Download Vertical migration for horizontal transport while

Journal of Plankton Research Vol.19 no.12 pp.1929-1947, 1997
Vertical migration for horizontal transport while avoiding
predators: I. A tidal/diel model
J.L.Manuel and Ronald K.O'Dor
Biology Department, Dalhousie University, Halifax, Nova Scotia B3H 4J1,
Canada
Abstract. Research into the vertical migration behavior of scallop (Placopecten magellanicus)
veligers has led us to examine whether these, and possibly other small zooplankters, may migrate in
response to a combination of tidal and diel stimuli. This paper uses Hill's (1991) model to evaluate
the horizontal transport effects of such migrations. We demonstrate that most types of vertical
migration behavior reported in the literature (e.g. nocturnal, twilight, midnight sink) appear at different phases of the lunar cycle. Moreover, migrating in response to both of these cues may provide horizontal transport advantages if the zooplankter is very small (unable to migrate the full water column
depth) and/or has difficulty determining its position in the water column (especially if the behavior
also holds it in regions of increased shear). Such behavior need not interfere with other advantages
of vertical migration, including avoiding predation, avoiding UV light, searching for patchy food, etc.
Tidal/diel migration may have distinct advantages for occupying new habitats or coping with local
changes associated with altered current regimes. Because averaging the results of several days, sampling less frequently than every 2 h or sequential sampling of different sites is likely to obscure the
tidal portion of a tidal/diel migration, such behavior could be common without being obvious to
researchers. Aliasing of the lunar and solar cycles (a 14.8 day period) may allow the detection of tidal
period migrations in long-term records with lower sampling frequencies.
Introduction
Vertical migration of planktonic organisms is found in all marine and freshwater
phyla (Huntley, 1985). Both the proximal (responses to stimuli such as light, temperature and salinity) and the ultimate (horizontal transport, avoiding predation,
finding patchy food) reasons for vertical migrations have been studied (Lampert,
1993). Enright and Hamner (1967) showed that even the physiology underlying
endogenous rhythms is by no means uniform. It is, therefore, reasonable to
assume that vertical migration behavior has evolved several times, for a variety
of purposes, and there may be several ultimate causes of this behavior.
Of the many proposed ultimate causes of vertical migration, two have been
widely substantiated in both laboratory and field studies. There has been considerable recent evidence that predator avoidance can be the ultimate cause of
vertical migration in crustacean zooplankton (reviewed by Lampert, 1993). The
strength of vertical migration of prey is positively correlated with predator abundance (Bollens and Frost, 1989b; Bollens et al., 1992), marine copepods migrate in
response to free-ranging, but not caged fish (Bollens and Frost, 1989a), and, in
fresh water, zooplankton responses to changes in light intensity are greatly
enhanced by the chemical exudates of a predator (Ringelberg, 1991). However,
Osgood and Frost (1994) found very few similarities in the behavior of similarsized individuals of two co-occurring species of copepod (Calanus pacificus and
Metridis lucens), and also that vertical distributions and behaviors vary between
species, among the different developmental stages of each species and between
© Oxford University Press
1929
J.LJManoel and R.K.O'Dor
dates. This argues for at least one other factor in vertical migration besides
predation pressure.
Vertical migration for the purpose of horizontal transport was first proposed
by Hardy and Gunther (1935), who suggested that oceanic plankton, too small to
make effective horizontal migrations by swimming, might migrate vertically to
allow currents to carry them horizontally. A related hypothesis proposed by
Rogers (1940), that vertical migration through the pycnocline prevents organisms
washing out of estuaries, has been widely substantiated (see Sinclair, 1988). Even
in bays where stratification is weak, shear created by bottom friction can affect
horizontal transport, and vertical migration has been shown to be used by resident (but not by non-resident) species to mitigate against washout (Kimmerer
and McKinnon, 1987). Even very small organisms such as dinoflagellates are able
to reduce losses from enclosed bays by exhibiting appropriate vertical migration
behavior (Anderson and Stolzenbach, 1985).
Scheltema (1986) suggested that, given the importance of vertical movement
for retention in estuaries, oyster (Crassostrea virginica) and clam (Mya arenaria)
veligers might exhibit different behavior in different types of estuaries. Manuel
et al. (1996a,b) demonstrated such differences in vertical migration behavior in
one offshore and two coastal populations of the sea scallop (Placopecten magellanicus). While they observed a fairly consistent diurnal pattern in the data,
careful examination of the data from that study and a previous one (Gallager et
al., 19%) also showed variations in the migration patterns (which were initially
considered noise) that showed evidence of a superimposed progressive, lunar
periodicity that was not consistent among populations (Manuel, 1996).
Open-ocean systems such as continental shelf regions are less well studied.
Benthic organisms with long pelagic larval stages (such as deep-sea scallops) may
suffer from over-dispersal in such habitats. Currents are virtually ubiquitous, and
any larva that acts as a passive particle will be carried some distance from spawning beds. If the adult organisms are patchy in distribution, these organisms face
problems similar to those of organisms that inhabit reefs or islands. The problem
is, however, less evident to an observer. Behavior is recognized as crucial for
retention near oceanic islands and reefs, because it is obvious to a researcher that
without a retention mechanism, organisms would be flushed seaward. In other
words, where physical discontinuities can be clearly visualized, evidence is substantial that behavior of the organism aids in retention in a suitable habitat (see
Sinclair, 1988).
The member/vagrant hypothesis (lies and Sinclair, 1982) proposes that retention of larval stages in aggregations is an important factor in the year-to-year variability of recruitment in species with pelagic larval stages, and that larval behavior
may play an important role in such retention (Sinclair, 1988). Members of a population return to the same location to spawn over many generations, and larvae
evolve behavior appropriate for survival at that location. From time to time,
vagrants may spawn at other locations, but the progeny of the vagrant exhibit
larval behavior inappropriate for that location, and even if they do survive may
not return to that site to spawn again. Thus, the vagrant is 'penalized' in two ways:
by inappropriate behavior both as a larva and as an adult. This would reduce the
1930
Vertical migration: a tidal/die) model
flow of genes between populations that have been selected for different behavior,
and create an effective barrier to geneflow.That this might indeed be the case is
seen in species such as Atlantic herring, which mix and feed together at most
times of the year, yet separate into subpopulations for the purpose of spawning
(Sinclair, 1988). If retention, and thus survival, depend on appropriate larval
behavior, then not all individuals in a species with more than one population are
able to restock locations where numbers have been decimated by over-fishing. In
fact, the influx of vagrants from other populations could 'swamp' the gene pool
and cause further damage. Thus, establishing the significance of larval behavior
in different populations is important for managing all stocks that contain more
than one breeding population.
Because of the complex nature of water movements in tidal areas, it is often
difficult to attribute changes in the vertical distribution of field-collected zooplankton to swimming alone. That is, it is difficult to sort out passive transport
from active swimming behavior. We looked to the freshwater literature because
in fresh water tides are not as large, and changes in vertical distribution due to
the active movement of zooplankton are more easily perceived. Harris (1963)
looked at endogenous rhythms in the freshwater cladoceran Daphnia magna.
When kept in continuous darkness, the crustaceans exhibited a diurnal migration
pattern with a 24 h period. However, there were secondary peaks upon which the
M
N
M
N
N
M
Fig. 1. Mean depth (arbitrary units equal to 1/10 of the water column) of a population of Daphnia
magna kept in continuous darkness. The black rectangles and shaded regions indicate ambient night.
M is midnight, N is noon. As indicate secondary mid-day rises and Bs indicate secondary mid-night
dips in mean depth. Modified from Harris (1963).
1931
J.UMinuel and R.K.O'Dor
author does not comment (Figure 1). Specifically, there is a mid-day rise and a
mid-night sink in mean depth. Both of those secondary peaks seem to drift later
each day, which is consistent with a lunar cycle. It is possible that these crustaceans were exhibiting both diel and tidal cycles in their vertical migratory
behavior.
While tidal flows are the most consistent and reliable feature of current fluctuation in offshore waters, and thus represent great potential for use in horizontal
transport, the biggest stumbling block in generating a hypothesis for their role in
transport outside of estuaries has been a mechanism for a zooplankter to perceive
the tidal phase. If zooplankton were able to interpret the tidal phase from the
time of moon rise and/or moon set, then they could use migration on a tidal cycle
for the purposes of horizontal transport. The time of moon rise varies during the
lunar cycle (Figure 2). Around the full moon, because the moon and sun are at
opposite sides of the Earth, the moon rises near dusk and sets near dawn. Around
the new moon, the moon rises near dawn, and is up, but not visible, during the
day. In the first quarter, the moon rises during the day and sets near midnight. In
the third quarter, the moonrisesaround midnight and sets during the day. Gliwicz
(1986) found, in a freshwater lake, a lunar cycle in the abundance of six species
of freshwater zooplankton that was the result of predation. Zooplankton avoided
surface waters in the first quarter and full moon (when the moon rose before the
sun set), and were closer to the surface during the new moon. During the third
quarter, when the moon did not rise until after the sun set (Figure 2), zooplankton were easy prey in near-surface waters after moon rise. This 'moon trap' might
easily drive selection for a lunar cycle in zooplankton behavior that would allow
zooplankton to avoid being caught at the surface when the moon rises. If the
'moon trap' is a common phenomenon, it could select for individuals with vertical migration behaviors that change with the lunar cycle. Such behaviors could
be controlled by a combination of endogenous rhythms and light cues that set
NrvMoon
Fig. 2. Time of moon rise and moon set relative to sunrise and sunset through a lunar month. Small
clear boxes indicate time when the sun is up, small gray boxes indicate when the moon is up. Large
boxes: clear indicates the sun is up, gray indicates night with the moon up, and black indicates night
with no moon.
1932
Vertical migration: a tidal/diel model
internal clocks, and might then be adapted for other purposes, such as using tidal
currents for horizontal transport.
Another problem for very small organisms, such as bivalve veligers, is that they
are unable, because of their small size, to migrate the full water column to obtain
horizontal transport. Therefore, the individual does not obtain all of the potential transport available from the shear in the water column. In addition to this,
the organisms may have greater difficulty using encounters with discontinuities,
such as the surface or the bottom, to locate themselves with respect to depth, since
responses of organisms seem to rely on the rate of change in a parameter such as
light (Forward, 1988) or temperature (Angel, 1968), rather than the absolute
intensity of the stimulation.
This paper examines the vertical distribution patterns and horizontal transport
produced if a small plankter migrates vertically both diurnally to avoid predation
and tidally for the purpose of horizontal transport. We have fitted simple behaviors incorporating both solar and lunar cycles to Hill's (1991) model of tidal currents. We use, as an example, the potential for altering transport that tidal/diel
migration would provide to a scallop (P.magellanicus) veliger on Georges Bank.
By examining the transport effects of variations in veliger behavior, we demonstrate the advantages of tidal/diel migration for small zooplankton. This has provided at least a partial explanation of the selective pressures that might lead to
the evolution of such behavior. The principles elucidated probably have broader
application beyond transport for scallop veligers on Georges Bank.
Method
The velocity of a water particle in a tidal current can be described by an equation
that accounts for the changes in water velocity with tidal phase and depth:
£/ M = sin(arf + <|>)£/o(z//i)»
(1)
where f/(u) is the velocity at depth z and time /, Uo is the velocity at the surface,
h is the total height of the water column and n is a power function describing the
shear (differences in water velocity) in turbulent boundary conditions.
The function wt + <}> describes the tidal phase at time t, where <»f is the tidal phase
(if <(> = 0, oscillation begins with (/(z^ = 0 at / = 0) and
co = 2ir/r
(2)
where T is the tidal period (Hill, 1991).
In equation (1), the function
sin((o/ + <j>)
describes the change in water velocity through time. This function is represented
in Figure 3. The sine curve in Figure 3A (solid line) describes the velocity of
diurnal tidal currents with a period of 2TT (or 360°) and a maximum velocity of
1933
J.LManuel and R.K.O'Dor
A
u.
/A\
/7\\
-V.
1142 h
TimeQi)
Velocity (m/i)
Fig. 3. (A) Change in current in a tidal regime over time. The solid line describes the velocity of diumal
tidal currents with a period of 2-ir (or 360°) and a maximum velocity of Uo The dotted line describes
a case where the tide is 4.5 h later. (B) A tidal ellipse, which also shows the direction of the current,
of the same tidal regimes. The open circle locates a common time (high tide for the solid ellipse), tick
marks indicate lunar hours, the center of the ellipse shows the position of the sample, the distance
from the center indicates the strength of the current, and a line drawn from the center to the edge of
the circle indicates current direction. (C) Left of center shows how current velocity changes with depth
where the relationship between current and depth is linear. Where the relationship between current
and depth is linear, n = 1. Right of center shows how current changes with depth where the relationship between current and depth is non-linear. Dotted line, n = 1/3; solid line, n = in.
Uo The phase of the current determines where in the curve the cycle begins: the
solid line represents the case where wt + <\> = 0. The dotted line in Figure 3A
describes a case where the tide is 4.5 h later (i.e. where t = 4.5 h). Assuming a tidal
period (T) of 12.42 h, from equations (1) and (2) we have:
uf = (2nlT)t
tor = (360/12.42)4.5 = 130°
Thus, the solid line and the dotted line represent two tidal functions that are 4.5
h (or 130°) out of phase with each other. We assumed a tidal period (T) of 12.42 h
1934
Vertical migration: a tidal/diel mode)
[the period of the moon (M2) tidal currents] for all tidal migrations except those
that calculate the effect of currents due to sun tides (S2), where we used a period
of 12 h.
The same current may be plotted as a tidal ellipse (Figure 3B), which also shows
the direction of the current. The center of the ellipse locates the station, and the
ellipse describes the current speed over a tidal cycle. The direction and speed of
the current are represented by the direction and distance from the center. Each
tick mark represents a lunar hour and there are 12 lunar hours in a tidal cycle.
Thus, with a tidal period of 12.42 h, each lunar hour would be 1.035 h.
In equation (1), the function
Uo (z/h)"
describes the change in current with depth. We began with a value of 1 m r 1 for
maximum velocity (Uo). We chose an arbitrary depth of 1 for the water column,
and z varied from 0 at the bottom to 1 at the surface. Hill suggests a value of 1/7
for n under turbulent boundary conditions. The exact value is not important for
demonstrating the principal effect of various migratory behaviors, but would be
important in calculating the real amount of transport obtained. What is important for the present purposes is that shear (effectively the difference in velocity
in different parts of the water column) is non-linear. The significance of this equation is that the change in velocity that occurs with depth is greater near the bottom
than higher in the water column, and thus the horizontal transport obtained by
vertical migration is greater deeper in the water column than near the surface
(Figure 3C). In fact, differences in velocity with depth may also be associated with
differences in tidal phase with depth, and shear created by differences in phase
may be more important than boundary effects for scallop veligers (see Manuel et
ai, this issue). At least on Georges Bank, this difference in phase results in significant shear in the region of the thermocline, with less shear in the upper parts
of the water column (Brown and Moody, 1987) (Figure 4). Differences in shear
can also be seen at different tidal phases on Georges Bank. For example, at lunar
hours 5 and 11, there is little difference in the speed of currents throughout the
water column. At lunar hours 1 and 7, the differences in velocity are more substantial (Figure 4).
We refer to migrations with a period of 12.42 h as 'tidal' and to the proximal
cue that induces the migrations as the 'tidal cue'. This implies neither that the
zooplankter must have the ability to perceive changes in current or the state of
the tide directly, nor that the migrations necessarily occur exactly at high or low
tide. Similarly, we refer to migration with a period of 24 h as 'diel' migration and
the proximal cue that induces the migration as the 'solar cue'. For simplicity we
have assumed a day length of 12 h, and equal, semi-diurnal tides.
We used simple numerical integration of this equation (every 10 min through
the lunar cycle) to calculate the net transport of a zooplankter migrating with
both diel and tidal periods. The term 'transport' means the horizontal movement
obtained by a veliger's vertical migration. Initial parameters for the tidal/diel
model were: (i) veligers are unable to migrate the full water column upon
1935
J.LManoel and R.K.O'Dor
Carront (mi*)
Fig. 4. A sequence of predicted M2 tidal current profiles on the northeastern part of Georges Bank.
The first and second halves of the tidal cycle (solid and dashed lines, respectively) are distinguished.
Time is indicated in lunar hours and is arbitrarily referred to the time of maximum on-bankflow.Note
that shear is high in the region of the thermocline (shaded) and opposite to the shear created by
bottom friction. Modified from Brown and Moody (1987).
receiving a cue; (ii) veligers moved only for 1 h following a cue; (iii) veligers could
not move above the surface or below z = 0; (iv) solar and tidal cues are additive.
Each tidaL/diel model starts one veliger at each of five depths (0, 0.25, 0.50, 0.75
and 1) and is evaluated by five criteria defined as follows.
(i) Power: the mean net velocity of transport at the five depths. More Power
confers a greater ability to move to another place or counter net flows.
(ii) Effort: a measure of how many times the zooplankter has to swim up the
equivalent of the height of the water column each day. We have assumed
that swimming upwards requires extra effort, while swimming downwards
does not. Assuming that effort was expended in swimming downward would
have doubled the absolute value of effort, but not altered the conclusions.
(iii) Gain: the distance (m) that a veliger would be transported in 1 day (Power
converted to m day"1). Gain is equivalent to the difference between net
transport without vertical migration and net transport with migration. Thus,
a veliger that migrated vertically to stay in place is considered to have gained
transport relative to net currents, even though actual movement was zero.
(iv) Efficiency: the amount of Gain for Effort expended.
(v) Reliability: the uniformity of transport at different depths. This is a measure
of how quickly and to what degree veligers at different depths end up at the
same depth. The more quickly this happens, the more veligers will tend to
aggregate and migrate in synchrony. Reliability can be represented visually
by plotting the paths of veligers that begin at different depths on the same
graph and noting the depth over which the different plots are distributed at
any given time (compare Figure 6B and C: the latter is more Reliable
1936
Table L Power, effort, gain and efficiency of variations on the tidal/diel model
Variation
on the
tidal/diel
model
Distance moved
on tidal signal
(% water column)
Distance moved
on light signal
(% water column)
Other restrictions
Power: mean net
speed (m s~')
Effort: distance
per day
(% total depth)
Gain: movement
(m day 1 )
Efficiency:
gain/effort
12.5
25
12.5
25
50
50
25
25
0.032
0.062
0.140
0.081
0.034
0.110
34
69
138
94
66
66
2753
5329
12 057
6985
2974
9508
8097
7723
8737
7431
4506
14 406
LIGHT:TIDE
25
25
0.049
68
4203
6180
TIDE.UGHT
25
25
0.043
69
3685
5340
TIDE»l/4
TIDE*l/2
SUNKJ.l
SUN:0.5
25
50
25
25
_
_
50
50
None
None
None
None
Cannot double speed
Accepts tide up signal
only if no light signal
Simultaneous signals,
light over-rides tide
Simultaneous signals,
tide over-rides light
None
None
S? current, 0.1 m s-1
S2 current, 0.5 m s"1
0.064
0.091
0.092
0.135
48
96
94
94
5547
7890
7925
11686
11556
8219
8430
12 432
BASE*l/8
BASE»l/4
BASE*l/2
BASE'A
BASE'B
BASE*C
25
50
25
25
J.L-Manuel and R.K.O'Dor
because all paths become the same in the first day), or measured by
comparing the net speed of transport of veligers that begin a lunar cycle at
different depths [e.g. in Table II, BASE*C is more Reliable than TIDE*l/4
because veligers in the lower three (versus one) depths obtain reasonable
transport]. This is important for small animals in turbulent environments
with few clues as to absolute depth. If a behavior is Reliable, then an individual that is moved from a depth is soon returned. Note that Reliability
does not require that the individual perceive the change in depth. Reliability
requires only that an individual responds consistently to stimuli (e.g encounters with discontinuities such as the surface, thermal gradients or the
bottom).
We began with veligers moving 1/4 of the water column on each cue, up at low
tide and at dusk, and down at high tide and dawn. The tidal/diel model is onedimensional in its calculation. However, in an elliptical tidal regime (Figure 3),
changes in the tidal phase, but not the period of migration, simply produce net
transport in different directions. Next, we examined the effect of a series of
altered assumptions about veliger behavior (Table I).
(i)
The distance traversed on each cue is 1/8 (BASE*l/8), 1/4 (BASE*l/4) or
1/2 (Base*l/2) of the water column. Zooplankton swim at a variety of
speeds and may encounter boundaries (e.g. pycnoclines or bottom) at a
variety of depths. We consider the depth moved as a portion of the water
column to encompass the variety of circumstances that might be encountered by a zooplankter.
(ii) The migration in response to the solar cue is double the migration in
response to the tidal cue (BASE*A). Crustacean zooplankton have been
shown to modify diel vertical migration in response to both starvation
(Huntley and Brooks, 1982) and predators (Bollens and Frost, 1989a,b).
This variation addresses the question of how increasing the diel migrations
might affect transport obtained by the tidal migrations.
(iii) Veligers cannot double their speed if they receive two cues at the same time
(BASE*B). The initial premise is that small zooplankton are unable (or it
is too energetically expensive) to migrate the full water column. This variation examines the concurrence of the two cues (solar and tidal) every 7.4
days, and considers what happens on days when the veliger receives two
cues to swim up or down at the same time.
(iv) Veligers only swim up in response to low tide if there has been no solar cue
in the past hour (BASE*C). Other tidal/diel models (e.g. BASE*B) exhibit
a penalty when the concurrence of cues moves veligers away from the
deepest water level. This variation shows the effect of removing that
penalty. There are a number of behaviors that might have the same effect
(e.g. responses to pressure, light or other parameters that might change with
depth, any response that results in greater downward than upward
migration, etc.), and we present this as an example.
1938
Vertical migration: a tidal/diel model
(v)
The solar cue over-rides the tidal cue (LIGHT:TIDE) or the tidal cue overrides the solar cue (TEDE:LIGHT). These variations explore the effects of
two different hierarchies of response when solar and tidal cues coincide.
(vi) Tidal migration alone, traversing 1/4 (TIDE* 1/4) or 1/2 (TIDE* 1/2) of the
water column. This variation is presented so that we can compare the transport effects of tidal/diel migration with migration in response to tidal cues
alone.
(vii) The effect of solar (S2) tidal currents of 0.1 m s"1 (SUN:0.1) and of 0.5 m s"1
(SUN:0.5) on the tidaL/diel model. Hill (1994) has shown that diel migration
may have a considerable effect on the zooplankton assemblages in the
North Sea. S2 tidal currents transport zooplankton making diel migrations
in a direction that is determined by the time of sunrise and sunset relative
to the phase of the currents at any given location (Hill, 1995). This variation
considers the maximum effect transport by solar tides could have for a
veliger that is migrating in response to both solar and tidal currents.
Results
The BASE*l/4 tidal/diel model exhibited patterns of migration that changed
through the lunar cycle, with a semi-lunar period (Figure 5). Days 1 and 16
produce a sharp rise at dusk, downward migration after midnight and the greatest depth in late afternoon. Days 4 and 19 saw veligers shallowest in the middle
of the night, with greatest depth just after dawn and before dusk. Sampling at
midnight, dawn and dusk would reveal only a diel migration, not the tidal component. At days 8 and 23, there is a precipitous drop to the greatest depth just
after dawn, a rise at dusk and a further rise around midnight. At days 16 and 29,
the greatest migration is upward at dusk and there is downward movement after
midnight. This pattern has also been reported in thefield.At days 12 and 27, midday is the deepest point, and there are dawn and dusk rises. This pattern is the
equivalent of 'twilight' migration often reported in the literature. An observer
collecting data for only a few sequential days would find a different pattern
depending on the time of the lunar cycle when the samples were collected.
Table I records the Power, Effort, Gain and Efficiency of the variations of the
tidal/diel model, and Table II compares the net speed of transport for veligers that
Table IL Net speed of transport of veligers, starting at various depths, after one lunar month
Variation
Net speed (ms-')
model
Depth 0
Depth 0.25
Depth 0.50
Depth 0.75
Depth 1
BASE* 1/4
TIDE* 1/4
TIDE*l/2
BASE*A
BASE*B
BASE*C
LIGHT:T1DE
T1DE:LIGHT
0.125
0.245
0.271
0.119
0.055
0.128
0.055
0.100
0.129
0.032
0.089
0.119
0.059
0.128
0.055
0.065
0.025
0.018
0.035
0.123
0.025
0.132
0.059
0.020
0.016
0.013
0.031
0.022
0.017
0.086
0.038
0.014
0.015
0.013
0.031
0.020
0.017
0.076
0.036
0.014
1939
J.L.Manuel and R.K.O'Dor
Fig. 5. Vertical migration of a veliger over a 30 day period that begins mid-water and responds to both
light (swimming up during the night and down during the day) and tidal (swimming up on incoming
and down on outgoing) cues. The cycle begins when tidal and solar cues coincide. The cycle of 14.8
days is repeated twice in one lunar cycle. Depth is expressed as the proportion of the water column
with 0 as the bottom and 1 as the surface. The solid line is the vertical position of the veliger. Black
boxes indicate darkness, numbered clear boxes indicate day.
begin at different depths (0, 0.25, 0.50, 0.75 and 1) in selected tidal/diel models.
The greatest transport in the BASE*l/4 tidal/diel model is in the two shallowest
depths, because animals that are in the upper water layers never move down to
the lower depths where shear is greatest (see Figure 3B). Effort is equivalent to
swimming up 69% of the water column each day (Tables I and II, BASE*l/4)
The thennocline acts as a stimulus to migration in scallop (P.magellanicus)
veligers, at times acting as a barrier and at other times aggregating veligers
1940
Vertical migration: a tidal/diel model
n
a
11 u
a
a
n
n
a
a
a ra
HI
in in ra ra
fi D: n
u
Tide* 1/4
EI
H
n
u
[•
u n:
IJI
fU
a
a
B: Bate* 1/4
n__n
C: Base* 1/2
n
a
in Qi
DJ
in in ra
.7J
JO
.25
ao
1.0
.75
30
.25
Oil
M
M
M
_n - i
•• n
H
ci u
u
u
ti
a
n
n
O:Bue*C
H
in in
a
t
IH
rn rn m
iu in
IM
in ra
.75
30
.25
ao
naananntmamuimiamm
Fig. 6. Depth of veligers as a portion of the water column when they respond to both solar and tidal
cues by making vertical migrations. The depth of five veligers that begin at five different depths (0,
0.25,0.50,0.75,1) is plotted every 10 min through one half of the lunar cycle (14.8 days). The pattern
will be repeated in the second half of the lunar cycle. Numbers in clear boxes indicate the day of the
lunar cycle, black filled boxes indicate night. (A) Position veligers determined using the BASE'1/8
tidal/diel model. (B) Position veligers determined using the BASE*l/4 tidal/diel model. (C) Position
veligers determined using the BASE*l/2 tidal/diel model. (D) Position veligers determined using the
TIDE*l/4 model. (E) Position veligers determined using the BASE*A tidal/diel model. (F) Position
veligers determined using the BASE*B tidal/diel model. (G) Position veligers determined using the
BASE'C tidal/diel model.
1941
J.LMannel and R.K.O'Dor
(Gallager et al., 1996). On the northeastern edge of Georges Bank, scallop
veligers are likely to encounter thermoclines at depths between 20 and 40 m.
However, veligers from the same species, in other locations (e.g. Mahone Bay or
Passamaquoddy Bay), are likely to encounter thermoclines at much shallower
depths (e.g. 3-8 m). To examine the effect of thermocline depth on migration
pattern, we sequentially used 1/8, 1/4 and 1/2 the water column for the distance
moved upon cue. All three tidal/diel models exhibit the same pattern, but the
further the veligers move, the more they aggregate, i.e. the greater the Reliability
(Figure 6, compare A, B and C). Greater distance moved also results in greater
Effort and Power, but the Efficiency is about the same (Table I, compare
BASE*l/8, BASE*l/4 and BASE*l/2).
How does the BASE*l/4 tidal/diel model compare with simply moving up and
down on a tidal cycle? Moving 1/4 of the water column on only a tidal cycle
(TIDE* 1/4) provides about the same mean transport, but almost all of the movement is in the veligers that start at the bottom and there is very little aggregation
of the veligers (Figure 6D). The BASE*l/4 tidal/diel model is more Reliable
(Table II, compare BASE*l/4 and TIDE*l/4 at various depths), even if slightly
less Efficient (Table I, compare BASE*l/4 and TIDE*l/4). Even moving 1/2 of
the water column with each tide does not make simple tidal migration Reliable
(Table II, TIDE*l/2 at various depths) while Effort is 40% greater than that of
the BASE*l/4 tidal/diel model (Table I, compare BASE*l/4 and TIDE*l/2).
The response to tidal and solar cues need not be equal. Assuming that veligers
move double the distance on a solar cue compared to a tidal cue (Figure 6E)
improves Power and Reliability, while Efficiency remains about the same (Table
I, compare BASE*l/4 and BASE* A). If we assume veligers move only 1/4 of the
water column on each cue because they are unable to swim any faster, then we
must assume they are unable to move twice as fast if they receive tidal and solar
cues to move in the same direction. This restriction reduces both the Power and
Efficiency (Table I, compare BASE*l/4 and BASE*B), primarily because
veligers move away from the thermocline every second week as a result of the
concurrence of the cues every 7.4 days (Figure 6F). If the solar cue over-rides the
tidal cue (LIGHT:TIDE), then again veligers move shallower when tide and solar
cues coincide, and there is loss of Power. However, this variation is very Reliable,
with equal effects further up in the water column (Table I, compare BASE* 1/4
and LIGHT:TIDE). If the tidal cue over-rides the solar cue (TIDE:LIGHT), the
model has less Power and is much less Reliable, while the Effort expended
remains about the same, resulting in a considerable reduction in Gain and
Efficiency (Tables I and II, compare BASE*l/4 and TIDE:LIGHT). When
veligers can only move up at low tide if there has not been a solar cue in the last
hour (BASE*C), the veligers remain down in the thermocline region (Figure 6G),
and Power, Reliability, Gain and Efficiency are all better than moving 1/2 the
water column on each tidal cue (Tables I and II, compare TIDE*l/2 and
BASE*C).
Since the tidal/diel model includes migration in response to solar cues, realistic assessment of the effects of this behavior must include transport obtained from
solar (S2) tidal currents. Contrary to common belief, diel vertical migration in
1942
Vertical migration; a tidal/diel model
response to solar cues (usually at dawn and dusk) can produce unidirectional
transport (Hill, 1994,1995). Diel (24 h period) migration will produce transport
by the S2 currents, with the direction of that transport depending on the phase of
the S2 currents. S2 currents are predictable and stable at any given location, but
vary among locations. Phase contours of 90 and 270° (i.e. where high tide occurs
1/4 or 3/4 of the way through the day, relative to local time) are regions of convergence and divergence, respectively, for organisms that migrate up into the
water column once each night. The S2 component of the tidal current can be
added to the tidal/diel model by using equation (1), with T = 12. For the northeastern part of Georges Bank, Uo = 0.1 m r 1 is a reasonable estimate of the S2
current (Moody et al, 1984). Since the S2 current in the frontal zone on the northeastern peak of Georges Bank has a phase of -360°, that region should see nearly
maximum transport when organisms migrate with a diurnal rhythm. Because the
S2 current is small relative to the M2 current on Georges Bank, including this
transport in the tidal/diel model improves total transport by only -12% (Table I,
compare BASE*A and SUN:0.1). However, in other locations where the S2
current is larger relative to the M2 current (Uo = 0.5 m s"1), transport would be
improved by -67% (Table I, compare BASE*A and SUN:0.5).
Discussion: Implications of the tidal/diel model
Migrating in response to both solar and tidal cues has several transport advantages for zooplankton that are too small to migrate the full height of the water
column on a tidal period. If variation in shear is non-linear with depth, then vertical migration in the water column produces much greater transport near the
bottom than near the top (Table I, TIDE*l/4) because changes in velocity are
greater near the bottom. Shear may also be greater within the thermocline than
in the bodies of water on either side of the thermocline. Baroclinic tides occur
where two bodies of water of different densities (usually due to temperature
and/or salinity differences) meet. Under these conditions, a zooplankter could
gain the most transport by migrating in the thermocline region. If a zooplankter
can perceive its depth in the water column and swim fast enough to overcome
turbulent conditions, then obviously the best transport is to be obtained by
migrating near the lower boundary of this model (which may be the thermocline
in real situations). However, if a zooplankter is small and subject to turbulence
that is strong relative to its swimming ability, moving with both solar and tidal
cues provides reasonable transport with greater Reliability, and with only slightly
more Effort than migrating at a tidal period alone (Table I, compare BASE*l/4
and TIDE*l/4). Expending more Effort in moving over a greater portion of the
water column at a tidal period does not increase Reliability and only slightly
improves Efficiency (Tables I and II, TIDE*l/2). Tidal/diel migration also takes
individual zooplankters through twice the water column depth, thus providing
veligers with a better opportunity to locate patchy food. Another important
feature of this type of migration is Reliability: i.e. the zooplankter is more likely
to return to the pattern if it is dispersed by turbulence or currents (even without
the ability to perceive position in the water column). Finally, in some areas, the
1943
J.LMannel and R.K.O'Dor
sun (S2) tidal current may provide significant transport in addition to that
obtained from the moon (M2) tidal current.
When looking at the proximal causes for migration in zooplankton, the fact that
a zooplankter responds more strongly to the solar cue, or the solar cue over-rides
the tidal cue, does not necessarily indicate that migration for the purpose of transport is less important than some other ultimate reason for diel migration. Migrating more in response to solar than tidal cues improves Reliability and Power
considerably with about the same Efficiency (Tables I and II, compare BASE*l/4
and BASE*A). Either being unable to move at double the speed when solar and
tidal cues coincide (BASE*B), or having the solar over-ride the tidal cue
(LIGHT:TIDE), moves the zooplankters shallower in the second and fourth
weeks, and thus much reduces Power in those weeks. This is primarily because
transport is much greater in the lower 1/4 of the water column. In real life, turbulence may lessen this effect, since movement shallower or deeper occurs only
on the particular days when solar and tidal cues coincide (once every 7.4 days).
Having tidal cues over-ride solar cues (TIDE:LIGHT) reduces Reliability
without improving Power or Efficiency (Tables I and II, compare BASE* 1/4 and
TIDE:LIGHT). Thus, the fact that a zooplankter responds more strongly to the
solar cue, or the solar cue over-rides the tidal cue, may simply indicate that
Reliability is a prime selective pressure, or that it is important for the zooplankter to remain down in the water column near the thermocline. It does not imply
that migration in response to solar cues is more important ecologically for a zooplankter.
Any behavior that tends to move animals preferentially closer to the lower
boundary will improve transport. If the dawn cue over-rides the tidal cue, but the
dusk cue does not, if the dawn cue moves the zooplankter further than the dusk
cue, or if zooplankters respond to depth by resisting movement into shallow
water, then the migrations will tend to occur just above the lower boundary and
net transport will be greatly improved (Table I, BASE*C). A number of other
behaviors that are commonly exhibited by zooplankton might have the same
effect: responding to environmental cues in such a manner as to keep the zooplankter near the region of greatest shear will improve both the Power and
Efficiency of migration. This could be accomplished by responding to such cues
as light intensity, salinity changes or depth, but the most obvious cue about the
location of the thermocline is, of course, the change in temperature with depth.
To perceive a signal, it is best to sample more often than the 'Nyquist frequency', i.e. more than twice per period. This means that one would need a
minimum of two samples per tidal cycle for several days to detect tidal migration
in zooplankters. It is possible, however, to detect vertical migration at tidal
periods in longer records with less frequent sampling. When two signals with
different periods are added together, they will coincide to produce secondary
peaks with a period that reflects the differences between the two periods. Solar
and tidal cues coincide with a 14.8 day period. When samples are collected every
day, the phase of the tide changes by -0.84 h between successive samples. This
means that vertical migration with a tidal period will appear in the record as a
change in depth with a period of 14.8 days. For a simple vertical migration down
1944
Vertical migration: a tidal/diel model
at low tide and up at high tide, one would expect to see deeper profiles alternate
with shallower profiles about every 7.4 days. A shorter period of more intensive
sampling (perhaps a day or two) will distinguish between simple migration at a
14.8 day period and the aliasing of the tidal and solar cycles. If tidal migration is
assumed a priori, it is possible to detect such tidal period migrations in data that
have gaps (Manuel et al., this issue). There are undoubtedly many data sets in
existence that could be examined to test for tidal period migrations.
A number of common sampling practices may have either obscured this pattern
of migration or left the tidal component unrecognized in many field studies. The
two most commonly reported patterns of vertical migration are nocturnal (up at
night and down during the day) and twilight (rise at dawn and dusk, deep during
the day). With the nocturnal pattern, descent may begin at any time, but increases
in intensity at dawn (Forward, 1988). In the tidal/diel model, these two types of
migration represent different parts of the lunar cycle. Tidal migration might also
be hidden by averaging the results of several days. Because of the change in the
time of the tide over successive days, averaging would probably remove most of
the tidal cue and leave only the diurnal pattern. The logistics of field sample collection might also interfere. Studies of marine migration frequently involve
sequential sampling of different sites. Researchers then try to interpret differences in vertical distribution in relation to stratification, chlorophyll density, wind
events, etc. Since tidal/diel migration has a period of 14.8 days, ignoring the possibility of a tidal period in migration could seriously inhibit the correct interpretation of observations of migratory behavior.
One of the problems with assuming that small oceanic zooplankters are using
tides for the purpose of horizontal transport is finding a means by which the zooplankter can determine the tidal phase. Even with an endogenous clock determining the period of migration, the direction of transport is still heavily
dependent on being in phase with local tides. If the time of moon rise can lead to
increased predation during certain phases of the moon, then it must be bright
enough to determine the phase of the moon (and thus the tide). Zooplankton
respond to the relative change in illumination, rather than the absolute magnitude of the change, when making diurnal migrations (Forward, 1988; Ringelberg,
1995), so the time of moon rise should be perceivable even under overcast conditions. As a zeitgeber to let planktonic organisms determine the tidal phase, the
time of moon rise is a more attractive option than variation in absolute lunar luminance, since there would be no need to 'remember' light conditions from other
nights and it should still work on overcast nights.
The tidal phase relative to moon rise is consistent at a given location, but varies
between locations. That is, at one location high tide may be at noon at the full
moon, but at another location high tide may be at noon 3 days after the full moon.
This requires selection for different phasing of the internal clock in different
areas, and selection for a particular phase in the endogenous clock can only be
stable if the species divides into self-recruiting populations in which the phase is
the same. Thus, the evolution of behavior that relied on the time of moon rise to
determine tidal phase would inevitably lead to the division of a species into
a number of self-recruiting populations, and sets the stage for the kind of
1945
J.L-Manael and R.K.O'Dor
population dynamics proposed by the member/vagrant hypothesis (see lies and
Sinclair, 1982; Sinclair, 1988). The phenomenon of populations that consistently
return to the same spawning ground when the adult and juvenile stages from
different populations mix at other times of the year (e.g. Atlantic herring) may
be a consequence of selection for retention at the larval stage.
In evolutionary terms, it may be a distinct advantage for a zooplankter that uses
tidal currents for horizontal transport to have the ability to track both sun and
moon periodicity. Tides are driven by the gravitational forces of both the sun and
the moon, and the relative importance of each varies in different places (and presumably in different historical epochs). Having the ability to track both may help
populations colonizing new habitats and/or coping with changes at the present
location. A history of changing current regimes may have left us with many zooplankters that use a vertical migration behavior that responds to both solar and
lunar cues. Migrating with lunar and solar periods also executes an elegant mathematical 'trick'. The timing of high/low and neap/spring tides (i.e. the aliasing of
the lunar and solar cycles required to predict the tides) can be perceived in a relatively straightforward manner. All this is done without the aid of tide tables
(which we humans are obliged to use).
There may be several ultimate reasons for vertical migration in zooplankton
(avoiding predation, efficient harvest of food, horizontal transport, etc.), as well
as several proximal causes [changes in light intensity, internal clock(s), responses
to discontinuities, predators, etc.] that elicit different behavior depending on the
'state' of the zooplankter (developmental stage, size of zooplankter, hunger, etc.).
The behavior that is most successful no doubt involves trade-offs among several
selective pressures. The two best substantiated ultimate causes of vertical
migration (predator avoidance and horizontal transport) are represented in the
tidal/diel model, but other ultimate reasons for vertical migration (locating food
in a patchy environment, avoiding UV light) would be provided as well. Thus, the
total selective advantage for a zooplankter performing tidal/diel migration could
be substantial.
Acknowledgement
This is a contribution to the program of OPEN (Ocean Production Enhancement
Network, one of the 15 Networks of Centers of Excellence supported by the
Government of Canada) and IFRP (Interim Funding Research Program).
References
Anderson.D.M. and StolzenbachJC.D. (1985) Selective retention of two dinoflagellates in a wellmixed estuarine embayment: the importance of diel vertical migration and surface avoidance. Mar.
EcoL Prog. Ser., 25, 39-50.
AngelAl.V. (1968) The thermocline as an ecological boundary. Sarsia, 34, 299-312.
Bollens.S.M. and Frost.B.W. (1989a) Predator-induced diel vertical migration in a planktonic
copepod. J. Plankton Res., 11,1047-1065.
Bollens,S.M. and FrosUBW. (1989b) Zooplanktivorous fish and variable diel vertical migration in the
marine planktonic copepod Calanus padficus. LJmnoL Oceanogr, 34, 1072-1083.
Bollens,S.M., Frost3-W., Thorenson.D.S. and Watts,S_J. (1992) Diel vertical migration in zooplankton: field evidence in support of the predator avoidance hypothesis. Hydrobiologia, 234, 33-39.
1946
Vertical migration: a tidal/die] model
Brown,W.S. and MoodyJ.A. (1987) Tides. In Backus,R.H. and BournsJJ.W. (eds), Georges Bank.
MIT Press, Cambridge, MA, pp. 100-107.
EnrightJ.T. and Hamner,W.M. (1967) Vertical diurnal migration and endogenous rhythmicity.
Science, 157, 937-941.
Forward,R.B.rIr (1988) Diel vertical migration: zooplankton photobiology and behaviour. Oceanogr.
Mar. Biol. Annu. Rev., 26, 361-393.
Gallager.S.M., ManuelJ.L., Manning,D.A. and O'Dor.R.K. (1996) Ontogenetic changes in the
vertical distribution of scallop larvae Placopecten magellanicus in 9 m deep mesocosms as a function
of light, food, and temperature. Mar. Biol., 124, 679-692.
Gliwicz^Z.M. (1986) Predation and the evolution of vertical migration in zooplankton. Nature, 320,
746-748.
Hardy^A.C. and Gunther.E. (1935) The plankton of the South Georgia whaling ground and adjacent
waters, 1926-27. Discovery Rep., 11,1-456.
HarrisJ.E. (1963) The role of endogenous rhythms in vertical migration. / Mar. Biol Assoc. UK, 43,
153-166.
HillA-E. (1991) Vertical migration in tidal currents. Mar. Ecol. Prog. Ser, 75, 39-54.
Hill,A.E. (1994) Horizontal zooplankton dispersal by diel vertical migration in S2 tidal currents on
the northwest European continental shelf. Cont. Shelf Res., 14, 491-506.
Hill,A.E. (1995) The kinematic principles governing horizontal transport induced by vertical
migration in tidalflows./ Mar. BioL Assoc UK, 75, 3-13.
Huntley,M. (1985) Experimental approaches to the study of vertical migration of zooplankton. In
Rankin,M.A. (ed.), Migration: Mechanisms and Adaptive Significance. University of Texas at
Austin, Port Arkansas, TX, pp. 71-90.
Huntley^M. and Brooks.E.R. (1982) Effects of age and food availability on diel vertical migration of
Calanuspacificus. Mar. Biol., 71, 23—31.
Iles.T.D. and Sinclair,M. (1982) Atlantic herring: Stock descreteness and abundance. Science (Wash.),
215,627-633.
Kimmerer.WJ. and McKinnon.A.D. (1987) Zooplankton in a marine bay. II. Vertical migration to
maintain horizontal distributions. Mar. Ecol. Prog. Sen, 41, 53-60.
Lampert.W. (1993) Ultimate causes of diel vertical migration of zooplankton: New evidence for the
predation-avoidance hypothesis. Arch. Hydrobwl. Bcih., 38, 79-88.
ManuelJ.L. (1996) Population and temporal variations in the vertical migrations of scallop (Placopecten magellanicus) veligers. PhD Thesis, Dalhousie University, Halifax, Nova Scotia, Canada.
ManuelJ.L., Gallager.S.M., Pearce,C.M., Manning.D.A. and O'Dor,R.K. (1996a) Veligers from
different populations of sea scallop {Placopecten magellanicus) have different vertical migration
patterns. Mar. Ecol Prog. Ser., 142, 147-163.
ManuelJ.L., Burbridge.S., Ball.M., KenchingtonJE. and O'Dor.R.K. (1996b) Veligers from different
populations of scallops (Placopecten magellanicus) have different distributions in the same
mesocosm. /. Shellfish Res., 15, 251-257.
Moody ,J.A. et al. (1984) Atlas of tidal elevation and current observations on the Northeast American
continental shelf and slope. US GeoL Surv. Bull., 1611,122 pp.
Osgood,K.E. and Frost.B.W. (1994) Ontogenetic diel vertical migration behaviors of the marine
planktonic copepods Calanus pacificus and Metridia lucens. Mar. Ecol. Prog. Ser., 104,13—25.
RingelbergJ. (1991) Enhancement of the phototactic reaction in Daphnia hyalina by a chemical
mediated by juvenile perch (Percafluviatilis).J. Plankton Res., 13,17-25.
RingelbergJ. (1995) Changes in light intensity and diel vertical migration: a comparison of marine
and freshwater environments. / Mar. Biol. Assoc. UK, 75,15-25.
Rogers,H. (1940) Occurrence and retention of plankton within the estuary. J. Fish. Res. Board Can.,
5,164-171.
Scheltema.R.S. (1986) On dispersal and planktonic larvae of benthic invertebrates: an eclectic
overview and summary of problems. Bull Mar. Set, 39,290-322.
Silva-Serra,M.A. (1995) Early life history traits of sea scallops (Placopecten magellanicus). PhD
Thesis, Dalhousie University, Halifax, Nova Scotia, Canada.
Sinclair,M. (1988) Marine populations: an essay on population regulation and speciation. Washington
University Press, Seattle.
Received on March 6, 1996; accepted on August 13, 1997
1947
Journal of Plankton Research Vol.19 no.12 pp.1949-1973, 1997
Vertical migration for horizontal transport while avoiding
predators: II. Evidence for the tidal/diel model from two
populations of scallop (Placopecten magellanicus) veligers
J.L.Manuel, Christopher M.Pearce1 and R.K.O'Dor
Biology Department, Dalhousie University, Halifax, Nova Scotia B3H 4J1 and
'GIROQ, Departement de Biologie, Universiti Laval, Sainte-Foy, Quebec
G1K 7P4, Canada
Abstract. We examined the vertical migration behavior of scallop (Placopecten magellanicus) veligers
in mesocosms and in previously reported field studies. Evidence suggests that these bivalve veligers
migTate in response to both tidal and diurnal stimuli in a manner similar to a proposed tidal/diel
model. Both populations have a diurnal response to solar cues. The response to tidal cues differs
between the Georges Bank and Passamaquoddy Bay populations. Georges Bank veligers appear to
utilize the differences in tidal phase that occur with depth to transport them in a northeasterly direction, thus maintaining the population on the bank. Passamaquoddy Bay veligers respond by swimming up at slack water (high and low tides) and down when currents are strongest. Such behavior
would minimize dispersal on the strong tidal currents in the Bay of Fundy and thus also tend to maintain a population within an area. Horizontal transport resulting from vertical migration is the most
likely selective pressure toCTeateand maintain these different behaviors against the homogenizing
effects of migration between the two populations. The implications of inherited differences in
behavior probably require consideration in the management of both wild and cultured populations.
Common sampling practices that obscure the tidal part of tidal/diel migration, including averaging
the results from several days of sampling, sampling too infrequently to perceive a tidal periodicity,
and assuming that only behavior that changes at high and low tides will affect horizontal transport,
need to be avoided in studies of vertical migration.
Introduction
Vertical migration of planktonic organisms is common in all waters of the world
and the behavior is found in all phyla from vertebrates to dinoflagellates and
ciliates (Huntley, 1985). Both the proximal causes of vertical migrations
(responses to stimuli such as light, temperature and salinity) and the ultimate
reasons for vertical migrations (predator avoidance, horizontal transport, energetic savings) have received considerable attention [see Lampert (1993) for a
good review]. Most hypotheses put forward to explain vertical migration have
tried to explain why organisms leave the more productive surface zone (presumably sacrificing the amount of food obtained) to migrate to less productive, often
cooler, waters at depth. Two hypotheses explaining vertical migration have been
widely substantiated. In estuaries, around oceanic reefs and islands, and in other
habitats where discontinuities are clearly delineated, there has been substantial
support for a hypothesis, proposed by Hardy and Gunther (1935) and Rogers
(1940), that zooplankton make vertical migrations to gain horizontal transport
from the differences in current strength and direction that often occur with depth
(Sinclair, 1988). Recently, there has also been substantial support for predator
avoidance causing diel vertical migration in crustacean holoplankton (Lampert,
© Oxford University Press
1949
J.L-Manuel, CM.Pearce and R.K.O'Dor
1993). A third widely tested hypothesis, bioenergetic savings suggested by
McLaren (1963) and Enright (1977), has not been well supported by field data
(Lampert, 1993). Other hypotheses have been presented as well, such as locating
phytoplankton distributed at different depths or in patches (Angel, 1985) and
avoidance of UV light (Siebeck, 1978; Damkaer, 1982).
Both field (Maru et al, 1972; Harding et al, 1986; Scrope-Howe and Jones,
1986; Tremblay and Sinclair, 1990b; Raby et al, 1994) and mesocosm (Kaarrvedt
et al, 1987; Silva-Serra and O'Dor, 1994; Gallager et al, 1996; Manuel et al,
1996a,b) studies have found diel changes in the vertical distribution of bivalve
veligers. Although dispersal is an important rationale for planktonic larvae in
benthic invertebrates (Scheltema, 1986), those with long pelagic larval stages can
have major problems with over-dispersal far from parental beds on offshore
banks. Currents are virtually ubiquitous, and any larva that acts as a passive particle will be carried some distance from the parental beds. In order for the population to persist, either larvae must be supplied from elsewhere (elsewhere then
being faced with the same problem) or the larva must have some strategy to
return it to the parental population. Manuel and O'Dor (this issue) showed that
in the real world, tidal/diel migration may confer real horizontal transport advantages to small zooplankton such as bivalve veligers. Except where the water
column is very shallow, the small size of veligers (<300 urn) and weak swimming
abilities probably preclude the possibility that they are able to migrate the full
distance from the surface to the thermocline with each tidal change. Their small
size also makes veligers susceptible to turbulence and net currents, leaving them
vulnerable to being moved from their present depth at any time. Added to this
are the difficulties for the zooplankter of perceiving the phase of the tide (i.e. the
time of tide relative to solar time), especially where tidal phase varies with depth.
Manuel and O'Dor (this issue) demonstrated that this combination makes vertical migration with both diurnal and tidal periods a more reliable and predictable
means for very small creatures to gain horizontal transport than simple migration
with a tidal period alone. When predator avoidance, locating patchy food or
avoidance of UV light add increased survival to these advantages, then the option
becomes very attractive indeed.
Among the complex factors that affect diel migrations in the marine environment, tidal cycles probably present the greatest difficulty, because tides have a
profound influence on sampling procedures and are almost ubiquitous. We
decided to test data from mesocosm experiments (Gallager et al., 1996; Manuel
et al, 1996a) and published field studies (Tremblay and Sinclair, 1990a,b) of
scallop {Placopecten magellanicus) veligers for evidence of lunar periodicity. As
no mechanism has been identified to sense tidal movement in passively carried
zooplankters, we assume this would have to reflect endogenous rhythms, probably entrained by a zeitgeber such as light changes (perhaps the time of moon rise;
see Manuel and O'Dor, this issue) and exhibiting lunar periodicity. We used two
methods to look for such patterns in the behavior of P.magellanicus veligers.
Surface concentrations of veligers in previous mesocosm experiments explained
>85% of the variance in mean depth of veligers (Manuel, 1996). In Part I, we conducted a mesocosm experiment to compare long-term patterns of veliger con1950
Vertical migration: evidence for the tidal/diel model
centrations near the surface of mesocosms to the pattern predicted by a model of
migration with both tidal and solar periods (Figure 5, Manuel and O'Dor, this
issue). In Part II, two previous tower tank experiments (Gallager et al., 1996;
Manuel et al, 1996a) and field studies of the vertical distribution of scallop
veligers (Tremblay and Sinclair, 1990a,b) are re-analysed for evidence of
tidal/diel migration.
Part I: New experiments with long-term record of surface aggregations in
mesocosms
Method
In this paper, we refer to migrations with a period of 12.42 h as 'tidal' and to the
proximal cue that induces the migrations as the 'tidal cue'. This implies neither
that the zooplankter must have the ability to perceive changes in current or the
state of the tide directly, nor that the migrations necessarily occur exactly at high
or low tide. We do not stipulate the means by which such migrations are induced,
since that may vary among populations and species. The tidal cue may involve
responses to direct stimuli, or may involve more complex stimuli such as internal
clocks set by external zeitgebers. Similarly, we refer to migration with a period of
24 h as 'dieP migration and the proximal cue that induces the migration as the
'solar cue'. We consider that diel migration may be a direct response to changes
in light levels, but this is not integral to the models or conclusions.
In the spring of 1994, we conducted a tower tank experiment to obtain detailed
information about the concentrations of P.magellanicus veligers near the surface
of replicated mesocosms. Previous experience with veligers in deep mesocosms
had demonstrated that the portion of veligers in the top 1 m of 9-m-deep mesocosms is strongly correlated with the mean depth of veligers, explaining >85% of
variation (Manuel, 1996). This strong correlation is probably due to the calm,
shallow (relative to many offshore habitats) nature of our experimental conditions allowing veligers to reach the surface with most upward migrations, so that
every migration was reflected in the number of veligers at the surface. Thus, our
observations may reflect trends in nature, but would not be duplicated in nature
due to surface turbulence.
The experiment was conducted in conjunction with another experiment
(Manuel et al, 1996) which established that genetically identified veligers
spawned from Georges Bank adults had different vertical distributions than
those spawned from Passamaquoddy Bay adults, even when raised in the same
mesocosm. Further details of the methods used can be found in that publication.
Briefly, adult scallops from Georges Bank and Passamaquoddy Bay (Figure 1)
were spawned in the fall of 1993, and then artificially reconditioned in the
laboratory. The eggs from several females from each population were mixed and
fertilized with the mixed sperm from several males from the same population.
The gametes from parents selected for use in the genetics experiment were
fertilized separately from the remainder. Fertilized embryos from each crossing
were held in separate 9-m-deep polyethylene mesocosms until 4 days of age (to
1951
J.L.Minuel, CM.Pearce and RJCO'Dor
Passamaquodd)
Bay
Fig. 1. Location of the Georges Bank and Passamaquoddy Bay populations, and tidal ellipses in the
region. Modified from Moody et al. (1984).
D-stage), and then screened, counted, sampled for size measurements and
redistributed in experimental tubes. At this time, progeny in excess of requirements for Manuel et al. (1996b) were mixed with the remainder, so that the
veligers used for this experiment encompass a slightly larger gene pool than
those used for Manuel et al. (1996b). In total, three females and five males parented the Georges Bank veligers, and five females and five males parented the
Passamaquoddy Bay veligers. Twelve experimental tubes were established, four
with Georges Bank veligers at a density of one veliger per 1.6 ml, and eight with
Passamaquoddy veligers at a density of one veliger per 2.5 ml. Veliger density
did not affect vertical migration patterns (unpublished observations).
Experimental polyethylene mesocosms (60 cm diameter, 9 m depth) were
placed in the tower tank at the Aquatron facility at Dalhousie University, filled
with 1 urn filtered seawater (double cartridge filters) and inoculated with enough
cultured hochrysis galbana (clone TISO) to bring the concentration to 1.0 X 104
cells ml"1. The mesocosms were sealed at the bottom by tying a knot, and suspended at the surface with a styrofoam flotation collar. Filling the mesocosm
slightly above the level of the tower tank produced a positive pressure that kept
the mesocosms firm. A gravity-fed perforated vinyl sprinkler hose was used to
1952
Vertical migration: evidence for the tidal/diel model
distribute food evenly from the top to the bottom of the mesocosm. Supplemental
TISO was added on nine occasions (days 6, 12, 14, 17, 19, 25, 33, 35 and 39),
resulting in particle levels that varied between 4.0 X 103 cells ml"1 and 1.40 X 104
cells ml"1 through the experiment. A thermocline of ~5°C over 1 m was established in the tower tank by circulating chilled water through a titanium ring
around the periphery of the tower tank at a depth of 6 m and heated water
through a PVC ring placed just above that. At 24 days of age, veligers from replicate mesocosms were concentrated on an 80 um nitex screen, sampled for size
measurements, mixed thoroughly and evenly redistributed in clean mesocosms.
All samples collected for length measurements were preserved in 1% formalin
buffered with sodium borate and measured with a calibrated OPTIMAS program
within a month of the end of the experiment.
Two video cameras, moved by an XY positioner, recorded the number of
veligers located at the surface of 12 of the tubes (four Georges Bank and eight
Passamaquoddy) from the ages of 8-16 days and 18-22 days. Light for the camera
was provided by a source permanently fixed in each tube (Figure 2). Each light
consisted of a 50 W halogen projection bulb in a glass bottle lined with aluminum
foil (except where the hght beam exited) and a mirror to direct the light beam
parallel to the surface. The light beam was relatively tight and undiffused, being
~5 cm wide X 2 cm deep at source, and about double that size at the opposite
side of the tube. Passage of the video camera over the row of tubes was computer
controlled to occur at 10 min intervals. Lights turned on only while the video
camera was passing over the row of tubes (-1.3 min) and diffusion down into the
tubes was minimal. A dim red light in the work area (Figure 2) provided visibility
at night without changes in illuminance associated with night sampling in previous
experiments (see Gallager et al., 1996; Manuel et ai, 1996a). Video tapes were
analyzed by counting the number of veligers in view in a sample area (~1 X 1 cm)
from a frame as close as possible to the center of the tube (the depth of the light
beam was ~3 cm).
Results
Veligers from the Georges Bank (GEO) and Passamaquoddy (PAS) populations were similar in size and growth rate. Both populations began with a mean
size of 107 um at 5 days of age (n - 100 veligers in each sample), and by 25 days
of age the mean size of GEO veligers was only slightly larger (168 ± 5 urn, n =
4 mesocosms) than that of PAS veligers (163 ± 10 um, n - 8 mesocosms). Differences in the quality of the recorded tapes from the two cameras meant that comparisons required extensive manipulations in OPTIMAS, so here we report the
surface record from only four representative mesocosms (two with GEO and
two with PAS veligers). Visual inspection of the other tapes indicated similar
results.
The pattern of abundance at the surface of the mesocosms containing Georges
Bank veligers from the evening at age 8 days to mid-day at age 22 days corresponded closely to the predictions for day 3 to day 17 in the tidal/diel model
1953
J.LMannel, CM.Pearce and R.K-O'Dor
Tower Tank
Mesocosms
Camera Paths
Work Area
B
^ Cable to VCR
Sampler
Styrofoam
Collar
Water
Mesocosm
Light Beam
Fig. Z Experimental set-up. (A) Overhead view of the position of treatment tubes in the tower tank.
(B) Side view of the camera and a mesocosm with light source.
1954
Vertical migration; evidence for the tidal/diel model
(Figure 3A). At age 9 days, we saw a pattern of abundance during the day that
was predicted by the tidal/diel model, but has not been reported in the literature.
There was a rise in numbers part way through the night and then a precipitous
drop in numbers immediately after dawn (TL), indicating that both tidal and light
cues had induced the veligers to move down at the same time. Veliger numbers
were lowest at the surface just after dawn and before dusk, and higher mid-day.
300 r
300 r
TL T
T
T T T L T T / L T L T T T L T T
Fig. 3. (A) Comparison of the tidal/diel model of the vertical movements of a veliger responding to
both solar and tidal cues (upper solid line) and the number of veligers at the surface of GEO mesocosms. The model results (upper solid line) correspond to model days 3-17, mesocosm data are for
veligers aged 8-22 days. The V-aris scale is the number of veligers counted in a sample. Age 17 days
had no data and was therefore omitted. Replicate mesocosms have been averaged (data points) and
smoothed (thin solid line) using a median and then a mean smooth (window of three points for each).
Some examples of good agreement between the two: T = response to tidal cue; T/L = tidal and solar
cues cancel each other, TL = tidal and solar cues coincide and produce a stronger response.
1955
J.L.Mannel, CJVf-Pearce and R.K.O'Dor
During the night of the 10th and the day of the 11th, numbers were very low at
the surface. It is possible that veliger migrations were not bringing them all the
way to the surface, that vertical migrations had ceased, or that the migrations
were no longer synchronized. The abundance of veligers at the surface from ages
12 to 16 days matched the tidal/diel model well. Tidal and light signals combined
to produce strong downward migration at dawn on the 13th, 14th and 15th. We
saw mid-day rises predicted by the model on the 14th, 15th and 16th, and perhaps
on the 12th. A rise in numbers after midnight on the 13th and 14th turned into
something similar to 'twilight' migration on the 14th and 15th. At dusk on the
13th, we saw opposing tidal and light signals combine to cancel each other out
and produce no noticeable change when the lights went out. Note that these
300 r
200
0
300
200
100
0
300
200
100
Fig. 3. (B) The number of veligers at the surface of PAS mesocosms. Replicate mesocosms have been
averaged (data points) and smoothed (solid line) using a median and then a mean smooth (window
of three points for each). Data could not be reasonably fitted to the tidal/diel model.
1956
Vertical migration: evidence for the tidal/diel model
patterns were in phase with the mid-day rise seen on the 9th. Note also that the
'tidal' cycle in these migrations during the night seemed to be shifted to the right,
relative to the 'tidal' cycle during the day. From age 18 to 22 days, veliger numbers
near the surface were low and variations did not correspond with the tidal/diel
model, except on day 19 when a steep increase in numbers started some hours
before dusk and continued after dusk. Possibly, the veligers were still migrating,
but the migrations took place lower in the water column and were not visible at
the surface except when the strong, dual light + tide signal on the 19th moved
them further up in the water column.
The pattern of abundance at the surface of mesocosms containing Passamaquoddy veligers (Figure 3B) was unlike the tidal/diel model, with the exception, perhaps, of the mid-day rise on day 9. This difference between the two
populations was consistent with previous work that indicated different vertical
distributions for these two populations (Manuel et al., 1996a,b).
Part II: Re-analysis
Evidence from tower tank experiments
If scallop veligers are indeed migrating at both tidal and solar periods, and if
endogenous cycles are presenting themselves in this tower tank experiment, then
we might expect evidence of an endogenous cycle in previous mesocosm experiments. If the results of this experiment are spurious, then there should be no evidence of tidal cycles in previous experiments. Since solar and tidal cues coincide
with a 14.8 day period, we should see modification of behavior around a 15 day
period. The tidal/diel model (Manuel and O'Dor, this issue) suggests that organisms should be deepest mid-day during the second and fourth week of the lunar
cycle. On the first and third week, veligers should be shallowest in the middle of
the night and during the day, and deepest at dawn and dusk (Figure 5; Manuel
and O'Dor, this issue). Therefore, by plotting the mean depth (ZCM) at mid-day
for our tower tank experiments, we would expect to see deeper profiles alternate
with shallower profiles about every 7.5 days. We re-examined the data from two
previous tower tank experiments (Gallager et al, 1996; Manuel et al., 1996a) to
see whether there was any evidence that P.magellanicus veligers were in fact
deeper on alternate weeks.
Our first experiment (Gallager et aL, 1996) was conducted in the fall of 1992
with P.magellanicus veligers whose parents came from Trinity Bay, Newfoundland (the 'Deep Site' of MacDonald and Thompson, 1985). Briefly, in replicated
polyethylene mesocosms (60 cm diameter, 9 m depth) we established a thermocline with a gradient of 11°C over 0.5 m about half-way down the tube (i.e. at
4.5 m depth). The second experiment (Manuel et al., 1996a) was conducted in the
winter of 1992-93. Briefly, that experiment compared the behavior of veligers
spawned by adults from three different populations: Georges Bank, Passamaquoddy Bay and Mahone Bay. There was a thermocline with a gradient of
1.5°C between 4 and 5 m depth, and a fourth treatment where Georges Bank
1957
J.LJVianueU CM.Pearee and R.K.O'Dor
l . , . , I , . . . l . .-. . I
10
IS
20
25
30
35
40
45
Age (d)
Fig. 4. Mean depth of veligers (ZCM) of the day profiles of PASS, MB, GEO and NO THERM treatments versus the age of the veligers. The solid line is data smoothed using a distance-weighted least
squares algorithm.
1958
Vertical migration: eridence for the tidal/diel model
veligers had no thermocline. From the mid-day video profiles, we recorded the
number of veligers at 5 cm depth intervals down to 9.0 m for each replicate. In
the first experiment, we used 13 sets of data from the MIXED and BOTTOMFED treatments when the veligers were between 19 and 49 days of age. In the
second experiment, we used 11 sets of data from each of the three populations
when the veligers were between 9 and 40 days of age. We calculated the proportion of veligers found below 4 m (i.e. within and below the thermocline) and the
ZCM for each profile, and plotted these two parameters against age. The trend
was assessed with a distance-weighted least squares smoothing algorithm.
Results
In both previous tower tank experiments (Gallager et al, 1996; Manuel et ai,
1996a), there was an alternation of deeper and shallower ZCM on a weekly basis
(Figures 4 and 5). We also saw a strong pattern of more veligers below the
thermocline alternating with fewer veligers below the thermocline in all except
the BOTTOM-FED treatment in 1992 (Figures 5 and 6). That this pattern initially
went unobserved and was not picked up until we looked specifically for it in the
data reflects the dangers of infrequent sampling. Increasing the sampling
frequency to once a day might have made this aspect of the behavior more apparent in the beginning, and strengthened the results here.
Evidence from field observations
Tremblay and Sinclair (1990b) reported the vertical distribution of P.magellanicus veligers every 2 h over a 50 h period from 3 to 5 October 1985 at afixedstation
near Grand Manan Island, in Passamaquoddy Bay. The same authors also published the vertical distribution of P.magellanicus veligers on the northeastern
edge of Georges Bank (Tremblay and Sinclair, 1990a). That study examined vertical distribution at 4 h intervals intermittently between 1 and 17 October 1987,
at several sites that were classified as well mixed, stratified or frontal. Data from
the well-mixed sites show nearly uniform distribution from surface to bottom.
That veligers were evenly distributed at the well-mixed site does not imply that
they are not migrating. Vertical distributions of zooplankton are not very aggregated if the distance that zooplankton move is small relative to the depth of the
water column (Manuel and O'Dor, this issue). On stratified and frontal sites,
downward migration would be limited by the pycnocline. On the well-mixed sites,
there was little temperature stratification to which the veligers could respond, so
the effective depth of the veligers would be greater than on the stratified and
frontal sites. The fact that the sites were well mixed also implies increased turbulence and mixing, which would probably overcome any aggregation that might
occur. At frontal and stratified sites, however, distribution is distinctly nonrandom, and frequently there were changes in distribution during the 4 h interval between samples.
We examined the results of the above studies for evidence that P.magellanicus
veligers in the field respond to both solar and tidal cues. For the Grand Manan
1959
J.L.Manuel, CM.Pearce and RJCO'Dor
-i—i—I—I—I—r-
"I—n—|
I |—i—i—i-
• i i i i i i i
I | I I II
I I I I | I
-~ 2
a
o
MIXED
-r-f
I | I I I I
I I I I | II
I I I I
I'
a
V
6 -
BOTTOM-FED
. i i i . i i
I i
. MIXED
§ 60
a
S 40
. BOTTOM-FED
0
sm
C 80
f
40
20
0
5
10
15
20
25
30
35
40
45
Agc(d)
Fig. 5. Mean depth (ZCM) and per cent of veligers below thermodine of the day profiles of
BOTTOM-FED and MIXED treatments versus the age of the veligers. The solid line is data
smoothed using a distance-weighted least squares algorithm.
1960
Vertical migration; evidence for the tWal/diel model
I
80
I
'
•
I
. PASS
60
elo>N Theirmocl ine
40
20
0
80
60
40
20
to
OEO
Vel
so 60
O 40
tc
20
<t-c
0
0
80
60
40
20
I
5
10
15
.
••,
.
I
20
.
,
,
.
I
.
.
25
,
.
I
30
.
,-.
.
35
40
45
Age (d)
Fig. 6. Proportion of veligers in or below the thermocline in the day profiles of PASS, MB, GEO and
NO THERM treatments versus the age of the veligers. The solid line is data smoothed using a distance-weighted least squares algorithm.
data, we examined a plot of vertical distribution, and considered non-diel
changes. For the Georges Bank data, we looked at plots from the stratified and
frontal sites, and recorded whether the vertical distribution had gone up or down
relative to the thermocline between sequential samples. Vertical distribution in
this case must be considered relative to the thermocline because the hydraulic
jump that occurs as the tides sweep over this part of Georges Bank and/or internal
1961
J.L-Manuel, CJvLPearce and R.ICO'Dor
waves (Loder et at, 1988,1992; Loder and Home, 1991) may passively transport
veligers up and down relative to the surface. We then determined on which days
in the tidal/diel model there would be movement (or non-movement) in the
appropriate direction. Finally, we determined which sequence of days in the
tidal/diel model would best accommodate all of the data.
Results
Veligers from the Grand Manan site (Figure 7) showed no evidence of migrating
in the pattern of the tidal/diel model proposed by Manuel and O'Dor (this issue).
However, these veligers seemed to be performing another version of migration
using both solar and tidal cues. The mean depth of veligers was compared with
the maximum and minimum currents predicted by Canadian Tide and Current
Tables for that location. Veligers at this site exhibited normal diel migration.
Superimposed on that behavior was a tendency to swim up when currents are
slack (high and low tide) and down when currents are strongest (both flood and
ebb).
1330
2130
0530
1330
2130
0530
1330
Time
Fig. 7. Modified from Tremblay and Sinclair (1990b). Larval scallop distribution (ZCM) versus time
at an anchor station near Grand Manan Island (Passamaquoddy region) in October 1985. Solid lines
are the times of maximum currents, dashed lines are times of minimum currents as predicted for
Grand Manas Channel on those dates (from Tidal Current Tables). S indicates where veligers seem
to have swum up in response to slack tide, F indicates where veligers appear to have swum down in
response to stronger flow, N indicates apparent upward swimming because of night fall, D indicates
apparent downward swimming at dawn, SN and FD show where solar and tidal cues combine to
stimulate strong upward movement, F/N and S/D indicate where solar and tidal cues have canceled
each other. Note also the placement of slack and flood tides, and the corresponding depth of veligers,
just after 1330 on the second day.
1962
Vertical migration: evidence for the tidal/diel model
The data from Georges Bank didfitthe tidal/diel model. Tremblay and Sinclair
(1990a) collected data at 4 h intervals (Table I). When vertical distribution
changed in that interval, the change in distribution could have occurred at any
time during that 4 h interval. It was, therefore, possible to match movement in
that 4 h interval with 4 days in the tidal/diel model. For example, on 1 October,
veligers at site S2a appear to move downwards between the sample taken at 10:05
and the sample taken at 13:54. In the tidal/diel model, there is downward movement between those times on days 13-17 (which is repeated with a period of 14.8
days). Likewise, if there is no change in vertical distribution, such as on 13
October between 20:04 and 00:15, that behavior can be matched with two consecutive days (Table I, possible days in model). After doing this for all pairs of
stratified and frontal zone profiles, we found that if we assumed veligers were
moving in response to both tidal and solar cues, the only sequence of days in the
tidal/diel model that matched the results was days 13-29.
Boston has approximately the same tidal phase as northeastern Georges Bank.
We compared the time of movement in the tidal/diel model with the time of high
tide at Boston, and found that changes in the vertical distribution of scallop
veligers occurred an average of 4 h 30 min after high tide (Table II). Magnell et
al. (1980) published tidal ellipses from two depths (above and below the pycnocline) at two sites on Georges Bank (Figure 8) very near the location of sampling
by Tremblay and Sinclair (1990a). Their ellipses showed differences in the speed
and direction of currents at different depths, as well as a phase difference in the
timing of tides at different depths. That is, there is opportunity for using migration
in phase with the tides to produce horizontal transport. The tidal/diel model suggests that local populations must employ migrations in phase with the local tides,
so one test is to show that vertical migration in a given population is phased with
the local tides. Imposing vertical migration 4.5 h after the high and low tide on
the tidal ellipses on Georges Bank showed that such behavior would take the
veligers down into the thermocline at a time that would make maximal use of the
Table L Fitting the direction of movement of veligers between profiles of stratified and frontal sites
to the tidal/diel model. Original profiles from Georges Bank in 1987 (modified from Tremblay and
Sinclair, 1990a)
Date
(October)
Site
Time
(h)
Movement
Possible days
in model
Best fit
to model
1
S2a
3
S2a
12
Fl
13
Fl
14
S2b
15
17
S2b
S3
1(W)5-13:54
13:54-18:32
12:07-15:54
15:54-19:43
19:43-00:14
12:06-16:07
16:07-20^)4
20#4-00:15
00:15-04:09
04:09-08:08
17:48-22:00
22:00-02:11
10:20-13:57
20:13-23:13
Down
Up
Down
Up
Down
Up
Down
None
Up
Down
None
Down
None
Down
13-17
8-15
16-20
13-17
11-15
23-27
21-25
23-25,30-32
24-28
21-25
23-27
28-32
18-20,26-27
26-29
13
13
15
15
15
24
24
24
25
25
26
26
27
29
1963
J.LJVItnuei, CM.Pearee and R.K.O'Dor
Table IL Comparing the time of movement in the fitted model to the time of high and low tides on
Georges Bank on the same day. (D) = downward movement, (U) = upward movement, (N) = no
movement, (H) = high tide, (L) = low tide. The time of movement in the model and tide times on
Georges Bank have both been rounded to the next 10 min interval. The average difference in time
between the change in tide and veliger movement is 4 h 30 min
Date/
model day
Profile times and
movement
Oct. 1/
Day 13
Oct. 3/
Day 15
10fl5-13:54 (D)
13:54-18:32 (U)
12:07-15:54 (D)«
15:54-19:43 (U)
19:43-00:14 (D)
12:06-16:10 (U)
16:07-20:04 (D)b
20:04-00:15 (N)
00:15-04:09 (U)
04:09-08:08 (D)
17:48-22:00 (N)c
22:00-02:11 (D)"
Oct. 12/
Day 24
Oct. 13/
Day 25
Oct. 14/
Day 26
Oct. 15/
Day 27
Oct. 17/
Day 29
Movement in
model (diel)
18:00 (U)
18:00 (U)
18.00 (U)
06:00 (D)
18:00 (U)
Movement in
model (tidal)
Time of tide Time difference
that day
model - tide
10:10 (D)
16:20 (U)
11:50 (D)
18:00 (U)
00:10 (D)
13:10 (U)
19:20 (D)
05:20 (H)
11:20 (L)
07:30 (H)
13:30 (L)
19:40 (H)
0830 (L)
14:40 (H)
4 h 50 min
ShOOmin
4 h 20 min
4h30min
4 h 30 min
4 h 50 min
4 h 40 min
01:30 (U)
07:50 (D)
21:00 (D)
03:10 (U)
09:30 (D)
21:10 (L)
03:30 (H)
16:30 (H)
23:00 (L)
05:20 (H)
4h20min
4 h 30 min
4 h 30 min
4 h 10 min
15:10 (U)
11:10 (L)
4 h 00 min
23:30 (D)
19:30 (H)
4h 00 min
10:20-13:57 (N)
20:13-23:13 (D)
•On day 15 in the model, the tide change is slightly outside of the sampling interval.
Tide and light signals are opposite, both within the same hour, downward movement.
•Tide and light signals are opposite, separated by 3 h, no movement.
"The second profile here (02:11) has very few veligers relative to the other profiles.
phase difference in tides to transport veligers in a generally northeasterly direction (Figure 8). Thus, the analysis indicated that vertical migration may be transporting veligers in a northeasterly direction by taking advantage of the difference
in tidal phase and direction between the upper and lower water layers.
Discussion
In both previous tower tank experiments (Figures 4-6), re-examination of the
data suggested that scallop veligers did indeed change mid-day depth at -7.5 day
intervals. This post hoc analysis would have been greatly strengthened had the
sampling frequency been daily. This period appeared both in differences in ZCM
(mean depth) and in the proportion of veligers in and below the thermocline
(which are, of course, somewhat autocorrelated). The only exception to this was
the BOTTOM-FED treatment in the first experiment (Gallager et al, 1996).
Huntley and Brooks (1982) found that diurnal vertical migration was reduced in
copepods that were food limited. In our experiments (Gallager et aL, 1996),
vehgers refused to go below the thermocline even when they were food limited
and there was food below the thermocline. This suggests that food limitation may
reduce migration down into the thermocline, which would reduce transport effectiveness considerably. Remaining near the surface when food levels are low may
increase the probability of encountering scarce resources. Thus part of the
1964
Vertical migration: evidence for the tidal/diel model
10
C: M2/44m
tO
CM/SEC
A: M1/77m
N
ft M1/192m
D: M2/77m
/
Fig. 8. Modified from Magnell et al (1980). Semi-diurnal tidal ellipses from moored instrument
stations on the north flank of Georges Bank in the spring of 1987. Two stations (Ml and M2) are each
represented at two depths. The station is located at the center of the ellipse, and the direction and
speed of current are represented by points on the ellipse. Tick marks indicate lunar hours. There are
12 lunar hours in one tidal cycle, so each lunar hour is slightly longer than 60 min. Open circles indicate the time of high water at station M2. Note that, at both stations, the current changes direction at
depth before it changes nearer the surface. In other words, the tides are at different phases at different depths. The shaded area represents the position (and'thus the direction and speed of movement)
of a hypothetical veliger that jumps to the lower depth 4.5 h after high tide and to the shallower depth
4.5 h after low tide at each station. Vectors at the bottom indicate net transport for such a veliger
when it occupies each tidal ellipse (A and B, C and D) and the net movement at each site (S).
1965
J-L-Manod, CMPearce and R.K.O'Dor
hierarchy of responses that veligers seem to have evolved might be summarized
as 'better lost than unfed', or even 'if it is this bad here, better take a chance on
elsewhere'. Abandoning vertical migration will almost certainly result in passive
transport to another location on net currents. Since the BOTTOM-FED treatment did exhibit differences at mid-day in the per cent of veligers below the
thermocline, it is possible that there are two separate mechanisms at work here:
(i) swimming behavior and (ii) response to thermocline. Scallop veligers tend to
swim upward more quickly when food limited (Silva-Serra, 1996), and this in turn
would reduce the probability that they would encounter the thermocline. If the
tidal migration required a thermal change in addition to an endogenous clock to
induce it, tidal migration would only occur in those few individuals that were
lower in the water column. However, it may also be argued that since the per cent
of veligers below the thermocline was always low in that treatment, this response
may represent the activity of a small portion of the population that was of larger
size, perhaps because they were by chance better fed. We cannot distinguish
between these hypotheses with the current data set.
In both years, on the second day we profiled we found veligers much deeper
than on the first day, even though we began profiling on day 19 in 1992 and day
10 in 1992-93 (Figures 4 and 5). In both cases, however, we made the first series
of night profiles around midnight of the night before the first day profiles shown
here. These night profiles may have provided a zeitgeber that suggested to the
veligers that the moon either rose at midnight or set an hour thereafter (either
might have acted as a cue). It is reasonable to assume something associated with
our first night sample (possibly the light of the camera or the light of the VCR
monitor) provided a zeitgeber that cued the veligers for their tidal cycle. That the
veligers maintained this cycle even though we continued to sample at night suggests that the initial zeitgeber sets the behavior and subsequent stimuli are less
important. The details of how this behavior is triggered are worthy of investigation. Many benthic invertebrates spawn at a particular phase of the moon, often
near the new or full moon. The Passamaquoddy population of P.magellanicus
spawns preferentially in the days just prior to the full moon (Parsons and
Dadswell, 1992), scallops on the Baie de Chaleur spawn near either the new or
full moon (Himmelman, 1996), and there is anecdotal evidence that other populations of this species, including the Georges Bank population, spawn near either
the new or full moon. A reasonable hypothesis would be that the veligers would
normally arrive at D-stage (4 days post-spawning) near the full or new moon, and
the internal clock may then be fine-tuned by changes in moonlight. This could
explain the fact that many invertebrate trochophores initially swim strongly
towards the surface, where the perception of changes in moonlight would be
much easier than at greater depths, since moonlight is probably not strong enough
to penetrate to very deep water. Whatever the initial zeitgeber, there is evidence
for a change in depth at mid-day with a period of -15 days, which is consistent
with veligers migrating with both tidal and diel periods.
The long-term record of abundance near the surface of mesocosms indicated
that veligers of the Georges Bank population could be migrating in response to
both tidal and solar cues (Figure 3A). Where the record did not exhibit this
1966
Vertical migration: evidence for the tidal/diel model
pattern (age 11, 18 and 20-22 days), veliger abundances were low. This could
reflect that veligers were not making synchronous migrations, that veligers were
generally deeper and not arriving at the surface in great numbers (and thus were
not recorded by our surface cameras) or that veligers were not making vertical
migrations during this period. Diel vertical migration in a vertically sheared tidal
current will result in horizontal transport. For that portion of the current that is
due to the lunar cycle, the direction and speed of transport will change with the
lunar cycle, with the result that the net transport will be nil (Hill, 1991a,b, 1994,
1995). Abandoning diel vertical migration for a part of the lunar cycle will result
in horizontal transport for only that part of the lunar cycle when the organism is
migrating, and thus there will be net transport. In addition to this, horizontal
transport from the tides induced by the sun (see Hill, 1994) will remain constant.
Therefore, migrating in phase with the tides on one week and not migrating the
next week could result in net transport, as long as the effects of the lunar tidal
and solar tidal (diel) migrations do not conflict. However, it is also possible that
migration continued, but was centered lower in the water column. The steep rise
at age 19 days may be an indication of this.
At times during this experiment, the record was remarkably similar to the
tidal/diel model, especially considering that there were, in fact, no real tides or
moon in the tower tank to which these veligers could respond. The diel response,
of course, was undoubtedly triggered by changes in light in the tower tank
(although not necessarily a direct response to light). We were able to offset the
time of 'dawn' and 'dusk' for these veligers simply by changing the time of day
when the lights were on. The 'tidal' responses during the night appear to be
shifted somewhat to the right, relative to the 'tidal' responses during the day. The
existence of two separate internal clocks (one for day and one for night migrations) is not inconsistent with other endogenous rhythms: internal clocks often
appear as if the day and night behaviors are controlled by separate clocks (see,
for example, Harris, 1963). A reasonable hypothesis, well worth investigating, is
that two endogenous clocks might be 'counting' from the time of dawn and dusk
each day. Placopecten magellanicus spawns near the equinox of each year. The
timing of the 'tidal' migrations may have been offset in our experiments. Our
experimental protocol had a controlled 12 h on/12 h off light cycle (which is
appropriate at the time of spawning) instead of the longer nights and shorter days
of the post-equinox period when the veligers would have been in the water
column. Thus, running a simple test of periodicity on vertical migratory behavior
may obscure valuable results in controlled laboratory experiments unless, when
setting up the experiment, the researcher has controlled for the expected time of
sunrise and sunset, for both the location of the population and at the time of year
when spawning occurs.
The field data from the same population were consistent with migration with
both solar and tidal periods (Tables I and II). When movement relative to the
thermocline was matched to the tidal/diel model, it was consistent with veligers
moving down -4.5 h after high tide, and up -4.5 h after low tide. Such migration
provides northeasterly transport on the northeastern edge of Georges Bank
(Figure 8). That the migration is in phase with the lower water column and not
1967
J.LJVfcuwel, CM.Pearce and RJCO'Dor
the surface tide suggests that it is transport in a northerly direction that is
important to the veligers. Had the migration been in phase with the upper tide,
transport would have been in a more westerly direction and movement to the
north would have been lessened.
The difference in depths of the tidal ellipses in Figure 8 is undoubtedly much
greater than the depth of vertical migration possible for a scallop veliger.
Maximum vertical swimming speed for a P.magellanicus veliger seldom exceeds
2 mm s-1, i.e. 7.2 m Ir 1 (Silva-Serra, 1996). If the veliger migrated for half the tidal
cycle, it would only move 22 m. We would also have to assume that the veliger
could feed while it was migrating (since no time would be left during the day to
feed otherwise). Thus, half that distance (10 m) would be a more conservative
and likely expectation. This is about the distance that the median moves in Tremblay and Sinclair's (1990a) field study. The rate of change in the speed and direction and of currents with depth will correlate with the sharpness of the
thermocline. A gradual thermocline will thus afford less opportunity for a small
zooplankter to gain horizontal transport from vertical migration than a sharper
one. The strength of stratification may therefore affect transportation, and thus
survival of scallop veligers. Similarly, if the depth of the thermocline varies over
the tidal cycle, this may affect the time at which veligers must perform migrations.
Sinclair (1988) has already proposed that the 'stratification parameter' is important to the location of spawning beds for herring. Perhaps vertical migration of
the larvae for the purpose of horizontal transport helps retain larvae in profitable
zones.
McGarvey et al. (1993) show that the scallop population on the northern edge
and northeast peak of Georges Bank is probably self-recruiting. Our results agTee
with this, and contrast with the predictions of a physical oceanography model that
assumed no active vertical migration for scallop veligers (Tremblay et al, 1994).
Tremblay et al. do not include baroclinic tides or vertical migration of individual
veligers in their model. That means that they a priori preclude the possibility that
vertical migration is affecting horizontal movement. We believe that assumption
may be incorrect. We also advise caution in assuming that the location of the
greatest density of veligers indicates the movement of recruits to the fishery. If
advection away from parental beds results in massive losses from the population,
the important individuals may be the few that, by a combination of good luck and
appropriate behavior, are not washed away. As an analogy, if predation were the
cause of greatest mortality in a population, one would not assume that being
eaten indicated the best strategy for parenting the next generation. Similarly, the
vast majority of veligers drifting away from parental beds may be those who have
already lost the fight for survival. The most important location may not necessarily be where the density of veligers is greatest.
Passamaquoddy veligers exhibit quite different behavior. In the long-term
surface record in the tower tank experiment, there is no indication of migration
with a pattern similar to the tidal/diel model (Figure 3B). Examination of the field
record (Figure 7) suggests that veligers from that population are swimming up at
slack water (high and low tide) and down when currents are strongest. Migration
upward at high and low tides, and downward during flood and ebb tides, would
1968
Vertical migration: evidence for the tJdaVdiel model
tend to minimize tidal transport in either direction. In this region, tidal excursions
are among the largest in the world (Figure 2), and the area of suitable habitat
relatively small. The result of their behavior would be to remain in the same area
as much as possible, minimizing dispersal by tidal excursions. This strategy would
only be effective where currents are sheared, and given the relatively small amplitude of the migrations (-1/6 of the water column), it may be that the strategy only
works in specific locations in the bay: perhaps in eddies in the lee of the many
islands that dot the bay, for example. In our experiments (Manuel et al, 1996a),
Passamaquoddy veligers remained much closer to the surface than any of the
other populations studied. This may be explained as those veligers always acting
as if the current were slack at all times (i.e. perpetual high or low tide). Since the
veligers from Passamaquoddy seemed to 'know' the tidal phase at Grand Manan,
but not in our tower tank mesocosms, this would suggest that the mechanism for
determining the tidal phase is different for Passamaquoddy veligers than for
Georges Bank veligers. In other words, we suggest that Passamaquoddy veligers
do not use the endogenous clodc/moonlight zeitgeber system used by Georges
Bank veligers to determine tidal phase. It would be useful to determine whether
these veligers were able to perceive and respond to some other more proximal
tide phase cue, such as turbulence, shear or direction/speed of water movement.
In fact, the mechanism suggested for veligers to determine the phase of the tide
on Georges Bank (time of moon rise combined with an endogenous clock) would
not be as effective in an estuarine environment. Although the tidal phase is
similar over large geographical areas in offshore regions, it is often modified by
geography in estuaries. The time of moon rise would not be a good predictor of
the tidal phase in an estuarine situation. The phase of tide will vary over short
geographic distances, and thus presumably within the same population. Thus, the
fact that Passamaquoddy Bay veligers respond differently than Georges Bank
veligers to the same stimuli strengthens the argument that veligers are using vertical migration for horizontal transport. The 'umbrella' hypothesis is that veligers
from populations with different hydrographic regimes would behave differently
to enhance retention under different conditions.
It is unlikely that the differences observed are the only ones present: other
differences not induced by our experimental protocol probably exist. In studying
organisms where behavior is inherited rather than learned, it should be noted that
although the ability to perceive and respond to stimuli such as changes in light or
temperature may exist over large groups (e.g. a genus), there may be differences
in the nature of the response not only among species, but also within species
among populations. Behaviors are among the most plastic of biological responses,
so caution should be exercised whenever generalizations are drawn about specific
behaviors. It would be unwise, for example, to assume that differences in the vertical distributions of bivalve veligers of different species (or even populations of
the same species) are solely the consequence of the level of food, predation or
degree of stratification in two separate areas.
Aquaculturalists should at least be aware of differences in vertical migration
behavior among different populations of potential aquaculture species. Selection
for different appropriate behaviors in different areas indicates that mortality (or
1969
J.LJMannel, CJVi.Pearce and R.K.O'Dor
loss from the breeding population) is the consequence of inappropriate behavior.
There must, therefore, be some selective disadvantage to any animal moved from
one type of hydrographic regime to another. Movement of animals from one area
to another where a local stock already exists may result in cross-bred individuals
that are not retained well enough in the system to sustain a breeding population.
The seriousness of this type of scenario depends on two factors: (i) the degree of
difference among behaviors in the populations and (ii) the heritability of those
behaviors. If the consequences of different behaviors are not complete (i.e. if
some individuals survive even if at a rate less than adequate to replace deaths in
the short term) and if heritability is high (i.e. if variability exists in the behaviors
exhibited and the variability is highly correlated with genetic differences), then
the consequences of introducing 'foreign' stocks might be minor. However, if
either is not the case, then the introduction of 'foreign' stocks may effectively
remove existing stocks or at the minimum severely hamper recruitment. As a
minimum, we suggest that introduced stocks should come from areas with similar
hydrography. On the other hand, with a little imagination, it might be possible to
place stocks where none presently exist (or numbers are low) because of poor
retention, but where the adult habitat is nonetheless suitable. By stocking the
location from a population with appropriate veliger behavior, veligers might
survive quite well, but be in an inappropriate place (too far from parental beds)
when it is time to settle. An enhancement program such as those currently being
tested in New Zealand (Bull, 1994; Morrison, 1994) could collect veligers on spat
bags and transport them back to adult beds.
It is possible that behaviors vary in their effect on survival or growth under the
controlled conditions in the hatchery. This would allow for choice of stock that
has behavior more appropriate for culture conditions, and also allow selection for
behavior that enhances survival. Indeed, simply culturing veligers for several
generations may alter behavior, and strong selective pressure could severely
increase inbreeding by removing all but the progeny that have inherited a particular chromosome from a particular individual. Placopecten magellanicus
makes extensive vertical migrations, aggregates in great concentrations at the
surface at night and may therefore suffer from depletion of food reserves while
at the surface. Individuals that lack such behavior may increase rapidly in the
population under culture conditions, providing initial short-term success.
However, selection for such behavior with a limited gene pool could also result
in rapid inbreeding depression and consequent failure to thrive.
In summary, the migratory behavior of P.magellanicus appears to be determined at least partially by the transport effects of that migration. Georges Bank
veligers migrate in response to both solar and tidal cues in a manner that will
result in transport (at least on the northeast edge and possibly over the whole
bank) northeastward. Transport in a northeastward direction would minimize
losses from the southern portion of Georges Bank, and tend to move veligers
towards the frontal region on the northern part of Georges Bank (possibly
placing veligers in the region of greatest food concentration). This pattern is consistent with real distributions and fisheries data that suggest the scallop population on the northern edge and northeast peak of Georges Bank is self-recruiting
1970
Vertical migration: evidence for the tidal/diel model
(McGarvey et al., 1993), and contrasts with the results of a physical oceanography
model that assumed no active vertical migration for scallop veligers (Tremblay
et al, 1994). Passamaquoddy veligers also appear to respond to both solar and
tidal cues, but the tidal migration is such that dispersal by tidal excursions would
be minimized. Horizontal transport resulting from vertical migration is the most
likely selective pressure to create and maintain these different behaviors against
the homogenizing effects of migration between the two populations. The implications of inherited differences in behavior for the management of both wild
stocks and for aquaculture are not quantified at the moment, but managers should
at least be aware of potential consequences (either problems or benefits) for both
wild and cultured populations.
We have now shown that inherited differences in veliger behavior exist in populations (Manuel et al, 1996a,b) that probably exchange enough individuals to
prevent random genetic drift. Persistence of different behavior therefore indicates selective pressures for different behaviors among the populations studied.
Considered individually, these results do not prove the hypothesis that veligers
are responding with vertical migrations to both solar and tidal cues, but collectively they certainly justify further analysis and/or experiments. It may be that the
behavior appears in other species, but has not been recognized because of
common sampling and/or analysis practices. Practices that would obscure this
type of behavior include averaging the results from several days of sampling, sampling too infrequently to perceive a tidal periodicity, and assuming that migration
for the purpose of horizontal transport requires that organisms change behavior
at high and low tides (in a sheared circular tidal current, migration with a tidal
period will result in horizontal transport in some direction no matter when the
migration begins). Given the wide range of consequences associated with the
hypothesis supported by these observations, we believe it is important that
careful consideration be given to experimental and sampling program designs to
ensure that it can be adequately tested or refuted for a range of species.
Acknowledgement
Contribution to the program of OPEN (Ocean Production Enhancement
Network, one of the 15 Networks of Centers of Excellence supported by the
Government of Canada) and IFRP (Interim Funding Research Program).
References
Angel,M.V. (1985) Vertical migrations in the oceanic realm: possible causes and probable effects. In
Rankin,M.A. (ed.), Migration: Mechanisms and Adaptive Significance. University of Texas at
Austin, Port Arkansas, TX, pp. 45-70.
Bull,M.F. (1994) Enhancement and management of New Zealand's 'southern scallop' fishery. In
BourneJN.F., Bunting.B.L. and TownsendJ-D. (eds), Proceedings of the 9th International Pectinid
Workshop, Nanaimo, BC, Canada, April 22-27,1993. Vol. 2, pp. 131-136.
Canadian Hydrographic Service (1993) Canadian Tide and Current Tables: Atlantic Coast and Bay of
Fundy. Department of Fish. Oceans, Ottawa, Canada.
DamkaerJD.M. (1982) Possible influences of solar UV radiation in the evolution of marine zooplankton. In CalkinsJ. (ed.), The Role of Solar Ultraviolet Radiation in Marine Ecosystems. Plenum
Press, New York, pp. 701-706.
1971
J.L.M«nuel, CM.Pearce and R.K-O'Dor
EnrightJ.T. (1977) Diurnal vertical migration: adaptive significance and timing. Part I. Selective
advantage: a metabolic model. Limnol. Oceanogr., 22, 856-872.
Gallager.S.M., ManuelJ.L., ManningJD.A. and O'Dor,R.K. (1996) Ontogenetic changes in the
vertical distribution of scallop larvae Placopecten magellanicus in 9 m deep mesocosms as a function
of light, food, and temperature. Mar. Biol., 124, 679-692.
Hardmg.G.C, Vass,W.P., Hargrave,B.T. and Pearre.S. J r (1986) Diel vertical movements and feeding
activity of zooplankton in St. George's Bay, N.S., using net tows and a newly developed passive
trap. Can. J. Fish. Aquat. ScL, 43,952-967.
Hardy,A.C. and Gunther,E. (1935) The plankton of the South Georgia whaling ground and adjacent
waters, 1926-27. Discovery Rep., 11,1-456.
Hill,A.E. (1991a) Vertical migration in tidal currents. Mar. EcoL Prog. Ser, 75, 39-54.
Hill,A.E. (1991b) A mechanisms for horizontal zooplankton transport by vertical migration in tidal
currents. Mar. Biol, 111, 485-492.
Hill.A.E. (1994) Horizontal zooplankton dispersal by diel vertical migration in S2 tidal currents on
the northwest European continental shelf. Com. Shelf Res., 14, 491-506.
Hill^A.E. (1995) The kinematic principles governing horizontal transport induced by vertical
migration in tidalflows.J. Mar. Biol. Assoc. UK, 75, 3-13.
Huntley,M. (1985) Experimental approaches to the study of vertical migration of zooplankton. In
Rankin,M.A. (ed.), Migration: Mechanisms and Adaptive Significance. University of Texas at
Austin, Port Arkansas, TX, pp. 71-90.
Huntley.M. and Brooks,E.R. (1982) Effects of age and food availability on diel vertical migration of
Calanus pacificus. Mar. BioL, 71, 23-31.
Kaartvedt.S., AksnesJD.L. and EggeJ.K. (1987) Effect of light on the vertical distribution of Pecten
maximus larvae. Mar. EcoL Prog. Ser., 40, 195-197.
Lampert.W. (1993) Ultimate causes of diel vertical migration of zooplankton: New evidence for the
predation-avoidance hypothesis. Arch. HydrobioL Beih., 38, 79-88.
LoderJ.W. and Horne,E.P.W. (1991) Skew eddyfluxesas signatures of non-linear tidal current interactions, with application to Georges Bank. Atmos. Ocean., 29, 517-546.
LoderJ.W., Ross.C.K. and Smith.P.C. (1988) A space- and time-scale characterization of circulation
and mixing over submarine banks, with application to the northwestern Atlantic continental shelf.
Can. J. Fish. Aquat ScL, 45, 1860-1885.
Loder^.W., Brickman.D. and Horne,E.P.W. (1992) Detailed structure of currents and hydrography
on the northern side of Georges Bank. /. Geophys. Res., 97,14331-14351.
MacDonald.B.A. and Thompson,RJ. (1985) Influence of temperature and food availability on the
ecological energetics of the giant scallop Placopecten magellanicus: II. Reproductive output and
total production. Mar. Ecol. Prog. Ser., 25,295-303.
Magnell.B.A., Spiegel.S.L., Scarlet.R.I. and AndrewsJ.B. (1980) The relationship of tidal and low
frequency currents on the north slope of Georges Bank. /. Phys. Oceanogr., 10,1200-1212.
Manuel J.L. (1996) Populaton and temporal variations in the vertical migrations of scallop {Placopecten
magellanicus) veligers. PhD Thesis, Dalhousie University, Halifax, Nova Scotia, Canada, 247 pp.
Manuel^.L. and O'Dor,R.K. (1997) Vertical migration for horizontal transport while avoiding
predators: I. A tidal/diel model. / Plankton Res., 19,1927-1945.
Manuel^T.L., Gallager.S.M., Pearce.C.M., ManningJJ.A. and O'Dor,R.K. (1996a) Veligers from
different populations of sea scallop (Placopecten magellanicus) have different vertical migration
patterns. Mar. EcoL Prog. Ser., 142, 147-163.
ManuelJ.L., Burbridge.S., Ball.M., Kenchington,E. and O'DorJt.K. (1996b) Veligers from different
populations of scallops (Placopecten magellanicus) have different distributions in the same
mesocosm. / Shellfish Res., 15, 251-257.
Maru,K., Obara^A., KikuchiJC and Okesaku,H. (1972) Studies on the ecology of the scallop,
Patinopecten yessoensis (Jay) 3. On the diurnal vertical migration of scallop larvae. Sri Rep.
Hokkaido Fish. Exp. Stn, 27,33-53.
McLaren,I.A. (1963) Effects of temperature on growth of zooplankton, and the adaptive value of
vertical migration. J. Fish. Res. Board Can., 20, 685-727.
MoodyJ.A. et al (1984) Atlas of tidal elevation and current observations on the Northeast American
continental shelf and slope. US GeoL Surv. Bull, 1611,122 pp.
Morrison.M.A. (1994) Recruitment to artificial collectors of the scallop Pecten novaezealandiae in
northeastern New Zealand. In Bourne.N.F., Bunting,BL. and Townsend,L.D. (eds), Proceedings
of the 9th International Pectinid Workshop, Nanaimo, BC, Canada, April 22-27,1993. Vol. 1, p. 187.
Parsons.GJ. and Dadswell^MJ. (1992) Inter-annual and long-term patterns in the reproductive cycle
of giant scallops Placopecten magellanicus (Bivalvia: Pectinidae) from Passamaquoddy Bay, New
Brunswick, Canada. Mar. EcoL Prog. Ser., 80, 203-214.
1972
Vertical migration: evidence for the ridal/diel model
Siebeck.O. (1978) UV-Toleranz und Photoreacktiviening bei Daphnien aus Biotopen verschiedener
HOhenregionen. Natunvissenschaften, 65, 390.
RabyJD., Lagadeuc.Y., Dodson,JJ. and Mingelbier,M. (1994) Relationship between feeding and
vertical distribution of bivalve larvae in stratified and mixed waters. Mar. Ecol. Prog. Ser, 103,
275-284.
Rogers.H. (1940) Occurrence and retention of plankton within the estuary. J. Fish. Res. Board Can.,
5,164-171.
Scheltema.R.S. (1986) On dispersal and planktonic larvae of benthic invertebrates: an eclectic
overview and summary of problems. Bull Mar. ScL, 39, 290-322.
Scrope-Howe.S. and JonesJ).A. (1986) The vertical distribution of zooplankton in the western Irish
Sea. Esutarine Coastal Shelf Sci., 22, 785-802.
Silva-Serra,M.A. (1995) Early life history traits of sea scallops (Placopeaen magellanicus). PhD
Thesis, Dalhousie University, Halifax, Nova Scotia, Canada, 148 pp.
Silva-Serra,M.A. and O'Dor.R.K. (1994) Early life history traits of sea scallops, Placopecten magellanicus, from the Georges Bank population: vertical distribution of larvae. In Proceedings of the 9th
International Pectmid Workshop, Nanaimo, BC, Canada, April 22-27,1993. Vol. 1, pp. 67-75.
Sinclair,M. (1988) Marine Populations: An Essay on Population Regulation and Speciation. Washington University Press, Seattle.
Tremblay,M J. and Sinclair,M. (1990a) Sea scallop larvae Placopecten magellanicus on Georges Bank:
vertical distribution in relation to water column stratification and food. Mar. Ecol Prog. Ser., 61,
19-25.
Tremblay.MJ. and Sinclair.M. (1990b) Diel vertical migration of sea scallop larvae in a shallow
embaymenL Mar. Ecol. Prog. Ser, 67, 19-25.
Tremblay,MJ., LoderJ.W., Werner.F.E., Naime,C.E., Page,F.H. and Sinclaii.M.M. (1994) Drift of sea
scallop larvae Placopecten magellanicus on Georges Bank: a model study of the roles of mean
advection, larval behavior and larval origin. Deep-Sea Res., 41, 7-49.
Received on March 6, 1996; accepted on August 13, 1997
1973
Journal of Plankton Research Vol.19 no.12 pp.1975-1991, 1997
The correlation of downwelling irradiance and staggered vertical
migration patterns of zooplankton in Wilkinson Basin, Gulf of
Maine
Tamara M.Frank and Edith A.Widder
Harbor Branch Oceanographic Institution, 5600 U.S. 1 N, Ft Pierce, FL 34946,
USA
Abstract Field studies on the characteristics of light that influence vertical migrations in the
mesopelagic realm are sparse, due to the difficulty in simultaneously monitoring changes in species
distributions with changes in downwelling irradiance. Using the Johnson-Sea-Link submersible as a
platform, in situ measurements of the changes in downwelling irradiance at sunset were made simultaneously with observations on changes in animal distribution patterns in Wilkinson Basin, Gun' of
Maine. The results indicate that the vertical migrations for several species of large zooplankton are
staggered, with euphausiids (Meganyctiphanes norvegica) migrating first, cydippid ctenophores
(Euplokamus) migrating next, and two species of caridean shrimp (Dichelopandalus lepiocerus and
Pasiphaea multidentata) migrating last Data collected on daytime dives indicate that the daytime
depth distribution is not solely responsible for the migration order, and that different species may be
responding to different cues, or have different thresholds for the same cue.
Introduction
Huge populations of mesopelagic organisms undergo vertical migrations in all the
world's oceans. In some areas, the abundance of organisms is so great that sonic
scattering layers are formed, which can be detected on shipboard sonar systems.
These migrating populations include organisms, such as copepods and larval
shrimp, which are primary sources of nutrition for pelagic larval and juvenile
stages of commercially important fish species, as well as some of their major
predators, such as fish, adult shrimp, squid and ctenophores.
This is one of the most widespread and well-documented phenomena in the
world, and considerable debate has taken place about the ultimate (long-term)
and proximate (immediate) factors controlling this behavior. Recent studies in
the shallow water realm have provided overwhelming evidence in support of the
predator-avoidance hypothesis (Murray and Hjort, 1912; Hardy and Gunther,
1935; Zaret and Suffern, 1976) as the ultimate cause of, or driving force behind,
this migratory behavior. These studies [see Ohman (1990), Lampert (1993) and
Bollens et al. (1994) for review] indicate that the adoption of migratory or nonmigratory behavior is based on the relative abundance of predators and food in
the water column. The proximate factors controlling the daily timing of the vertical migrations have not been as intensively studied, and there is no overriding
support in favor of any particular cue, although it is generally accepted that light
plays some role in controlling the daily timing of the migrations (for reviews, see
Banse, 1964; Forward, 1988; Ringelberg, 1995), since most migrations occur at
sunrise and sunset. However, the specific characteristics) of the changing downwelling light field that acts as the trigger to cue these migrations in mesopelagic
species remains unknown.
© Oxford University Press
1975
T.M.Frank and E.A.WMder
There are several aspects of the environmental light field which could serve as
cues, including: (i) changes in underwater spectra; (ii) changes in the polarization
pattern; (iii) changes in absolute light intensity; (iv) the relative rate of change in
intensity [see Forward (1988) for a review]. Changes in the polarization pattern
and spectral distribution of light are not significant factors to consider for
mesopelagic organisms because (i) at the depths they inhabit, there are no longer
any changes in polarization distribution with changes in depth (Jerlov, 1976) and
(ii) they live below the depth at which spectral changes in the downwelling light
field can be detected (Frank and Widder, 1996). The most viable potential cues
for mesopelagic species are a threshold value of absolute light intensity, and/or a
relative rate of change in intensity. There is good evidence that some shallow
water species are cued solely by exposure to a particular light intensity (Pearre,
1973; Forward et al, 1984; Sweatt and Forward, 1985; Swift and Forward, 1988),
while in other species the relative rate of change in irradiance appears to be the
primary trigger (Clarke, 1930; Ringelberg, 1964; Daan and Ringelberg, 1969;
Strickler, 1969; Buchanan and Haney, 1980; Stearns and Forward, 1984; Haney et
al, 1990; Wagner-Dobler, 1990; Ringelberg and Flik, 1994). There are also excellent field studies indicating that some shallow water species are responding to a
threshold value of the relative rate of change in light intensity for cueing their
migrations (Buchanan and Haney, 1980; Stearns and Forward, 1984; Forward,
1985; Haney et al., 1990; Ringelberg et al., 1991).
Field studies in the mesopelagic realm have been rare, however, due to the
difficulty in measuring rapid changes in light intensity simultaneously with
changes in animal distributions, and most of these involved measuring responses
of scattering layers to changes in downwelling light (Kampa and Boden, 1954;
Boden and Kampa, 1958,1965,1967; Currie et al, 1969; Kampa, 1970,1975,1976;
Tont and Wick, 1973). Several of these studies indicate that organisms present in
these layers are probably not influenced solely by the absolute change in light
intensity, as the layers do respond to long periods of cloud cover (Boden and
Kampa, 1967; Kampa, 1976), but do not appear to respond to rapid changes in
intensity produced by passing clouds (Roe and Harris, 1980). However, studies
of the movements of scattering layers during solar eclipses produced equivocal
results. In some studies, the scattering layers began to ascend at the beginning of
the eclipse (Backus et al., 1965; Tont and Wick, 1973; Kampa, 1975), while in other
studies the scattering layers did not respond at all to the eclipse (Caruthers et al,
1970; Franceschini et al, 1970). A drawback of these aforementioned studies is
that the species composition of the layers was unknown, and some of these equivocal results could be due to different responses by different species. This contention is supported by the work of Bright et al. (1972), who studied the
movements of individual species into surface waters during an eclipse. They
found that four species of vertical migrators ascended during the eclipse, while
seven did not.
Roe and his colleagues (Roe, 1983, 1984a,b; Roe et al, 1984) have made the
only attempts in the mesopelagic realm to correlate in situ measurements of
downwelling light levels with daily movements of single species, by quantifying
samples collected with a net system equipped with an underwater light meter.
1976
DownweUing Irradiance and vertical migration patterns
These studies suffer from an unavoidable drawback associated with net sampling,
in that the net had to be continuously retrieved and deployed during a time when
light intensities and animal distribution patterns were changing rapidly. Additionally, there was no way of ensuring that the light meter attached to the net was at
the same angle for each trawl. The variable results of these studies, which showed
that significant differences in light levels could sometimes be associated with significant differences in the numbers caught, and sometimes could not, may be due
to real differences in animal distribution patterns, or may be due to experimental
errors in light measurements due to the way the light meter was oriented as the
net was towed. So, for mesopelagic organisms, very little has been established
about the light cues which trigger their migrations, primarily due to the difficulty
in measuring rapid changes in light intensity simultaneously with changes in
animal distributions.
The present study used the Johnson-Sea-Link submersible as a platform from
which to conduct in situ measurements of downwelling irradiance concurrently
with in situ observations of changes in animal distribution patterns. Our results
indicate that the migrations of several species of zooplankton in Wilkinson Basin
are staggered, and are separated by short time spans that might be missed by traditional net sampling techniques.
Method
In situ observations of the animal distribution patterns and measurements of
downwelling light were conducted in Wilkinson Basin, Gulf of Maine, during
June of 1995, with the Johnson-Sea-Link submersible. Two dives per day were
conducted whenever possible. The morning dives were conducted between 10:00
and 13:00 h, and data were collected on animal distribution patterns as well as
downwelling irradiance at various depths. The sunset dives were initiated 1.5 h
prior to sunset. The submersible was positioned at 122 m depth, which was well
above the depth at which the migrators had been observed during day dives, and
data were collected for the next 2.5-3 h. Animal distribution patterns were quantified via visual and bioluminescence transects, and in situ measurements of
downwelling irradiance were made before and after each transect.
Visual transects
The lights on the overhead work platform and the lower work platform on the
submersible were turned on, and the submersible moved forward at 0.5 knots for
4 min. The transect area, demarcated by the outermost bars on the upper and
lower work platforms, the starboard robotic arm (locked into the upright position) and the upright on the port side, encompassed 2.65 m2. Organisms entering
the transect area during the 4 min transect were identified and verbally recorded
to tape by the scientist (T.M.F. or E.A.W.) seated in the front chamber of the
Johnson-Sea-Link, which is a 5' diameter Plexiglas sphere that offers a panoramic
view of the water column. This sampling technique was utilized for the quantification of shrimp and gelatinous zooplankton.
1977
T.M.Frank and E.A.WWder
Bioluminescence transects
All lights were extinguished on the submersible, which again moved forward at
0.5 knots for 4 min. Luminescent organisms contacting a 0.2 m2 transect screen
were recorded to video tape with an intensified silicon intensified target (ISIT)
camera (Simrad Osprey) for later analysis. This sampling technique was utilized
for the quantification of gelatinous zooplankton (Widder et al., 1992b) and proved
to be particularly useful at several dive sites where the densities of ctenophores
were so high that they could not be accurately quantified during visual transects.
Light measurements
Light measurements were made with the Low Light Autoradiometer (LoLAR),
mounted on the submersible. This is a photomultiplier tube (PMT)-based autocalibrating radiometer, with a sensitivity range of 10~2 to 10"8 uW cm~2 (Widder
et al., 1992a). The PMT input was filtered with a 480 nm interference filter (fullwidth half-maximum intensity = 10 nm). A 480 nm interference filter was chosen
because this is close to the wavelength of light that penetrates best in clear seawater (475 nm), as well as the peak spectral sensitivities for those crustaceans for
which these data are available (480-490 nm; Frank and Case, 1988). The LoLAR
has a 1000 m depth capability and includes sensors for temperature, depth and
tilt. The input optics were designed around a lens system, rather than the conventional optical diffuser, in order to optimize sensitivity. The incident photon
flux is integrated over a 2 IT steradian solid angle and the measured angular
response of the system closely matches that of an ideal cosine collector. The error
between the measured and ideal response was 7.5%, calculated using Tyler's
(1960) data for radiance distribution at 66.1 m, according to the method of Smith
(1969). The experimentally measured (based on Smith, 1969) immersion-effect
factor was 0.57. The lower responsivity of the wet compared to the dry collector
is primarily a result of the increased angular divergence of the light as it passes
from the water to the air inside the dome and secondarily a result of decreased
internal reflections within the dome. The initial calibration of the system, with a
National Institute of Standards and Technology (NIST) referenced standard
(Optronics Laboratory Model 310 multi-filter calibration source), is maintained
by referencing the PMT output to the known quantum flux from a radiophosphorescent source. The flux from the radiophosphorescent source is monitored
with a high-stability silicon photodiode. During operation, periodic internal calibration sequences adjust the PMT responsivity calibration factor to correct for
changes in PMT responsivity. The stability and reliability of the LoLAR are
therefore based on the known temperature independence and excellent longterm stability of the silicon photodiode.
In situ irradiance measurements were made with the LoLAR before and after
each transect. Data were transmitted via a thru-hull penetrator from the detector head to a laptop computer in the front chamber of the submersible. Measurements of surface irradiance were collected with a ship-mounted LI-COR
quantum sensor (400-700 nm).
1978
Dowmreumg irradiance and vertical migration patterns
Data analysis
Owing to the patchy nature of the ocean environment, the distribution of organisms varied from dive to dive (Table I). To ensure that data from dives with abundant organisms did not bias the results, transect data for each species enumerated
on a dive were normalized to the peak abundance for that species on that dive.
When multiple transects were conducted at one depth or light level on a dive, the
data from these replicates (for that dive) were averaged so that only one data
point at each depth/light level per dive was incorporated into the final analysis.
This was done because the same numbers of replicates were not conducted at
every depth/light level on every dive. Therefore, for daytime dives, even though
multiple transects were conducted, each dive contributed three data points (one
in each depth range) to thefinalanalysis. Similarly, for night dives, transects were
conducted continuously for 3-3.5 h, but data from all the transects conducted at
the same light level on one dive were averaged, so that each dive contributed six
or seven data points (one at each light level) to the final analysis. The data were
analyzed with a single-factor ANOVA, and if statistical significance was
Table I. Numbers of organisms per cubic meter observed on each dive
Dive
#1
6/17/95
Day
#2
6/17/95
Sunset
#3
6/18/95
Day
#4
6/18/95
Sunset
#5
6/19/95
Day
#6
6/19/95
Sunset
#7
6/20/95
Sunset
Bottom
depth (m)
Transect
depth (m)
Number m~3
Euplokamis
M.norvegica
Carideans
42°32'N
69°33'W
279
215
245
260
0.188
11.8
11
0.03981
0.012
0
0
0
0.024
42°31'N
69°32'W
302
122
01
0.046
0.015
42°32'N
69°32'W
280
200
215
230
245
260
0
0
1.17
8.91
5.44
0
0
0.061
0.015
0.008
0
0
0.018
0
0.12
42°32'N
69°32'W
280
122
0.891
0.168
0.006
42°32'N
69°32'W
280
200
215
230
245
0.094
0.281
1.079
9.897
0
0.115
0.061
0.031
0
0
0.015
0.012
42°32'N
69°32'W
280
122
0.26
0.191
0.01
42°32'N
69°32'W
280
122
0.298
0.336
0.023
Location
Values for sunset dives are maximum numbers observed on any one transect during that dive.
•No Euplokamis were observed during the entire dive, so this value was not factored into the analysis. Note that the daytime dive on that date is not at the same location as the sunset dive.
1979
T-M.Frank and E-A.Widder
established with this test, the Newman-Keuls multiple range test for unequal
sample sizes was used to determine where the statistical significance occurred
among the sample means (Zar, 1974).
Rate of change (re) values were calculated using the following formula (Ringelberg, 1964):
re = [ln(/,//0)]/df
where /, and Io are irradiance at time t and 0, respectively, and dt is the time difference in seconds.
Results
Four species of vertical migrators were present in sufficient abundance to allow
for quantifiable observations of their distributions. These were: (i) Euplokamus
sp., a cydippid ctenophore; (ii) Meganyctiphan.es norvegica, a euphausiid crustacean; (iii) Dichelopandalus leptocerus, a pandalid shrimp; (iv) Pasiphaea multidentata, a pasiphaeid shrimp. Based on samples recovered with the submersible
and the midwater trawl, Euplokamus was the only species of cydippid ctenophore
present at the study sites during the study. The two species of caridean shrimp
(P.multidentata and D.leptocerus) could not be reliably distinguished from one
another during the visual transects, and are therefore grouped together in the
analysis as caridean shrimp. Based on net samples taken with a modified Tucker
Trawl (Frank and Widder, 1994a) during the cruise, P.multidentata outnumbered
D.leptocerus by 20 to 1.
Daytime depth distributions
A series of visual and bioluminescence transects were conducted between the
hours of 10:00 and 13:00, to determine the daytime depth distributions of the
species under examination. A vertical profile of the water column was conducted
from the surface down to the depth at which the first individual (of one of the
four species of interest) was observed. The first horizontal transect was then conducted at this depth. Subsequent transects were conducted 20 m above this depth,
and in 20 m intervals below this depth until the submersible was within 20 m of
the bottom. Data were collected on three dives with bottom depths ranging from
276 to 300 m (Table I). No individuals of any of the four species of interest were
observed shallower than 200 m on any of these dives, and the data displayed in
Figure 1 are therefore only from transects conducted at depths of 200 m and
deeper. At these sites, more M.norvegica were present between 220 and 240 m
than at any other depth, but statistically there were no significant differences
between the numbers present in any of the three depth ranges (Figure 1A).
Euplokamus, on the other hand, had a statistically significant population
maximum between 240 and 260 m water depth, and virtually no individuals were
found shallower than 220 m depth (Figure IB). The two species of carideans were
also always deeper than 220 m depth, and a statistically significant population
1980
Downweffing irradiance and vertical migration patterns
maximum was found between 220 and 240 m depth (Newman-Keuls, P <, 0.05;
Figure 1C). So, in summary, M.norvegica had a fairly broad daytime depth distribution, being equally abundant between 200 and 260 m depth, while the carideans
were primarily distributed between 220 and 240 m depth, and Euplokamus
remained primarily between 240 and 260 m depth.
Looking at population distributions relative to downwelling irradiance
(Figure 1), Euplokamus, whose population peak was deeper than those of the
other species, was also found to have a statistically significant population
maximum at the dimmest light level (105 photons cm""2 s"1 nnr 1 ), and virtually no
individuals were seen at the highest light level (107 photons cm"2 s"1 nm 4 ;
Figure IE). Meganyctiphanes norvegica (Figure ID), which had a shallower depth
distribution than Euplokamus, had a statistically significant population maximum
at 106 photons cur 2 s"1 nm"1, and no individuals were found at irradiances of 107
photons and above. The carideans (Figure IE) were statistically equally distributed between 105 and 106 photons cm"2 s"1 nm"1, and no individuals were found
at the highest light level (107). These light level divisions do not correspond precisely with the depth divisions, i.e. the depth range of 200-220 m contained transects corresponding to light levels of both 106 and 107 photons cm"2 s"1 nm"1.
While it is tempting to conclude that M.norvegica preferred a light level of 106
photons cm"2 s"1 nm"1 (at 480 nm), since it has a statistically significant population
maximum at this light level (Figure ID), but no apparent preferred depth range
(Figure IE), there are not enough data to support this contention. Light
measurements taken with the LI-COR indicated that all dives were conducted on
bright sunny days with superimposable surface irradiance values (an unusual, and
for our purposes, untimely, occurrence in the normally foggy Gulf of Maine) and
measurements taken with the LoLAR indicated that the attenuation coefficient
for 480 nm light was 0.08 for the water at all the dive sites. With these identical
conditions in both water clarity and surface irradiance on all three dives, it is not
possible to use these data to determine whether the daytime depth distributions
of these species indicate a preference for a specific light level. In order to determine whether a species actually prefers this light level, it is necessary to ascertain
whether the preferred daytime depth changes on a cloudy day versus a bright
sunny day, as heavy cloud cover on one day might result in less light being present
at, for example, 200 m depth on this day than at 230 m depth on a clear sunny
day. Heavy storm clouds can reduce irradiance by 1 log unit (Lythgoe, 1979),
which is comparable to a 29 m depth differential in the type of water present in
Wilkinson Basin.
Sunset migration patterns
The submersible was situated at 122 m depth 1-1.5 h before sunset, and transects
were taken continuously for the next 3 h. Data were collected on four sunset dives
(Table I). The euphausiid M.norvegica was the first species for which a substantial increase in abundance was observed. The first individuals were seen when the
downwelling irradiance was -10 7 photons cm"2 s"1 nm"1 (at 480 nm; Figure 2A).
The peak of the population passed by at 106 photons cm"2 s"1 nm"1, and after this
1981
TJHJrank and EA.Widder
U. norvegica
U. norvegica
1.00
T
0.800.60
0.400.200.00
107
200-220 220-240 240-260
B
106
10 s
Euplokamis sp.
Euplokamis sp
1.00 X
0.80-
0.00
200-220
10 7
220-240 240-260
1.00
T
10 5
Carkteans
Carkteans
1.00
10 6
T
200-220 220-240 240-260
Water Depth (m)
1982
Irradiance (photons c m V n m ' )
at480nm
Downweffing irradiance and vertical migration patterns
peak the numbers decreased significantly. Statistical analysis indicates that significantly more M.norvegica were seen at 106 photons cm"2 s"1 nnr 1 than at any
other light level (Newman-Keuls, P < 0.05). The next species in the migration
sequence was the ctenophore Euplokamus. The bulk of its migrating population
was observed when the light level was between 10s and 104 photons, and occurred
-15-30 min after the peak of the euphausiid migration (Figure 2B). The last
organisms to migrate were the caridean shrimp, P.multidentata and D.leptocenis.
The greatest number of individuals from this final migrating population passed
by the submersible when the light levels were between 104 and 103 photons cm"2
s"1 nm"1, -10-20 min after the peak of the Euplokamus migration (Figure 2C).
Migration timing versus rate of change in light intensity
The rate of change in irradiance was calculated both at the surface (using LI-COR
data) and at 122 m depth (using in situ data measured with the LoLAR) for four
sunset dives. As seen in Figure 3, which shows data for the sunset dive on 19 June
1995, the maximum rate of change at 122 m depth (where the submersible was
positioned) occurred shortly after sunset (20:20 h), around 20:35 h. This was consistently seen on all the sunset dives: the maximal rate of change was -15 min
after sunset. The times when the peak migrations of M.norvegica, Euplokamus
and the carideans at 122 m depth occurred are also shown on the figure. These
data indicate that the euphausiids must have started their migration well before
the maximum rate of change in irradiance occurred, since most of the migrating
population passed by the submersible before the maximum rate of change took
place (see Discussion).
Discussion
These data clearly demonstrate that there is a staggered migration pattern among
some of the dominant vertically migrating species in Wilkinson Basin, Gulf of
Maine. Among the organisms we were studying, the peak of the euphausiid
Fig. L Daytime distributions of vertical migrators in Wilkinson Basin, Gulf of Maine. (A-C) Distribution of denoted species versus water depth. (A) Meganyctiphanes norvegica: there are no statistically significant differences between the number of individuals present at the three depth ranges
(one-factor ANOVA, P £ 0.05). (B) Euplokamus: significantly more individuals were present between
240 and 260 m water depth than at any other depth (Newman-Keuls multiple range test, P <, 0.05).
(C) Carideans (P.multidentata and D.leptocerus): significantly more individuals were present between
220 and 240 m water depth than at any other depth (Newman-Keuls, P £ 0.05). (D-E) Distribution
of denoted species versus light level. Each jr-axis label denotes a range of irradiances, such that 107
stands for 1 x 107 to 9.9 x 107 photons cm"2 s~' nm"1 at 480 nm, etc. Light levels do not correspond
to depth ranges, as one depth range encompassed several light levels. (D) Meganyctiphanes norvegica:
significantly more individuals were present at 10* photons cm"2 s"1 nm"1 than at any other light level
(Newman-Keuls, PS 0.05). (E) Euplokamus: significantly more individuals were present at the lowest
light level (105 photons) than at the higher light levels (Newman-Keuls, P £ 0.05). (F) Carideans
(P.multidentata and D.leptocerus): significantly more individuals were present at 105 and 106 photons
than at 107, but differences between numbers present at 105 versus 10* photons were not significant
(Newman-Keuls, P £ 0.05).
1983
TJUFrank and E^V.WHder
M. norveglca
10 8
10 7
B
10 6
105
10 4
10 3
10 2
Euplokamls sp.
10°
10 7
10 8
10°
Cari deans
o
c 1.00 -r
a
•o
c 0.803
Si 0.60 <
•o
I"5
0.400.20-
E 0.00
o
z
108
107
20:15)
(20:1520:30)
106
(20:3020:45)
105
(20:4521«>)
104
103
102
(21KW21:10)
(21:102120)
(20-.2021:30)
Irradiance (photons cm"2s'1 nm'1) at 480 nm
(Time)
1984
Downwelling hradiance and vertical migration patterns
0.000
-. -0.001
o
a
Carl deans
6 -0.002
1
o
S
«a -0.003
ac
-0.004
18:45
19:15
19:45
20:15
20:45
21:15
Time
Fig. 3. Time versus rate of change in downwelling irradiance calculated utilizing data collected at the
surface (with the LI-COR) and 122 m depth (with the LoLAR) during a sunset dive on 19 June 1995.
The rate of change was calculated according to the formula: re = [ln(///0)]/d/. Times of migrating
population maxima at 122 m depth for M.norvegica, Euplokamus and the carideans are also indicated
on the graph.
Fig. 2. Migration patterns versus downwelling irradiance at sunset. All observations were made from
the submersible at 122 m depth, jr-axis labels indicate both light levels (at 480 nm) and times correlated with those light levels. (A) Migration pattern of M.norvegica. Significantly more individuals
were observed when the downwelling irradiance was 106 photons cm"2 s~' nm"1, between 20-30 and
20:45 h, than at any other light level (Newman-Keuls, P £ 0.05.) (B) Migration pattern of Euplokamus. The number of individuals recorded when the downwelling irradiance was between 105 and 10*
photons cm"2 s~' nm"1 was significantly greater than the quantities present at any other light level, but
they were not significantly different from each other (Newman-Keuls, P S 0.05). (C) Migration
pattern of the carideans, D.leptocerus and P.multidentata. The peak of the migrating population was
observed at downwelling irradiances of lOMO3 photons cm"2 s"r nnr1, and the numbers observed at
these two light levels, while not significantly different from each other, were significantly greater than
the numbers present at the other light levels. (Newman-Keuls, P <, 0.05).
1985
XMJVank and E^V.Widder
migration occurred first, followed 15-30 min later by that of the ctenophores, and
finally, 10-20 min after that, the peak of the caridean migration was seen. Data
from Hardy and Gunther (1935) and Cushing (1951) suggest that the order in
which animals appear during a migration is related to their day depth, which was
found to be true for several species of crustaceans (Anderson and Sardou, 1992;
Wiebe et al, 1992). However, Roe (1974) found that while the crustaceans in his
study maintained their sequential migrations with respect to day depth, the fish
did not, with deeper-living species arriving at the surface sooner than shallowerliving species. In the present study, the daytime depth range of the leaders in the
migration order, M.norvegica, was from 200 to 260 m water depth, while Euplokamus, which followed them in the migration order, was found to occur almost
exclusively between 240 and 260 m depth. Therefore, for these two species, differences in day depth might provide an explanation for the order of their migration
patterns (assuming equal migration speeds, which will be studied in the future).
However, the carideans, P.multidentata and D.leptocerus, share a similar depth
distribution as M.norvegica, and a substantial fraction of the population is actually found 20 m shallower than Euplokamus during the day, yet their migration
peak occurred -30-40 min after the peak of the euphausiid migration, and 10-20
min after the peak of the Euplokamus migration. The migration speed of
M.norvegica is -100 m h"1 (Hardy and Bainbridge, 1954), and while those of
P.multidentata and D.leptocerus have never been measured, observations on individuals in aquaria indicate that they are robust and more powerful swimmers than
M.norvegica. However, even erring on the side of caution and assuming equivalent migration rates as those of the euphausiids and ctenophores, the result is that
they should appear at approximately the same time in the migration pattern,
based on similar water depths, as the euphausiids, and earlier than the
ctenophores. The fact that they appeared later suggests that the carideans initiate their migration at a later time than both M.norvegica and Euplokamus, and
that they are using a different cue, or have a different threshold to the same cue.
It is not known what structure might serve a photoreceptive function in
Euplokamus (although an apical organ possessing tissue that resembles photoreceptive structures has been described in a ctenophore; Horridge, 1964), so little
can be said about their photosensitivity. However, with respect to the crustaceans,
there is some indication that the photoreceptor sensitivity of the carideans may
be different than that of Meganyctiphanes. The photosensitivities of the crustacean species in this study were determined electrophysiologically by means of
the electroretinogram (ERG), and while all three species share a similar spectral
sensitivity, with maximum sensitivity between 480 and 490 nm, there is some indication that M.norvegica is less sensitive to light than either of the carideans
(Frank and Widder, in preparation). The ERG is not the best way to compare
photoreceptor sensitivity between species, as the sensitivity of the eye depends
to some extent on the location of the electrode, but individuals of M.norvegica
were consistently 1-2 log units less sensitive than individuals of P.multidentata
and D.leptocerus, which might correlate with a higher threshold to a light cue. If
M.norvegica is indeed less sensitive to light than the carideans, then from
M.norvegica's point of view, at sunset it would appear to get darker sooner than
1986
Downweffing irradiance and vertical migration patterns
for the carideans, and therefore the euphausiids might start their migration
earlier. These preliminary indications of differences in photoreceptor sensitivity
will be examined in the future by conducting behavioral experiments to determine sensitivity thresholds in the reflexive light response, as has been done with
other species of deep-sea crustaceans (Frank and Widder, 1994a,b).
Vertical migrations of ctenophores have been observed in both the open sea
and nearshore environments (Hirota, 1974; Mackie, 1985; Vinogradov etai, 1985;
Wang et al., 1995). We were not able to study the migration of copepods (a major
prey item of this species) in our study, but neither Vinogradov et al. (1985) nor
Wang et al. (1995), in their study of an estuarine species, were able to demonstrate any correlation between the copepod migration and the ctenophore
migration. This brings up the interesting question of whether light is cueing the
migration in Euplokamus, as a functional photoreceptor has never been
described in any species of ctenophore.
0.000
-0.001
&
§ -0.002
o
s
•/ •
J5
-0.003
"5
\
-0.004
1/
V
- - • - • Surface
—D—300m
-0.005
17:30
Sunset
i
-
18:00
18:30
Time
19:00
Fig. 4. Time versus rate of change in irradiance calculated utilizing data from Qarke and Kelly (1965).
1987
T.M.Frank and E.A.WMder
One of the light cues that might be triggering the vertical migrations of various
species is the relative rate of change in irradiance. The data presented here indicate that the maximum rate of change in irradiance at 122 m depth actually occurs
-15 min after sunset. Our data corroborate those of Clarke and Kelly (1965), who
made simultaneous measurements of downwelling irradiance at the surface and
at 300 m. They did not calculate the rate of change, but one can use the data given
in their paper to do so, and from these calculations (Figure 4) it is clear that the
maximum rate of change at depth also occurs -15 min after sunset. Moreover,
what is very clear from their data, and suggested by our data, is that the maximum
rate of change at depth also precedes that at the surface by 10-15 min. Our own
data do not show this later peak at the surface as clearly as those of Clarke and
Kelly, because our surface light meter was not a photomultiplier, such as they
used, but a silicon photodiode with considerably less sensitivity. However, both
these data sets emphasize the importance of conducting in situ light measurements versus extrapolating surface data.
From the data presented here, it appears that the maximum rate of change in
irradiance is not triggering the migration of these M.norvegica. The maximum
rate of change occurred at -20:35 h at 122 m water depth, but the shallowest
daytime depth of this species was 95 m deeper than this and it is not known
whether the maximum rate of change at this deeper depth would occur at the
same time as, or earlier than, at the measured depth. However, it is clear that the
maximum rate of change would not have occurred before sunset at 20:20 h, and
therefore one can safely assume that this is the earliest possible time that the
maximum rate of change would have occurred at the shallowest daytime depth
of the euphausiids. The majority of the migrating M.norvegica passed by the submersible between 20:30 and 20:45 h, and must have covered a distance of at least
95 m before reaching the submersible. This would have required migrating speeds
of 200-500 m h"1, which is substantially higher than their measured maximum
migrating speed of 100 m h"1 (Hardy and Bainbridge, 1954). Therefore, it appears
that M.norvegica is not utilizing the maximum rate of change in intensity to
trigger its migration. Similar conclusions cannot be drawn for Euplokamus or the
carideans, as they were later in the migration order, and the earliest possible
maximum rate of change time point (20:20 h) would lead to a different conclusion
than later time points. The fact that the maximum rate of change in irradiance is
not the cue that triggers the migrations of M.norvegica does not preclude the
possibility, however, that they are responding to some submaximal relative rate
of change in intensity, as has been shown for shallow water Chaoborus flavicans
larvae (Wagner-Dobler, 1990), Chaoborus punctipennis larvae (Haney et al,
1990) and sessile coral reef invertebrates (McFarland, 1986).
In summary, the results of this study indicate that the migration patterns of
some of the dominant large vertical migrators in Wilkinson Basin are staggered,
and are separated by time intervals (10-30 min) that may have been missed by
traditional net sampling methods. The sequence of the migration order cannot be
completely explained by differences in daytime depth distributions, but appears
to be a result of different species responding to different cues, or having different thresholds to the same cue. The fact that the maximum rate of change in
1988
DownweQing irradiance and vertical migration patterns
irradiance occurred earlier at depth than at the surface underscores the need for
conducting simultaneous in situ measurements of downwelling light and animal
distribution patterns in the mesopelagic realm.
Acknowledgements
We would like to thank the captain and crew of the RV 'Edwin Link' and the
Johnson-Sea-Link submersible for their invaluable assistance with data collection. This research was supported in part by NSF grant no. #OCE-9313972 to
T.M.F. and E.A.W., and NOAA subgrant UCAP-95-02, University of Connecticut (award no. NA46RU0146) to T.M.F. and E.A.W. This is contribution #1208
of the Harbor Branch Oceanographic Institution.
References
Anderson.V. and SardouJ. (1992) The diel migrations and vertical distributions of zooplankton and
micronekton in the Northwestern Mediterranean Sea. 1. Euphausiids, mysids, decapods and fishes.
J. Plankton Res., 14,1129-1154.
Backus,R.H., Clark,R.C. and Wing^A.S. (1965) Behaviour of certain marine organisms during the
solar eclipse of July 20,1963. Nature, 205, 989-991.
Banse.K. (1964) On the vertical distribution of zooplankton in the sea. Prog. Oceanogr, 2, 53-125.
Boden3-P- and Kampa.E.M. (1958) Lumie're, bioluminescence et migrations de la couche diffusanta
profonde en M&literranee occidentale. Vie Milieu, 9,1-10.
Boden3-P- and KampaJEL.M. (1965) An aspect of cuphausiid ecology revealed by echo-sounding in
a fjord. Crustaceana, 9, 155-173.
Boden3-P and Kampa,E.M. (1967) The influence of natural light on the vertical migrations of an
animal community in the sea. Symp. ZooL Soc. London, 19, 15-26.
Bollens,S.M., Frost.B.W. and CordetU.R. (1994) Chemical, mechanical and visual cues in the vertical
migration behavior of the marine planktonic copepod Acartia hudsonica. J. Plankton Res., 16,
555-564.
Bright.T., Ferrari J1., Martin.D. and Franceschini,G.A. (1972) Effects of a total solar eclipse on the
vertical distribution of certain oceanic zooplankters. LimnoL Oceanogr., 17, 296-301.
Buchanan.C.L. and HaneyJ.F. (1980) Vertical migrations of zooplankton in the arctic: a test of
environmental controls. In Kerfoot,W.C. (ed.), Evolution and Ecology of Zooplankton Communities. Special Symp. Am. Soc. LimnoL Oceanogr. 3. University Press of New England, Hanover, pp.
69-79.
CaruthersJ.W., Thompson.R.C. and NovariniJ.C. (1970) The response of deep scattering layers in
the Gulf of Mexico to a total solar eclipse. Deep-Sea Res., 337, 337-338.
Clarke.G.L. (1930) Change of phototropic and geotropic signs in Daphnia induced by changes of light
intensity. / Exp. BioL, 2,109-131.
Clarke.G.L. and Kelly,M.G. (1965) Measurements of diurnal changes in bioluminescence from the
sea surface to 2000 m using a new photometric device. LimnoL Oceanogr., 10, R54-R66.
Currie,R.I., Boden3-P- and Kampa^E.M. (1969) An investigation on sonic-scattering layers: the RRS
'Discovery' SOND cruise, 1965. / Mar. BioL Assoc UK, 49, 489-513.
CushingJD.H. (1951) The vertical migration of planktonic Crustacea. BioL Rev., 26,158-192.
DaanJJ. and RingelbergJ. (1969) Further studies on the positive and negative phototactic reaction
of Daphnia magna Straus. Neth. J. ZooL, 19, 525-540.
Forward.R-B.Jr (1985) Behavioral responses of larvae of the CTab Rhithropanopeus harrisii
(Brachyura: Xanthidae) during diel vertical migration. Mar. BioL, 90, 9-18.
Forward.R.B.Jr (1988) Diel vertical migration: zooplankton photobiology and behavior. Oceanogr.
Mar. BioL Annu. Rev., 26, 361-393.
Forward.R-B.Jr, Cronin.T.W. and SteamsJD.E. (1984) Control of diel vertical migration: photoresponses of a larval crustacean. LimnoL Oceanogr., 29,146-154.
Franceschini.G.A., Bright.TJ., CaruthersJ.W., El-Sayed,S.Z. and Vastano^A.C. (1970) Effects (of the
total solar eclipse, 7 March 1970) on migration of marine organisms in the Gulf of Mexico. Science,
226,1155-1156.
1989
XM.Fr.nk and E.A-WIdder
Frank.T.M. and Case J.F. (1988) Visual spectral sensitivities of bioluminescent deep-sea crustaceans.
BioL Bull., 175, 261-273.
Frank,T.M and Widder,E.A. (1994a) Evidence for behavioral sensitivity to near-UV light in the deepsea crustacean Systellaspis debilis. Mar. BioL, 118, 279-284.
Frank.T.M. and Widder.E.A. (1994b) Comparative study of behavioral-sensitivity thresholds to nearUV and blue-green light in deep-sea crustaceans. Mar. BioL, 121,229-235.
Frank,T.M. and WidderJE. A. (1996) UV light in the deep-sea: in situ measurements of downwelling
irradiance in relation to the visual threshold sensitivity of UV-sensitive crustaceans. Mar. Freshwater Bchav. PhysioL, 27,189-197.
HaneyJ.F., Craggy,A., Kimball^K. and Weeks,F. (1990) Light control of evening vertical migrations
by Chaoborus punctipennis larvae. UmnoL Oceanogr., 35,1068-1078.
HardyA-C. and Bainbridge.R. (1954) Experimental observations on the vertical migrations of
plankton animals. / Mar. BioL Assoc UK, 33, 409-448.
HardyA-C. and Gunther,E.R. (1935) The plankton of the South Georgia whaling grounds and
adjacent waters 1926-27. Discovery Rep., 11, 456 pp.
HirotaJ. (1974) Quantitative natural history of Pleurobrachia bachei in La Jolla Bight. Fish. Bull., 72,
295-335.
Horridge.G.A. (1964) Presumed photoreceptive cells in a ctenophore. Q. J. Microsc. So., 105,
311-317.
Jerlov,N.G. (1976) Marine Optics. Elsevier, Amsterdam.
Kampa.E.M. (1970) Underwater daylight and moonlight measurements in the Eastern North
Atlantic. / Mar. BioL Assoc. UK, 50, 397-420.
KampaJE.M. (1975) Observations of a sonic-scattering layer during the total solar eclipse, 30 June,
1973. Deep-Sea Res., 22,411-423.
Kampa.E.M. (1976) Photoenvironment and vertical migration of mesopelagic marine animal communities. In Decoursey,PJ. (ed.), Biological Rhythms in the Marine Environment. University of
South Carolina Press, Columbia, pp. 257-272.
Kampa^.M. and Boden.B.P. (1954) Submarine illumination and the twilight movements of a sonic
scattering layer. Nature, 174, 867-873.
Lampert.W. (1993) Ultimate causes of diel vertical migration of zooplankton: new evidence for the
predator-avoidance hypothesis. Arch. Hydrobiol. Beih. Ergebn. LimnoL, 39, 79-88.
LythgoeJ.N. (1979) The Ecology of Vision. Clarendon Press, Oxford.
Mackie.G.O. (1985) Midwater macroplankton of British Columbia studied by submersible PISCES
IV. J. Plankton Res., 7, 753-777.
McFarland,W.N. (1986) Light in the sea—correlations with behaviors of fishes and invertebrates. Am.
ZooL, 26, 389-401.
MurrayJ. and HjortJ. (1912) The Depth of the Ocean. Macmillan, London.
Ohman,M.D. (1990) The demographic benefits of diel vertical migration by zooplankton. Ecol.
Monogr., 60, 257-281.
Pearre.S.Jr (1973) Vertical migration and feeding in Sagitta elegans VerriU. Ecology, 54,300-314.
Ringelberg J. (1964) The positive phototactic reaction of Daphnia magna Straus: a contribution to the
understanding of diurnal migration. Neth, J. Sea Res., 2, 319-406.
RingelbergJ. (1995) Changes in light intensity and diel vertical migration: a comparison of marine
and freshwater environments. /. Mar. BioL Assoc. UK, 75,15-25.
RingelbergJ. and Flik3-JG. (1994) Increased phototaxis in the field leads to enhanced diel vertical
migration. LimnoL Oceanogr., 39,1855-1864.
RingelbergJ., Flik,BJ.G., LindenaarJ). and RoyackersJC. (1991) Diel vertical migration in Daphnia
hyalina (sensu latiori) in Lake Maarsseveen. Part 1. Aspects of seasonal and daily timing. Arch.
HydrobioL, 121,129-145.
Roe.H.SJ. (1974) Observations on the diurnal vertical migrations of an oceanic animal community.
Mar. BioL, 28, 99-113.
Roe.H.SJ. (1983) Vertical distributions of euphausiids and fish in relation to light intensity in the
Northeastern Atlantic. Mar. BioL, Tl, 287-298.
Roe^HSJ. (1984a) The diel migrations and distributions within a mesopelagic community in the north
east Atlantic. 2. Vertical migrations and feeding of mysids and decapod Crustacea. Prog. Oceanogr.,
13,269-318.
RoeJrl.S J. (1984b) The diel migrations and distributions within a mesopelagic community in the north
east Atlantic. 4. The copepods. Prog. Oceanogr., 13, 353-388.
RoeJrl.SJ. and Harris,M J. (1980) A new acoustically telemetering deep-sea photometer with some
observations on underwater light in the northeast Atlantic. Deep-Sea Res., 27A, 181-195.
Roe.H.SJ., Angel,M.V., BadcockJ., Domanski.P., JamesJ^.T, PughJ'.R. and Thurston,M.H. (1984)
1990
DownweQing irradiance and vertical migration patterns
The diel migrations and distributions within a mesopelagic community in the north east Atlantic.
1. Introduction and sampling procedures. Prog. Oceanogr, 13, 245-268.
Smith,R.C. (1969) An underwater spectral irradiance collector. /. Mar. Res., 27, 341-351.
StearnsJD.E. and Forward.R.B.Jr (1984) Copepod behavior in a simulated natural light environment
and its relation to nocturnal vertical migration. Mar. Bioi, 82,91-100.
StricklerJ.R. (1969) Ueber das schwimmverhalten von Cyclopoiden bei verminderungen der
Bestrahlungsstaerke. Schweiz. Z. HydroL, 31,150-180.
Sweatt, A J. and Forward^R.B. Ji (1985) Spectral sensitivity of the chaetognath Sagitta hispida Conant.
BioL Bull., 168, 32-38.
Swift,M.C. and Forward^R.B.Jr (1988) Absolute light intensity vs. rate of relative change in light
intensity: the role of light in the vertical migration of Chaoborus punctipennis larvae. BulL Mar.
ScL, 43, 604-606.
Tont.S.A. and Wick.G.L. (1973) Response of a deep scattering layer to the 1972 total eclipse. DeepSea Res., 20, 769-771.
TylerJ.E. (1960) Radiance distribution as a function of depth in an underwater environment. BulL
Scripps Inst. Oceanogr., 7, 363-412.
Vinogradov,M.E., Flint.M.V. and ShushkinaJE.A. (1985) Vertical distribution of mesoplankton in the
open area of the Black Sea. Mar. BioL, 89, 95-107.
Wagner-Dobler,I. (1990) Vertical migration of Chaoborusflavicans(Diptera, Chaoboridae): Control
of onset of migration and migration velocity by environmental stimuli. Arch. Hydrobiol., 117,
279-307.
Wang,Z., Thi6baut,E. and DauvinJ.C. (1995) Spring abundance and distribution of the ctenophore
Pleurobrachiapileus in the Seine estuary: advective transport and diel vertical migration. Mar. BioL,
124, 313-324.
Widder,E.A., Cainu\F.M., Taylor.L.D. and Tusting,R.F. (1992a) Design and development of an autocalibrating radiometer for deep-sea bio-optical studies. Ocean. Eng. Soc. IEEE Oceans '92, Vol. 1,
525-530.
WidderJE.A., Green.C.H. and Youngbluth.MJ. (1992b) Bioluminescence of sound-scattering layers
in the Gulf of Maine. / Plankton Res., 14,1607-1624.
Wiebe,P.H., Copley.NJ. and Boyd.S.H. (1992) Coarse-scale horizontal patchiness and vertical
migration of zooplankton in Gulf Stream warm-core ring 82-H. Deep-Sea Res., 39, S247-S278.
ZarJ.H. (1974) Biostatistical Analysis. Prentice-Hall, Inc., Englewood Cliffs, NJ.
Zaret.T.M. and SuffernJ.S. (1976) Vertical migration in zooplankton as a predator avoidance
mechanism. LimnoL Oceanogr, 21, 804—813.
Received on April 14, 1997; accepted on August 15, 1997
1991