312 MR. FARADAY ON A PECULIAR CLASS OF ACOUSTICAL FIGURES. 
In fact, the filings proceeded from the quiescent to the moving parts, and there 
were gathered together ; not only forming the cloud of particles over the places 
of intense vibration, but also settling down, when the vibrations were weaker, 
into the same involving groups, and in every respect imitating the action of 
light powders in air. Sand was affected exactly in the same manner; and 
even grains of platina could be in this way collected by the currents formed in 
so dense a medium as water. 
39. The experiments were then made under water with the membranes 
stretched over funnels (22) and thrown into vibration by horse-hairs drawn 
between the fingers. The space beneath the membrane could be retained, 
filled with air, whilst the upper surface was covered two or three inches deep 
with water ; or the space below could also be filled with water, or the force 
applied to the membrane by the horse-hair could be upwards or downwards 
at pleasure. In all these experiments the sand or filings could be made to 
pass with the utmost facility to the most powerfully vibrating part, that being 
either at the centre only, or in addition, in circular lines, according to the mode 
in which the membrane vibrated. The edge of the funnel was always a line of 
rest ; but circular nodal lines were also formed, which were indicated, not by 
the accumulation of filings upon them, but by the tranquil state of those filings 
which happened to be there, and also by being between those parts where 
the filings, by their accumulation and violent agitation, indicated the parts in 
the most powerful vibratory state. 
40. Even when by the relaxation of the parchment from moisture, and the 
force upwards applied by the horse-hair, the central part of the membrane was 
raised the eighth of an inch or more above the edges, the circle not being four 
inches in diameter, still the filings would collect there. 
4 1 . When in place of parchment common linen was used, as becoming tighter 
rather than looser when wetted, the same effects were obtained. 
42. Both the reasoning adopted and the effects described were such as to 
lead to the expectation that if the plate vibrating in air was covered with a 
layer of liquid instead of sand or lycopodium, that liquid ought to be deter- 
mined from the quiescent to the vibrating parts and be accumulated there. 
A square plate was therefore covered with water, and vibrated as in the former 
experiments (2. 6.) ; but all endeavours to ascertain whether accumulation 
