FOOD AND FEEDING OF OYSTERS. 1 299 



As the economic importance of the subject merits, it has frequently been the 

 matter of investigation and has probably attracted more attention from biolo- 

 gists than has any other direct correlation between marine plants and animals. 

 The nature of the oyster's food was long ago determined and the work of the last 

 twenty years has been hardly more than confirmatory of that which preceded it. 



Dean appears to have been the first to attempt the quantitative determination 

 of the oyster food available in the water. He employed a chemical analysis 

 of the water to determine the albuminoid ammonia content, assuming that the 

 results would indicate the comparative food values of different regions. 



Subsequent investigators have recognized the grave defects in this method, 

 and, including myself, have all followed the general method of Rafter. Water 

 specimens of definite volume, usually i liter, have been collected either by means 

 of a stoppered bottle or jug, from which the cork is pulled after it has been sunk 

 to the bottom, or by a specially designed metal cylinder constructed on essen- 

 tially the same principle. The suspended matter in the specimen, a large part 

 of which often consists of sand, mud, and debris, is then concentrated in, say, 

 10 cubic centimeters of water by filtration through sand or precipitation in an 

 Erlenmeyer flask after the addition of a few drops of formalin. A definite quan- 

 tity of the filtrate is then removed after agitation and the food organisms counted 

 in a Rafter cell, the calculated number of such organisms per liter being regarded 

 as an expression of the food value of the water. 



This method has two defects, the first of which is that the water specimen is 

 not drawn from the stratum tenanted by the oyster, but solely from a height of 

 about 12 inches above the bottom. It would be possible to correct this defect 

 by using a shorter, broader bottle or specimen cup, but as the water flows rather 

 slowly into the necessarily narrow inlet, there would enter with it a considerable 

 quantity of material stirred up when the instrument strikes the bottom. As 

 the amount of this material would vary with the bottom, the impact, and the 

 currents, a more serious source of error would arise and the results would become 

 worthless. 



To obviate these difficulties I have designed the type of bottle illustrated in 

 text figures i to 5. It consists essentially of a brass barrel of a capacity somewhat 

 over I liter, two conical valves, and a tripping device. The lower valve is fixed 

 at a height of 2 inches above a broad base, which prevents the instrument from 

 sinking in soft mud, but the barrel and upper valve slide freely on a central 

 column or rod. The instrument is set by engaging the lug (F) over the inclined 

 surface (G) of the stirrup or tripping device (CDEG), which suspends the 

 upper valve (B) and the barrel (A) so as to leave a gap of 2 inches between the 

 two valves and their respective seats, the stirrup being maintained in position 

 by tension on the cord by which the instrument is lowered. By rotating the 

 cam (H) so as to pinch the cord between it and the collar on top of the upper 



