Studies in Phcenology 35 
In regard to other conditions it appears that “the rainfall was 
decidedly less than normal except in west Scotland.” Indeed the 
summer deficiency in rainfall is 5-42 inches, and “the total is only 
19 per cent, of the average.” The character of the rainfall is strikingly 
expressed as being “ the smallest in the last 105 years.” It is interest¬ 
ing to find that the next lowest rainfall occurred in 1918 when the 
flowering dates were decidedly eaily but not as strikingly so as in 1921. 
It is not surprising that, with a temperature above normal, the 
flowering dates of 1921 should be early. The connexion of low rainfall 
with early flowering is probably to be ascribed partly to the quicker 
warming up of the drier soil in the spring, so that early growth would 
be promoted through the higher soil temperature, provided sufficient 
water is present for the needs of growth. In regard to early flowering, 
as such, it is well known that individual plants growing in spots where 
the supply of water is soon exhausted produce a smaller bulk of vege¬ 
tative organs and flower correspondingly earlier, though of course less 
profusely, than those growing in spots where water is available for a 
longer period. This can be seen very clearly by comparing the indi¬ 
viduals of Dr aba verna, Cerastium, Myosotis, etc., growing on the very 
thin soil on an old wall with those growing on deeper soil. By flower¬ 
ing early such plants seize the opportunity of setting some seed, how¬ 
ever little, before the thin soil on which they grow has quite dried up. 
This state of things applies primarity to annuals, but if relative lack 
of water (and mineral salts) has a direct effect not only in checking 
vegetative growth, but in promoting flowering as such, we should ex¬ 
pect to find it tending to produce early flowering in perennials also. 
When we analyse the flowering dates for 1921 1 we find that from 
January 1 to July 1, i.e. in the first 198 flowering-dates, as nearly as 
possible three-fifths of the dates are “early.” From July 3 to 
August 6 the dates are almost all “late,” but since in this period a 
preponderating number cannot be described as either late or early 
(from want of observations with which to compare them) we cannot 
include them in the inquiry. From August 6 to September 16, the 
late flowerings decidedly preponderate, but the total number in this 
period is small, viz. 13 in number. 
If we include the whole number of dates, viz. 264 in number, we 
find that o-6i are early. 
1 When the name of a plant is preceded by the letter C, such a species was 
found near Cambridge. In the same way S means that the species was noted 
in Surrey (near Gomshall). The remainder, whose names are not marked either 
C or S, were found at Brookthorpc near Gloucester. One plant, Colchicum 
autumnale, was found near Hereford. 
3—2 
