556 
NATURE 
[APRIL 11, 1907 
making Norwegian kings speak ‘‘the Danish 
tongue.’’? These instances show how readily the 
name of the Angles might efface that of the Saxons 
even at an early date. 
While such objections may be made to some of Mr. 
Chadwick’s arguments, the method he has followed 
in tracing the origins of the English people is a 
sound one. He begins with what can be learned 
of the invading nations immediately after their settle- 
ment in Britain, and from this works back as far 
as possible into their previous history. A necessary 
result of the method, however, is that as the inquiry 
advances the evidence becomes more scanty, and the 
use of conjecture more and more obvious. For this 
there is no help, but it seems a little disproportionate 
to give only ninety pages to the English period and 
two hundred and fifty to the Continental, of which 
so little is known. These ninety pages contain four 
chapters, of which the first gives a survey of England 
in the sixth century, showing the extent of the con- 
quest at that date. The West Saxon invasion, and 
that of Kent, are specially discussed in the following 
chapters, and the fourth is occupied with the ques- 
tion of the three nations referred to above. It in- 
cludes some useful tables of early linguistic 
variations, and remarks on these, together with an 
account of the difference between Wessex and Kent 
in respect of the various classes of the community 
and their wergelds. 
The very hypothetical character of Mr. Chadwick’s 
inquiry does not do full justice to the great mass of 
interesting matter which he has brought together. 
A very wide range of reading and research underlies 
every chapter of it, and each point has evidently 
been the subject of much study and consideration. 
Many of his views are highly suggestive, and may 
yet lead to more certain results. In the meantime, 
the evidence produced does not seem sufficient to 
convict Bede of any essential error, or to modify in 
any important way the usual views on the subject. 
W. A. Craicir. 
THE RAINFALL OF NORTH GERMANY. 
Die Niederschlige in den mnorddeutschen Strom- 
gebieten. By Prof. G. Hellmann. In _ three 
volumes. Vol. i., pp. vit+386+140; vol. ii., pp. 
Viii+ 722; vol. iii., pp. vili+872. 
Reimer, 1906.) Price 60 marks. 
R. HELLMANN’S three volumes contain a 
wealth of information relating to the rainfall 
and allied phenomena in the North German river 
basins. The principal observations are elaborately re- 
duced, and in many aspects very fully discussed. 
The significance of this rainfall in its wider relation 
as part of the world weather, and, as such, its prob- 
able correlation with solar changes, are investigated 
with the advantage of well-marshalled data. 
The area specifically dealt with in the volume is 
extensive, consisting practically of the great plain 
which extends without interruption from the chain 
of mountain ranges in south Germany to the North 
Sea and the Baltic. The direction of the river flow 
NO. 1954, VOL. 75 | 
(Berlin: Dietrich 
and the precipitation of the country are very largely 
determined by this chain of mountains, which is part 
of the great water-shed of Europe. The conditions 
of the rainfall problem over such an area would seem 
to be fairly simple, and capable of being dealt with 
in general terms. The local conditions, however, as 
is usual, exercise a considerable influence, the 
actual’? varying widely from any ‘“‘ mean.” 
The three volumes may be taken as a summary of 
the meteorological work of many years in the de- 
partment of rainfall measurement within the district 
named. Its fulness and painstaking completeness is 
such as is expected from the efficient State-supported 
meteorological organisation of Germany. Much of 
the data is from the numerous and evenly distributed 
stations, daily returns from which are made immediate 
use of for short-date forecasting. 
The first volurne is general, describing and dis- 
cussing the data and results. This volume is divided 
into five sections, of which the first, in dealing 
generally with the observation material and the 
manner in which it has been obtained, discusses the 
distribution of stations and the quality of the observ- 
ations themselves. The recognition of the influence 
on these of the type and position of the gauges used 
is of value. Such considerations affect the credentials 
of the older observations, a knowledge of the stand- 
ing of which allows the full length of the record to 
be used safely or to be rejected where untrustworthy. 
A long meteorological record is sometimes, like the 
curate’s egg, merely good in parts. The ease of ap- 
proximate rainfall measurement conduced to its early 
commencement, and very old records exist. Observ- 
ations made at Breslau (1717-1727) gave an annual 
mean of 576 mm., which does not differ greatly from 
the modern value of 567 mm. A valuable bibliography 
of the history of rainfall measurement concludes this 
first part. 
The amounts of rain and their reduction and inter- 
comparison are next dealt with. The influence and 
value of smoothing curves by taking means is illus- 
trated both by actual curves and by tables. Means 
for several stations, for periods varying from five to 
forty-five years, together with the ‘‘ greatest differ- 
ences’? in each set of means, are obtained and com- 
pared. The standing of short-period means and the 
necessity of taking a long period to obtain a normal 
value become clear. A valuable table of monthly 
seasonal and annual means, both actual and per- 
centage of mean year, is given in the text for nearly 
too stations. The distribution of rainfall in the year, 
from ten- and twenty-year means, is discussed and 
Ulustrated by curves for Konigsberg and Stettin. 
Abnormal rains and thunderstorms are considered at 
some length, while material for further discussion is 
given in tables of great detail. 
The reduction of the data is further extended in 
the next section of the volume to the problem of the 
determination of the expectancy of greatest rainfall 
and the probability of the number of rainy days of 
definite intensity. Various mean curves are used 
depending on periods of observation of from nineteen 
to forty-three years. Snowfalls are dealt with in re- 
