Diel vertical migration (DVM) is a ubiquitous phenomenon in marine and freshwater plankton communities. Most commonly, plankton migrate to surface waters at dusk and return to deeper waters at dawn. Up until recently, it was thought that DVM was triggered by a relative change in visible light intensity. However, evidence has shown that DVM also occurs in the deep sea where no direct and background sunlight penetrates. To identify whether such DVM is associated with latitudinal and seasonal day light variation, one and a half years of recorded acoustic data, a measure of zooplankton abundance and movement, were examined. Acoustic Doppler current profilers, moored at eight different sub-tropical latitudes in the North-Atlantic Ocean, measured in the vertical range of 500–1600 m. DVM was observed to follow day length variation with a change in season and latitude at all depths. DVM followed the rhythm of local sunrise and sunset precisely between 500 and 650 m. It continued below 650 m, where the deepest penetrable irradiance level are <10−7 times their near-surface values, but plankton shortened their time at depth by up to about 63% at 1600 m. This suggests light was no longer a cue for DVM. This trend stayed consistent both across latitudes and between the different seasons. It is hypothesized that another mechanism, rather than light, viz. a precise biochemical clock could maintain the solar diurnal and seasonal rhythms in deep sea plankton motions. In accordance with this hypothesis, the deepest plankton were consistently the first to migrate upwards.
Citation: van Haren H, Compton TJ (2013) Diel Vertical Migration in Deep Sea Plankton Is Finely Tuned to Latitudinal and Seasonal Day Length. PLoS ONE 8(5): e64435. https://doi.org/10.1371/journal.pone.0064435
Editor: Andrew Davies, Bangor University, United Kingdom
Received: November 22, 2012; Accepted: April 15, 2013; Published: May 22, 2013
Copyright: © 2013 van Haren, Compton. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The funding of instrumentation by the Netherlands Organization for the advancement of scientific research (N.W.O.) large investment program Long-term Ocean Current Observations (LOCO) is gratefully acknowledged. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The author has declared that no competing interests exist.
Diel vertical migration (DVM) in ocean zooplankton is likely to represent the largest daily migration of animals on earth, in terms of biomass . Most commonly, plankton migrate to surface waters at dusk and return to deeper waters at dawn. This pattern of migration is considered to reflect a trade-off between to need to feed versus predator avoidance, especially in the sunlit photic zone , .
Knowledge of DVM in plankton goes back centuries (e.g., , ). But since the advance of acoustic methods since the early 1980s, considerable details about DVM have been revealed; and have also been verified with plankton net-tows (e.g., , ). Acoustics provide numerous advantages over plankton net-tows. They are non-invasive, they can detect the larger zooplankton capable of swimming out of the way of nets, they can be used for continuous monitoring  and they can directly measure the vertical speed of zooplankton . Consequently, acoustic backscattering is now a common method used to monitor relative zooplankton abundance and their vertical migrations (e.g., , –). But acoustics are limited in that they cannot provide taxonomic resolution and they have difficulty in quantifying plankton biomass with a single frequency instrument .
In the photic zone, down to a few 100 m from the surface, temporal high-resolution acoustic data, representing plankton migration, identified different stages of dusk ascent and dawn ascent . Ascent started slowly well before sunset but the fast portion ended before twilight. The first slow descent started hours before twilight, while the fast descent ended just before sunrise. This confirmed earlier conclusions that near-surface DVM was triggered by local solar variations. Furthermore, near-surface evidence using annual time series from acoustics in high latitude sites proved seasonal variations in day length estimated from DVM , .
Initially, DVM seemed limited to the photic zone. But in the arctic, synchronized DVM behaviour of zooplankton was found in both open and ice-covered waters of 30–60 m, where light was below the detection level of many standard irradiance meters . Interestingly, this migration was linked with background light intensity levels at less than half the minimum detection level for the human eye. Deep acoustic observations down to 1200 m  and down to 1400 m  have also demonstrated DVM. But by contrast with the Arctic, deep sea migration is below both visible and background light intensities. At these depths, in the subtropics, a 1% light level reaches ∼100 m deep, whereas the deepest penetrable 480 nm irradiance level is <10−7 times the near-surface value at 600 m . A 1.5 years long time series from the Canary Basin  even demonstrated a lunar monthly modulation of DVM, exactly in phase with the moon. Moonlight reaches its <1‰ light level at about 125 m in that area . Thus a gap in our knowledge remains on whether DVM in the deep sea is pervasive across latitudes and whether the timing of migration is sensitive to changes in seasonal day length.
To identify whether DVM from the deep-sea is associated with seasonal day light variation across latitudes, records of one and a half years of acoustic data, a measure of zooplankton migration, were examined. The acoustic data was derived from Acoustic Doppler current profilers (ADCP) at eight different sub-tropical latitudes ranged between 500 and 1650 m vertically. More than half of this range is not reached by any detectable daylight .
Materials and Methods
Over a period of three years, up- and downward looking 75 kHz Teledyne-RDI ADCPs have been moored at eight sites between the equator and 33°N in the North-Atlantic Ocean (Table 1 for details). Record lengths varied between 17.5 and 19.5 months. The ADCPs were mounted around 1000–1600 m in the top-buoy of 3000–3700 m long moorings located in 4500–5200 m water depth in the Canary, Cape Verde and Ceará (equatorial Brazil) Basins. No specific permissions were required for these moorings. The field studies did not involve endangered or protected species, as no samples were collected.
The ADCPs have four beams that are slanted at an angle of θ = 20° to the vertical. They were set to range 500–600 m from the head, in 10 m vertical intervals. ADCPs' pressure and tilt sensor information showed that the moorings did not move much in the Canary and Cape Verde Basins, <1.5° tilt angle implying mooring motions across <1.2 m in the vertical z and <100 m in horizontal x, y directions. In the equatorial region, a single strong current event caused deflection of about 5°. These small mooring motions are due to efforts to minimize drag using two large, elliptically shaped, 300–500 kg net buoyancy elements and a thin (0.007 m diameter) nylon-coated steel cable.
Determination of zooplankton migration from ADCP-data
The current components (u, v, w) in the associated Cartesian coordinates (x-East, y-North, z-up) are measured at different depth levels z (usually 10 m apart) per time step t ( = 900 or 1800 s) as averages over the horizontal acoustic beam spread. This is due the θ = 20° vertical slant angle. The spread measures 20–440 m, depending on the range from the ADCP. In contrast, the acoustic reflections ‘echo intensities’ (I) are obtained per beam. They are thus averaged over much smaller horizontal scales of typically 10 m. As the I-data are dominated by the attenuation of sound through the water column, the suspended particle signal dI is obtained by subtracting the time mean, denoted by <…>, from the raw data for each z, dI(z, t) = I(z, t)−<I>(z) [dB]. Far from bottom boundaries, as found in the open North-Atlantic, the dominant source for 75 kHz ADCP dI are macroplankton, including zooplankton and ichtyoplankton, and fish species that have sizes larger than about 0.01 m , , . In fact, without ocean life of this size, (75 kHz) ADCPs simply do not function in clear ocean waters.
Variations in dI and w will represent variations in the DVM of plankton. These variations consist of changes in plankton abundance, size and form (dI, ), and in plankton motions (w). The existence of macroplankton at 500–2000 m in the NE Atlantic Ocean have been confirmed between 15 and 53°N using net-tows and manned submersible , .
For monitoring daily variations in the relative amount of plankton (dI), composites are computed per selected monthly period (N = 30 days),(1a)where dIn represents dI at a given day n . Similarly, for monitoring DVM-movements the composite wc is computed from w,(1b)These vertical currents are dominated by internal (tidal) wave motions and by plankton migrations, although at clearly distinguishable different frequencies .
We thus have two independent and different measures from a single acoustic instrument for remotely studying vertical plankton migration. The daily migrations, in raw dI- and w-data or in averages over a month, as in equation (1), are used for monitoring day length variations in DVM. This is done by searching for the times of daily minima and maxima in w and in the derivative of dI at given depths. The time series of such minima and maxima are then harmonic analyzed  with a period of exactly one year, to determine amplitude, phase and mean. As a measure for “error”, the normalized residual variance is used. It is noted that this “error” is not a random error, as it contains all other possible harmonic motions including tides. Especially for the equatorial region the “error” is expected to be large as seasonal variations will be negligible.
Estimation of day length
The DVM-variations in day length will be compared below with exact sunrise/sunset data for the given sites, as computed using http://aa.usno.navy.mil/data/docs/RS_OneYear.php. Latitudinal variations were computed using the model ,(2)where D is day length (in hours), J is day of the year and L is the latitude (in radians). This model included the effects of declination of the earth (23.5°) and the atmospheric refraction. It deviated from true solar day length by less than 1 minute for |L|<40°.
All acoustic observations investigated here showed a seasonal variation, see the example from 22.5°N in Figure 1a. Superimposed on this seasonal cycle were monthly and shorter period variations. Typical variations in dI were 10 dB (Fig. 1b,c), or one order of magnitude in back-scattered sound energy, and in w 0.03 m s−1 (Fig. 1d,e). These speeds were maintained for about 3 hours, resulting in a vertical range of DVM of typically 300 m. The zooplankton studied here never reached the surface and most, those deeper than about 1000 m, never reached depths to where sunlight variations could be sensed.
Below −1150 m noise is increased, due to lower amounts of plankton once they have migrated upwards after dusk. Time in 2007 is yearday +365. (a) Entire 18 months of depth versus local solar time series of plankton, or echo intensity, relative to its time mean. (b) Monthly averaged summer time data from white rectangle in a. Symbols ‘s’ indicate local sunrise and sunset, ‘t’ nautical twilight and ‘a’ astronomical twilight. (c) Month of July raw data (rectangle in a.) of which b. is the average. (d) as c., but for w. (e) as b., but for w.
Zooplankton at subtropical latitudes were observed over nearly the entire acoustic range, even though noise levels increased, and thus amounts of acoustic scatterers decreased, below z = −1150 m. The time series of dI did not exhibit a daily sinusoidal signal, but rather a rectangular shaped periodicity (Fig. 1b,c). In July, the range of maximum daily dI-values (dI>5 dB) had a duration of 9–11 hours within the ADCP-range (Fig. 1b). This was about 25% shorter than the sun's cycle at 22.5°N. The same was found for the daily variation in down- and upward plankton motions, when they exceeded noise levels (Fig. 1e, |w|>0.005 m s−1). The period of ascent (near dusk) was shorter than descent (near dawn), but associated peak values in w-amplitude, or the vertical movement of plankton, were larger for up- than for downward motion. Consequently, the vertical displacement of plankton was the same in both directions; indicating DVM.
The sharply defined peak-w near sunset, as determined from daily (Fig. 2b) or monthly (Fig. 3b) averaged periods, was used to construct a time series of plankton DVM that followed seasonal sunset variation. The time series describing plankton DVM even followed the difference between spring (slow change) and autumn (rapid change), thereby explaining 75% of the variance in sunset variation at sub-tropical latitudes like 22.5°N in Figure 3. Similar observations of seasonal variation in DVM were independently made using the first derivative of dI (Fig. 3, Fig. 2a). The opposite slow/fast changes during sunrise were less well mimicked, apparently because the plankton moved in a more diffusive manner during descent. The net result of seasonal DVM-variations in day length (Fig. 3a) showed slightly larger amplitudes of plankton migration, as compared with the amplitudes of the sun's variations. However, at 1000 m the plankton seasonal day length amplitudes reflected by plankton DVM were around a mean that equaled 9.3 hours, or about 77% of the mean solar daylight period of 12.1 hours (Fig. 3a).
(a) Relative echo intensity. (b) Vertical current; note the much clearer upward migration of the plankton, confined in a shorter period, compared with the less clear downward migrations. In both panels, the black lines denote times of local sunrise and sunset.
Thick solid lines indicate the times of true local day length determined from the sun. (a) Total day length; dashed line is the solid line multiplied by a factor of 77%. This factor is the average day length determined from plankton DVM, being slightly larger for dI-determined day length (o) and slightly smaller for w-determined day length (*). (b) Near sunrise and sunset.
At the equator, seasonal solar day length did not vary, as was also confirmed by plankton DVM around −587 m (Fig. 4a). Below −650 m, the plankton determined day length decreased steadily with increasing depth, down to about 7.5 hours or 63% of the solar day length at −1600 m. It was thus seen that deepest plankton were the first to rise. The composite plots of Figures 4b,c also showed that plankton migrated over ranges of about 300 m.
(a) As Figure 3a, but for z = −587 m at the equator (0°N, 37°W, mooring LOC144 (Table 1). (b) Composite mean relative echo intensity from upward looking ADCP at (0°N, 37°W) between −973 and −483 m and downward looking ADCP at (1°N, 38°W, mooring LOC164 (Table 1) between −1075 and −1565 m. (c) The associated sunrise and sunset profiles with depth with respect to the minimum morning and maximum afternoon values in dI, or plankton, variation (first derivative of, black lines). The filled blue fields indicate errors. For reference, similar profiles have been computed from minimum morning and maximum afternoon w (red dashed lines). Times of local sunrise and sunset are followed (to within errors) down to about −650 m; for z<−650 m the times between minima and maxima for both d(dI) and w decrease roughly linearly with increasing depth; at large depths (z<−1400 m) errors become large, especially for w for which values become unrealistic.
When we include all 8 available 1.5 years long time series, the day length variation adopts a latitudinal dependence, which follows model (2) relatively well (Fig. 5a). The main discrepancy between the present observations and the model (2) was a shift by about 0.6 hours for w-data, with the larger amplitude for the latter. This discrepancy could not be explained by equation (2) and it was entirely due to different peak-downward-speeds in winter and in summer (see Fig. 6). In winter, the relatively long period of descent peaks late (day 377.38 in Fig. 6a≈9 o'clock in the day), whereas in summer it peaks early (560.29 in Fig. 6b≈7 o'clock in the day) in the group of downward migrating plankton. No such seasonal variation was found in the group of the more clearly defined ascending plankton (Fig. 6). As sunrise and sunset at 22.5°N both vary over the season by about an hour, the observed zooplankton w-peak in winter was later, by an hour, in the example of Figure 6. Averaged over a year, this extra shift in time amounted to 0.6 hours (Fig. 5). Such extra seasonal differences were not found for both the down- and upward motion periods in dI-data. Note that twice the above extra shift (2×0.6 = 1.2 hours) in w-data did not explain the deviation from solar day length by the acoustic day length means, which were on average decreased to 77% of 12.1 hours (a difference of 2.8 hours) at −1000 m. The fit to the day length model (2) with the present acoustic data was remarkable, as the residual variance was quite high (Fig. 5b). Hence the explained variance by the harmonic daily variation was quite low. This was expected for the equator, where the seasonal amplitude was negligible. However, also at other latitudes away from the equator daily variations in plankton constitution and effects by, e.g., internal wave motions, induce considerable “noise” variations.
(a) Day length amplitudes. The red graph indicates the suns variation in day length according to equation (2). The dashed red line indicates the same graph shifted by 0.6 hours (see text). Note that at 1000 m the average acoustics data residual is 77% (9.3 hours) of the sun's day length. (b) Normalized residual variance, following harmonic analysis. High values are expected near the equator, where a yearlong harmonic is expected to be small. For those data, vertical averages over 200 m were computed and an expected reduced “error” is indicated by purple symbols. Similarly, averages over 4 bins are computed for data from 15°N, which showed unexpected high normalized residual variance.
In winter, the more diffuse downward migration peaks later than expected as compared with the summer-data. The discrepancy between the absolute time difference between plankton movement (w-peaks) in summer and winter (dashed lines in lower panel) is larger in the morning than in the evening. This explains the larger amplitude of DVM-variations with day length observed in Figure 5 for peak-w (*) by about 40 minutes on average.
The present observations show that DVM of zooplankton in the deep sea follows precise solar variations in day length, across different latitudes and thus matches the sun's latitudinal influx variation. Contrary to the conventional paradigm that DVM would cease in areas where visible light does not penetrate, this study, as well as others , , thus show that plankton continue to migrate where visible light does not penetrate. Interestingly, deep sea plankton always formed dense aggregations during the day but became diffuse closer to the surface during the night. The dense aggregations formed by zooplankton at depth during the day might suggest a form of anti-predation strategy , , similar to flocking in birds (e.g., ).
Up until recently, it was thought that zooplankton migrations were triggered by the absolute and relative changes in downwelling irradiance , . But the maximum depth of sunlight penetration that can be sensed by plankton was estimated to be 700 m at sub-tropical latitudes . This was close to the depth of 650 m at which the plankton (from acoustic dI and w) showed a change in mean day length variation, independent of latitude. Above 650 m, here verified acoustically up to 500 m, mean day length for the plankton was equal to the mean solar day length. Below 650 m, verified acoustically down to 1600 m, mean day length variation for the plankton decreased with depth. These acoustical observations suggest that plankton deeper than 650 m do not sense sunlight.
At all latitudes throughout the vertical range studied, the observed vertical migratory speed of the plankton was ∼0.03 m s−1, which was maintained over 3 hours (104 s). Hence, their vertical migratory distance was ∼300 m. These values confirmed previous acoustic observations of DVM (e.g., , , ). As a result, plankton living at a daytime depth of 1000 m or deeper will certainly never directly sense daylight and background solar irradiance. Such individuals thus need a trigger other than sunlight to control their vertical migration.
The question remains as to what the precise trigger mechanism of the seasonal (and lunar ) modulated DVM is, especially because in this study and others it has been shown that the deepest plankton groups move first, apparently without solar or lunar light ‘Zeitgeber’ (e.g., ).
On the one hand, acoustic noise by moving plankton could be a trigger for the descent, whereby plankton at levels higher-up in the water column, those triggered by the sun, are a signal for deeper plankton to migrate downwards. But such a trigger cannot work for the ascent, as the deepest plankton, that are deprived of sunlight, are the first to migrate upwards. On the other hand, deep-sea DVM could be induced by the temporal variation in free-falling food objects like faecal pellets, but it is unlikely, as their drop speeds are too slow (∼0.005 m s−1 (400 m day−1); –). Such sinking speeds would cause large phase delays in DVM, i.e. by several days.
Alternatively, DVM could be controlled by precise internal clocks, as has been speculated for lunar modulation in DVM . However, this would require a clock imprint or clock learning in earlier life stages when plankton live closer to the surface, like life stages of a number of zooplankton species (e.g., ). As follows from the observations, either the daily variation has to be memorized in the case of signaling by plankton higher up. Or when no such signals are received, a complete seasonal cycle needs to be learned, which might seem unlikely. On the contrary, it is known that biochemical oscillators exist that can maintain stable rhythms for months or even years in the absence of a daily trigger , . They do so in many organisms including mammals, migratory birds  and unicellular [marine] organisms like cyanobacteria and dinoflagellates , . If these results are transferrable to zooplankton, it would be challenging to perform experimental work on zooplankton, starting with studying species in flume tanks under varying light conditions. Subsequently, to see if DVM is genetically programmed, one would need to study different life stages.
We thank the crew of the R/V Pelagia and NIOZ-MTM for the deployment and recovery of the moorings. The funding of instrumentation by N.W.O. large investment program Long-term Ocean Current Observations (LOCO) is gratefully acknowledged.
Conceived and designed the experiments: HvH. Performed the experiments: HvH. Analyzed the data: HvH. Contributed reagents/materials/analysis tools: HvH. Wrote the paper: HvH TC.
- 1. Heys GC (2003) A review of the adaptive significance and ecosystem consequences of zooplankton diel vertical migrations. Hydrobiol 503: 163–170.
- 2. Fortier M, Fortier L, Hattori H, Saito H, Legendre L (2001) Visual predators and the diel vertical migration of copepods under Arctic sea ice during the midnight sun. J Plankt Res 23: 1263–1278.
- 3. Pearre S Jr (2003) Eat and run? The hunger satiation hypothesis in vertical migration: history, evidence and consequences. Biol Rev 78: 1–79.
- 4. Latreille PA (1829) in Le règne animal distribué d'après son organisation, pour servir de base a l'histoire naturelle des animaux et d'introduction a l'anatomie compare, ed Cuvier JLNF, Déterville, Paris, F: 1–584.
- 5. Murray J, Hjort J (1912) The depths of the ocean. London: MacMillan, 821 p.
- 6. Flagg CN, Smith SL (1989) On the use of the acoustic Doppler current profiler to measure zooplankton abundance. Deep-Sea Res 36: 455–474.
- 7. Schott F, Johns W (1987) Half-year long measurements with a buoy-mounted acoustic Doppler current profiler in the Somali current. J Geophys Res 92: 5169–5176.
- 8. Haney JF (1988) Diel patterns of zooplankton behaviour. Bull Mar Sci 43: 583–603.
- 9. Plueddemann AJ, Pinkel R (1989) Characterization of the patterns of diel migration using a Doppler sonar. Deep-Sea Res 36: 509–530.
- 10. Clay CS, Medwin H (1977) Acoustical oceanography: principles and applications. New York: John Wiley & Sons, 544 p.
- 11. Jiang S, Dickey TD, Steinberg DK, Madin LP (2007) Temporal variability of zooplankton biomass from ADCP backscatter time series data at the Bermuda testbed mooring site. Deep-Sea Res I 54: 608–636.
- 12. Fielding S, Griffiths G, Roe HSJ (2004) The biological validation of ADCP acoustic backscatter through direct comparison with net samples and model predictions based on acoustic-scattering models. ICES J Mar Res 61: 184–200.
- 13. Thomson RE, Allen SE (2000) Time series acoustic observations of macrozooplankton diel migration and associated pelagic fish abundance. Can J Fish Aquat Sci 57: 1919–1931.
- 14. Fischer J, Visbeck M (1993) Seasonal variation of the daily zooplankton migration in the Greenland Sea. Deep-Sea Res I 40: 1547–1557.
- 15. Berge J, et al. (2009) Diel vertical migration of Arctic zooplankton during the polar night. Biol Lett 5: 69–72.
- 16. van Haren H (2007) Monthly periodicity in acoustic reflections and vertical motions in the deep ocean. Geophys Res Lett 34: L12603 .
- 17. Kampa EM (1970) Underwater daylight and moonlight measurements in the eastern North Atlantic. J Mar Biol Ass UK 50: 397–420.
- 18. RDI (1996) A practical primer. San Diego: Teledyne-RDI, San Diego, 32 p.
- 19. Vinogradov GM (2005) Vertical distribution of macroplankton at the Charlie-Gibbs Fracture Zone (North Atlantic), as observed from the manned submersible “Mir-1”. Mar Biol 146: 325–331.
- 20. Vinogradov ME, Vereshchaka AL, Vinogradov GM, Musaeva EI (2000) Vertical distribution of zooplankton at the periphery of the North Atlantic subtropical gyre (in Russian). Izv AN Ser Biol 4: 496–510 English translation: Biol Bull (Trans Izv AN Ser Biol) 27:417–429.
- 21. Dronkers JJ (1964) Tidal computations in rivers and coastal waters. Amsterdam: North Holland Publishing Company, 518 p.
- 22. Forsythe WC, Rykiel EJ Jr, Stahl RS, Wu H-i, Schoolfield RM (1995) A model comparison for day length as a function of latitude and day of the year. Ecol Mod 80: 87–95.
- 23. Krebs JR, Davies NB (1993) An introduction to behavioural ecology. Oxford: Blackwell, 420 p.
- 24. Krause J, Ruxton GD (2002) Living in groups. Oxford: University Press, 211 p.
- 25. Piersma T, et al. (1993) Interactions between stomach structure and diet choice in shorebirds. Auk 110: 552–564.
- 26. Ringelberg J (1999) The photobehaviour of Daphnia spp. as a model to explain diel vertical migration in zooplank- ton. Biol Rev 74: 397–423.
- 27. Ringelberg J, Van Gool E (2003) On the combined analysis of proximate and ultimate aspects in diel vertical migration (DVM) research. Hydrobiol 491: 85–90.
- 28. Roenneberg T, Merrow M (2005) Circadian clocks – the fall and rise of physiology. Nature rev 6: 965–971.
- 29. Bijma J, Hemleben C, Wellnitz K (1994) Lunar-influenced carbonate flux of the planktic foraminifer Globigerinoides sacculifer (Brady) from the central Red Sea. Deep-Sea Res I 41: 511–530.
- 30. Khripounoff A, Vangriesheim A, Crassous P (1998) Vertical and temporal variations of particle fluxes in the deep tropical atlantic. Deep-Sea Res I 45: 193–216.
- 31. Lončarić N, Brummer G-J, Kroon D (2005) Lunar cycles and seasonal variations in deposition fluxes of planktic foraminiferal shell carbonate to the deep South Atlantic (central Walvis Ridge). Deep-Sea Res I 52: 1178–1188.
- 32. Żmijewska MI, Bielecka L, Grabowska A (2000) Seasonal and diel changes in the vertical distribution in relation to the age structure of Microcalanus pygmaeus Sars and Ctenocalanus citer Bowman & Heron (Pseudocalanidae, Copepoda) from Croker Passage (Antarctic Peninsula). Oceanologia 42: 89–103.
- 33. Zwicker D, Lubemsky DK, ten Wolde PR (2010) Robust circadian clocks from coupled protein-modification and transcription-translation cycles. Proc Natl Acad Sci USA 107: 22540–22545.
- 34. Piersma T (2002) When a year takes 18 months: evidence for a strong circannual clock in a shorebird. Naturwiss 89: 278–279.
- 35. Mihalcescu I, Hsing W, Leibler S (2004) Resilient circadian oscillator revealed in individual cyanobacteria. Nature 430: 81–85.