94 BULLETIN OF THE 



Disregarding the rest of tlie apparatus for the present, we start with 

 the trough sufficiently filled with water to cover the eggs. The water 

 in tlie faucet is started, the filter-box becomes filled, and water now flows 

 through the tube D and falls into the first trough. This trough now 

 becomes filled to the level F G, whereupon tlie siphon is filled and the 

 water begins to siphon out of the trough. By regulating the faucet, the 

 supply from tube D must be made considerably smaller than the stream 

 issuing from the siphon. As a small stream from an ordinary hydrant 

 faucet is apt to be very irregular, the addition of a fine glass tube in- 

 serted into a cork, and this tightly fitted to the faucet, will be found very 

 useful in insuring a more uniform supply. The water in the first trough 

 is under these circumstances siphoned oft' more rapidly than it enters, 

 and is soon lowered to the level H I, when the siphon is emptied and the 

 trough begins to fill again as before, this continuing until the level F G 

 is again reached, when the water is again siphoned oif, and so on. 



Thus we have the water alternately rising and fixlling, washing the 

 eggs constantly in a gentle way, that does not heap them together nor 

 disturb them to any considerable extent. 



The water that is siphoned from the first trough falls into another 

 similar trough supplied with jar and eggs, and placed far enough below 

 the first trough to prevent interference with the siphoning. The second 

 trough should be provided with a siphon slightly larger than that of the 

 first trough, and the amount of water required to fill it and start the 

 siphon should be less than that required for the first. 



Although still other troughs could be added if necessary, the water 

 would be apt to get warm and thus become unserviceable. An increase 

 of surface could be gained to a certain extent by increasing the size of 

 the troughs to accommodate a number of bell jars. The only limit 

 in this direction would be determined by the available flow of water and 

 the desired frequency of the oscillation of the water in siphoning. 



Milwaukee, Wisconsin, August, 1892. 



Note. — Circumstances be^'ond the control of tlie author have prevented tlie 

 publication of his paper at an earlier date. — E. L. M., June, 1894. 



