108 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



at the height of four feet above the ground. But the great majority of 

 plants do not attain this height ; many rise but a few inches above the 

 ground, whilst, on the other hand, the foliage of trees may be fifty feet oi 

 more above it. Now the fact that dew and hoar-frost are frequently 

 observed on grass and low-growing plants when the temperature by the 

 thermometer in the screen is too high for their formation shows us that 

 the temperature near the ground is at times much lower than it is at four 

 feet above it, and the thermometer on the grass confirms the inference. 

 In the same way we know from direct observation that under certain 

 conditions the air fifty feet above the ground may be warmer than it is at 

 the level of the screen, w^hilst under other conditions it will be colder ; 

 and, therefore, we have to ask how far we can accept temperatures at the 

 height named as suitable for any inquiry into the influence of temperature 

 upon vegetation. 



But notwithstanding the fact that a tree, a shrub, and an herbaceous 

 plant growing near one another may none of them receive exactly the 

 amount of heat which is indicated at a given moment by the thermometer 

 in the screen, yet on the whole it is most probable that the error likely to 

 arise from this cause would be very slight ; and since for comparative 

 purposes it is necessary to adopt a uniform height for all observers the 

 approximate level of the eye is a convenient one to select. Probably it is 

 only under very special conditions that these differences would be consider- 

 able, and by the aid of the grass minimum thermometer it should not be 

 difficult in such cases to correct the screen temperature so as to meet the 

 circumstances of any plant under consideration. 



Another question of greater importance, when considering the influ- 

 ence of temperature upon vegetation, is to determine what temperatures 

 are really useful to plants, and how we are to deal with the observations 

 so as to separate those temperatures at which plant life is active from 

 those at which it becomes dormant, and which may therefore be regarded 

 as of no value to the plant. 



The old plan was to deal exclusively with means, and it was a step in 

 advance w^hen means were prepared for seasons as well as for years and 

 months, and when the extremes of temperature were considered as well 

 as the averages. It is now a long while ago (1750) that Adanson first 

 suggested using the sum of the daily mean temperatures for throwing 

 light upon the relation of temperature to the development of plants. 



Adanson thought the unfolding of the bud was determined by the 

 sum of the daily means of temperature from the beginning of the year, 

 and that view found support for a considerable time in spite of the 

 opposition it received from some well-known botanical physicists. 



Boussingault modified it considerably by calculating the temperatures, 

 not from the commencement of the year, as Adanson had done, but 

 from the time at which the plant under investigation began to grow ; 

 and having found the mean daily temperature for the period between the 

 commencement of growth and its completion, he multiplied it by the 

 number of days in the period to get the required sum of the temperature. 



The conclusion at which Boussingault arrived as the result of his 

 investigations was that the period required for the growth of a plant 

 was inversely proportional to the mean temperature. 



