under the most satisfactory conditions, because warm weather made 

 it necessary to employ a freezing mixture of snow and salt, which 

 resulted in a much lower temperature than that usually found in 

 the field. A further difference from field conditions was caused by 

 the aeration which necessarily results from the ordinary method of 

 filling a pot with soil. 



The results of this pot experiment are summarized in Table II. 

 They are inconclusive. An increase in germ content is shown while 

 the soil is frozen ; but it is a much smaller increase than that observed 

 under field conditions, and is not followed by a decrease after thawing. 



8 The soil nomenclature of the Bureau of Soils has been used throughout this work. 

 The soils mentioned are described in the Soil Surveys of Ontario and of Tompkins 

 Counties, New York, published by this Bureau. They are all glacial lake-bottom 

 soils; see Fairchild, New York State Museum Bui. 127, pp. 06. 1909. 



