26 THE ROYAL SOCIETY OF CANADA 



gives the depth of sampHng, and the shaded area showing the depth 

 of frost below this level' of sampling gives a very clear idea to what 

 extent the samples were frozen. For example, where the shaded 

 area below the level of sampling is very extensive the samples were 

 frozen much more solidly and the soil temperature was much lower 

 than in places where the level of sampling touched the frost line or 

 came close to it. 



The field in which thfs experiment was conducted was manured 

 with well rotted manure on November 25th 1916 and plowed about 

 4 inches deep on November 28th 1916. The plot on which the first 

 samples were taken had a very light snow covering and the severe 

 frost during the greater part of December 1916 penetrated easily. 

 As large areas of this field were deeply covered with snow comparative 

 samples were taken in these unfrozen sections which showed the 

 same large numbers of bacteria and were therefore not included in the 

 tables. 



The February samples were taken from a plot which had a 

 heavy snow covering from the first, and where the frost had penetrated 

 only three inches. The reduced depth of frost as shown in the diagram 

 for February is therefore not due to milder weather conditions but 

 to a change in locality. The heavy snowfall during the first weeks 

 of February provided an ample snow protection for the whole field 

 and caused the soil temperature to be uniform throughout the field. 



The outstanding feature of this experiment was the exceptional 

 bacterial count recorded about the middle of January as well in frozen 

 as in unfrozen soil. From this date bacterial activity decreased rapidly 

 and about March 1st the lowest numbers of bacteria were observed at 

 all four levels of sampling. During this period the moisture content 

 increased at the 4" and 12" levels and decreased at the 8" and 16" 

 levels. 



During March whilst the soil was frozen at all four levels bacterial 

 numbers increased. The increases were far from regular. Por 

 example, on March 22nd 1917 a sudden increase in moisture, due 

 to the penetration of snow water in the frozen soil layer was accom- 

 panied by a decrease in the bacterial count, but samples taken a week 

 later in exactly the same spot gave again the higher number of bacteria. 

 As soon as the spring thaw had definitely thawed the soil a sudden 

 decrease in the number of bacteria was observed, which decrease is 

 still in progress at this late date (20th May 1917). 



In this experiment so many influences acted simultaneously 

 that it is difficult to draw any definite conclusions regarding the influ- 

 ence of changes in moisture content of the soil on the numbers of 

 bacteria. On January 11th at the 12-inch level of sampling an 



