I 



432 University of California Fublications in Agricultural Sciences [Vol.3 



Greenhouse Data 



There are objections to all greenhouse work due to somewhat un- 

 natural conditions for the usual indicator crops, the lack of a normal 

 water supply, the small amount of root space, etc. Crowding of the 

 pots is also apt to cause variations. Even the slight change in the loca- 

 tion of a pot on the bench will affect the growth of plants, as some of 

 the elaborate precautions for moving the pots daily, and in a given 

 order, testify. The outstanding advantage of greenhouse work is that 

 with a given indicator crop a group of soils, or soil conditions, may be 

 compared under very similar conditions. 



In the present case, the leaks in the sash allowed rain water to fall 

 into some of the pots to a considerable extent. The pots so affected 

 showed a poorer growth in the cases of the heavy Altamont and 

 Diablo samples, where the soil was readily compacted, while in the 

 poor Hanford and San Joaquin soils the pots receiving leakage water 

 showed markedly better growth. 



To minimize such errors, as much as possible, triplicates were 

 used, as above explained, besides repeating the series. In working 

 out the final averages of the crop it was suggested that a selection be 

 made of the crop dry weights, in case that there was a marked varia- 

 tion between the triplicates, using the two weights close together, and 

 excluding the third if it were widely divergent. However, when one 

 begins to select certain figures from a series, and bases comparisons 

 upon these alone, there is apt to be the tendency to select those figures 

 that will prove the point in question, unless there is some known dis- 

 turbing factor causing the divergence and which warrants the exclu- 

 sion of certain figures. 



Other cases that are rather hard to deal with are those in which 

 the number of plants reaching maturity was not up to the standard to 

 which the series was thinned when the plants were young. This fail- 

 ure may have been due to poor germination, or to accidental destruc- 

 tion of the plants during growth. Sometimes less than the standard 

 number of plants will give a much greater dry weight per plant than 

 the normal number. It was not deemed advisable to use the weight 

 per plant, but rather to use the total dry weight of the crop, and only 

 consider of value the series in which the number of plants per pot 

 was practically constant. 



In the greenhouse work the Diablo clay adobe, the Altamont clay 

 loam, and the Hanford fine sandy loam samples were compared by 



