356 
Kepoet of the Botanist of the 
changes of temperature in the room in which the apparatus stood. It 
is believed that if it were placed in a room having at the most only- 
five to ten degrees of fluctuation, there would be no further hindrance 
in securing absolutely satisfactory results. 
Weed Statistics. 
The problem of the weeds is one that every cultivator finds suffi- 
ciently perj)lexing at all times. It is a question that must have an 
answer of some sort, and the answer that can stand the test of practi- 
cal application, giving upon trial the least loss and the greatest gain, 
is the best one. The statistics which are given in this connection are 
presented with the hope of adding a little to the .knowledge of the 
subject, and to the elucidation of the problem. 
The number of weeds of different kinds growing upon one-twentieth 
of an acre was determined by the same methods used last year and the 
year before.* This year, as in the preceding years (although not so 
stated in former reports), the land was plowed and harrowed in 
May, and then left undisturbed for the remainder of the season, except 
the small disturbance of the surface in pulling up the weeds at the 
time of counting. The plowing did not take place until after many 
seeds had begun to grow. This probably does not materially affect 
the results, as the seeds which finally grew are those turned up by the 
plow, while those that were at or near the surface were mostly buried 
too deep to grow during the current season. 
* Report for 1885, p. 262 ; 1886, p. 298. 
