Special scientific report

;452:

A TOWED PUMP AND SHIPBOARD
FILTERING SYSTEM FOR SAMPLING
SMALL ZOOPLANKTERS

,.^^»»r.j<^.!.lJlM^.«Ji»Jl V*

WOODS HOLE, MASS.

^MtttmTir*iik-x

SPECIAL SQENTIFK REPORT-FISHERIES Na452

UNITED STATES DEPARTMENT OF THE INTERIOR, Stewart L. Udall, Secretary

FISH AND WILDLIFE SERVICE, Clarence F. Pautzke, Commissioner

BUREAU OF Commercial Fisheries, Donald L. McKernan, Director

A TOWED PUMP AND SHIPBOARD

FILTERING SYSTEM FOR SAMPLING

SMALL ZOOPLANKTERS

by
Charles P. O'Connell and Roderick J. H. Leong

United States Fish and Wildlife Service
Special Scientific Report— Fisheries No. 452

Washington 25, D.C.
May 1963

CONTENTS

Page

Introduction 1

Description 2

The collector 3

The winch 6

The filtering unit 9

Operation 11

Performance 13

Discussion 17

Summary 18

Literature cited 19

111

A TOWED PUMP AND SHIPBOARD
FILTERING SYSTEM FOR SAMPLING
SMALL ZOOPLANKTERS

by

Charles P. O'Connell, Fishery Biologist

and

Roderick J. H. Leong, Fishery Biologist

Bureau of Commercial Fisheries Biological Laboratory

U.S. Fish and Wildlife Service

La Jolla, California

ABSTRACT

The construction, operation and performance of a towed pump and shipboard
filtering system for sampling small zooplankters is described. The system is com-
posed of (1) a collector containing a pump towed by a suspension unit consisting of a
hose through which runs a steel cable for support and an electric line to power
the pump, (2) a winch, and (3) a filtering unit composed of a watermeter, a double-
throw valve and two filtering funnels. The towed collector and its electrically driven
pump operate satisfactorily to a depth of 5 or 6 meters at a vessel speed of 9 knots.
The winch is inadequate for the task it is meant to perform, but demonstrates that it
is practical to use a winch with a hose and cable suspension. The filtering unit, easily
operated by one man, is highly satisfactory. Discrete samples can be taken at
intervals of a few minutes while traveling at vessel cruising speed by using the two
funnels alternately. Special samples and tests indicate that errors due to escapement
and entrapment of zooplankters and to mixing of zooplankters between consecutive
samples are negligible. It is concluded that the towed pump and shipboard filtering
system is a practical sampling tool that needs further modification for the full
utilization of its capabilities.

INTRODUCTION

In 1958 the Bureau of Commercial Fisheries
Biological Laboratory, La Jolla, California,
initiated a study of the relation between the
behavior of the Pacific sardine (Sardinops
caerulea) and the density distribution of its
planktonic food. The first phase of the pro-
gram was largely devoted to the development
of a towed pump and shipboard filtering unit
for quantitative sampling of the small zoo-
plankters that constitute the bulk pf the sar-

dine's diet. Sampling surveys were carried
out with this apparatus in the fall of 1961 by
the Bureau of Commercial Fisheries research
vessel Black Douglas. The apparatus will be
described and evaluated in this report, and the
results of the 1961 surveys will be presented
in later reports as the samples are processed.

A study of the density distribution of sardine
food organisms requires that the smallest
possible zooplankters be collected quantita-
tively. Hand and Berner (1959) found that

small copepods supplied, on the average, 74
percent of the total organic matter in the
stomach contents of sardines and that all
crustaceans supplied nearly 89 percent. The
size range of small copepods is not explicitly
stated, but it may be surmised from the data
in the above report that the small category
includes organisms up to about 1 mm. in
length. To collect organisms of the size
indicated, filtering screens with mesh open-
ings of 100 microns (/j) or perhaps even less
must be used.

The study also requires that estimates of
known precision be obtainable for areas as
small as 20 square miles or as large as
several hundred square miles. This can be
accomplished efficiently by subsampling, i.e.,
taking several small, discrete samples within
the specified area, while the vessel is
traveling at cruising speed.

The study further requires that areal esti-
mates be made for more than one depth or
that they represent some vertical range. To
accomplish this the sampling instrument must
be capable of collecting samples over some
range of depth rather than at a single depth.

A towed pump and shipboard filtering unit
appeared to offer the best possibility of
satisfying these requirements. Filtering rate
and filtering efficiency can be independent of
filter mesh size, and they can be directly
measured. With the proper mechanical ar-
rangement, discrete samples representing
very short segments of the vessel track can
be taken at vessel cruising speed. If the
pump is suspended at the end of a hose it
can be lowered to different depths.

Although self-contained high speed samplers
such as the Isaacs high speed sampler (Ahl-
strom, Isaacs, Thrailkill, and Kidd, 1958), the
Gulf III sampler (Gehringer, 1952), and the
Hardy plankton recorder (Hardy, 1939) can be
towed at different levels, they do not adequately
retain zooplankters less than 1 mm. in length.
Filtering screens with meshes fine enough to
collect these smaller organisms would ad-
versely effect filtering rate and filtering effi-
ciency by intensifying clogging. The Isaacs sam-
pler and the Gulf III sampler, furthermore, are
unsuitable for the collection of short interval

samples because they must be retrieved and
serviced to terminate each sample. The Hardy
recorder does resolve a continuous strip col-
lection into an intergrade series of small sam-
ples, but there is a practical limit to how small
these samples can be and it is necessary to
make assumptions about filtering rate and fil-
tering efficiency.

Other workers have resorted to pumps for
plankton sampling from time to time, because
this seemed to be the best way to collect
numerous small samples in time and/or space
to study the variability of plankton density.
Aron (1958) lists 17 investigators in addition
to himself who have used pumps to collect
plankton. Cassie (1959) has also made ex-
tensive use of a pump to collect plankton.
However, none of the systems used by these
investigators would have fulfilled the require-
ments outlined above. Although several of
them have been used to sample from various
and considerable depths, this was possible
only from vessels that were drifting or moving
very slowly. Collier (1957) is the only in-
vestigator listed by Aron who sampled with
a pump while traveling at vessel cruising
speed. He achieved this by using an inboard
pump with the intake protruding through the
hull of the vessel, an arrangement that offers
no potential for sampling at different depths.

The towed pump and shipboard filtering unit
described in this report was designed to retain
zooplankters, particularly crustaceans, as
small as 100 or 200 /J in length and to take
samples that would be discrete for time
intervals as short as a few minutes. It was
designed also with the potential for collecting
over some vertical range, though no attempt
has yet been made to realize this capability.

DESCRIPTION

The entire system is shown schematically in
figure 1 as it is installed on the Black Douglas.
It is composed of (1) a collector containing a
pump and motor towed by a suspension unit
consisting of a hose through which runs a
steel cable for support and an electric line
to power the pump, (2) a winch and auxiliary
guide wheel, and (3) a shipboard filtering unit.
The towed hose is a 150- foot length of 2-inch
internal diameter (I.D.) single- jacket firehose.

FILTERING UNIT

WINCH

TOWED UNIT

Watermeter
PVC Funnel P^ — p
Plastic Ttash Can-"^ -^ ]l

^

'

i ' ■ i

^__^^.

1" I.D. Hose

Depressor'

Figure 1. --Schematic drawing of the towed pump and shipboard filtering unit as it is installed on the research
vessel Black Douglas. The arrows indicate the path of water through the system. From the funnels it flows
through drainpipes to the scuppers of the vessel. The collector is drawn larger than scale size.

Both the support cable (1/4-inch stainless
steel aircraft cord) and the powerline (#14,
3-conductor, type S.O. neoprene cord) run
unattached inside the hose. All three elements
extend from a lower terminal connected to
the bridle in which the collector is mounted
to an upper terminal on the winch. Approxi-
mately 100 feet of hose is beyond the guide-
wheel when the system is set for towing.

The winch is about 10 feet forward of the
fantail, and the filtering unit is about 60 feet
forward of the winch. They are connected by a
70-foot length of 1 1/2-inch I.D. thick-walled
polyethylene hose. The filtering funnel inlets,
about 11 feet above sea level, are the highest
points in the system. The swinging davit with
block and tackle, located just inboard of the
guidewheel, is used to set and retrieve the
collector.

A detailed description of the system fol-
lows.

The Collector

Figure 2 is a photograph of the collector
mounted in the bridle and connected to the
lower terminal assembly of the hose-cable
suspension. The collector is secured in a heavy,
steel ring clamp with a pin on each side. The
pins project through holes in vertical plates
welded to the inside of the rigid diamond-
shaped bridle and function as an axle. The
bridle is suspended from a connecting rod
projecting from the bottom of the hose-cable
terminal assembly, and a 43-pound homo-
geneous depressor (California, State of, Marine
Research Committee, 1950) is in turn sus-
pended from the bottom of the bridle on a
short length of chain. The pump outlet is
coupled to the terminal assembly by a trailing
loop of 1-inch I.D. thick-walled rubber hose.
The powerline for the electric motor also
forms a loop between the collector and the
point where it enters the hose-cable terminal
assembly. Though not visible in the photograph,

Figure 2.--The collector mounted in the diamond -shaped bridle with the hose-cable
term.inal assembly above and the depressor below.

there is a Joy plug in the powerline about
halfway between the collector and the ter-
minal assembly so that the powerline can be

disconnected at this point. The trailing loop of
powerline is taped securely to the 1-inch I.D.
hose before the collector is set for towing.

Elbo

Centrifugal Pump
PVC Nozzle
V I.D. Orifice

Pump Nose

Nipple ,Powerline

Rubber Shaft Coupling

Opening

Micarto Discs

Threaded Rod

-Micarto Discs

SCALE

6"

Figure 3. --Schematic drawing of the collector with the outer casing cut away.

Construction details of the collector are
shown in figure 3. The outer casing is a 24-
inch length of 6-inch I.D. polyvinyl chloride
(PVC) tubing slotted on top at the front edge
to admit the powerline for the motor. A 1/2-
inch bronze centrifugal pump (1/2-inch dis-
charge, 3/4-inch suction) is bolted to a
micarta (1-inch linen phenolic sheet) disc
that fits in the front opening of the PVC
casing. The impeller in this pump is the
semi-open type, which, according to speci-
fications, permits pumping a percentage of
solids. A capacitor starting, 1/2-horsepower
(hp.) A.C. motor (3,450 r.p.m.) is friction
mounted in two micarta discs that slide
easily into the casing. The motor is an
hermatically sealed, water-cooled unit. The
pump and motor shafts are connected by a
flexible rubber coupling and the entire pump-
motor assembly is rigidly aligned by four
3/8-inch threaded rods through the micarta
discs. The assembly is secured in the casing
by tightening another micarta disc against the
back edge with a nut on each of the threaded
rods. To insure proper cooling of the motor
there are a few small slots at the forward
edge of the casing and also on the outer edges
of the motor mounting discs. The disc against
the back edge of the casing has a large central
opening.

The conical nose of the collector is com-
posed of a truncated aluminum funnel held
against the face of the pump mounting disc

by a specially machined PVC nozzle screwed
into the axial intake of the pump. The funnel
is strengthened by a coating of fiberglas
and has one small hole to admit water so that
it will not collapse under external hydrostatic
pressure. The back edge of the funnel is cut
away on one side to accommodate the outlet of
the pump, A street elbow brazed to the outlet
contains a "close" pipe nipple to which the
end of the 1-inch I.D. delivery hose is coupled.

The one vertical and two horizontal stabi-
lizers, which are not shown in the drawing of
the collector, are made of 1/32-inch sheet
stainless steel.

The lower hose-cable terminal assembly
(fig. 4), which serves tobring water, electrical
and support elements together inside the 2-inch
I.D. hose, is made of standard galvanized pipe
fittings. The hose couplings are secured in the
ends of the hoses with heavy-duty hose clamps.
The heavy steel rod to which the cable is con-
nected by a swivel and shackle was specially
formed and welded through the center of the
pipe plug at the bottom of the assembly.

The hose and powerline are a few feet
longer than the support cable to insure that
the latter bears the entire load of the towed
unit. The strain exerted by the towed unit
should be approximately 800 pounds, the down-
ward force that the depressor is designed to

Cable (%" Stainless Steel Aircraft Cord)
2" I.D. Firehose

Hose Coupling {

Pipe Nipple-

2" X Vi" X 2" Pipe Tee-

Pipe Nipple-

Jaw-end Swivel'

Round-pin Shackle -

2"X1"X2" Pipe Tee —
Pipe Plug

SCALE

2"

Stuffing Box

^ / Packing Glond

-Powerline
(#14 type S.O., 3 Conductor Neoprene Cord)

Pipe Nipple

Hose Coupling

^1" I.D. Rubber Hose

— Welded SteeJ Rod

Figure 4. --Schematic drawing of the lower hose-cable terminal assembly. Pipe threads are not

indicated.

produce at a speed of 10 knots. The support
system has an ample safety factor for this
load. The average ultimate breaking strength
is 6,000 pounds for the cable and 8,000
pounds for the shackles and the swivels. The
recommended safe working load is about 1.600
pounds for all three.

The upper half of the support cable is en-
cased in yiny tubing to prevent the inner
surface of the hose from being damaged
where it rides on the guidewheel when set
for towing.

The Winch

The winch (fig. 5) is a steel drum 3 feet in
diameter and 4 feet in length with a 1/2-inch
I.D. pipe axle supported at each end by a
channel iron stand. The base of the stand is
bolted through the deck of the vessel. The
hose is spooled onto the lower surface of the
drum to minimize strain on the deck bolts,
and its upper end is connected to a 2-inch
plumber's pipe cross mounted near one edge
of the drum.

Figure 5. — The winch, showing the plumber's cross and axle delivery and the power drive for the drum.

The plumber's cross (fig. 6) is the upper
hose-cable terminal assembly where electrical
and support elements are separated from the
water. The hose is secured to a close nipple
in one horizontal opening of the cross with a
standard hose coupling, while the cable is
shackled to a ring welded to a pipe plug in the
opposite opening. A close nipple in the bottom
opening extends through a hole in the surface
of the drum. A 2-inch to 1 1/2-inch bell
reducer tightened on the nipple against a
broad washer facing the underside of the
drum surface holds the plumber's cross rigidly
in place. A 1 1/2-inch pipe extends from the
reducer to a union projecting from an opening
in the axle of the drum. The end plate of the
drum is set back 6 inches from the edge of the
cylinder so that the reducer and connecting
pipe are on the outside. The axle is plugged

immediately behind the point where the radial
pipe joins it so that water flowing from the
plumber's cross is forced towards the outer
end of the axle. A right- angle 1 1/2- inch
swivel coupling turned onto the end of the axle
allows the axle to rotate without turning the
pipe screwed into the other opening of the
swivel coupling. The hose leading to the
filtering unit is clampled to this pipe.

The powerline, which passes out of the
plumber's cross through a packing gland and
stuffing box in the top opening, is connected
with a watertight electrical coupling to another
line that is bent sharply towards the axle and
secured to the radial pipe between the bell
reducer and the axle. Free cable beyond this
point is spooled onto the axle as the drum is
rotated to let the hose out, and reeled off

Powerline

2" I.D. Firehose

Cable 1

Edge of Drum

SCALE

2"

Stuffing Box

2" Plumber's Pipe Cross
Round Pin Shackle

Welded Steel Rod

Pipe Plug

Jaw End Swivel

-2" to V/i" Bell Reducer
■VA" Pipe

To Axle

Figure 6. --Schematic drawing of the upper hose-cable terminal assembly mounted near one edge

of the winch drum.

again as it is rotated in the opposite direction
to pull the hose in. Thus the electrical con-
nection need not be interrupted when the drum
is rotated. The end of the powerline is con-
nected to the capacitor starter, which is
located near the winch, and the capacitor in
turn is plugged into an A.C. outlet on the deck
of the vessel.

Rotation of the drum is controlled by a
reversible 3/4-hp. A.C. gear motor mounted
on the base of the winch stand. The combina-
tion reduction gear and chain- sprocket drive
to the axle rotates the drum at about 2 r.p.m.,
which reels the hose in or out at a constant
rate of 20 feet per minute. Braking is auto-
matic. The motor shaft is always locked by a
built-in brake when the control lever is in
the off position.

The guide wheel (fig. 7) is actually a roller
cage designed and built specifically for this
apparatus. It consists of two horizontal and
two vertical rollers made of 2-inch diameter
PVC tubing with sintered nylon bushings. The

Figure 7. --The guide wheel, composed of 2 horizontal
rollers and2 vertical rollers. It is mounted on a short
vertical post and rotates freely in the horizontal
plane.

clearance between the two vertical rollers is
3 inches. The top plate of the cage is hinged
on one side and secured with wing nuts on the
other so that it can be swung out of the way for
inserting or removing the hose. The unit is
mounted on a short vertical post projecting
from the deck of the vessel and rotates freely
in the horizontal plane.

The Filtering Unit

The filtering unit (fig. 8) consists of a
watermeter, a quick acting double- throw valve
and two filtering funnels. These elements are
arranged so that water can be directed to
either of the funnels after it passes through
the meter. A thermister sensing unit mounted

in a pipe tee between the watermeter and the
double-throw valve is connected to a Rustrak
recording unit to provide a continuous record
of temperature.

Each funnel is set in a large trash can with
two openings in the front. A 4-inch I.D. hose
connected to a pipe nipple in the lower opening
carries filtered water to the scuppers of the
vessel. The upper opening permits the operator
to reach the collecting bucket at the bottom of
the funnel. The funnels and trash cans are
lashed in place.

When desired, water can be diverted to the
scuppers before it reaches the filtering unit

Figure 8.--The filtering unit composed of watermeter, double-throw valve and two filtering funnels mounted
in trash cans. The thermister just above the watermeter is connected to a Rustrak recording unit at the far
right. The smaller hose is the incoming waterline and the larger hoses are drainagelines.

by opening a valve inserted in the polyethylene
hose between the winch and the filtering unit.

The watermeter is a 1 1/2-inch bronze
wobble plate type commonly used by utility
companies. It registers up to one hundred
million liters in hundreds of liters. Tens
and units are read from a rotating needle
dial. Calibration of the meter by the Helix
Irrigation District Laboratory showed that it
registers 2.37 percent higher than the true
volume at rates of 75 to 98 liters per minute,
the delivery rate range within which the
towed pump system operates.

Each funnel (fig. 9) is formed of 1/1 6- inch
PVC sheet stock. A short length of 1 1/2-inch
PVC pipe glued tangentially into a hole near
the top Serves as the inlet, and a specially
machined PVC neckpiece glued to the bottom
serves as the mounting for the collecting
bucket. The inlet is connected to a pipe from
the double-throw valve by a short length of
plastic tubing secured with hose clamps. The

greatest portion of the funnel surface between
the inlet and the neck consists of windows
covered by 105 /i -mesh stainless steel cloth
(the diagonals of the mesh openings are
I50fl). The stainless steel cloth is glued
along the margins of each window with PVC
solvent. All internal areas of glue were made
as smooth as possible.

Each collecting bucket (fig. 10) is a 3-inch
length of 2-inch I.D. PVC pipe that is bayonet
mounted on the neck of the funnel. Two "J"
slots in the upper edge of the bucket fit
over a pair of pins projecting from a steel
clamp around the neck. The steel clamp is
set in a groove around the neck and a rubber
"O" ring is set in a lower groove to seal the

m

Figure y.--The filtering funnel with the bayonet mounted
collecting bucket and the friction fitted cap in place.

Figure 10.--The collecting bucket with the friction cap
detached. The edge of the bucket is recessed to ac-
commodate the tension screw of the steel clamp on
the neck of the funnel. The rubber "O" ring on the
neck of the funnel seals the bucket below the edge of
the large recess.

10

bucket. The "O" ring is just below the broad
slot in the bucket that allows clearance for
the tension screw of the steel clamp. As an
added precaution, a rubber baffle is attached
to the funnel just above the neck to keep the
"O" ring seal from being unnecessarily de-
luged by filtered water coming down the
outer surface of the funnel.

The bottom of the bucket is covered by
105/i-mesh stainless steel cloth. When the fil-
ter is operating, this screen is covered by a
friction fitted cap so that all water filters
through the windows of the funnel. The cap is
twisted off at the end of a filtering period to
allow the small amount of water trapped in
the bucket and neck of the funnel to drain
out before the bucket is removed.

OPERATION

The collector is set for towing with the
vessel running slow ahead. The hook sus-
pended from the davit is placed in the top of
the diamond bridle frame, and the collector
is hoisted from its cradle on the deck and
lowered over the fantail until the hose-cable
terminal assembly is just below the guide
wheel. The hose is set in the guide wheel, and
slack is taken up by the winch. The hook is
then removed from the bridle, and the collec-
tor is lowered to a depth of a few meters below
the surface. The pump is turned on and, after
water is flowing satisfactorily through the
hose to the laboratory, the hose is reeled out
slowly. When the desired length of hose is
out, vessel speed is increased to cruising.

The collector is retrieved by the reverse
procedure. When the unit is to remain aboard
for any length of time the entire system is
flushed with fresh water. The nose cone and
nozzle are removed from the collector, and a
fresh-water hose from the vessel is coupled
directly to the pump intake. Water is allowed
to run through the systemfor about 15 minutes.

Once set, the collector is towed and the
pump operated continuously for the duration of
a survey. The longest continuous runs so far
have been approximately 72 hours. When
samples are not being collected during such a

run, water is diverted to the scuppers through
the valve between the winch and the filtering
unit.

The sampling patterns carried out in the
fall of 1961 required that sequences of 1-mile
samples be taken for one to a few hours with
interim nonsampling periods of one-half to
several hours. The 1-mile sample intervals
were taken to be equivalent to 6.5-minute
time intervals at vessel cruising speed, which
was estimated to average 9 knots. Samples are
easily collected by one operator under this
regime. The procedure followed is outlined
below:

1. A few minutes before arriving at the
starting point of a sampling sequence, re-
move the collecting bucket from one funnel
(B), and close the diversion valve so that the
flow of water is directed to the filtering unit.
Set the double-throw valve so that the water
flows through the open funnel.

2. Upon arrival at the starting point, switch
on the interval timer (an electric device
that gives an audible signal every 6.5 minutes),
record the watermeter reading and set the
double-throw valve to direct water into
filter A.

3. Put the collecting bucket back on funnel B.

4. When the timer rings again, switch the
valve to funnel B and record the watermeter
reading.

5. Twist the friction cap clockwise off the
bottom of the collecting bucket on funnel A
and allow the residual water in the funnel to
drain through the bottom screen.

6. Twist the bucket counterclockwise off
its bayonet mounting.

7. Wash the sample from the bucket into a
jar with a gentle stream of water and a spray
gun.

8. Fill the jar with water and the necessary
amount of formalin to make a 4-percent
solution. Label and cap the jar.

11

9. Put the friction cap back on the bucket
and put the bucket back on funnel A.

10. When the timer rings again, switch the
valve back to funnel A, record the water-
meter reading and remove the sample from
funnel B as above.

11. Repeat the procedure to the end of the
sampling sequence, at which point the timer
can be switched off and the water diverted
to the scuppers of the vessel.

This procedure generally requires 2 or 3
minutes work each time a sample is taken,
leaving a few minutes for the operator to take

care of odd chores, such as making notations
on the temperature recorder and preparing
labels for the sample jars.

The spray gun referred to in step 7 was
found to be very effective in removing the
sample from the bucket. Actually the bulk
of the plankton is first removed by tipping
the bucket over a small powder funnel set
in the sample jar and washing the inside
down with a gentle stream of water from a
small hose connected to a laboratory tap.
Organisms remaining on the screen are re-
moved by placing the bucket upside down in
the powder funnel (fig. 11) and "blasting"

Figure ll.--The collecting bucket being cleaned with the spray gun.

12

through the bottom screen with the spray gun.
Two or three short blasts are usually sufficient.
There is no back blast out of the funnel or
disturbance to the water in the jar because
the screen itself dissipates the force of the
spray markedly. The adjustment of the spray
gun is not particularly critical, but the instru-
ment seems most effective when the atomizer
is set to emit a relatively heavy or "wet"
spray under moderate pressure. The pressure
for the gun is set by the valve at the outlet
of the vessel's compressed air system to which
a hose from the gun is attached.

The spray gun is also used to clean the main
filtering funnels when they become heavily
clogged with phytoplankton. This has to be
done at least every 3 or 4 hours and in actual
practice it is usually done at intervals of 1 or
2 hours during convenient breaks between
sampling sequences. Water entering the fun-
nels forms a rapidly swirling vortex near
the lower edge of the screened windows when
the funnels are clean. Over a period of a
few hours phytoplankton clogs the screens
progressively from the bottom, raising the
level of the vortex. A funnel is always cleaned
before the vortex level rises to two-thirds the
height of the screened windows.

To be cleaned, a funnel is uncoupled from
the double-throw valve and removed from
the trash can. The screens are gone over
thoroughly from the outside with the spray
gun while the funnel is held over a drain
trough. Spray-gun pressure is increased con-
siderably for this operation. After spraying is
completed, loose, flocculent clumps of phyto-
plankton remaining on the inner surfaces of
the screens are washed down from the inside
with a moderately forceful stream of water.
On some occasions the phytoplankton was
washed directly into jars for later inspection.

It takes 10 or 15 minutes to clean each
filter in the manner described above, thus
making it impossible to clean them after
every 6.5-minute sampling interval. For
present purposes such frequent cleaning is
not necessary, but should it be desirable for
future operations, a quick-cleaning mechanism
could undoubtedly be incorporated into an
improved model of the filtering unit.

PERFORMANCE

With approximately 100 feet of hose out the
collector tows at a depth of 5 or 6 meters.
This was determined early in the fall of 1961
by signals telemetered from a Bourns pres-
sure potentimeter mounted on the lower hose-
cable terminal assembly. The device was
removed because of frequent malfunction, but
since vessel speed (9 knots) and length of hose
out (approximately 100 feet) were kept con-
stant during the later surveys, it is assumed
that all samples were collected at the same
depth. It should be possible to achieve greater
depth in the future by increasing the depressing
force and length of the hose.

Water is delivered to the filtering unit at
the rate of 9215 liters per minute which is
about 16 percent greater than the rate at which
water would freely enter a 3/4-inch diameter
orifice moving through the sea at a speed of
9 knots. The free flow rate of any orifice is

TT r^d k

■ liters per minute

where r is the radius of the orifice, d the
number of feet traveled per minute at speed x,
and k the factor for conversion from cubic
feet to liters. Since the pump orifice is
three-quarters of an inch in diameter and the
cruising speed of the vessel is 9 knots

2
3.1416 X 0.0312 X 912 x 28.32 = 78.95 Uters
per minute •

Thus there should be a field of suction ahead
of the collecting orifice, and the diameter of
this field would be the diameter of the core
of water being sampled.

Tests carried out on the collector in a
laboratory trough with a flow of 3 knots
suggests that the diameter of the core is
well under 2 inches at a speed of 9 knots.
Dyes and particulate matter released in the
tT'rough clearly defined a bulbous zone about 2
inches in diameter ahead of the orifice. Partic-
ulate matter was instantly pulled into the pump
if it drifted into this zone. If it missed the zone,
it drifted on past the collector undisturbed.

13

Whether the size of the zone of suction
fluctuates during actual towing due to agitation
of the water by the movement of the vessel
itself is not known, but it is at least reason-
able to assume that the orifice of the collector
is always preceded by a field of suction
rather than a field of back pressure. Back
pressure, which is often produced by clogging
in plankton nets, is undesirable because it
may delay the passage of some organisms into
the orifice or even deflect them from it.
Suction, on the other hand, will insure that
organisms are transported instantly from
the orifice through the hose to the filtering
unit.

Wiborg (1948) suggests that the faster mov-
ing organisms will succeed in avoiding the
currents at the mouth of the suction hose of
pump collectors. He concluded, after com-
parative tests with the Clarke-Bumpus sam-
pler and the Nansen closing net, that a pumping
rate of at least 200 liters per minute is
necessary for adequate sampling. Though this
figure may have some merit for pumps that
are virtually stationary, as was the one used
by Wiborg, it cannot be arbitrarily applied
to a pump intake that is moving through the
water at a good rate of speed. The size of
the suction zone, which depends on the ratio
of the rate of intake to the rate of travel,
would be less for a moving collector than for
a stationary one. Undoubtedly the movement
of the collector, rather than the small zone
of suction, would be the major cause of avoid-
ance with an instrument towed at 9 knots.

A second reason given by Wiborg (1948)
for the 200 liter-per-minute minimum pump-
ing rate is that at lesser rates it takes too
long to collect sufficiently large numbers of
some kinds of organisms. During the 1961
fall surveys the towed pump and shipboard
filtering unit, with a delivery rate of 92
liters per minute, collected early stages of
copepods and even the adult stages of
Calanus heligolandicus in numbers sufficient for
density' estimates of good precision.
Euphausiids occur in much lower, but still
sufficient numbers. The same is true of
chaetognaths. Fish eggs and larvae, on the
other hand, are among the organisms that
occur too rarely in the samples to yield

abundance estimates with a satisfactory de-
gree of precision. It is possible that these
would occur in greater numbers in other
areas and during other seasons of the year.

The upper size limit of organisms col-
lected by the towed pump and shipboard
filtering unit is indicated by the condition
of the euphausiids and the fish larvae in the
samples. Both are mutilated if they are
more than a half inch in length, and virtually
no organisms longer than three-quarters of
an inch occur in the samples. This may be
directly related to the physical characteristics
of the pump, i.e., diameter of orifices and
spacing of impeller blades.

The lower size limit of organisms collected
is suggested by the size composition of or-
ganisms escaping from the filtering unit. The
greatest detectable escapement was through
the filtering screen of the collecting bucket.
On five occasions the water remaining in the
funnel at the end of a 6.5-minute sampling
interval was collected in a jar when the
friction cap was removed from the bucket
screen. The number of organisms in each
escapement sample was estimated by vol-
umetric subsampling with replacement. Each
sample was stirred in 1,000 cc. of water and
12 aliquots of 25 cc. were removed, examined
and returned successively. The mean num-
bers per aliquot were multiplied by 40 to
obtain estimates of the numbers in the sam-
ples.

The estimated average numbers of copepods
escaping for five samples are tabulated in
three size categories in table 1 and illustrated
for eight size categories in figure 12. They
are not adjusted for differences in the volume
of water filtered in each of the five samples,
but volume varied very little between samples.

Percentage escapement and along with it
the minimum size for quantitative collection
cannot be definitely established until the re-
lation between the size composition of es-
caping copepods can be compared to the size
composition of the copepods retained by the
filters during the five pertinent sampling
intervals. These samples have not yet been
counted. However, the numbers of copepods
present in 129 samples collected in the same

14

TABLE 1. — The estimated average numbers of copepods escaping through the bucket screen per 6.5-
mlnute sampling interval in 1961 compared to the average numbers in the collecting bucket per
sampling interval in 1958

Copepod length

Average escaping per interval ■""

Average accumulated
per interval^

Escapement

Mm.

<0.2

0.2 - 0.5.
0.5 - 1.0.

Numbei

71

31

2

Number

2,266
612

Percent

1.3

0.3

^ From 5 sampling intervals taken at various times during 1961 surveys by collecting all the
water filtering through the bucket screen after removal of the friction cap at the end of each
interval.

^ From 129 sampling intervals taken on two grid patterns in the fall of 1958 in the same manner
but with a different arrangement of hose and pump.

<

Of

o
u.
o
q;

HI

CQ

S
ZD

100

200

300 400 500

LENGTH IN MICRONS

600

Figure 12. --The average length frequency distribution of copepods escaping through the bucket screen per
6.5 -minute sample interval. The light shading denotes nauplii and the hea\'y shading denotes copepodites.
No adult copepods were found in the escapement samples.

15

manner but with a different arrangement of
hose and pump in the fall of 1958 suggest the
probable order of escapement (table 1). The
average number of copepods between 200 and
500// in these earlier samples is 2,266, and
the average number between 500// and 1 mm.
is 612. Thus escapement of copepods in any
size category above 200// is not likely to be
more than a few percent. If necessary, es-
capement can probably be reduced by using
a filtering screen of smaller mesh size in
the collecting bucket.

No other zooplankters appeared in the es-
capement samples except larvaceans. The
estimate for the first sample is 23, and none
were found in the other 4. Since larvaceans
are present in considerable numbers in the
samples collected during the fall of 1961, it
is assumed that the above represents a
negligible escapement.

The main filtering screens in the funnels
are the only other obvious site of escapement.
Examination of 20 half-liter samples of water
taken from these screens at various times
during the 1961 fall surveys has revealed no
zooplankters. Though this result is encourag-
ing, the observations are too few to permit a
conclusive appraisal. The half-liter samples
are so small a fraction of the 600 liters or
so that are filtered during each 6.5-minute
interval, that the absence of zooplankters in
20 such samples could easily have occurred
by chance alone. Much more filtered water
will have to be examined.

A simple calculation and a laboratory test
indicate that organisms are not displaced to
any significant degree as they are trans-
ported from the collector to the filtering
unit. The length of the entire hose is 210
feet. Assuming that 1/2 inch of the 2-inch
diameter of the hose on the winch is occupied
by the support and electrical cables, the
diameter can be considered 1.5 inches for the
entire length. The volume of the hose, there-
fore, is

2
TTr Ik

72.8 liters

where r is the radius, 1 is the length, and k is
the factor for conversion of cubic feet to
liters. At the average delivery rate of 92

liters pe." minute the volume of the hose is
turned over every 48 seconds, or eight times
during the course of each 6.5-minute sam-
pling interval. This was verified by the fact
that spurts of dye introduced into the pump
intake in a laboratory trough took about 50
seconds to reach the other end of the hose.
The spurts of dye, furthermore, showed very
little diffusion as they moved through the hose
(a transparent vinyl hose was being used at
the time of these tests), suggesting that sus-
pended organisms would approximately main-
tain their relative positions during transport.

Zooplankters that do not descend into the
collecting bucket at the end of a sampling
interval are either deposited freely on the
inner surface of the filtering funnel or trapped
in the phytoplankton film clogging the main
filtering screens. Those deposited freely on
the surface could be washed into some sub-
sequent sample, thus decreasing sampling
accuracy. Those entrapped in the phytoplankton
would constitute a fraction of the sample that
is lost, also decreasing sampling accuracy.
Virtually no zooplankters could be detected
on the inner surface of the funnels by visual
inspection following 6.5-minute sampling
intervals, indicating that there is essentially
no loss of accuracy due to organisms remaining
freely on the filter surface. Considerable
numbers did remain on the surface, however,
during special sampling sequences where the
sampling interval was 2 hours rather than
6.5 minutes. Drainage rate was greatly re-
duced by heavy clogging of the bucket screens,
and many zooplankters settled out and remained
on the inner surface of the funnel as the
water receded after the friction cap was
removed from the bucket screen. Thus rapid
drainage is necessary to insure that all or
most of the zooplankters captured during a
sampling interval are deposited in the collect-
ing bucket. If sampling intervals are to be
longer than a few minutes, or if the rate of
clogging is increased by a greater rate of
water delivery or the use of a finer mesh
screen in the bucket, the area of the bucket
screen would probably have to be enlarged
to achieve rapid drainage.

Examination of the phytoplankton film, which
was removed from the funnels and preserved

16

TABLE 2. The estimated average numbers of zooplankters per 6.5-minute sampling interval entrapped

in phytoplankton on the filter screens in 1961 compared to the average numbers retained in the
collecting bucket per interval in 1958

Type organism

Sample periods

L

Average entrapped
per interval, 1961

Average collected
per interval, 1958^

Suggested

1

2

3

Total

entrapment

All copepods

Larvaceans

Doliolids

Number

1,380
1,680

180

Number Number

540 930

1,080 1,020

30 30

Number Number

2,850 29

3,780 38

240 2

Number

2,907.0

88.0

0.1

Percent

1.0
30.2

^ Period 1 had 45 sampling intervals, period 2, 18, and period 3, 36 for a total of 99. At end of
each period all material adhering to main filter screens was removed and preserved for later enumer-
ation.

^ 129 sampling intervals.

on several occasions, showed that a small
portion of the zooplankters entering the filter-
ing funnels are trapped in this material. The
numbers of zooplankters in three such phy-
toplankton samples, as estimated by volu-
metric subsampling with replacement, are
given in table 2. Since the phytoplankton was
accumulated over different numbers of 6.5-
minute sampling intervals in each case, the
estimates are summed and divided by the total
number of intervals for the three samples
and expressed as average numbers per sam-
pling interval.

Copepods, larvaceans, and doliolids were
the only zooplankters found in the phyto-
plankton. The larvaceans are somewhat more
abundant than the copepods, while the doliolids
are far less abundant. Comparison to the
average numbers of such organisms present
in the 129 samples collected in 1958 suggests
that the percentage loss of copepods is negli-
gible, but that^the loss of larvaceans may be
higher than 4» percent. If entrapment of
larvaceans in the phytoplankton proves to
be this high in terms of the matching 1961
samples, the towed pump and shipboard filter-
ing unit would have to be considered unsuitable
for making quantitative collections of this
group. No worthwhile assessment can be
made of entrapment of doliolids because they
are virtually nonexistent in the 1958 samples.

Except for larvaceans, then, it appears that
the loss of zooplankters through escapement
and entrapment is probably not more than a
few percent, and that loss of accuracy through

mixing between consecutive samples is prob-
ably negligible. For most purposes such
minor losses can be disregarded, but if
necessary the losses from escapement and
entrapment can be measured directly so that
corrections can be applied to sample esti-
mates.

DISCUSSION

The towed pump and filtering unit was
designed for the purpose of making estimates
of known precision of small zooplankters for
areas of specified sizes. Operations during
the fall of 1961 indicate that the system
performs this task satisfactorily. Zooplankters
between the lengths of 0.2 mm. and 12.0 mm.
(approximately 1/2 inch) are collected quanti-
tatively and the collection of samples that
are discrete for consecutive 6.5-minute inter-
vals of vessel travel at cruising speed is a
simple procedure. A number of such small
samples can be taken as subsamples from a
given area.

The present operational status of the towed
pump and shipboard filtering unit, particularly
of the collector and the hose-cable assembly,
was achieved only after considerable experi-
mentation and modification. The towed pump
and hose-cable suspension are patterned after
a model originally designed and built for the
Bureau of Commercial Fisheries Biological
Laboratory, La Jolla, by the Fisheries Instru-
mentation Laboratory of the Bureau of Com-
mercial Fisheries Biological Laboratory,

17

Seattle, Washington. The original collector
was lost due to failure of the support cable
during early sea trials, but the trials showed
that certain changes were needed to make the
unit durable and reliable. Among these were
a heavier cable and linkage elements, heavier
hoses, enlarged hose-cable terminal assem-
blies, a rigid bridle, and a smaller pump. The
earlier hose, which was 2-inch I.D. thin- walled
vinyl tubing had the advantage of transparency,
but it stretched excessively with internal
pressure and was also susceptible to the
development of weakpoints and subsequent
small leaks at sites of folding and abrasion.
The earlier chain bridle, though simple to
fabricate, fouled too easily during setting of
the collector. The earlier pump, which had a
1-inch diameter intake, frequently overloaded
the electrical system and stopped operating.
It required power in excess of that obtainable
from the 1/2-hp. motor, and a larger motor
would have required power in excess of that
obtainable from the vessel.

Though the above changes have made the
collector and hose-cable assembly reliable
and durable, they do not make the instrument
capable of sampling at various depths. This
will require, in addition to further changes
in the collector and hose-cable assembly,
very considerable improvements in the winch.
The winch in its present configuration does
demonstrate that a hose of considerable length
can be reeled in and out while water is being
pumped, but the unit is not strong enough to
sustain the load that a longer hose and greater
depressing force would produce, nor could it
accommodate a hose much longer than that
now being used.

Sampling at various depths will also require
continuous depth-of-tow information. The in-
clusion of reliable depth sensing and tele-
metering apparatus will be a priority item in
the design of future models of the system.

The only difficulty with the filtering unit
was the appearance of pinhead rust spots on
the filtering screens after the two funnels
together had been exposed to nearly 600
cubic meters of sea water (approximately 100
hours of sampling). These were patched with
epoxy to forestall the development of holes.
It is probably unreasonable to expect longer

service than this from stainless steel screen-
ing as delicate as that used here. Future im-
provement in the filters should include some
simple way of replacing these surfaces at
intervals. Monel screening might well elimi-
nate this problem, but it is not available in
mesh sizes as small as that used here.

The temperature sensing and recording
apparatus incorporated into the filtering unit
should also be improved in the future. The
electronic components do not perform reliably
enough for sustained operations, and the loca-
tion of the sensing element at the filters
rather than in the collector is questionable.
It may well be that the temperature increases
as the water moves from the collector through
a few hundred feet of hose to the sensing ele-
ment. Submersible temperature sensing units
and various kinds of telemetering equipment
are available so that a satisfactory solution
should be a matter of integrating a unit with
the desired performance characteristics into
the existing system or some later model of it.

SUMMARY

1. A towed plankton pump and shipboard
filtering system for sampling small zooplank-
ters has been designed for use aboard the
Bureau of Commercial fisheries vessel
Black Douglas.

2. The system consists of (1) a collector
towed by a suspension unit consisting of a
hose through which runs a steel cable for
support and an electric line to power the
pump, (2) a winch, and (3) a filtering unit
composed of a watermeter, a double-throw
valve and two filtering funnels.

3. The collector, which contains a 1/2-inch
bronze centrifugal pump and an hermatically
sealed, capacitor starting, 1/2-hp. A.C. motor,
is mounted in a frame bridle suspended from
a terminal hose-cable assembly made of
galvanized pipe and fittings. A 43-pound homo-
geneous depressor is suspended from the
bottom of the bridle.

4. The hose is a 2-inch I.D. single- jacket
firehose and both the 1/4-inch stainless steel
support cable and the neoprene covered power-
line extend to the deck of the vessel through
the hose.

18

5. The winch is a steel drum 4 feet long and
3 feet in diameter with a 1,5- inch pipe axle.
The water stream is dissociated from the
support cable and the powerline at a terminal
assembly mounted on one edge of the drum.
Rotation is controlled by a reversible 3/4-hp.
A.C. gear motor with an automatic brake.

6. A 1.5-inch I.D. polyethylene hose ex-
tends from the winch to the filtering unit,
which consists of a wobble plate type water-
meter, a quick acting double-throw valve and
two filtering funnels set in trash cans. Fil-
tered water flows from the cans through
outlet hoses to the scuppers of the vessel.

7. Each filtering funnel has windows
screened with 105 //-mesh stainless steel
filtering cloth. A collecting bucket bayonet
mounted on the neck of the funnel has a bottom
screen of the same cloth which is covered
by a friction-fitted cap while the funnel is
filtering.

8. Removing and preserving a sample takes
2 or 3 minutes, so the filtering unit is easily
operated by one man.

9. With 100 feet of hose out, the collector
tows at a depth of 5 or 6 meters at a vessel
speed of 9 knots. It has been operated con-
tinuously for periods of about 72 hours.

10. E scapement through the screen of the col-
lecting bucket and the size of mutilated zoo-
plankters indicate the lower and upper size
limits of quantitative collection to be 0.2 mm.
and 12.0 mm. respectively.

11. The error due to escapement through
the filtering surfaces appears to be negligible
between the above sizes.

12. Better sensing and recording apparatus
is needed for depth-of-tow and temperature
information and considerable improvement is
needed in the winch if the system is to be
used for towing at various depths.

ARON, WILLIAM.

1958. The use of a large capacity portable
pump for plankton sampling, with notes
on plankton patchiness. Journal of
Marine Research, vol. 16, no. 2, p. 158-
173.

CALIFORNIA, STATE OF, MARINE RE-
SEARCH COMMITTEE,
1950. California Cooperative Sardine Re-
search Program, Progress Report,
1950. Sacramento, State Printer, 54 p.

CASSIE, R.M.

1959. An experimental study of factors
inducing aggregation in marine plankton.
New Zealand Journal of Science, vol. 2,
no. 3, p. 339-365.

COLLIER, ALBERT.

1957, Gulf-II semiautomatic plankton sam-
pler for inboard use. U.S. Fish and
Wildlife Service, Special Scientific
Report — Fisheries No. 199, lip.

GEHRINGER, JACK W.

1952, An all-metal plankton sampler (model
Gulf III). In High speed plankton sam-
plers, U.S. Fish and Wildlife Service,
Special Scientific Report — Fisheries
No. 88, p, 7-12.

HAND, CADET H., and LEO BERNER, JR.
1959. Food of the Pacific sardine (Sardinops
caerulea). U.S. Fish and Wildlife Serv-
ice, Fishery Bulletin No. 164, vol. 60,
p. 175-184.

HARDY, A. C.

1939. Ecological investigations with the
continuous plankton recorder: object,
plan and methods. Hull Bulletins of
Marine Ecology, vol. 1, no. 1, p. 1-57.

LITERATURE CITED

AHLSTROM, ELBERT H., JOHN D. ISAACS,
JAMES R. THRAILKILL, and LEWIS W.
KIDD.
1958. High-speed plankton sampler. U.S.
Fish and Wildlife Service Fishery Bulle-
tin No. 132, Vol. 58, p. 187-214.

19

WIBORG, K. F.

1948. Experiments with the Clarke- Bumpus
plankton sampler and with a plankton
pump in the Lofoten area in northern
Norway. Fiskeridirektoratets Skrifter,
Serie Havunders0kelser (Reports on
Norwegian fishery and marine investi-
gations), vol. 9, no. 2, 32 p.

MS #1269

GPO 939988

"BL WHOI Library - Serials

5 WHSE 01579

Created in 1849, the Department of the Interior — America's
Department of Natural Resources — is concerned with tlie man-
agement, conservation, and development of the Nation's water,
fish, wildlife, mineral, forest, and park and recreational re-
sources. It also lias major responsibilities for Indian and
Territorial affairs.

As the Nation's principal conservation agency, the Dej^art-
nient works to assure that nonrenewable resoiu'ces are developed
and used wisely, that park and recreational resources are con-
served for the future, and that renewable resources make their
full contribution to the progress, prosperity, and security of
tlie United States — now and in the future.