266 



Mvorao-o of ten years there is no appreciable difference between the 

 results. 



Duration, in dayft, fmm sowing to heading of winter wheat, at Montsouris, France. 



This table shows that on the average of ten years the seed that was 

 sown, e. g., on the 27th of October and required one hundred and 

 fifty-five days to head, is that which took the longest time; for sow- 

 ings before that date, as well as after it. the durations steadily 

 diminish; in other w^ords, this sowing is that whose development 

 was the most retarded by the winter cold. If we compare this table 

 with those given by Marie-Davy, showing the frosts, we find a com- 

 plete inversion in the chances of injury from frost ; wheat as a green 

 plant has as little to fear from frost as has the dry grain. But 

 during and after the formation of the embryo seed, as well as during 

 germination, on the contrary, frost is very injurious, and if the 

 embryo is seized by frost it perishes. If this accident occurs it is 

 possible that the progress of heading may permit a new formation 

 of embryo to replace those which have perished. Such accidents 

 must have occurred to the seed sown in the hope of reaping an early 

 harvest in 1874, 1875, 1876, 1877, 1878, and 1881, but did not occur 

 in 1882. This accident is not incompatible with an excellent harvest, 

 as we see in the case of 1874, but it causes a decided retardation of 

 the harvest, as in 1877. The mean of the ten years shows that the 

 heading occurs at an epoch in the spring when the mean temperature 

 of the air is between ()° and 13° C, and when the rainfall is generally 

 iibundant. so that at this epoch damage does not generally occur to 

 the grain; only in case of the sowing of September 20, 1878, did the 



