208 
periments are constantly going on, looking 
towards both the improvement in these dress- 
ings and also in the extension of the useful- 
ness of sphagnum along other lines. 
J. W. Horson 
THE UNIVERSITY OF WASHINGTON, 
SEATTLE, WASH. 
A NEW SEASONAL PRECIPITATION 
FACTOR OF INTEREST TO GEOG 
RAPHERS AND AGRICUL- 
TURISTS 
Most persons who have attempted to corre- 
late soils and vegetation with atmospheric 
precipitation have directed their attention 
chiefly to the total annual rainfall and its va- 
riations from place to place.t It has long been 
recognized that seasonal distribution also needs 
to be taken into consideration, but there has 
been no unanimity as to how seasonal varia- 
tions of rainfall should be treated. Some have 
simply mapped the total precipitation for each 
month or season separately, or indicated the 
months of maximum rainfall in different 
regions; but a more common method has been 
to map the percentage of the total occurring 
in the six middle months of the calendar year, 
April to September inclusive.2 Such an arbi- 
trary division is rather unscientific, however, 
for in the eastern United States, if not through- 
out the northern hemisphere, the warmer half 
of the year usually extends from the latter part 
of April to the latter part of October, so that 
May to October inclusive would more nearly 
represent it.2 The use of the earlier period 
has been defended on the ground that April 
rain is more beneficial to crops than October 
rain, which is probably true (and so would 
March rain be better than September rain) ; 
1 For such correlations between soil and rainfall 
see Bulletin 3 of the U. S. Weather Bureau, by E. 
W. Hilgard, 1892. A review of this, which may be 
more accessible than the original, can be found in 
the Experiment Station Record, 4, 276-282, Oc- 
tober, 1892. 
2 For a map of the United States on this prin- 
ciple see Plate 2 in U. 8. Geological Survey Water 
Supply Paper 234, 1909. 
3See Geol. Surv. Ala. Monog. 8, 24; Bull. 
Torrey Bot. Club, 40, 395, 1913. 
SCIENCE 
[N. 8. Von, XLVIII. No. 1235 
but the type of rainfall that is best for crops, 
other things being equal, is not necessarily 
best for soil in the long run. A warm rain 
presumably has a greater leaching effect than 
cold rain or snow, and regions subject to heavy 
summer rains, like most of Florida, generally 
have poorer soils and more swamps than where 
the summers are dry, as in California.* 
In recent years the writer has calculated the 
rainfall percentages for May to October and 
also for June to September for numerous sta- 
tions in the southeastern states, and thereby 
shown some interesting correlations with soil 
and vegetation.” But when these factors are 
mapped for the whole United States the corre- 
lation does not work out so well. For the 
northern part of the Great Plains has about 
the same proportion of its total rain in summer 
as peninsular Florida, but very different soils 
and vegetation. Of course part of the differ- 
ence is due to the fact that the total rainfall 
and average temperature are much less on the 
Plains than in Florida. But there is another 
important climatic difference. 
In the Great Plains and much: neighboring 
territory the bulk of the rain falls in early 
summer, while along the Atlantic and Gulf 
coasts there is generally more rain in late 
summer than in any other equal period. Con- 
sideration of this fact recently led to some 
comparisons between early and late summer 
rainfall for the whole United States. After 
some experimenting it was found that the most 
striking results were obtained by taking the 
difference between the rainfall for April to 
June inclusive and that for August to Octo- 
ber inclusive,* the former being good for the 
4Dr. A. D. Hall, of Rothamsted, in an address 
on agricultural extension problems published in the 
Popular Science Monthly for October, 1914 (p. 
349), says: ‘‘ Winter rain is more valuable than 
summer. ’? 
5See Bull. Torrey Bot. Club, 37, 415-416; 40, 
395; 41, 556-557; Geol. Surv. Ala. Monog. 8, 19 
24, 36, 1913; Fla. Geol. Surv. Ann. Rep., 6, 182— 
184, 1914; also Ward in Bull. Am. Geog. Soc., 46, 
47, January, 1914. 
6If climatological data for fractions of months 
were available we could include the first half of 
July in the early summer period and the second 
