DECEMBER 3, 1897.] 
same means, care being exercised in re- 
moving the plankton quickly and thor- 
oughly to reduce the loss occasioned by its 
adherence to the paper. This method is 
very simple, rapid, and, in my experience, 
more efficient than the sand filter, yielding 
from 75% to 85% of the planktonts. 
The centrifuge at first employed was a 
small one having a capacity of 60 c. cm. 
Later, we had constructed for this work a 
large machine geared to give 3,000 to 4,000 
revolutions per minute and arranged to act 
upon a continuous stream of water, all of 
which was subjected to the maximum and 
uniform action of the centrifugal force. 
This machine is more efficient than the 
filter paper, securing in some instances 
98% of the planktonts. It is, however, 
subject to a selective error, in that the in- 
dividuals and species whose specific gravity 
is the same as or less than that of the water 
are notremoved by the action of the centrifu- 
gal force. Samples rich in water blooms 
proved to be most troublesome. Anabena 
and Clathrocystis, as a rule, and many in- 
dividuals of other genera, as Huglena and 
Chlamydomonas, readily pass through the 
machine. Water kept in the dark, or at 
low temperature, for some hours yields up 
such plankton more readily. The addition 
of alcohol to the water also facilitates pre- 
cipitation. The mostaccurate results, how- 
ever, were obtained after adding chloral 
hydrate to the water in quantities sufficient 
to kill the plankton. The selective char- 
acter of this error, and its consequent un- 
even distribution in plankton varying 
greatly in the abundance of water blooms 
with the season and situation, render the 
use of the centrifuge of questionable utility 
as a basis for a complete analysis of the 
biologic contents of water. 
The cylinders of the Berkefeld filter are 
made of ‘ infusorial earth’ of such fineness 
as to remove effectually all solid matter 
-from the water passing through them. The 
SCIENCE. 
831 
smallest cylinders were encased in a suit- 
able mantle and attached to the centrifugal 
machine. The catch obtained in this man- 
ner contained from one-fourth to one-third 
more organisms than that of the filter paper, 
and was not subject to the irregularities re- 
sulting from the use of the centrifuge. This 
preliminary test was so promising that a 
larger form of the Berkefeld filter, known 
as an ‘army filter, System Bruckner,’ which 
has of late been introduced in the German 
and Austrian armies, was tried as a means of 
collecting plankton. It has the following 
advantages: its maximum capacity under 
favorable conditions is about two liters per 
minute; it is portable and can be used in 
the field; its simple construction favors the 
removal of the catch, and its capacity for 
filtration can be quickly renewed when it 
becomes clogged. At present it seems to 
offer the most effective method for the col- 
lection of the plankton which eludes the 
silk. Itis, however, subject to one serious 
drawback; the removal of the catch from 
the filtering cylinder is accompanied by the 
addition of a considerable amount of the 
infusorial earth to the plankton. This 
renders the ‘Danaid task’ of counting 
doubly difficult and precludes volumetric 
determination. The desideratum for a fil- 
tering cylinder for this work is yet to be 
found. It should be an inexpensive porous 
earthenware cylinder whose outer surface 
is of sufficient fineness to preclude the 
penetration of minute organisms, and of a 
firmness sufficient to permit the removal of 
the catch with a stiff brush without disin- 
tegration. Experiments in this direction 
are now in progress. 
This leakage of plankton through the silk 
is a matter of fundamental importance. A 
considerable volume of the contents of the 
water is lost at all seasons of the year, and in 
some instances the actual catch of the silk 
net is but a small fraction of the total plank- 
ton present. Filter-paper catches from a 
