INTEODUCTION. XXXVll 



observations during the years 1813-1819, the mean rainfall was 

 25-46 inches. The greatest fall was 32-37 inches in 1816. (Howard, 

 Climate of London, vol. i. pp. 100-101.) At Greenwich the rainfall, 

 deduced from 28 years' observations, was 25-4 inches. 



^ One-third, or 33 per cent, above or below the average, is usually the 

 extreme difference. All years may fairly be set down as wet or 

 dry in which the fall is 20 per cent, either in excess or in defect 

 (Symons, p. 55). 



The two preceding tables were kindly supplied by Mr. Glaisher. 

 They give the most complete information obtainable up to the 

 present time for the climate of London. The table of annual mean 

 temperatures, extending over all but a century, is the result of the 

 elaborate labours of Mr. Glaisher in combining the registers kept 

 at the Royal Society's apartments in Somerset House, at Epping, and 

 at Lyndon in Kutlandshire, with those kept at the Royal Observatory. 



III.— DIVISIONS OF MIDDLESEX FOE BOTANICAL PURPOSES. 



Although the county is not extensive it is convenient to divide it. 

 The various localities are more easily seen when grouped under 

 different divisions, and thus the flora of any particular part more 

 readily ascertained. But what is of more consequence, such a 

 method is the only way to ensure a systematic examination of the 

 whole surface, and to express with any accuracy the rarity or com- 

 monness of each species. 



It is evident that a plant occurring in all the districts into which 

 a given tract of country may be divided is more widely diffused in that 

 tract than another found in any number of such districts less than the 

 whole ; so far such a plant may be said to be more common than the 

 other. But it may happen that the former species occurs in but one 

 or two spots in every district, whilst the latter is 'found in very 

 numerous places in but one district ; so far, then, as that district is 

 concerned, the latter species is the more common. The difference 

 between general commonness and local abundance can be readily 

 expressed by the use of district-divisions, and it will be perceived 

 that the more numerous the divisions, the more accurately (supposing 

 the examination to be complete) can we express comparative fre- 

 quency ; so that if a tract of country could be completely examined 

 throughout its whole area, and as many districts made as there are 

 square yards in it, a statement of the districts in which any species 

 occurred would almost perfectly express its true distribution. In this 



