OF PERIODIC PHENOMENA. 357 



seemed to require ; for, in reference to this last matter, 

 it may still be of some service. Though it cannot be 

 directly compared with the results of any one year's 

 observations made upon the plan which it is proposed 

 hereafter to adopt, yet, according to what was 

 before stated (12.), it must always have a certain 

 value, so far as it assists in determining the mean 

 date of any periodic phenomena in a given district. 

 To fix the date of an occurrence in different locali- 

 ties with reference to a given season, the observa- 

 tions must be made simultaneously, or during the 

 same year ; but to fix it in reference to climate 

 generally considered, and independent of the varia- 

 tion caused by seasons of different character, they 

 may be made at any time, provided only they are 

 continued a sufficient number of years to eliminate 

 the error arising from this source. A mean date 

 thus obtained will always be available in any ques- 

 tions of climatological research with which it is con- 

 nected. It stands ready for use when wanted ; and 

 against the time when dates of the same phenomena, 

 similarly determined in other places, may be mar- 

 shalled along with it. Thus, for example : M. Que- 

 telet has observed that the time of the lilac's flower- 

 ing at Brussels is the 1st of May :* here then is a given 

 fact always ready for embodying with others, when ob- 

 tained, to unite in forming a synchronic line in refer- 

 ence to the flowering of that particular plant, or, as 

 we have called it, an isanthesic line (5). In like 

 manner, the mean time of any other periodic pheno- 

 menon, when once determined for a given locality, 

 * Instructions, &c. p. 2. 



