298 
the illuminated ripples which are then pro- 
jected by the second grating. In the figure 
sunlight arriving at L is reflected from the 
large mirror M (one foot square), passing 
thence to and from the shallow rectangular 
trough 7, containing mercury, to be again 
reflected by M, to the screen S. Ifalensis 
placed at B, so that Tis the conjugate focus 
of S, magnificent ripple patterns are seen on 
the screen whenever the table is slightly 
jarred by drumming on it with the fingers. 
These are stationary capillary waves origi- 
nating at each of the straight edges of the 
trough. They may be obtained in different 
wave lengths in the ratio of 1:2, according 
as the table is more or less sharply 
rapped. Gravitational waves, though al- 
ways markedly present, do not appear in 
the picture and may be ignored. To pro- 
duce them observably it is necessary to 
have some special device for starting them. 
Two or three straight iron or steel wires, 
lying stably at w, on the capillary edge of 
the trough, are an excellent wave producer 
if controlled by the electromagnet £. 
Among interesting experiments of this kind 
I will only mention the reflection of waves, 
which advance with a crest and return with 
a trough, from a fixed obstacle, the lines in 
the image being respectively light and dark. 
It takes some practise, however, to see this, 
for with troughs as shallow as convenient 
the velocity will not lie much within six 
inches or a foot per second. 
Returning from this digression, let the 
lens be removed and replaced by a grating, 
G, capable of rotation about a vertical 
axis. When the proper angle is obtained, 
a fairly sharp image of the ripples may be 
seen whenever the table is jarred. Care 
must be taken to avoid errors due to the 
diffraction by the grating of the light is- 
suing from the round capillary edge of the 
mercury in 7. These lines of light give rise 
to extensive streamers on the screen, in- 
tersected by sharp diffraction cross bands, 
SCLENCE. 
[N. S. Vou. XIII. No. 321. 
the streamers intersecting each other at 
right angles in the patch of light due to the 
mercury surface proper. Hence an annular 
screen, AA, is added to blot out the convex 
edges. 
If « is the (broken) distance between - 
mercury and grating, G, if y is the distance 
between grating and screen, 0, the azimuth 
angle of G for which the image appears 
sharpest, finally if a is the distance apart 
of the ripples and 6} the grating space, the 
relation a=bcos@(1+2/y) is available. 
To give an example of experiments tried in 
this way, I found roughly, «= 105 em., ¥ 
= 670 cm., b= .21 cm., = 30°, whence a 
= .2 cm. 
If the trough 7 is long (say two feet 
long and three inches broad), the waves 
from the end are dampened out before they 
reach the middle, and the cross waves are 
alone in presence there. These lines of 
light, if the long edge is parallel to the di- 
rection of projection, give sharper images. 
I have not yet, however, been able to de- 
velop this method toa degree useful for 
measurement, and I merely communicate it 
here as an interesting experiment. I may 
note, in conclusion, that if progressive in- 
stead of stationary ripples be produced, 
and if the grating move in a direction op- 
posite to the ripples, with a velocity increas- 
ing until the shadow bands moving in the 
first instance become stationary, the veloc- 
ity of the ripples would be deducible as 
well as their wave length. 
i C. Barus. 
BRowN UNIVERSITY, PROVIDENCE, R. I. 
REVIEWS OF CURRENT BOTANICAL LITERA- 
TURE. 
A LITTLE more than a year ago at the an- 
nual meeting of the Society for Plant Mor- 
phology and Physiology, held in New Ha- 
ven, a committee, consisting of Dr. Farlow 
of Harvard University, Dr. MacDougal of 
the New York Botanical Garden and Dr. 
