However important the quantitative treatment of a problem may be, the quality of the initial data remains solely decisive for the validity of the results, and these data must never be at variance with the facts observed in Nature.

Volcanoes and Their Activity, Conclusion (p. 290), John Wiley & Sons. 1962