26 FREDM. UBER 



to statistical factors may indicate in advance that the execution of the 

 program as originally planned is not likely to yield meaningful results. 

 Assistance in designing experiments may be obtained from the several 

 books listed in the bibliography (7-11) and from professional stat- 

 isticians. 



Experimental work should be planned on a scale adequate to yield 

 unambiguous results. Halfway attempts are uneconomical, often 

 quite futile, and may be misleading. In exploratory investigations, 

 the planning usually is not elaborate, but even preliminary experi- 

 ments should be conducted in an efficient manner. 



2. Control Environmental Factors 



Adequate quantitative control of the various environmental con- 

 ditions constitutes the most difficult part of many biological experi- 

 ments. While it may be relatively simple to hold one or two condi- 

 tions constant at any one time, the problem rapidly grows in com- 

 plexity when several factors must be controlled simultaneously. 

 Since the potentially variable factors are numerous, it is rarely pos- 

 sible to regulate all of them in a completely satisfactory manner. The 

 alternative procedure of employing a statistical method of control, 

 discussed in some detail by Fisher (7), should not be overlooked; it 

 may enable a simultaneous evaluation not only of the various en- 

 vironmental factors separately but also of their mutual interactions. 



Since it is almost never possible to control all the physical and 

 chemical factors to the extent that one would like, it may prove de- 

 sirable to select and control the particular environmental influences 

 likely to exert the greatest effect on the results of a given investigation. 

 Examples of physical factors could include such items as tempera- 

 ture, humidity, light intensity, electrical and acoustic fields, static 

 pressure, and perhaps still others. With regard to chemical influences, 

 it will probably be necessary to control the chemical composition of 

 ,the medium including its hydrogen ion concentration. Since it is 

 well known that biological systems may be very sensitive to exceed- 

 ingly small quantities of essential elements or compounds, the control 

 of the chemical environment often constitutes a limiting factor in an 

 investigation. 



Great care is required to insure the point-to-point uniformity of a 

 particular environmental condition throughout a system containing 

 biological materials. The mere installation of a humidistatically 

 regulated container, for example, does not necessarily guarantee that 



