Introduction

Nature is Full of Rhythmic Changes. The environment in which we live changes from day to day, season to season, and, perhaps, from year to year in a rhythmic manner that is more or less predictable. The behaviour and abundance of animals often reflect these rhythmic changes in the environment; in fact, cycles in nature are universal in occurrence and of profound importance.

Cyclic changes in the earth's environment are induced primarily by the sun, e.g., the short-term variations in light, temperature, and humidity in the solar daily cycle, or long-term flux in these variables in the longer-term annual solar cycle. These changes profoundly modify the lives of organisms, and we find that animal behavioural patterns are highly adapted to avoiding harmful parts of the environmental cycles and utilising beneficial parts.

Although we naturally think first of the sun as responsible for changes in the earth's environment, the moon also has important effects, and information collected about many organisms indicates that certain parts of their activities recur at regular periods related to parts of the lunar cycle. In this general review the reader is introduced to some of the wide variety of animal responses to lunar-related environmental cycles, mostly in the sea.

The Lunar Cycle

The lunar orbit around the earth takes 27.29 days, but because of the earth's rotation and the moon's movement relative to the earth, the moon rises 50.5 minutes later each day, i.e., the lunar day is 24.84 hours long. The moon, sun, and earth are in the same positions relative to each other every 29.5 days, the lunar or synodic month.

Movements of the sun and moon interact in such a way that at full moon the moon rises at dusk and sets at dawn, and at new moon the moon and sun rise and set together. These relationships produce a lunar monthly cycle in the duration of moonshine at night.

The light intensity of the moon also varies with the lunar phase: at full moon 1.83 × 10^-1 micro-watts/sq cm, at the quarters 2.12×10^-3, and at new moon (i.e., no moon, clear sky) 1.8×10^-4. Heavy cloud cover reduces these values about 10 times. The intensity of moonlight is about 1/500,000 that of sunlight (Moore, 1958).

Tides are caused by gravitational pull of the moon and sun on the earth, although because of the sun's great distance from the earth, despite its much greater mass, its gravitational effect is less than half that of the moon. The lunar cycle of tide-producing forces rotates around the earth each 24.84 hours, producing two pairs of tidal maxima and minima at any place each revolution. However, many factors affect the occurrence of tides, so that in some places there are no tides and in others up to four tides per day. Because the moon's orbit is at an angle to the earth's equatorial plane, there is an oscillating tidal asymmetry which varies 28° north and south of the equator. Since the lunar tide has a 24.84-hour cycle and the solar tide a 24-hour one, the effects of the moon and sun alternately amplify and oppose each other every 14.8 days, causing spring and neap tides (see Smith, 1968; Sverdrup, Johnson, and Fleming, 1942, for discussions of tidal phenomena).

Several secondary effects of the lunar cycle occur in the aquatic environment. Water pressure is a function of depth and so fluctuates with tides, varying with the daily and lunar-monthly tidal cycles. Similarly air exposure or water immersion varies on similar cycles. Brown (1962) hypothesised that forces like magnetism exhibit lunar cyclic changes, and he claimed that these forces are perceived in some fashion by organisms; but the evidence is extremely scanty.

The Occurrence of Lunar Rhythms in Animals

Environmental variables affected by the moon may have lunar-daily (24.84 hours), semilunar (14.8 days), or lunar-monthly (29.5 days) periodicity. The most obvious environmental variables affected by the moon are light and tide, and of these the tide appears to have the greatest impact. For this reason lunar rhythms seem to occur mostly in marine species. In fact, Allee et. al. (1949) suggested that lunar periodicity ‘is of relatively little consequence to terrestrial communities as far as our present knowledge is concerned’. Cloudsley-Thompson (1961) concurred: ‘With the exception of marine animals, very few organisms are known that show lunar rhythms of activity.’

Species in diverse phyla exhibit lunar rhythms, mostly in their reproductive behaviour. In many littoral animals increases in activity occur with the rising tide; tide pool fishes leave rock pools to forage more widely, sea anemones expand as they are immersed by the rising tide, etc. But such examples are of little interest here, unless it is found that rhythmic activity patterns persist in the unvarying environment of the laboratory. Then we have the problem of determining how the continued activity is timed and regulated — the problem of ‘biological clocks’. Somewhat more complex is the problem of how animals time their activity and physiology on a long-term basis to specific parts of the lunar cycle and how they perceive and respond to lunar stimuli. Some of the examples discussed below have defied explanation up to the present time.

Protozoa: Ray and Chakraverty (1934) reported that the ciliate Conchophthirius lamellidens, parasitic in a freshwater mussel, conjugates most freely after full moon. This example, although cited in subsequent literature, has not been studied further, and no causal mechanism is known.

Coelenterata: Moore (1958) noted that a species of Pocillipora breeds throughout the year, but with a lunar rhythm. In winter, breeding is related to full moon and in summer to new moon. There is no correlation with tides, since in the winter the lowest neap tides are related to full moon and in summer to new moon. Response of the animal appears to be to tidal amplitude, water pressure, or air exposure. The rhythm is an irregular lunar-monthly one.

A sea anemone, Actinia equina, has a tidal rhythm in its cycle of expansion and contraction; although this would seem to be a simple case of direct response to tidal stimuli, the rhythm persisted in the laboratory for several days (Cloudsley-Thompson, 1961, after Pieron, 1958).

Platyhelminthes: Gamble and Keeble (1903) found that the intertidal flatworm Convoluta roscoffensis rises to the surface when the tide is low and retreats when the tide rises. The movements do not occur at night. This apparently simple tidal rhythm persists in the laboratory and so has some internal timing mechanism.

Annelida: Many examples of varying complexity are known in which annelids, all marine polychaetes, respond to lunar influences. Their periodicities and causation are variable and often poorly understood.

Odontosyllis enopla, the Bermuda fireworm, breeds about 55 minutes after sunset, for about half an hour, only on nights when there is no moon in the sky during the early part of the night, i.e., from two or three days after full moon until new moon. On such nights the worms swim to the sea surface from their tubes on the sea bed, apparently induced to do so by the decrease in light at dusk. The lunar rhythm seems to be a response to decline in light intensity, from sunlight to night sky without moon. The gonads mature with a lunar-monthly cycle which is in phase with the lunar-monthly cycle of a night sky without moonlight at dusk (Huntsman, 1948). Other species of Odontosyllis have similar cyclic breeding behaviour; in all these species breeding is thought to be a response to the lunar cycle of illumination (Korringa, 1957). Korringa listed similar periodicities in other annelids, including species of Platynereis and Ceratocephale.

One of the best known annelids with lunar spawning rhythms is the Pacific palolo, Eunice viridis, which spawns in immense concentrations on Pacific island coral reefs. The reproductive part of the worm (epitoke) detaches from the benthic, tubicolous, vegetative part (atoke) and swims to the surface. Swarming is reported seven to nine days after the full moon, at low tide, when the coral reefs page 136 are awash, and usually at dawn, regardless of moon and cloud conditions. Since tidal amplitude is quite small, and since the new and full moon spring tides alternate in relative magnitude over a cycle of several years, it is thought that the causal stimulus is not a tidal one. Because moonlight is not essential for the swarming, the response does not seem to be directly related to changes in light, as occurs in Odontosyllis. It is believed that breeding is related to the lunar cycle in such a way that the gonads become ripe at the third quarter of the moon, perhaps in response to the lengthening dark period after full moon, as the moon rises later and later in the night with the approach of the time of new moon (Korringa, 1957). Eunice viridis differs from Odontosyllis in that gonad maturation has an annual cycle which is independent of the lunar cycle except that the gonads finally ripen apparently in response to changes in lunar illumination.

The Atlantic palolo, Eunice fucata, is reported to have similar breeding patterns to E. viridis (Clark, 1941).

Spirorbis borealis, another benthic, tubicolous polychaete, releases its larvae at the first and third quarters of the moon. But Knight-Jones (1951) showed that this is due to the occurrence of breeding at new and full moons. Thus breeding which occurs at the spring tides manifests itself by neap tide release of larvae. Environmental variables to which the worms respond are undescribed.

Lunar periodicity in the settlement of another tubicolous polychaete, Hydroides norvegica, has interesting economic implications. Dew and Wood (1955) found that it is better not to clean the hulls of ships near the spring tides, because this worm exhibits spring tide (semilunar) periodicity in larval settlement. Ships cleaned at other parts of the tidal cycle were found to remain unfouled longer than those cleaned during spring tides.

Other invertebrate phyla—Mollusca, Echinodermata, Arthropoda: Instances have been reported of lunar rhythms in all these groups, although most are poorly documented. A chiton, Chaetopleura apiculata, exhibits lunar periodicity in breeding activity (Grave, 1922); Mytilus edulis matures during the new moon period and spawns at the following neap tide; the oyster Ostrea edulis spawns at about the spring tides (Korringa, 1957) and so do Chlamys opercularis (Wilson, 1951) and Pecten maximus (Mason, 1958). Littorina neritoides lives at about high-water mark of high spring tides and releases its larvae when immersed by the sea. It thus exhibits semilunar breeding rhythm (Moore, 1958). Clarke (1965) recorded a Red Sea echinoid the gonads of which reached maximum volume at certain parts of the lunar tidal cycle. Lunar rhythms of various types have been claimed for the crab Uca pugnax by Brown (1962, and numerous papers in Biological Bulletin, Marine Biological Laboratory, Woods Hole, not listed here).

Naylor (1963) showed that the chilling of Carcinus maenas will lead to replacement of a circadian (solar-daily) rhythm by a tidal rhythm which ‘though not phased with external tides … indicates that the ability to show tidal rhythmicity is deep seated in British Carcinus’ (Williams and Naylor, 1967). Williams and Naylor found that chilling of crabs reared from eggs in the laboratory will cause ‘spontaneous’ establishment of a similar tidal rhythm, confirming that ‘tidal rhythmicity is deep seated’ and suggesting that ‘it may be inherited’. They postulated that in nature, phasing of activity is corrected to harmonise with tides by tidal variables like hydrostatic pressure, temperature changes and periodic immersion.

One of the better documented, and therefore most informative, instances of lunar rhythm is that of Clunio marinus, a marine, tidal chironomid (Insecta, Chironomidae). The larval insects live among intertidal algae, and the adults emerge from their pupae when the water level is low. As the female is short lived and wingless, emergence at low tide facilitates the males reaching the females to copulate, and allows the females to deposit their eggs among shore algae without interference from the water. In Heligoland it was found that adult emergence occurs with semilunar periodicity, at the spring low tide. But populations in the Black Sea were found to emerge and copulate when offshore winds drive the water level down and expose the algae. The chironomids would seem to be responding directly to ‘tidal’ fluctuations, but it was found that animals in the laboratory maintained their rhythmic emergence in synchrony with their parental populations (Korringa, 1957, after Caspers, 1951).

Vertebrata Pisces: Perhaps the best known example of lunar periodicity is that of the Californian grunion, Leuresthes tenuis. This small, silvery, atherinid fish leaves the sea to deposit its eggs in the sand of several Californian beaches. Spawning lasts from late February until early September, but takes place only on three or four nights after full or new moon. It commences from one to three hours after the peak of high spring tide and lasts for about an hour and a half. The female swims on to the beach with an ingoing wave, usually accompanied by several males. She digs her tail into the wet sand and releases the eggs about 2 in. beneath the sand surface. Accumulation of sand on the beach as the tide falls leaves the eggs buried by 8 to 10 in. of wet sand. The eggs are washed out of the sand again at the next spring tide cycle and hatch rapidly under the stimulus of wave agitation.

This cycle is highly adapted to the tidal cycle. Since spawning follows the highest of the spring tide series, the spawning bed is not again disturbed by the tides until the next spring tide cycle, about 12 days later. At this time development is complete and the larvae are ready to hatch. In the meantime the eggs have a safe refuge buried in moist sand. On the particular tide at which spawning occurs the page 138 eggs are deposited as the tide is falling, so that they are unlikely to be washed out by later waves on that tide; in fact the beach is being built up over the eggs as the tide recedes (Walker, 1952).

Korringa (1957) could not ascertain environmental variables to which the grunion responded to time its spawning movements. He rejected light intensity, since spawning is semilunar, and he claimed that it is not related to setting of the sun. He suggested that the tidal rhythm may be the causal factor, but posed the question ‘How do the animals know which is the maximum tide of the spring series, and when the particular tide has passed its peak?’ Walker (1952) has shown that there is a semilunar cycle in ovarian maturation, but the basis for the cycle and the precise spawning response remain undetermined.

Hubbsiella sardina is another atherinid species which spawns in similar fashion on beaches of the Gulf of California. It differs from L. tenuis by spawning during the day as well as at night (Rechnitzer, 1952). Hypomesus pretiosus (family Osmeridae), the Pacific surf smelt, is not closely related to the above fishes, but it was found to have a spawning peak at high tide, the higher the tide the higher the spawning peak. This appears to be a simple response to water volume on the beach. Like those of the above atherinids, the eggs are deposited in beach sand (Loosanoff, 1937).

Although the spawning of Galaxias maculatus is less widely publicised than the grunion's it is equally fascinating, and even more difficult to understand. Larval G. maculatus live in the sea, but juveniles migrate into fresh water, and juvenile and adult life is spent in lowland streams and rivers, usually beyond tidal influence, often many miles upstream from the sea. It is one of very few well authenticated examples of lunar rhythms in non-marine species (although actual spawning is estuarine and tidally controlled). The gonads mature mostly in summer and breeding occurs mostly in early autumn. Burnet (1965) found that the mature fish migrate downstream to spawning grounds at full moon, but other workers (see Hefford, 1932) have found fishes spawning at both full and new moon spring tides. Ripe adults move up on to the grassy flats of river estuaries in immense shoals, penetrating to areas covered by water only at the spring tides. The eggs are deposited among grasses and are eventually washed down among the bases of grass clumps, where air humidity remains high and temperatures are stable and low. The eggs develop and usually hatch at the next spring tides that cover the grasses, although if low temperatures retard development, hatching will be delayed until later tides (McDowall, 1968).

This spawning pattern exhibits some of the adaptive features of the grunion. The most notable difference is that lunar-related spawning migration begins when the fish are far, perhaps many miles, from the page 139 tidal environment. The manner in which the fish are able to perceive and respond to the lunar cycle remains a complete mystery.

Deelder (1954) discussed migration in the ‘silver eel’ of Anguilla anguilla and concluded that it is influenced by the moon ‘to a high degree’. He found that in the Upper Rhine Valley the peak of migration was prior to the last quarter and comprised mostly females. In the Baltic and in Dutch waters migration was after the last quarter and involved mostly males. Bertin (1956) reported that females live mostly in inland waters and males in coastal waters, so that this apparently sexual difference in migratory pattern may in some way function to co-ordinate the sexes in their seaward migration. Deelder concluded that lunar influence is not exerted by light occurring at migration, as the eels exhibit their lunar rhythm regardless of moonlight conditions, and heavy migrations are known on stormy nights. He suggested that the lunar light cycle may establish an endogenous rhythm so that migration occurs at the appropriate time, regardless of night sky conditions prevailing on the night of migration. Lowe (1952), on the other hand, found that eels will migrate at full moon only when water is turbid, but that lunar influence on the silver eel migration is disrupted by cloud. Lowe suggested that light quantity may be critical. It seems that eel migration does exhibit a lunar rhythm, but present information is confusing and at times contradictory.

Gibson (1965, 1967) described a tidal activity rhythm in Blennius pholis, an intertidal, rock-pool species, peak activity occurring about the time of high tide (lunar-daily rhythm). This rhythm was found to persist in the laboratory, disappearing over a period of several weeks. Gibson suggested that this rhythm may be connected with feeding migrations outside tidal pools in which the fishes were found at low tide. Gibson (1965) also found a very short-lived rhythm in another species, Acanthocottus bubalis.

Savage and Hodgson (1934) detected a ‘definite rhythm’ in the quantity of herring caught off East Anglia, a monthly rhythm which reached its peak at about full moon. More recently Blaxter and Holliday (1963) have suggested that this periodicity, and those in other clupeid fisheries, may reflect behaviour of either fish or fishermen, and they have found periodicity less evident in recent years; this they attributed to improvement in fishing methods.

Lunar rhythms in other commercial fisheries have been reported, e.g., hake and bream (Moore, 1958). L. J. Paul (pers. comm.) has found that commercial trawl fishermen in north-eastern New Zealand believe that at certain phases of the moon, fish are more abundant, or, at least, are more easily caught, and so they fish more intensively at these phases. Shark long-line fishermen work during the daytime and during nights with sufficient moonlight to enable them to locate their fishing ground and work their gear. Long-lines are fished in page 140 the harbours only during slack water at the turn of the tide. Fishermen using set nets for flounders avoid fishing at spring tides because tidal rips make work hazardous and difficult. Here, we have cases of lunar periodicity in man's activities, which, without careful analysis, could be interpreted as lunar periodicity in fish behaviour. As Blaxter and Holliday (1963) showed, it is difficult to establish whether lunar rhythms in commercial fisheries are real expressions of fish activity rhythms.

Vertebrata Reptilia: Carr (1967), in his search for understanding of the breeding migrations of the Atlantic ridley turtle, Lepidochelys kempi, on to beaches of the Gulf of Mexico, suggested that they may be related to the full moon. This seemed in part the product of local Mexican folklore and in part educated conjecture by Carr.

Causal Factors

It has been found that animals can synchronise their activities to a great diversity of natural, geophysical rhythms, e.g., diurnal (24 hours), tidal (12.4 hours), semilunar (14.8 days), lunar (29.5 days), annual, seasonal, or photoperiodic (365 days), etc. The environmental variables that in complex and various ways stimulate and control the activities of animals are poorly understood and subject to much conjecture. Opinions about how these activities are timed and maintained vary widely and really are not understood. In some instances behavioural patterns depend on receipt of direct stimuli from the environment, in others the repeated pattern of environmental stimuli results in short-term imprinting of a behavioural sequence in the animal. This may persist temporarily in the laboratory; sometimes it is found to persist indefinitely, because the repeated pattern of stimuli in the past has led to a recurrent and persistent pattern of behaviour in the species. Thus some animals exhibit rhythmic behaviour dependent on environmental variables, while others are to some extent independent, but display cyclic behavioural patterns which nevertheless correlate with cyclic fluctuation in environmental variables. Sweeney and Hastings (1962) used these differences to define exogenous (dependent) and endogenous (independent) rhythms; Brown (1962) called the latter ‘biological clocks’.

Endogenous rhythms apparently arose by long-term adaptation and synchronisation to persistent cycles in the physical environment, the cycles induced almost exclusively by the sun and moon. Brown (1962) considered that since much of the evolution of life occurred in the intertidal zone, over hundreds of millions of years the ancestors of many organisms may have been subjected to the rhythmic fluctuations of the tide. He suggested that this may explain the origin of animal rhythms of tidal nature. With some animals this may be so, but it is clear that this generalisation is not broadly applicable. Some animals are found to respond to lunar rather than tidal stimuli. page 141 Although numerous polychaetes exhibit lunar rhythms, the rhythms are very variable in character; apart from polychaetes, lunar rhythms occur haphazardly in the various phyla. The ‘tidal’ periodicity of Clunio marinus discussed above has clearly not resulted from exposure to millions of years of tidal fluctuations, since one population has an endogenous rhythm correlating with a wind-induced ‘tidal’ rhythm. This seems to be a product of relatively short-term adaptation to prevailing ‘tidal’ conditions.

The experiments of Williams and Naylor (1967), demonstrating spontaneous establishment of rhythms in laboratory-reared animals, show that tidal rhythms in activity may be genetically fixed and are not necessarily experimentally induced: but at present we cannot generalise from Williams and Naylor's results. Brown (1962) suggested that timing of ‘biological clocks’ does not depend on obvious environmental stimuli like light and temperature; he stated this because there is no obvious response to these stimuli in some species and because the animal cycle sometimes continues if environmental variables are artificially modified. Brown has resorted to subtle geophysical forces to explain rhythm, claiming, though not demonstrating, that these forces correlate with other, more obvious environmental variables; e.g., he claims to have found responses to a magnetic field in the planarian Dugesia and also in Ilyanassa (Gastropoda), Drosophila (Diptera), and Paramecium (Protozoa). Careful analysis of Brown's work on the crab Uca pugnax shows that some of the lunar responses he has reported are mutually exclusive; e.g., in the sea Uca was reported to have tidal rhythm out of phase with the local lunar rhythm; in the laboratory at Woods Hole it was found to transfer to a strictly lunar rhythm, but when transferred 3000 miles west to California it maintained a lunar rhythm in synchrony with the Woods Hole lunar rhythm, not responding to the Californian lunar rhythm.

Korringa (1957), in contrast with Brown's view believed that we do not need to utilise factors like periodic changes in the moon's declination, gravitational influences, air ionisation, and other subtle geophysical forces. Korringa maintained that moonlight and tide are most probably of primary importance and is able to uphold this view in a number of instances. But others do exist where there is no apparent relationship between rhythms in animal activity and obvious environmental variables. The types of influences that Brown has investigated may need continued recognition and study.

Korringa believed that periodicity is called forth by the sequence of neap and spring tides in those localities where tidal amplitude is large, but that where tides are small, other factors can be held to be the cause of the animal's cycle, e.g. in those animals sensitive to light, the alternation of dark and moonlit nights appears to correlate sometimes with sexual maturation. Close relationship to the page 142 environment is demonstrated very nicely in the life cycle of Dictyota, which, although not an animal, but a marine alga, is an informative example. Hoyt (1927) found that it releases its gametes on a semilunar cycle at the spring tides in Europe, where there are large and regular semilunar tides; but in North Carolina, where one of the monthly spring tides is much greater than the other, gametes are released only once each month. And in Jamaica, where tides are irregular, there is no lunar periodicity in gamete release. Obviously the algae are responding rather directly to existing tidal conditions and are exhibiting profound adaptability.

However, Korringa's view does not explain the spontaneous establishment of lunar rhythms in laboratory-reared crabs (Williams and Naylor, 1967) or the persistence of endogenous rhythms in animals held under stable laboratory conditions.

Selective Advantage of Lunar Rhythms

Adaptations to cyclic variations in conditions of the environment are of obvious value to animals, enabling prediction of imminent events in their habitat, of either catastrophic or beneficial character, and allowing compensatory behavioural changes to be made. This is particularly true of endogenous rhythms, with their great predictive value in producing behaviour harmonic with environmental rhythms. The very existence of endogenous rhythms in many diverse animals shows that there are ecological and evolutionary advantages. Some of these advantages can be observed readily; others can be deduced, but of many we are probably largely unaware.

Convoluta receives an obvious and simple benefit by emerging from the sand at low tide and retreating again as the tide rises. Exposure allows the receipt of sunlight for phytosynthesis (Convoluta carries symbiotic green algae), and burrowing into the sand prevents the worms from being washed away by wave action. Copulation and oviposition among intertidal algae by the chironomid Clunio are obviously facilitated by occurring at low tides. Foraging by Blennius at high hides is more profitable, because the foraging area accessible to the fishes becomes larger. Leuresthes tenuis and Galaxias maculatus share the advantage of being able to deposit their eggs in the supratidal environment, where they are protected from egg predators.

In addition to such advantages in specific cases, there are generalised advantages in the occurrence of lunar rhythms. Most animals that react to the lunar cycle do so in a part of their reproductive cycle. Especially in species that broadcast their reproductive products into the sea in haphazard fashion, periodicity is of great value in concentrating spawning activity in both space and time. When spawning occurs at very low tides, especially in shallow water, there is a further advantage in having gametes concentrated in a reduced volume of water, so that chances of fertilisation are further increased.

At the basis of lunar cycles, as in all cycles in nature, there is the fundamental advantage of responding to sthmuli which correlate with recurring environmental conditions. Thus the animals respond to these stimuli with a background of ‘species experience’ that such stimuli are associated with near optimal conditions for the imminent activity; or the activity may increase survival. In simple terms, lunar periodicity in breeding tends to produce concentrated breeding activity in stable breeding conditions year after year.

Acknowledgments

I am grateful to Professor G. L. Clarke, Dr. P. M. Ralph, G. D. Waugh, and L. J. Paul for comments on this review.