MICHIGAN ACADEMY OF SCIENCE. 
251 
any point was uniform (as to the angle of the point) with any other 
point. The points were finished on an oil-stone, until a microscopical test 
showed them to be all of the same degree of sharpness. The angle of the 
cone was 45°, which was found to give the best results. Screwed into 
the sides of the socket at angles of 120° with each other were three 
iron radius rods terminating in the heavy iron band at the rim. By 
means of lock nuts and gaskets on either side of the tank these rods 
could be shifted, allowing the point to be centered with ease and ac- 
curacy. This was essential, for, with the point located centrally, it 
was possible, at least theoretically, to make correct readings whether 
the tank was vertical or tilted slightly. 
Utilizing as it does the principles of the well-known Hooke gauge 
quite remarkable results were realized with this apparatus. Suppose 
that water is poured into a tank until it reaches the extremity of the 
point. By placing the eye close to the level of the water a position is 
found where the entire surface of the water appears like a bright mir- 
ror. If, now, a small quantity of water is removed, there appears in the 
center of the mirrored surface a minute black spot, which, though small, 
is readily distinguishable. Upon removing a little more water this 
dark spot rapidly enlarges, spreading out in a circle around the ex- 
posed point. Upon slowly replacing the water which was removed the 
dark circle diminishes in size, and at length a point is reached where 
it disappears entirely. If the finger is touched lightly to the water so 
that a tiny wave is sent across the surface, the spot bobs into view 
momentarily, showing that the point is just beneath the surface. Ex- 
perience showed that this phenomenon was most readily observed when 
the sun was low and the light was less intense than at midday, that is, 
near sunrise and sunset. At noon the sun on a clear day is so intense 
that the eye is flooded with light and the dark spot is not readily seen. 
Again, at daybreak and sunset the air is almost calm, thus facilitating 
the operation of measuring and rendering the measurement more exact. 
Repeated observations under ali conditions showed that on the average 
a change in the height caused by the adding or subtracting of 25 c.c. 
of water could be distinguished. Since the area of water surface in 
the tank approximated 4,575 sq. cm., this volume amounts to a theoreti- 
cal difference in height of l-20th of a millimeter. 
The pontoons which supported the tanks in the water consisted of 
two boxes 14 feet long, 12 inches wide and 14 inches deep, constructed of 
rough hemlock lumber and covered with a good grade of asphalt roof- 
ing felt. The top was left loose so that any water which might get 
into the boxes might be bailed out and rocks might be put in to sink 
the tanks to the required depth. It was found to be an advantage to 
have these boxes made extra large in order that the necessary work of 
filling the tanks with soil and plants might be done from the pontoons, 
rather than from a boat. Cross members held the pontoons rigidly in 
place and the tanks were spiked to these. Poultry netting of 2 inch 
mesh was placed around the finished raft to a height of a foot, and also 
in the spaces not filled by the tanks. The purpose of the netting was 
to keep out the turtles which sunned themselves by the score on the 
pontoons. The station was sheltered from the prevailing westerly winds 
by the close proximity of a narrow neck of land which shut olf the 
high waves. There was no natural protection on the south and east, 
