470 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 3 8 



installation by the addition of heating devices. Anyone who desires 

 to test the influence of heating the culture solution should make com- 

 parisons of plants grown under exactly similar conditions, except for 

 the difference of temperature in the solutions. 



COMPARISONS OF YIELDS BY SOIL AND WATER CULTURE 



The impression conveyed by most of the popular discussions of the 

 water-culture method is that the inherent productive capacity of a 

 given surface of nutrient solution far surpasses that of an equivalent 

 surface of soil, even under the best soil conditions feasible to maintain. 

 Often quoted is the yield of tomato plants grown for a 12 months' 

 period in a greenhouse water-culture experiment in Berkeley. 9 This 

 yield is compared with average yields of tomatoes under ordinary field 

 conditions, and the yield from the water-culture plants is computed 

 to be many times greater. But closer analysis shows that erroneous 

 inferences may be drawn from this comparison. Predictions con- 

 cerning yields in large-scale production are of doubtful validity when 

 based on yields obtained in small-scale experiments under laboratory 

 control. In any event, there is little profit in comparing an average 

 yield from unstaked tomato plants grown during a limited season 

 under all types of soil and climatic conditions in the field, with yields 

 from staked plants grown in the protection of a greenhouse for a full 

 year. Evidence has long been available that yields of tomatoes 

 grown in a greenhouse, in soil, can far exceed yields obtained in the 

 field. It is true that in one series of outdoor experiments, the yields 

 of tomatoes under water-culture conditions were reported to be much 

 higher than under ordinary field conditions, on a unit-surface basis; 

 but again, the general cultural treatment of the plants (especially 

 with regard to spacing and staking) was so different that comparisons 

 of yield are of very limited value. Furthermore, the equipment for 

 an acre of water-culture plants would be very costly, and technical 

 supervision of the cultures and the labor of staking vines would neces- 

 sitate large and as yet unpredictable expenditures. 



A real test of the relative capacities of soil and water-culture media 

 for crop production requires that the two types of culture be carried 

 on side by side, with similar spacing of plants and with the same 

 cultural treatment for plants grown in soil and water culture. The 

 soil should be of suitable depth and have its nutrient supplying power 

 and physical condition as favorable for plant growth as possible. 

 We initiated an experiment of this kind in Berkeley late last summer, 

 with the tomato as the test plant. The experiment has now been 

 carried on over a full year, and several of the conclusions derived 

 from it warrant emphasis. The yield of tomatoes grown by the usual 



» Qericke, W. F., Crop production without soil. Nature, vol. 141, pp. 636-540, 1938. See also the article 

 cited in footnote 8, p. 469. 



