330 METEOROLOGY AND ALLIED SUBJECTS. 



cision for various phases of many species of plants, so that the present 

 method of investigation can be safely recommended. 



2. Very improbable values of this epoch are recognized at once by 

 large limits of error; only such normal data as correspond to the small- 

 est probable errors have a claim upon our acceptance. 



3. In the deduction of normal data the computations lead to nearly 

 the same value for the dates of budding for all the fifty- five plants. Dif- 

 ferences of two or three days occurred only when few observations were 

 at hand. 



4. Even in the case of fragmentary but longer series of observations, 

 Oettingen's method of computation can be applied, and it is immaterial 

 whether the phase under consideration occurs in the extreme coldest or 

 warmest portion of the year. 



5. The computation of a mean date from a sexies of observations af- 

 fords opportunity for determining the variability in time or the probable 

 error of the average date. 



6. The probable error of the total sums of heat, as deduced from 

 observations of many years, must depend upon the variability of the 

 weather. 



7. The present method of investigating the conditions of budding can 

 perhaps be checked by exiDeriments upon certain species that will de- 

 velop in water at different temperatures. {Z. 0. G. M., XIY, p. 326.) 



In a further review of Yon Oettingen's work on this subject Karl 

 Fritsch states that in 1857 he first compared all the formulae that were 

 then known with the best observations available in the development of 

 the growth of plants and showed that the Boussingault formula has 

 the greatest probability although it was intended especially for annuals. 

 In 1861 he extended this work to a computation of the constants for 

 1,889 different species of plants, for which ten consecutive years of ob- 

 servation were available, the result of which showed that his method of 

 computing his thermal constants must be very near the truth. New 

 formulcTB were, however, proposed by Tomaschek in 1862, and by Kabsch 

 in 1863, Hoffman in 1865, Ziegler in 1867. 



In 1867 and 1869 Linsser attempted a further elaboration of the sub- 

 ject, which, however, was not brought to a satisfactory end at the time 

 of his death. Krassan in 1868 and Koppen in 1870 spoke against the 

 general principle of the temperature summations ; but the important 

 work of Oettingen brings us back to the firm conclusion that the law of 

 a constant quantity of heat is necessary to a given stage of develop- 

 ment has a high degree of probability, and his method of determining 

 the base temperature from which the sums are to be counted seems to 

 be the best at present available. {Z. 0. G. M., XIV, p. 376.) 



Hoffmann, of Giessen, has published additional confirmation of his 

 method of determining "thermal-constants" for plants. He sums up 

 the daily maxima of a thermometer exposed to the full sunlight from 

 January 1, the time of the minimum of plant activity up to the date of 



