740 THE AMERICAN NATURALIST. [VoL. XXXII. 
avoided ; (@) there is an actual loss of small individuals through 
the meshes of the net ; (e) the long and complicated process 
must necessarily give varied results due to personal variations 
in methods of work, and to changes in the local conditions, ¢.g., 
the rate of currents may vary from day to day, the quantity of 
silt may modify the filtering capacity of the net, etc. Any one 
of these sources of error is sufficient to invalidate the entire 
method, rendering the results worthless for comparison with 
the results of similar processes in different localities. 
Under (2) (passing known quantities of water through a filter 
of fine bolting cloth) the sources of error are reduced but not 
eliminated ; (2) the pressure of water forces certain small 
forms, e.g., certain species of bacteria, through the meshes. 
Many of the very delicate forms may be broken up and 
destroyed ; (0) failure to wash out all the individuals from the 
net. The method of pumping known volumes employed by 
Kofoid (97) is particularly good. The most apparent source 
of error is the control of the quantity of water pumped, and 
the possibility that the strong suction of the pump used may 
draw mud when the water is taken within a foot of the bottom. 
In the method of filtration through sand, as employed by 
Calkins (91) and as improved by Jackson ('96) and by Whipple 
(96), the possible sources of error are several, varying with the 
characteristics of the sand used, with the shape of the fun- 
nel, and with the nature of the organic matter in the water. 
Calkins says: “ The sloping sides of the glass funnel offer a 
surface for the settling of organisms, and the error arising in 
this way may be considerable. A water free from amorphous 
matter and zoogleea will filter very accurately, but a water con- 
taining these gives opportunity for error.” Jackson ('96) adds : 
“This is undoubtedly due to the jelly-like character of the 
zoogloea, and to the fact that while adhering to the funnel sides 
itself, it also retains with it other organisms.” ... “Not 
only do amorphous matter and zoogloea readily adhere to the 
sides of the ordinary glass filter funnel, but the same is true of 
the gelatinous growths of the Cyanophycez and of the floccu- 
lent threads of Crenothrix.” Even Jackson’s ('96) improve- 
ments in the sand filtration method which reduce to a minimum 
