568 Reviews—Level of Ground-water, Wisconsin. 
soda-felspars, and the third those containing no felspar. These are 
further subdivided according to the presence or absence of quartz 
and the nature of the felspathic and ferro-magnesian constituents. 
Two main divisions—the granular and the porphyritic (including 
glassy) — based on structure are recognised. These correspond 
approximately, but not exactly, with the plutonic and volcanic 
divisions of other authors. Geological age is introduced, as in the 
Rosenbusch classification, only with reference to the volcanic rocks. 
The limitation of the terms rhyolite, andesite and basalt to rocks of 
the Tertiary and post-Tertiary periods will strike anyone acquainted 
with British rocks as very strange and unnatural. 
The value of this work, however, does not depend on the classifi- 
cation which is adopted or on the terminology employed; but on its 
being an encyclopaedic record of the present state of petrographical 
knowledge. In many cases the student will find here all that he 
wants, but if not he will have no difficulty in obtaining references to 
the original communications bearing on the subject in which he is 
immediately interested. The second and third volumes are promised 
within a month or two. 
V.— OBsERVATIONS AND EXPERIMENTS ON THE FLUCTUATIONS IN 
THE Leven anp Rate or Movement or GROUND-WATER ON 
THE Wisconsin AGRICULTURAL ExpErRiIment Sration Farm, 
AND aT Wuuirewater, Wisconsin. By Prof. Frayxun H. 
Kine, United States Department of Agriculture, Weather Bureau, 
Bulletin No. 5 (1892). 75 pp., six plates. 
HE experiment farm lies on the south side of Lake Mendota, 
in Wisconsin. The ground rises from the level of the lake to 
about fifty feet above it, within the area in which the wells are 
situated, and just outside that area toa height of 111 feet. The farm 
is situated on Till, resting on an uneven surface of Madison sand- 
stone. Allover the surface of the ground there is a layer of reddish 
clay, 23 to 4 feet thick, containing pebbles and boulders sparsely 
scattered, which passes rather rapidly downwards into a nearly 
uniform sand. The wells are 54 in number, and the majority vary 
in depth from 5 to 26 feet and are made wholly in the Till, but a | 
small number reach a greater depth and penetrate a few feet into the 
rock. Most of them are tubed with 5-inch drainage tile, surmounted 
at the surface with one length of 8-inch glazed sewer pipe with a 
galvanised iron cover. The height of the water in the wells was at 
first ascertained by lowering a weight through a distance which 
could be measured by a micrometer on the surface, the base of the 
weight being a hemispherical button which produced waves in the 
water at the instant of contact. Afterwards a simple recording 
instrument was employed, consisting of a copper float connected 
with the short arm of a lever, the long arm of which carried a pen 
which, on a scale three times the natural size, traced the fluctuations 
in the water-level on paper moved by clockwork. 
The level at which the water stands in the wells is everywhere 
above that of the water in Lake Mendota. The general tendency, 
