178 Plant Tissue Culture 



in making dry weight determinations. In order 

 to reduce the random error, each experiment 

 should include at least 20 replications in each 

 environmental complex, and with this number a 

 difference of at least 30 per cent of the control 

 value must be produced by any given variable 

 before that variable can safely be assumed to be 

 responsible for the difference. Records are kept 

 in standard notebooks of a size to be slipped into 

 the coat pocket. Final results may be presented 

 as tables, graphs, histograms, or models. In order 

 that experiments carried out at different times 

 may be compared, it is desirable to express results 

 in terms of percentages of a control maintained 

 under as nearly as possible standardized con- 

 ditions. 



