Upstream and downstream migrations in relation
to the reproductive cycle and to environmental
factors in the amphipod, Gammarus zaddachi
by
H.G. Dennert A.L. Dennert P. Kant S. Pinkster &
J.H. Stock (coordinating editor)
Zoologisch Museum, University of Amsterdam and Laboratoire Ch. Maurice de la Faculté libre
des Sciences de Lille
Summary 8. Per night, the maximum number of animals caught
net (= 30 X 50 cm) was 1,500 downstreamersand 6,000 1. The real has been studied of per identity juvenile gamma- upstreamers of G. zaddachi, which may roughly correspond rids, that were found in the fresh-water tidal region of the with 30,000 downstreamers and 120,000 upstreamers per River Slack (France), 3 to 4 km inland of the mouth of its section of the river. No other in the At these given gammarid species estuary. the outset, juveniles were (on basis of river Slack shows migratory behavior. their morphology) considered to belong to Gammarus that salinus. 9. Tagging experiments showed, animals migrating
up-river cover distances of 40 to 60 m per night. Down- 2. Rearing demonstrated that the juveniles in question most river migrants cover at least 50 m and at 80 m developed into adult Gammarus zaddachi. This indicates per night. that the morphological criteria used in discriminating 10. Several indications make it that, between G. salinus and G. zaddachi fail in juvenile mate- plausible though for the G. zaddachi uses the water currents its migration, rial. Several morphological criteria are discussed and
It up-river and down-river movements are not merely acci- tested. appears that a 100% certain key character does dental transportation. Animals in the (physiological) phase not exist, meaning that it is not possible to identify with of down-river continue to do of the absolute certainty all individuals in mixed populations of moving so, irrespective
direction of the stream. The holds true for animals the two species. same in the of in direction. 3. At the the resemblance of phase moving up-river G.same time, juvenile zad-
of is in to G. with 11. The reproductive cycle G. zaddachi completed dachi adult salinus provides us a morphological
one The in in autumn these feature allowing rapid distinction between adult and year. young appear spring, adult, the first females in juvenile G. zaddachi in studies with large numbers of in- juveniles grow ovigerous appear
dividuals. December, the maximum production of eggs falls in the
early spring. All adults die after the reproduction period. 4. In studies of the reproduction and migration of G. zad- 12. The reproductive cycle is coupled with the migration dachi, several types of nets were developed, both for cycle. The females in down-river migration anadromous and catadromous migrants. These nets agreed participating
dependent on and salinity, successfully with one another in having the same catch capacity. can, temperature
produce an offspring. At temperatures above 7°5C this 5. The activity of G. zaddachi shows a diurnalperiodicity, of the happens only in the mixohaline parts river Slack. the activity peak falling at night. the 7°5 When temperature drops below C, egg production This behavior results in a down-river migration(“drift”), is also possible in the limnic reaches of the river. The which reaches its maximum 2 to 3 hours after sunset, in the estuarine In juveniles the summer region. this slowing down later in the night, until the drift virtually pass the distribution season an important portion of up-river stops at sunrise. No influence of the light intensity during
area of G. zaddachi is depopulated. Repopulation of the the night (moon phase, clouds) was found. limnic region by juveniles takes place in the fall, aided by 6. During the periods with H.W.S., the direction of the the reversed water currents at H.W.S., thus completing stream in the entire of the reverses study area. In the parts the cycle. river adjacent to the estuary, this stream reversal is ac-
companied by a rapid increase in salinity. In the more I. INTRODUCTION up-river reaches, no salinity changes occur, so that we find
there a fresh-water tidal region. Downstream movements, often called “organic drift”, 7. All round certain of year a part of the population are known for various benthic aquatic invertebrates, G. zaddachi living at a given place (the so-called standing especially insects (Dendy, 1944; Müller, 1954a; Ta- crop) migrates down-river. During the autumn equinoxal
spring-tides (and — though much less — during the spring naka, 1960; Elliot, 1967; McLay, 1968), but also for
takes in equinox) mass migration place up-river direction. gammarids (Waters, 1962a; Müller, 1963a, 1963b; This up-river migration takes place only at nights in which Minckley, 1964; Lehmann, 1967). Various explana- the stream reverses under the influence of the H.W.S. tions have been given for this phenomenon (ably sum- Practically exclusively juvenile animals participate in the marized such up-river migration. by Lehmann, 1967), as: 12 H. G. DENNERT ET AL.
(1) Accidental transport. This is particularly plausible stream migrations. Such catadromous movements have been demonstrated for hatching insect larvae, that are transported at various occasions: for
random the lower reaches Cole (“drifted”) more or less at to Gammarus bousfieldi & Minckley by Minckley,
of brook where the hatch. The for a or river, imagines 1964; G.pulex (Linnaeus) by Müller, 1966a, and
adult females then fly upstream for the ovoposition, by Hultin, 1968; for G.fossarum Koch by Lehmann,
thus completing the “Besiedlungskreislauf” (“settle- 1967.
In with G.zaddachi ment cycle”, Müller, 1954b). our experiments Sexton, a
whose entire relation of the and downstream For a gammarid, life-cycle is com- complex upstream back” with the with pleted in the water, the “fly possibility is ex- movements life-cycle (point 3) and
environmental factors cluded, of course. various (light intensity, salinity,
temperature) has been demonstrated. We have found (2) The drift rate corresponds with the reproductive indications that the drift no is merely a compensation rate. Waters (1961, 1964, 1965) showed this to be for of the reaches. Our overpopulation upstream re- limnic the case in various American invertebrates, in- sults contradict the supposition of accidental trans- limnaeus Smith. The drift cluding Gammarus is con- port. sidered in these cases as a compensatory mechanism It must be remarked that the four Gammarus preventing overpopulation. species used in the experiments by other authors, The drift correlated with the of de- (3) rate is stage cited above, are all exclusively limnic, whereas G. velopment and not with the population density (Mül- zaddachi is a strongly euryhaline species, ranging ler, 1966b). from practically fresh to almost marine waters (0.04 -
is cf. (4) The (downstream) drift compensated by up- 18.5°/oo Cl, Stock, Nijssen & Kant, 1966).
Fig. 1. The seaward part of the stream system of the river Slack. In this part of the river Gammarus zaddachi is found
(X = farm houses). MIGRATION CYCLE IN GAMMARUS 13
II. STUDY AREA AND ENVIRONMENT Table I. Current speed in cm/sec. in the River Slack at
5 1967. Station I on August
Our carried out from Positive figures indicate currents in seaward direction, investigations were May 1966 negative figures indicate inland currents. Tidal difference to August 1968 in the river Slack, a stream running
37 cm. practically from East to West in the northwestern
part of France (département Pas-de-Calais). It runs local time speed remarks chiefly through a calcareous region and debouches tide in 11.40 -10 high open sea into the Straits of Dover just south of the village of 12.15 0 highest water level at Station I Ambleteuse. Feeded by springs, but principally by 12.30 + 20 temp. 17.1° C. the neither the the rain, course of river, nor its water 17.6° C. 13.30 + 53 temp. composition have been subject to great human in- 14.30 +41 temp. 17.8° C. 18.1° C. 15.30 +43 temp. fluences. In a direct line, the distance between the 16.30 +40 temp. 18.5° C. of the and the 17.5 km. main spring river sea is In 17.30 +40 temp. 18.6° C. this 15 of riversystem species Amphipoda belonging 18.30 + 39 temp. 18.6° C., lowest water to the family Gammaridae have been found (Stock level at Station I
18.49 37 18.5° low tide in et al., 1966); of these, Gammarus zaddachi occupies + temp. C.,
open sea a stretch of the river of some 4200 m length, which 19.00 +40 fresh is practically in its most upstream portion
(chlorinity 0.04 - 0.10°/oo), mixohaline in the greater
part, to almost marine at its most downstream end the observation period. Admittedly, these measure- (chlorinity 18.5°/oo at high tide). ments are rather incidental, but they nevertheless The part of the river Slack and its estuary in clearly give a rough picture of the yearly temperature which G.zaddachi illustrated The occurs is in fig. 1. the convinced that cycle in area. However, we are tides the wide of can enter freely into estuarine part higher and lower temperatures are not infrequent in the bor- river, west of highroad N 40. The is estuary the when area, observed over a longer period. dered there by intertidal flats and salt marshes. The For further details concerning the topography of tides are rather considerable, the tidal differences the Slack and the distribution of the various species being 8 to 9 meters at equinoxal spring tides, and of gammarids, the reader is referred to the publica- 5 normal tides. The wide is meters at neap estuary tion of Stock et al., 1966. separated by a small sluice, in highroad N 40, from In the area occupied by G.zaddachi, four fixed the narrower river bed. This sluice is of an auto- sampling stations have been chosen (stations I to IV of matic type, consisting an excentrically suspended in fig. 1). In Station I, tidal influence is still strong, door that on the force of the river drain, and opens and reflected both and by a sharp strong rise in that closes the of the tides. on force incoming Since and reversal of the direction salinity at high tide, a the sluice's rather tidal performances are poor, in- of the current. fluence is perceptible quite a distance upstream of the but the sluice. In fact the entire stretch of the Slack in- In Station II, current reverses at high tide,
from the normal habited by G.zaddachi is influenced by the tides, only a slight rise in chlorinity (rising
either noticeable river value of 0.03 - through a rise in salinity (which is 0.06°/oo to 0.1°/oo) occurs at
far the confluence of the Slack and as upstream as springtides only.
the Ruisseau de Laronville at the place marked Station situated of the III is in a tributary Slack, "Station II" in fig. 1, where the chorinity at H H W S called Rivière de Bazinghen; here, the current still kind of fresh-water rises to 0.1%o), or through a of kind of fresh-water tidal reverses (one can speak a tides, upstream of Station II. During high tide in the wave), but the water stays fresh all the time. the normal flow of the river is obstructed, estuary, the confined At Stations I to III, rivers are to a the water rises, and — especially during spring tides — rather bed with about narrow steep banks, being the direction of the current (which is normally to- 5 to 6 m wide, and being 40 cm (in the dry season ward the that the water flows in- sea), reverses, so low the at tide) to 160 cm (at wet season at high land for a short period (cf. fig. 15). tide) deep. During the observation period, the surface temper-
is ature of the water in the river Slack varied between Station IV finally situated near the mouth of a
18.6° C (Aug. 5, 1967) and 3.5° C (Feb. 27, 1968). small brook, the Ruisseau de Warem, where this de-
Table bouches the Ruisseau XII summarizes temperature measurements in into a slightly larger rivulet, 14 H. G. DENNERT ET AL.
de affluent of the Table Laronville, which in turn is an II. Current speed in cm/sec. in the River Slack at
Station I on 11 August 1967. Slack. From the mouth of the Ruisseau de Warem to
the bed there see table I. Tidal difference of the river Slack, is a distance of 28 m Explanation 41 cm.
brook about wide only. At Station IV, the is 40 cm local time speed remarks and 10 tot 20 cm deep.
10.35 50 Most localities mentioned in this article can be + low tide in open sea,
16.2° C. found in fig. 1. More information gives the Carte temp. 16.4" 11.00 +49 temp. C. de France 1/50000, feuille XXI-3, Marquise, edited 11.30 +48 temp. 16.5° C. by the Institut Géographique National, Paris (édition 12.30 +49 temp. 16.7° C. 1963). 17.4° C. 13.30 +41 temp.
14.30 +31 temp. 17.7° C.
14.53 0 highest water level at Station I
III. METHODS 15.10 in -23 high tide open sea,
temp. 18.2° C.
15.30 -27 temp. 18.1° C. III. 1) Nets 15.55 -25 temp. 18.2° C.
17.03 0 Conical with _.— nets, a rectangular aperture of 30 X 17.08 + 10 -•- 50 in several used. in- cm variations were In every the stance, conical net could be fixed with poles or with desired and wire at every place at every depth
the At the end of the in river. net, a rapidly change- able fixed of bag was by a screw clip on a piece Table III. Current speed in cm/sec. in the River Slack at hard The plastic tubing. illustrated in 2 7 1967. type fig. Station I on September used for downstream animals. was drifting For up- Explanation see table I. the but stream moving animals, same net was used, instead of a simple bag at the end, a recurved part local time speed remarks of the constructed net was to prevent upstream mi- 8.45 +45 grants of washed back the stream 3). being by (fig. 8.55 +40 low tide in open sea this Since special type of net for upstream migrants 9.55 +44 did catch 10.55 42 not significantly more animals than the + 11.55 + 45 normal drift net in in placed opposite sense, most 12.45 +21 later experiments a combination of the two nets was 12.55 0
used, as illustrated in 4-7. In this combination figs. 13.00 - 8
both and downstream 13.03 -15 net, upstream migrants at any catched the 13.10 -21 given place were in same net. 13.20 -25 In Stations I III the blocked small to nets only a 13.25 -29
of the riverbed. to block the entire part Attempts 13.40 -28 bed by nets (see, for instance, Needham, 1928) al- 13.55 -24
in ways resulted in a failure, since the strong current 14.05 -25 high tide open sea 14.20 -33 torn the net, and since floating debris obstructed the 14.35 -33 meshes. 14.50 -24 In Station where the rivulet much less IV, was 15.05 -25
the wide, entire stream could be fished out. 15.20 -13
-11 The interchangeable bags at the end of the nets 15.25 15.30 Station I —.— highest water level at were lifted hours, other every hour, every two or at 15.37 0 intervals. It if fished regular was supposed that, we 15.45 + 16 the and lifted the nets at the always on same spot 15.50 +36 the same time interval, contents of the bag would 15.53 + 63 15.58 69 consist of a representative sample of the migrating + 16.03 +57 Of the of random population. contents every bag a 16.08 +59 was at once in alcohol 70°. sample preserved ethyl 16.13 +50 The data obtained the Stations at I to IV were com- 16.23 +57 pleted by dip-net catches at various places all along 16.35 +58 the 17.00 +49 stream system. MIGRATION CYCLE IN GAMMAHUS 15
Figs. 2-3. Nets used.
2, net used for catching downstream migrants; 3, net used
for catching upstream migrants. Later, a combination of
4-7. these nets was used, see figs.
Figs. 4-5. Combination net (to be used for simultaneous catches of upstream and downstream migrants). 4, simplified diagram; 5, type of net finally used (see also figs. 6 and 7). 16 H. G. DENNERT ET AL
6-7. Combination combination Figs. net (see figs. 4-5). 6, exposed on the bank; 7, station I, with the net
(bags lacking) on the northern bank of the river Slack. MIGRATION CYCLE IN GAMMARUS 17
III. 2) Physico-chemical determinations bent cephalic segment was measured, along the dor-
sal from the distal end of the the aid of side, rostrum to Chlorinity was determined in mg/I with the articulation with the first pedigerous segment. an E.E.L. electric equipment. Both surface and bot- These measurements could be made rapidly under tom chlorinities were determined. with the aid of a compound microscope an eye-piece Stream velocity was estimated at the surface, with micrometer. the aid of a floating object and a stopwatch. linear correlation As is shown in fig. 8, there is a Temperature readings have been made with the between the cephalic length and the body length. aid of a mercurial thermometer (scale in 0.1° C.).
The various size classes obtained by measurements
were finally plotted and the difference found be- III. 3) Further treatment of the material tween catches of various dates were tested statisti-
The net catches were preserved and later studied in cally with the t-test (according to Wijvekate, 1966). the Under the laboratory. a dissecting microscope
animals were divided into three juveniles groups: III. 5) Tagging (that remained unsexed), males, and females. Ac- It proved to be possible to place a paint mark on cording to a method developed by Kant (see § V), the dorsum of live gammarids, which remained vis- animals which the the in spines on posterior margin ible until the moult. This done with next tagging was of the 7th leg are longer than the accompagnying brand Fluorart a synthetic luminescent paint, (man- classified The adults setae, were as juveniles. were ufactured by Winsor & Newton Ltd., London). This to the absence of cal- sexed, according presence or of which paint is manufactured in 16 colours, five second and the ceoli on the antenna, to characteris- were currently used in our experiments, viz., white, tics of the medial palmar spine of the second gna- and blue. yellow, red, green, thopod (cf. Sexton, 1912; Spooner, 1947; Segerstrâle, Live gammarids are placed on a sheet of filter pa- of 1947). Usually 100 animals each sample were for per, where they are left to dry 10 to 20 seconds. measured and sexed.
Then, with a thin, obtuse probe, one or more dots
of colours to the paint in one or more are applied
III. 4) Measurements dorsal surface of the while taken metasome, care is
be to the not to touch the At after the It proved not to easy measure body length appendages. once tag-
of the animals collected, because of the comma- ging, the animals are placed back in a container with
the the handiness of the shape of the body. Instead, the length of hardly river water. According to ex-
and zaddachi Fig. 8. The relation between body length (plotted on the x-axis) cephalic length (y-axis) in Gammarus
The calculated = 0.14 — 0.13. (n=40). points found, group clearly along a straight regression line, as y x 18 H. G. DENNERT ET AL.
individuals that The distribution this perimentator, the percentage of suf- pattern resulting from proce- fers noticeably from this treatment varies between dure in the Slack is essentially that published later
0 and 10%. Finally, the tagged animals that appear by Stock, Nijssen & Kant, 1966, fig. 1. However, a
released a and curious resulted from initial in good condition are at given place very inconsistency these
the noted. the distance the mouth of the percentage of recaptures is So, investigations: estuary was inhabited, well of the and direction of the migration, as as part among others, by G.salinus; this species was replaced,
day during which the migration took place, could be towards the head of the estuary, by G.zaddachi; the
studied. latter also distance species penetrated a certain up-
river into the fresh-water tidal reaches of the Slack;
ACKNOWLEDGEMENTS AND still farther G.salinus showed IV. up-river, however, up
again, living at chlorinities as low as 0.04°/oo. Such RESPONSABILISES distribution was contradictory to the pattern found The research started and coordinated the was by elsewhere along the European coasts, where G.salinus senior author assisted in various of (J.H.S.), phases is found only in those parts of the estuaries with the the fieldwork B.Sc. students for their by working highest salinities (cf. Segerstrâle, 1947; Spooner, the and the characters degree. Thus, growth sequence, 1947; Kinne, 1952).
to identify the various growth stages, were studied
by Mr. P. Kant. The downstream drift at Station IV
in V. 2) Trickery — salinus by Drs. S. Pinkster. The distance of migration one
was night (studied through experimental tagging) It was thought worth while to investigate the fresh-
worked out Dr. J. H. Stock for the downstream by water occurrence of the alleged G.salinus more closely. and Mr. H. G. Dennert and Mrs. A. L. migration by In the next discussion, these fresh-water gammarids, for the The last-named Dennert upstream migration. having the key characters (according to Kinne, 1954)
two also took of the at Stations I to care year-cycle of G.salinus, but living up-river instead of down-river
III, sometimes assisted the others, but of the inhabited be called by mostly area by G.zaddachi, will and of of the alone, the preparation graphs published “trickery-salinus”, in order to distinguish them from
in this paper. both the real salinus and zaddachi. indebted Professor H. The authors are greatly to For the distinction of G.salinus and G.zaddachi,
of the Faculté Libre des Sciences de Lille, used the classical bv Boulangé, we chiefly method, developed of the fieldstation called "Laboratoire Ch. Director Sexton (1912, 1913, 1942), Spooner (1947), Seger-
Maurice" at Ambleteuse-sur-Mer Pas- (département as a strâle (1947), and used key character by Kinne who has the facilities de-Calais), France, at our put (1954), according to which salin us is characterized
entire winter and summer, and disposal, day night, by the presence of short setae (shorter than the base for the fieldwork. as a the of 4 of 7, spines) on posterior margin segment leg The fieldwork has been made possible by grants whereas in zaddachi the setae on this in place, par- of the of Amsterdam (to some of the B.Sc. University ticular those of the apical and subapical groups of and of the Foundation students) Beijerinck-Popping elements, are much longer than the spines. This dif- of the Netherlands' of Sciences, Am- salinus and zad- Royal Academy ference seems trifling, and indeed, sterdam (to the senior author). We want to express dachi considered forms of the were a long time two for this financial our form from gratitude support. same species, a "spiny" higher salinities,
and form from lower salinities a "hairy" (Sexton, RESULTS 1942). But since Spooner (1947) showed that both
forms were a barrier, V. GROWTH OF GAMMARUS ZADDACHI separated by reproductive Kinne (1954) rightly took the step of considering V. 1) of G.zaddachi and G.salinus Range them (sibling) species.
the of the dis- to the At start our investigations concerning Trickery-salinus belongs morphologically
characterized as it is a tribution of the various species of Gammaridae in the salinus-zaddachi-group, by
Slack and used Kinne's setiferous of its first antenna, and dis- river its estuary, we nearly peduncle by a classical key (1954) to identify the members of the tinctive armature of the lower margin of the third
Kinne's has the that segment of the mandible On the posterior mar- genus Gammarus. key advantage palp. characters taken of the 4th of 7 of both adult and juvenile are into gin segment leg trickery-salinus,
and that all characters the shorter than the which classifies consideration, nearly key are setae are spines, illustrated. it consequently with G.salinus. MIGRATION CYCLE IN GAMMARUS 19
is is than salinus. in It true that Kinne's key based primarily upon to However, the light of the great adult but the stated shown all clear decision males, in same publication it is overlap by characters, no that shorter is little could be reached to the taxonomie of trick- or longer setation age dependent, as status and that the salient setal length-spine length ratio ery-salinus. The only thing that emerged from these
could be used with animals smaller that success in even morphological comparisons was, Itrickery-salinus than had distinctive feature of 4 to 6 mm body length. not a single absolutely its
had from and that all its characters fell within the Since our trickery-salinus body lengths own, range
4 11 the be- of variation of the tot mm, no difficulty was expected in zaddachi-salinus-complex.
ginning. Certain indications pointed in the direction that
It first find the of G.zaddachi: was attempted to in nature ovigerous trickery-salinus was juvenile phase
the of the a number of details showed this specimens or couples in precopula stage morphological ten- -salinus. The and the distributional with trickery These attempts were all in vain. dency, range (overlapping
could have three from that of formed clue problem possible answers: trickery- zaddachi, disjunct salinus) a salinus are juveniles of as well.
In order to down all incertitudes, trickery- a) the species G.salinus put salinus was brought in culture in the laboratory and b) the species G.zaddachi followed. its morphological changes were c) another species at the moment unknown to us.
of Possibility b) was rejected in earlier stages our V. Rearing experiments contradiction Kinne's 4) study, as it was in to statement
G.zaddachi that the morphology of segment 4 in leg 7 was al- On 4 August 1966, 140 specimens of were
the of the Slack. The ready typically developed in 4 mm juveniles. caught in estuary chlorinity at
As described 1 fitted the of was to possibility c), no species ) moment collecting 11.0°/oo. the material of trickery-salinus. On 5 August 1966, 840 specimens of trickery-
salinus were caught in a tributary of the river Slack,
the Rivière de off the farms La Parthe V. 3) Morphological comparison of trickery-salinus Bazinghen, The the of with salinus and zaddachi and Otove. chlorinity at moment collecting
was 0.05°/00. -salin the Preserved samples of trickery us (from river All these animals 500 were transported by car over Slack), of salinus (from Germany, The Netherlands, km to Amsterdam, in insolated cooled jugs. In the and the river Slack), and of zaddachi (from Germany, surface laboratory they were placed in fingerbowls, and the The Netherlands, Wales, river Slack) were 2 Each 40 cm 5 0.8°/oo. , waterdepth cm, chlorinity examination. subjected to a scrutinous, morphological bowl contained 15 gammarids. The temperature in have been The following seven characteristics com- the culture room was buffered to all too sudden pared: (1) the setation of the lower margin of the changes, but followed otherwise the more periodic pedunculus of the first antenna; (2) the number of changes of the air. Food was given in accord- the of open segments in the accessory flagellum; (3) shape ance with Sexton's (1928) directions. the propodus of the two gnathopods; (4) the armature Although the initial mortality was high, the culture of leg 7; (5) the shape and armature of the 2nd and was continued until 3 March 1967. In this period, the 3rd epimeral plates; (6) the dorsal armature of the raised After chlorinity was two times: 110 days, and the armature of the telson lobes. urosome; (7) the starting on 23 Nov. 1966, chlorinity was gradual- It appeared that these characters in G.salinus and ly raised to 3.47°/oo-This value was attained on 2 Dec. showed considerable but that G.zaddachi a overlap, the Starting on 12 Dec., chlorinity was gradually the the statistical means usually were different. On raised to 6.8°/oo, a value attained on 30 Dec. decision could basis of characters (1), (2), and (5), no number of Every week, a animals, by preference be reached whether trickery -salinus showed more re- 15, from the culture were taken out and preserved semblance to zaddachi or to salinus. On the basis of alcohol. the the in So, morphological changes as time the remaining characters — (3), (4), (6) and (7) — passed by, could be followed. trickery -salinus showed closer relations to zaddachi
V. 5) Changes in morphology during the growth
Gammarus sarsi Reid, ') With the exception perhaps of the ------In 9 ratio seta/length - ■ - - - fig. length longest longest 1943 (= G. ochlos Reid, 1945) which is,- however, synony- spine is plotted. In the left column, the spines and zaddachi and Stock mous with G. (see S'egerstrâle, 1947, & setae of 4 of 7 in the Kant, 1966). segment leg are incorporated, 20 H. G. DENNERT ET AL.
The and zaddachi from Fig. 9. ratio length longest seta/length longest spine on segments 4 5 of leg 7 in Gammarus
different localities and reared in the laboratory. This ratio is currently used to discriminate taxonomically between Gam- zaddachi and marus Gammarus salinus. Explanation see text (§ V. 5). MIGRATION CYCLE IN GAMMARUS 21
mentioned under right column those of segment 5 of leg 7. Since seg- (3) Members of the population (2),
ment 4 has of elements 1 reared in the marked 13 1966 usually 4 groups (bunches) ), laboratory (rows Aug. the left column is subdivided into 4 vertical rows of to 3 March 1967) show a gradual shift in the seta/
block The vertical of block dia- and still diagrams. 3 rows spine ratio. On 5 Aug. setae spines are
in the column with the 3 but later in the in the of grams right correspond subequal, year majority of elements illustration of the groups on segment 5. (See the specimens (in particular the males), setae out-
7 at the of both left and The the October the males have attained leg top right column). grow spines. In
females indicated solid the zaddachi which was maintained are by (black) blocks, a typical habit,
males blocks in the the After October by open (white) diagrams. during rest of the experiment.
The the absciss indicate how times the of the the figures on many length growth setae continues, highest the longest seta is longer than the longest spine (thus ratio being reached in March (at the end of the
- that the - 4.0 4.4 means longest seta is 4.0 4.4 times experiment).
than the longer longest spine). The of the (4) same increase in length setae during On the of animals at ordinate, groups preserved is observed for the period the animals were in culture
different dates are indicated. The lowest row (marked segment 5 of leg 7 (right column in fig. 9), although the data the animals "Slack, estuary"), contains on distinct zaddachi-hke moqihology is reached in this
collected in nature in the on 4 estuary Aug. 1966, in in segment earlier the year than segment 4.
identified as true G.zaddachi; the next lowest hori-
the zontal row (marked "Slack, limnic") contains data V. 6) Conclusion of the rearing experiments the animals collected 5 on in nature up-river on Aug. The the setae on posterior margin of segment 4 in 1966, called provisionally trickery -salinus. 7 used discriminate be- leg (which are currently to In the top row (marked "North Sea Canal"), a tween G.zaddachi and G.salinus) are still short in number of animals from this Canal (situated in The in summer, but increase length during autumn and Netherlands, chlorinity in nature 1.4°/oo) has been in- winter. Since summer animals are sexually inactive cluded for comparison. (see § V-2) and since they are of smaller size than The remaining horizontal rows represent animals winter (see § XIV-2, XIV-3), it is the specimens justified from rearing stock, preserved at regular intervals
to conclude that to a certain size still of time. juveniles up have short setae (thus resemble morphologically the A vertical line (division line) in each block diagram adults of G.salinus, and hence called “trickery-sali- indicates the key value in the seta-spine ratio, the nus” in the previous treatment), but that their mor- key value used in the literature to discriminate G. in phology develops later the year to that of real zaddachi from G.salinus. This key value is 1.0, be- G.zaddachi. if the than the the cause spines are longer setae, ratio
is < 1.0, and the animal in question classifies with
If the than the V. 7) Correlation between absolute size and G.salinus. setae are longer spines,
> the animal classifies with G. morphology giving a ratio 1.0,
zaddachi. Kinne, 1952, states that females of G.zaddachi attain
conclusions about The following can be drawn from fig. 9: sexual maturity at a body length of 7 mm.
Since the and the in Material collected the with body length cephalic length are (1) in areas a higher linear another proportion to one (see § III-4), it can salinity (lowest row, marked "Slack Estuary, 4 Aug. be deduced from that of 7 for the of animals fig. 8 a body length mm 1986") consists greater part in
corresponds with a cephalic length of about 0.9 mm. which the setae are longer than the spines (this means The shows the that of animals the diagram (fig. 10) seta/spine ratio in the greater part are plotted to leg 7 for two size classes of Gammarus zaddachi, viz. right of the line separating typical zaddachi from
of 0.60-0.79 mm and of 0.80-0.99 mm typical salinus). This is especially true for the males; (cephalic In the left vertical the ratio of the females fall length). row, a certain percentage turns out to longest is for each of the three to the left of the division line. seta/longest spine given
of elements in the groups on segment 4 of leg 7, right Material collected in with low (2) areas a salinity vertical for 5 of 7. the smaller row segment leg In (row marked "Slack, limnic, 5 Aug. 1966") chiefly is the 4 rule size class, setae on segment are as a around the division which grouped closely line, means shorter than the spines, in the larger size class, about that setae and spines are about equal in length. equal numbers of animals have the spines longer
than the setae and the setae than the ') Elements = spines and setae together. longer spines. 22 H. G. DENNERT ET AL.
Fig. 10. The ratio length longest seta/length longest spine onsegments 4 (left column) and 5 (right column) of leg 7 in
size classes zaddachi. - - two of Gammarus Lower row: cephalic length 0.60 0.79 mm. Upper row: cephalic length 0.80
0.99 mm. Explanation see § V. 7.
The key value of 1 (meaning setae and spines of tually, sexual maturity is reached by females having the of the the same length) is reached, as is apparent from fig. 10, completed development setae on ooste-
in the class of — but this character lends size 0.80 0.99 mm cephalic length. gites (cf. Sexton, 1924), it-
with value lower than self less for Specimens a key 1 (= setae mass treatment. shorter than the have spines) been considered by us
This as juveniles. assumption corresponds very well V. The distinction of the two indeed with Kinne's observation that animals smaller 8) sibling species,
in than 7 total G. zaddachi and G. sal us mm length (= 0.9 mm cephalic length)
have are immature. We seen in a preceding paragraph (see § V-6)
Of the that of G.zaddachi resemble the course, lengths (total length or cephalic juveniles in setation
rather length) are imperfect expedients, but con- of their seventh leg (the character usually employed venient in handling large numbers of animals. Ac- for the distinction of zaddachi and salinus) the adult MIGRATION CYCLE IN GAMMARUS 23 stage of G.salinus. The question then arises, whether has been calculated, in which W is the maximal width there could be found an observable of the second of 7, and S is the mean easily morpho- segment leg m
character that is workable also for the of the three This ratio is logical younger length longest setae. plotted stages. in fig. 11, right vertical column. The advantage of
this method is that the intersetal interval (which is a) It was shown (figs. 9, 10) that especially segment is used. The of 7 less suitable salient age-dependent) no longer disadvantage 4 leg was as a character.
is that the ratio W/S cannot be estimated but must From these figures, it emerges that the ratio longest m
be based on measurements with an micro- useful 5 of eye-piece seta/longest spine is more in segment meter. leg 7. We tested 430 specimens of G.zaddachi, with Also in the W/S ratio, some between from 0.4 for this lu overlap cephalic lengths ranging to 2.0 mm,
zaddachi and salinus is found, but the means for character. Only 24 out of these (= 5.5%) showed each mixed species are more widely separated. In setae shorter than the spines. These 24 exceptions individuals of both with a W/S the smallest their populations species, m were not necessarily animals, ce- ratio > 10 belong to G.salinus, those having this phalic lengths ranged from 0.4 to 1.3 mm. No ex- ratio < G.zaddachi. for this character found 52 6.5 belong to ceptions were among
we that it is not of Concludingly, may say always easy specimens G.salinus, cephalic lengths 0.6— 1.6 mm. to distinguish between G.salinus and G.zaddachi. Al- Both the zaddachi and the salinus material was col- though the adults (body length > 7 mm) are well- lected at a number of different localities in The characterized (by the armature of 4, but Netherlands, Denmark, Germany, and France. segment of of of particularly segment 5, leg 7, younger stages b) Spooner, 1947, mentions that in G.zaddachi the G.zaddachi difficult to from G.sali- are more separate posterior margin of segment 2 of leg 7 is armed with nus. The length of the setules on the posterior margin setae — though variable in — of which at least length of 2 of 7 relation the intersetal segment leg (in to in- 50% is clearly longer than the intervals between suc- terval or in relation to the width of the segment)
these — cessive setae. In salinus, setae are according forms in most cases a useful additional clue. to Spooner — shorter than the intervals separating adjacent setae.
have tested this feature VI. DAY - NIGHT PERIODICITY We on its possible taxo-
value. The left vertical nomie row in fig. 11 shows, VI. in 1) Introduction almost complete analogy with Spooner, the quo-
s Several authors (Harker, 1964; Müller, 1983a, 1963b; A = ÜL. In this S stands for the tient, quotient, U1 have Dm Waters, 1962a, 1965) emphasized a day-night
in several limnic mean of the 3 setae, D for the mean of 3 in- rhythm invertebrates, including re- longest m
of the Gammaridae. In it tersetal intervals chosen at random. presentatives our study, was
Per class of the block from the that the size 0.1 mm cephalic length, apparent beginning only during numbers of taken and diagrams for G.zaddachi <3 and 5, and for G.salinus night large gammarids were
and but At the never Of the 15 of Gam- (3 $, are usually separate adjacent. during daytime. species
size maridae known to occur in the river Slack and its top of the left row, the totals for all classes and both here also observe different Gammarus is found in sexes are plotted; we estuary, only one, zaddachi, statistical for zaddachi and and numbers in the stream nets. means salinus, only large Echinogammarus
berilloni is found in the downstream a narrow zone of overlap. Complete separation of the (Catta) regularly
for zaddachi and salinus found the nets, but never in numbers most 3% of the curves only is in great (at
total and never in the It is al- cephalic length classes larger than 1.30 mm. catch) upstream nets. that animals involved Although this character gives satisfactory results most certain then, feeding are in is here. these were carried several cases, and although it easily observable, Presumeably accidentally
the stream and into the net. by estimation and without taking precise measure- along by got
suffers from draw-back. The number of ments, it a
in VI. 2) Results setae increases with increasing age. This increase number is achieved subdivision of the in- by existing Like Müller, 1966b, and Jansson & Källander, 1968, tervals between the setae, in smaller and resulting showed for other species, the greatest activity for smaller intervals. G.zaddachi also is at night. Figs. 12-14 give the 24- W hours cycles, as shown in the net catches for down- = B -— c) To cope with this difficulty, the ratio, stream migrants, for 3 different dates and stations. m 24 H. G. DENNERT ET AL MIGRATION CYCLE IN GAMMARUS 25
The observations stations but at other and dates cor- place at night, however, not during every night,
with the data the around the respond closely presented in figure. only during a narrow period spring-tides.
The following regularities are clear at once:
VII. INFLUENCE OF THE SPRING-TIDE (1) Downstream migration always takes place in the ON THE MIGRATORY ACTIVITY dark; during daytime, migration activity is vir- tually absent. VII. 1) Stream
(2) One to two hours after sunset, the migration As shown in fig. 16, daily observations have proved starts all the sudden and reaches its maximum marked difference between the catches a during two to three hours after sunset. nights with spring-tides and other nights. As de- the of the the (3) During rest night, con- scribed and migration above, in § II ("Study area environment"),
but the number of animals in tinues, participating the to from the Stations I IV are separated open it At sunrise, the activity stops sluice. gradually drops. estuary by an imperfectly operating automatic
entirely. The tidal difference, which is 8 to 9 meters at spring-
ride in the off the open sea estuary, may cause a VI. 3) Discussion rise of the river level of 0.30 to 1.0 m at Station I.
In with these would This with reversal of the context observations, one expect rise is usually coupled a
that there would relation between the direction of the Under "normal" exist a noc- current (fig. 15).
turnal of the the activity animals, as expressed in conditions (i.e., no excessive rainfall, no H.W.S.) the
number of individuals unit of and river flow at amounts to to caught per time, Station I 40 50 cm/sec., the light intensity. Müller (1966a) and Jansson & in seaward direction, of course. Fig. 15 shows, that Källander have the hours before (1968) proved, in laboratory experi- stream velocity drops some two ments, that a dropping light intensity increases the H.W.S. One hour before H.W.S., the water stands
in activity G.pulex (Linnaeus) and G.oceanicus Seger- still, and from then on, till approximately an hour
strâle. Also Waters drew the conclusions after the direction of the (1962a) same H.W.S., current reverses
nocturnal from his field observations on whereas — after short second stand- on activity 180°, finally a
G. limnaeus Smith. still — the flow normal period river re-assumes its
While with these authors that G.zad- direction and Tables show the reversal we agree in speed. I to III dachi shows a marked day-night rhythm, with a noc- at the stream at various dates in Station I. turnal could activity period, we not clearly prove a In Stations II to IV, the stream reverses as well, correlation between the activity on dark nights and but the phenomenon is less marked, due to the on clear nights. As is shown in table IV and fig. 16, greater distance from the sea of these stations. there correlation between the Table and demonstrate that is no logic brightness IV fig. 16 upstream of the and the number of in the of is limited to night gammarids migration any importance nights catches. The after shown with which the activity peak just sunset is spring-tides during current reverses. also other such and 17 and 18 demonstrate that by organisms, as Ephemeroptera Figs. upstream migra-
tion is Chironomidae (see McLay, 1968, who postulates a limited moreover to the very hours during the mechanism for this activity pattern). night that the current is reversed. During the short
There is a clear correlation, however, with the period of current reversal (2 to 3 hours), the down-
and the whereas migratory activity spring-tides, irrespective stream migration practically comes to a stop, of the fact whether full the numbers of animals the a new moon or a moon spring- participating in upstream much than those tide is involved. migration are larger migrating down-
the facts the later this Anticipating on exposed in sequel, it stream. In a paragraph (§ IX) phenomenon might be remarked that upstream migration also takes will be worked out.
Fig. 11. Graphic representationof the lengthof the setules at the posterior margin of segment 2 of leg 7 in
zaddachi. = in Gammarus In the left column this length is expressed as A Sm/Dm (see § V. 8. b), the right
column as B = W/Sm (see § V. 8. c). Each horizontal row represents a size class (expressed in cephalic length).
At the the values for each column have been added top, up. 26 H. G. DENNERT ET AL
Table IV. Numbers of animals caught per net at night (Station I, August 1968) and environmental factors during
the catch.
numbers caught
of sunset tide low tide moonrise moonset remarks night down- high up-
stream stream
2 20.26h 2/3 ) 3 160 6.02h 0.39h 13.04h 23.38h d.n./r. n.c.r.
4/5 2) 0 200 20.23h 19.52h 3.14h 17.52h 0.02h d.n./r. n.c.r.
210 20.20h 22.32h 4.38h 19.50h 6/7 9 3.28h d.n./r. n.c.r.
l.n. n.c.r. 8/9 2 230 20.17h 0.30h 19.13h 20.50h 6.19h much plant debris
10/11 ') 1600 1130 20.14h 1.50h 20.42h 21.25h 8.56h l.n. c.r.
12/13 1) 950 20.08h 2.57h 21.52h 21.52h 11.20h v.l.n.
3.29h 22.55h 13/14 0 360 20.06h 22.06h 12.31h l.n. n.c.r.
15/16 0 290 20.05h 4.43h 23.29h 22.40h 14.52h d.n. n.c.r.
2 6.04h 23.37h 17.04h v.l.n. 17/18 ) 3 530 20.01h 1.16h n.c.r.
2 20.48h 4.10h 1.14h 18.41h l.n. 19/20 ) 3 100 19.57h n.c.r.
21/22 2 95 19.52h 22.55h 6.07h 3.34h d.n. n.c.r. 6.09h l.n. 23/24 7 135 19.50h 0.21h 19.07h n.c.r.
25/26 1) 3500 300 19.46h 1.28h 20.19h 20.31h d.n. c.r.
26/27 l ) 4000 400 19.44h 2.00h 20.54h 20.46h l.n. c.r.
announced in the tide tables. The hours of high and low tide are based on the open sea data, as governmental
Numbers of animals caught (if greater than 100) have been rounded off to nearest multiple of 5.
— = dark l.n. = v.l.n. = = current reversal; Abbreviations used: d.n. night; light night; very light night; r. rain; n.c.r. no
2 = level. c.r. = current reversal; —= no data available; ') = highest water level; ) lowest water
the reversal of the direction the first Figs. 17 and 18 also show that upstream migration it is current in
limited but that the that for the is not to Station I it occurs at place is responsible upstream migratoiy other Stations well. It is be that the as movements. to clearly understood,
is in "upstream migration" meant a geographic sense,
and in the that the animals in not sense moving up- VII. 2) Salinity stream direction also move against the direction of
As shown in the previous paragraph, estuarine waters the normal river flow. On the contrary, they use the the the short of reversal for their penetrate in lower reaches of river during periods stream migratory spring-tides. This fact is of the greatest importance activities.
results sudden of the environ- since it in a change (2) Downstream migration is present every night, pro- ment. vided the river flow is in seaward direction, irrespec- the shows and At our Station I, salinity a strong of tides. tive moon phase and/or sudden rise at H.W.S. (cf. tables V-VII). At Station II, The that G.zaddachi makes of the phenomenon use salinity changes occur, but are barely perceivable, direction of the water currents, both for its up-river Stations III and IV have current reversal and level and for its down-river migrations, is not unique at changes correlated with the H.W.S., but do not show all in the animal kingdom. Several cases in which the salinity changes (they remain fresh). The decreasing currents are used for horizontal migrations have re- influence of the the from H.W.S. on salinity West cently been summarized by Verwey (1966). (=the sea) to East (= the river) is shown in table VIII.
VIII. THE ANNUAL MIGRATORY ACTIVITY
VII. 3) Conclusions CYCLE
shows the number of (1) Upstream migrations are correlated at all Stations Fig. 19 specimens caught in
the the and downstream with the H.W.S. Since H.W.S. causes at all upstream (hatched) (black) nets
Stations Station I in the of 13 1967 to 29 Stations a current reversal, but only at two at period July April
and in safe that 1968. (I II) a rise salinity, it is to assume MIGRATION CYCLE IN GAMMARUS 27
Not included in fig. 19 is the period of 12 May
to 13 July 1967, although sampling was done in this
The the period. reason for leaving out data on this
that the period is, sampling was performed not on
veiy spot of Station I but at two other places, some 300 and 800 of These m upstream Station I. two abandoned sampling places were later, since they
situated the were in parts of river attracting too
curious disturbance of many spectators, provoking the the of the experiments. At any rate, picture
months of May through July is thus, that a moderate
downstream migration occurred during these months
2000 individuals net and in (up to per per night
May, and up to 350 in June and early July); no up- occurred stream migration of any importance during
this period (maximum number of upstream migrants
observed 40 per net and per night).
From fig. 19 it is clear that downstream migration, Fig. 12. Numbers of downstream drifters in one night at
in though fluctuating numbers, occurs throughout Station IV. The arrows indicate the time of sunset (arrow
sunrise on the is down) and (arrow up). the year. Upstream migrating contrary con- pointing pointing centrated in the period of September through No-
vember, that is during the equinoxal autumn spring-
tides. The April equinoxal tides do show a slight rise
in the but the numbers As shown the that upstream migratory activity, it was in previous paragraph, involved low with the are so as compared autumn migratory activity was chiefly concentrated in nights
that be with all done around the migration, they can neglected. H.W.S., sampling was
13-14. of the total 24-hours of at Figs. Percentages catch per hour downstream drifters Station I. Meaning of the ar- rows as in fig. 12. 13, on 27 July 1967; 14, on 2 Aug. 1967. 28 H. G. DENNEHT ET AL
measurements at Table V. Chlorinity Station I on 5/6
September 1967.
local time chlorinity in °/00 remarks
19.45 0.046
20.45 0.050
21.45 0.048
22.45 0.058
23.45 0.083
0.45 11.2 high tide in open sea
1.45 16.9 at 1.07 h.
10.45 0.049
Table VI. Chlorinity measurements at Station I on 7/8
September 1967.
local time chlorinity in °/oo remarks
20.45 0.049
21.45 0.050
22.45 0.046
23.45 0.076
0.45 0.044
1.45 0.058
14.8 tide 2.45 high in open sea
3.45 17.8 at 2.21 h.
4.45 5.80
5.45 1.20
Table VII. Chlorinity measurements at Station I on 19/20
September 1967. Fig. 15. Stream velocity (broken line) and chlorinity (solid
line) during H.W.S. at Station I (7 Sep. 1967). local time ehlorinity in °/oo remarks
22.00 0.057
7.80 0.30 high tide in open sea
1.30 16.1 at 1.00 h. with the of the data of springtides, exception August 2.30 13.1
1967 that were collected at narrower intervals. 3.30 0.089
4.30 0.071
5.30 0.049 IX. NUMBER OF MIGRANTS PER NIGHT
IX. 1) Downstream migrants
The small brook, the Ruisseau de Warem (Station IV resulted since (1928) always in a failure, they were could be blocked with in fig. 1), totally our nets obstructed too soon by floating debris, plants, etc. without causing serious obstruction in the flow of Since the nets used at Stations I to III were only the brook.
30 cm wide, and since the river at these stations was Under these circumstances it is easy to get a good approximately 20 times as wide (viz., 6 m wide), we notion of the total number of animals participating
— that takes the can assuming migration place over in the The follow- downstream migration per night. front the and that entire at same rate, assuming it data Station available: ing (from IV) are takes the 50 of place entirely in cm water nearest
2208 individuals — 12-13 May 1967 night of — to the bottom (50 cm being the height our net)
1112 individuals — night 17-18 June 1967 estimate that the total migration rate approximately
2761 individuals — night 23-24 June 1967 amounts to 20 times the catch per net. Of course,
At Stations I, II, and the wider and this is since the waterlevel III, river was a very rough calculation, could found consequently not be blocked completely (see changes from 40 to 160 cm. Waters (1965)
such Needham abundance of § III. 1). Larger nets, as proposed by no noteworthy difference in vertical MIGRATION CYCLE IN GAMMARUS 29
Fig. 16. Nocturnal catches of upstream and downstreammigrants in Aug. 1968. Note that downstream migration takes
number to place every night (and that the of downstreamers apparently has little relation the moon-phase or light- tides intensity), whereas upstream movements are limited to the periods of spring (i.e., about 2 days after new and full moon).
but he worked rivulet of table Numbers drifting animals, in a only figs. 16—19, IV). as high as 6,000
25 cm deep. specimens per net per night have been observed
The number of animals found the calculations in a "good" night (corresponding, according to ex- in the drift Station I between 1000 and the with net on varies plained in previous paragraph, some 120,000 This would that individuals 1500. mean at least 20,000 to 30,000 crossing, in upstream direction, the imag- individuals, migrating downstream, crossed the imag- inary section of the river at Station I). The number
section of the river that Station. of in hour be inary at migrants one net per may as high as
4,000.
IX. 2) Upstream migrants
X. COMPOSITION OF THE SAMPLES During the few days that upstream migration takes
(viz., around H.W.S.), the number of animals All collected in the driftnet have been place samples sep- that in this kind of is arated and females participate activity many into juveniles, males, and all times larger than that migrating downstream (cf. specimens have been measured. With the aid of these 30 H. G. DENNEHT ET AL.
the be answered: data, following three questions can
(1) what is the composition of the samples captured;
there difference between and down- (2) is a upstream
in station stream migrants one sampling at any
particular sampling time;
there between collected (3) is a difference samples
at different sampling stations.
From the composition of the samples throughout
the it is that year, apparent one complete reproduc-
tive is cycle completed per year. It is clear (see table IX and fig. 20) that ovigerous
females in the in appear population early December,
and disappear in April and May. In the same two
months, \pril and May, there is a strong rise in the of the catches. These percentage juveniles in net
data will be discussed more fully in a following
chapter, § XIV.
The answer to question 2 is more complex. Statis-
tically, the composition of upstream and downstream
migrants caught in the nets at Station I have been Fig. 17. Hourly catches of upstream and downstream mi- tested. (For of and see way measuring testing, § III.4). grants in the night of 21 to 22 Sep. 1967. Sunset at 18.51 h, The has been collected test applied on samples on sunrise at 06.16 h. Moonrise at 19.52 h. Station I.
5 3 Oct. and Dec. 1967 and 29 1968. Sep., 1 on Apr. of In two these, 5 Sep. and 1 Dec., the upstreamers differed significantly in size from the downstreamers.
(significance 99%, infinite number of degrees of free-
dom). In the remaining samples (3 Oct. and 29 Apr.)
differences existed between no significant upstream
and downstream migrants (cf.fig. 21).
Table IX others the contains amongst mean ce-
phalic length in mm of animals caught at three dif-
ferent Stations (the material from Stations II and IV,
that has been for this are only 30 m apart, pooled
purpose in one column of the table; the greater part
the of samples were collected at Station II, only
those marked with an asterisk came from Station IV).
in Feb. the Only one instance (28 1988), up-
than the downstreamers. all streamers are larger In
other both have the the cases, same length, or up-
streamers are smaller than the downstreamers. The
situation is exceptional of 28 Feb. puzzling, but may
be the side statistical explained on one by a uncer-
tainty since the number of individuals caught and
measured was small the other hand fairly (n=46), on
influenced doubt the it is no by high percentage of females downstream that ovigerous migrating on day. the smaller For, size of females, when present in
have large numbers, a lowering effect on the mean
length of the total sample.
this However may be, the most interesting fact Fig. 18. Hourly catches of upstream and downstream mi- emerging from table IX is that during autumn and grants in the night of 4 to 5 Oct. 1967. Sunset at 18.22 h, when the spring equinoxal spring-tides, just upstream sunrise at 06.37 h. Setting of the moon 18.44 h. Station I. MIGRATION CYCLE IN GAMMARUS 31
River migration is most pronounced, the animals involved Table VIII. Advance of salty bottom water in the
in this have the smallest size. Slack migration and its estuary at H.W.S. on 8 September 1967.
When we consider table conclude that IX, we can Localities arranged from West (= most seaward) to East
there is clear difference between the downstreamers a (= most inland). of Station III and those of the other Stations. The locality chlorinityin u/oo downstream migrants of Station III (the Rivière de
thus the have in 15.8 Bazinghen, most upstream Station) a pools estuary
Slack at Station I 16.9 smaller size and the percentage of ovigerous females
Slack nearbridge in village Slack 1.68 is usually lower than at the other Stations. •300 m East of former locality 0.050 It is clear from table X, that the more particularly junction of rivers Slack and Laronville 0.075 the the of in upstream, higher percentage juveniles 200 m East of this junction 0.046
each This table shows the junction of rivers Warem and Laronville 0.076 population sample. com-
of catches taken position dip-net at some 300 m in-
terval along a stretch of nearly 2 km long of the Rivière de Bazinghen. Although the size shows rather XI. DEPOPULATION OF UPSTREAM AREAS
considerable the The with fluctuations, percentages juveniles only period large upstream migration is, as the participating in total catch rises rapidly from we have shown before, the period of September
38% the of the river with the December. the down- (at junction Slack) to through During the rest of year
93% 2 km of the That the stream drift dominates. Of when downstream (nearly upstream mouth). course,
size of the may vaiy irrespective juvenile morphology migration takes place eveiy night, without sufficient
is shown also in of the reaches of one preceding chapters, § V. repopulation by upstreamers, the upper
small size is an indication of sexual the distributional of G.zaddachi must Although always area depopulate
individuals still be That this theoretical immaturity, medium-large can im- to a certain extent. assumption
holds be shown mature. true in nature can easily by dip-net
taken at Station net Fig. 19. Numbers of upstream and downstream migrants I per per night in the period 13 July
1967 to 29 April 1968. composition of downstreamers the The at Fig. 20. Population at Station I during year. mean cephalic length every sampling date is indicated, as well as the standard deviation from that mean. MIGRATION CYCLE IN GAMMAIiUS 33
Table IX. and and months’ observation at Cephalic lengths, percentages of females juveniles, during an eight period
three stations in the Slack river system.
Station I Station II Station III date
(day, month, year) upstream downstream upstream downstream upstream downstream
5-IX-1967 1 0.764 (258) 0.860 (90) no catch no catch
2 0 0
3 84 84
19-IX-1967 1 0.870(112) < 0.755 (100) 0.947 (105) < 0.825 (100)
2 0 < 0 4 < 3
3 95 < 94 70 < 88
3-X-1967 1 0.963 (99) 0.901 (49) 0.860(110) 1.113 (50) 0.843 (110) 0.869 (100)
2 0 0 0 0 0 0
3 80 81 87 54 89 91
0 l-XII-1967 1 0.990 (74) 1.347 (65) 1.091 (100) 1.455 (65)« no catch
• ° 2 0 17 0 0
* • 3 62 3 41 0
' catch 6-1-1968 1 1.239(113) no < 1.308 (87)
2 < 30 < 46
3 < 2 < 1
28-1-1968 1 < 1.409(100) 1.365 (91) 1.332 (99) 1.212 (105) 1.265(100)
2 < 49 8 50 15 39
3 < 0 3 0 U 1
28-11-1968 1 1.450 (46) 1.241 (99) < 1.391 (90) < 1.257 (88)
2 23 24 < 31 < 25
3 4 0 < 0 < 0
* * 28-111-1968 1 < 1.305(105) 1.097 (82) 0.767 (100) no catch
° ° 2 < 43 37 1
" ° 3 < 11 32 62
29-IV-1968 1 0.533 (90) 0.620 (100) < < < <
2 3 2 < < < <
3 96 84 < < < <
in size mentioned in 1 = mean cephalic length mm (sample parentheses)
2 = percentage of ovigerous females.
3 = percentage of juveniles.
* text = data from Station IV (see § X).
< = sample size too small for statistical treatment.
in before the be- The is colonized sampling: late summer, thus just depopulated area temporarily by ginning of the autumn upstream migration wave, the the fresh-water species, Gammarus pulex (L.) and the inhabited berilloni uppermost 1500 m of area in winter by Echinogammarus (Catta).
G.zaddachi is entirely depopulated. Thus, in the win- This results in a picture in which G.zaddachi aban-
dons in the months the ter, G.zaddachi populates the Rivière de Bazinghen to warmer summer most up-
1150 of its with Fausse reaches of that is in m upstream junction La river its area, a picture good
in the distribution of with Kinne's observation that Rivière; summer, boundary agreement (1952, 1953)
stand lower salinities better lower G.zaddachi lies 350 m downstream of that junction. Gammarus can at 34 H. G. DENNERT ET AL.
temperatures (in other words, that its osmoregulatory over 5.00 and more than 50 m, respectively, in
lower performances are better at temperatures). one night.
(3) The percentage of recaptures drops sharply, when
the individuals released at than XII. MIGRATION DISTANCE PER NIGHT tagged are more
50 m from the net.
XII. Downstream 1) migration individual has (4) No travelled 81 m per night.
numbers of individuals taken at in a Large night No individual has 1 (5) migrated 10 m upstream. ) These downstream driftnetat StationIV were tagged. Of course, the nets contained, in addition to the released 1967. were at 4 consecutive days in June tagged individuals, also considerable numbers of un- A first released 5 m dis- group was at position A, at marked specimens (ratio marked-unmarked specimens the tance upstream of a driftnet, blocking entire about 1 : 4.1). brook. second was released at B, A group position The current velocity at Station IV during these 50.50 third at 66.50 at m, a group position C, at m, experiments was 40—50 cm/sec. (this is 24 m/min., fourth and a group at position D, at 81.00 m up- or 1440 m/h.). If the transport was purely passive stream of the net (see fig. 22). The distances between much distances than (= accidental), greater 67 m the points of release and the net were determined would have been covered. only by the nature of the river. Thus, the somewhat Conclusion. — The downstream migration during of 5.00 50.50 66.50 and illogical figures m, m, m, one night often amounts to more than 50 m, but never 81.00 m have been obtained.
reaches a distance of 81 m. In addition to a net for catching downstream mi-
the released also for grants among animals, a net up- XII. 2) Upstream migration stream migrants was placed, always 10 m upstream of the place of release. An experiment comparable with that described in the
The results are shown in table XI. also executed Station previous paragraph, was at I.
This experiment is described in detail in the next
From the several conclusions can be drawn: From this be table, section (§ XIII). experiment it can con-
cluded that certain of the mi- in- a upstream (1) If we total all the experiments, 548 tagged percentage distances of least grants can cover at 60 m in one dividuals (or nearly 35%) have been recaptured.
These numbers sufficient are important to war- night.
Thus, there is no difference in the rant also the conclusions (2) to (5). great upstream and downstream migration distance, both being in (2) Nearly 50% of the tagged animals have migrated the order of 50-60 for least of the m at a major part
animals.
and Table X. Percentages of juveniles mean cephalic XIII. ACCIDENTAL VERSUS ACTIVE length in the Rivière de Bazinghen, beginning with the TRANSPORT
most seaward ending with the most inlandlocality both The fact that upstreamers and downstreamers (n = sample size). make use of the water currents (see § VII) and that
be- there are not always significant size differences
mean be tween the two groups of migrants (see § X) could % locality juv. cephalic n
an indication that the observed movements of mem- length bers of the population of G.zaddachi support entirely
Junction Slack/ Rivière de with upon accidental, passive, transportation along Bazinghen 38 1.016 100 the currents during their nocturnal activity (feeding) Riv. de Bazinghen, S. of period. La Parthe 49 1.010 100 There of facts that Riv. de Bazinghen, SE. of are a couple prove convincingly
La Parthe 54 1.085 103 that accidental, passive, transport is out of question, Riv. de Bazinghen, S. of
Otove 70 1.117 50
Riv. de Bazinghen, junction *) It must be borne in mind that the individuals used in with Ruisseau d'Otove 89 0.922 44
this tagging experiment were originally caught during Riv. de Bazinghen, junction downstream migratoryactivity, thus hardly show any ten- with Fausse Rivière 93 0.963 m dency to migrate upstream (see also § XIII). MIGRATION CYCLE IN GAMMARUS 35
Table covered individuals. XI. Distances in one night by tagged
DOWNSTREAMERS UPSTREAMERS
Number of distance to number percentage distance to number
downstream individuals recaptured recaptured upstream recaptured tagged net (in m) net (in m)
550 5.00 212 47 10.00 0
550 50.50 290 52 10.00 0
389 66.50 46 12 10.00 0
94 81.00 0 0 10.00 0
Table XII. Surface water temperatures in the Slack system.
locality date temp. remarks
(day, month, year) in °C.
Ruisseau de Warem (Stat. IV) 13-V-1967 13.1 day
Ruisseau de Warem (Stat. IV) 20-VII-1967 18.2 day Slack (Stat. I) 5-VIII-1967 18.6 day
do. do. 17.1 night
Slack (Stat. I) ll-VIII-1967 18.2 day
do. do. 16.2 night
Riv. de Bazinghen (Stat. Ill) 22-IX-1967 14.5 night
Slack (Stat. I) 2-XI-1967 10.9 night
Slack (Stat. I) l-XII-1967 8.1 night Riv. de Bazinghen (Stat. Ill) 5-1-1968 7.5 day
Junction Slack/Laronville 30-1-1968 6.5 day
Ruisseau de Warem (Stat. IV) 30-1-1968 6.5 day
Junction Slack/Laronville 27-11-1968 3.5 day
Riv. de Bazinghen (Stat. Ill) 31-111-1968 12.0 night
Junction Slack/Laronville l-V-1968 13.0 day
Slack (Stat. I) 20-VIII-1968 17.3 day
thus that the observed movements on ac- of and support an (physiological) phase migrating up-river —, marked tive, behaviour pattern, being part of the population white 116 specimens caught the night be- dynamics of G.zaddachi. fore in a downstream net (thus in a "down-river
The These proofs are the following: mood"). red specimens were released at 17.00 h. the the Slack and 5 the reflux of the (1) Of 15 species inhabiting river on Oct., during tide, thus during
all show nocturnal but which there seaward its estuary, activity, only a period in is a strong current in direction. The white G.zaddachi participates in migratory movements. specimens were released at
of G.zaddachi 23.15 h. 5 the reversal. (2) Upstream migration is important on Oct., during current Con- during the equinoxal H.W.S. of the autumn, but sequently the red specimens (those "in the mood" of when also released of down- during the spring equinoxal H.W.S., cur- migrating up-river) were at a period
reversal is river the white "in the rent observed, hardly any upstream migra- currents, specimens (those mood" of tion takes place (see fig. 19). drifting down-river) during a period of up-
the distance river currents. nets (3) If transport was purely passive, Two upstream were placed, up- covered downstreamers would of the release 23 for by during one night stream places (see fig. position be release Both lifted much larger than it actually is (see § XII. 1). of nets and places). nets were at
obtained the h. Both contained a number of (4) Final proof was by following exper- 02.00 on 6 Oct. nets iments: red specimens (25 specimens, or 2%% recaptured in
1967. Mark- both of the red had Tagging experiment 5-6 Oct. Station I. nets together); part specimens
red the before distance of 60 reach ed were 1000 specimens caught night migrated upstream over a m to
— thus in the the No white were in an upstream net specimens being net. specimens recaptured. 36 H. G. DENNERT ET AL.
different dates and downstreamers at Station I. (N = size of Fig. 21. The cephalic lengths at in upstreamers caught sample) MIGRATION CYCLE IN GAMMARUS 37
migrating up-river, whereas animals with an ana-
dromous physiological orientation continue migrating
down-river.
The had upstreamers have more than 6 hours,
the mostly moreover in dark, to move (actively or
passively) down-river, but they did move actually
40 to 60 m up-river. Not a single upstreamer ever
entered in a down-river net. Fig. 22. Position of drift net and points of release of
tagged individuals at Station IV. Distances indicated in
meters. Arrows indicate the direction of the stream. Ex- XIV. THE ANNUAL CYCLE planation see § XII. 1.
XIV. 1) Introduction
Kinne showed that the of Gammarus A less (1952) lifespan second, more or comparable, experiment was zaddachi is to restricted a period of 12 to 16 months. done on 1-2 Dec. 1967. Marked red were 1730 ani-
check Kinne's observation the situation found mals To for caught in the upstream net the night before, in river the Slack, a number of individuals from each marked blue 570 the were specimens caught in sorted samples was measured (see: § III.4), into juve- downstream net the night before. Upstream of the niles and adults, whereas the last-named release a combination category place, net (for both upstream was sexed. The maximum and minimum sizes in each and downstream captures, see figs. 4-5) was installed, the for each downstream of the release downstream sample, mean cephalic length category place, a net for each the installed of (<3, $, or juv.) sample, percentage was (see fig. 24). During a period nor- each of per sample, and the standard de- mal, seaward, river flow, at 16.45 h. on 1 Dec., the category
have been recorded. Lists these red released. viations, containing specimens were During a period of cur- data have been in the Museum rent h. the blue deposited Zoological reversal, at 22.45 on 1 Dec., speci- of the University of Amsterdam and are available for mens were released.
scientist interested in them. the combination thus of the inspection to every Some In net, upstream re- of the data have been used also in lease 21 red and 2 blue more pertinent place, specimens were re- table IX. captured. In the downstream net, placed down-river
of the release blue and red individuals place, one no XIV. 2) The situation at Station I have been recaptured.
The conclusion from these ob- two experiments is In fig. 20, the data for down-river migrants have
vious: animals physiologically in the catadromous been plotted, caught at Station I. Station I has been
phase, continue after tagging, and irrespective of the selected purposely for this subject, since it is the
direction of the the and stream at moment of release, most down-river Station, one might expect in the
in and described Figs. 23-24. Position of nets, etc. the tagging recapture experiments, in § XIII (4). The arrows indicate the direction of the normal (= seaward) stream in the river. 38 H. G. DENNERT ET AL.
catches not only those animals that lived near Sta- Two possible explanations of this observation (too
but also down-riverfrom low of females fresh tion I, those migrating more a percentage in autumn at the
upstream localities. Stations) can be brought forward:
On 27 July 1967 (fig. 20, top) the down-river mi- (1) Males grow faster than females, obtain earlier of blocks the grants consist entirely juveniles (open in their adult morphology, thus their sex could be prop- graphs). On 7 July 1967, the first adult females (black erly recognized earlier than that of the females which blocks) and adult males (stippled blocks) appear. The were classified longer with the unsexed first females in the catches De- category ovigerous appear on 1 juveniles. cember 1967 at this date (hatched blocks); hardly any This explanation fails to explain why at Station I, juveniles are present any more. Juveniles are entirely where the same method of classifying into the cate- absent in the catches from 6 January to 28 February gories juvenile-female-male was used, no anomaly in 1968, they reappear again at 28 March 1968. On the autumn sex-ratio was found. 29 April 1968 the juveniles form already the greater of the catch and 2 the This conclusion leads modification part on August 1968 entire (2) to a of state-
catch of Thus that that males fresh-water is composed juveniles. we see ment (1), in in attain ear-
ovigerous females occurred at Station I from 1 De- lier the adult morphology than fresh-water females.
cember 1967 to from 27 shows the 29 April 1968, juveniles July Fig. 25 appearance of ovigerous females
1967 (start of observation period) to 1 December at Stations I, II, and III, and the simultaneous reduc-
1967, and again, from 28 March 1968 to 2 August tion of the percentage of juveniles. From this figure,
1968 (end of observation period). Along with this it is clear that in the period from July to November,
shift of the the the of the in composition catches, mean length percentage juveniles in population gradual-
mean of the entire decreases from (= cephalic length sample meas- ly 100% to near 0%. At Station I,
about shows in ured, n being usually 100) an increase ovigerous females appear already the population
from July to November 1967, remains fairly constant in November, before the last juveniles disappeared
and decreases in 1968 moulted adult and during winter, suddenly April (— into an morph). At Stations II
the and standard deviation there is of 1 2 months between the dis- (see fig. 20, mean length III, a gap to
indicated under the of of the last being graph eveiy date). appearance juvenile and the appearance of
The reproduction period (December till April) is the first ovigerous females. In this lapse of time, the with Kinne's in good agreement data, as well as a population consists of morphologically adult, but sex-
of demonstrated the absence of lifespan 1 year (as by ually inactive individuals. In the two fresh Stations
adults in summer). Also in good agreement with Kin- (II and III), the reproduction period does not only
ne's report (1952) is the sex-ratio, which — in the start later, but also lasts shorter.
— in the These in with Kinne's reproduction period was 1 $ per 1.92 Ç phenomena are agreement
versus 1 <5 2 in the of Ger- work that in Slack, per Ç Bay Kiel, experimental (1952), showing Gammarus
duebeni and of > many. in G.zaddachi, at a temperature
the best a of 8°C, eggs developed at salinity 10°/oo, XIV. 3) The situation at the Stations II and III whereas lower the did at temperatures salinity not
At Station I, the sex-ratio and the influence the The lower the reproduction period any longer development. reasonable obtained were in agreement with the data temperature, the lower also the salinity at which
in showed the females still by Kinne Germany. Station I at same can produce eggs. In Stations II and III,
the resemblance to the German females do in the time greatest sampling ovigerous not appear population un-
in area, that both localities lie well within mixohaline til the temperature drops under a certain value. This the value which waters (classification according to Venice System). limitating temperature (above no females
Stations II and III, however, lie in the fresh-water in get sexually mature any longer fresh-waters) ap- of the region stream. pears to lie in the Slack roughly at 7°5 to 8°0 C. This
The these fresh value obtained of sex-ratio in Stations differs in is through comparison the temper-
autumn from that in Station in all values recorded in table XII and the of I: Station I year ature time
round there 1.5 to 2 times females of 25. are as many as appearance ovigerous females, as plotted in fig.
males; in autumn at the fresh Stations there This value of 7°5 — 8°0 is are more limitating temperature males 6.5 than females. both the of the (sometimes X so many) Later working at beginning reproduction in the when the the well year, reproduction period starts, period (retarding its start in fresh-water) as as at
ratio shifts again in the advantage of the females, the end of the reproduction period (accelerating the
and close to the : $ the in stays Ç ratio of 2 : 1. end of egg production fresh-water). MIGRATION CYCLE IN GAMMARUS 39
stations is Fig. 25. Composition of downstreamers at three different (Station I mixohaline, the other two are limnic).
Explanation see § XIV. 3 (2).
XIV. 4) Conclusion drifters (as reflected by the length distribution) on
27 1967 almost identical to in July (fig. 26, top) was Ovigerous females show up first the poikilohaline that on 2 August 1968 (fig. 26, bottom). This in- parts of the river in November. More up-river, in the the dicates that growth in 1968 is one week slow as permanently fresh region, ovigerous females appear compared with that of 1967. It is probable that this 1 to 2 months later, probably only then when the be the fact that 1967 had backlog can explained by water temperature falls under 7°5 C. In the fresh- favourable a fine, warm summer (probably for a more water region, the reproduction period not only starts whereas the summer in later, but also lasts shorter. rapid growth), temperatures
1968 were below the mean.
XV. COMPARISON OF POPULATION HISTIO-
GRAMS OF DIFFERENT YEARS XVI. SYNTHESIS OF THE ANNUAL REPRO-
DUCTIVE AND MIGRATORY CYCLE In figs. 26-27 down-river migrants from the same dates in 1967 and in 1968 have been compared. The In the river Slack, the reproductive cycle is correlated dates used for this 2 1967 with the In comparison are August, migratory cycle. autumn (September to and 1968, and 12-13 August, 1967 and 1968. In both November) when during the equinoxal tides current
the for 1967 show reversal of the instances, histiograms a greater is strong, mass repopulation up-river and of reaches takes this mean length, a higher percentage morpho- place (see § VIII). During period,
the the smaller logical adults in entire population than those for up-river migrants are of significantly size
1968. On the other hand, the composition of the than the down-river migrants (§ X), their morphology 40 H. G. DENNERT ET AL.
part of the river, later also — when the watertemper-
atures drop below roughly 7°5 C. — in the fresh-
of the The first water part river (§ XIV-3). juveniles
are found late March, and during the rest of spring,
the remaining adults gradually disappear to make
for the These place new generation (§ XIV.2). juve-
niles stay down-river till next fall, when they migrate
up-river again.
In agreement with previous observations by other
authors (§ VI.1), the migratory activity is subject to
a day-night-rhythm. Migration takes only place at
night and reaches its maximum shortly after sunset
for down-river the migrants (see § VI.2) or during
period of current reversal for up-river migrants
(§ VII. 1). The migration is not accidental (§ XIII)
animals that are physiologically in a catadromous
do of the direction phase move up-river, irrespective
of the current, and vice-versa. Moreover, other gam
of the of down-river marid the do show Fig. 26. Histiograms cephalic length species in area (14 species) not
at three different dates. The lower two histio- migrants migratory activity. interval. grams are based on catches at precisely one year Of the animals that live far do not course, up-river Explanation see § XV. necessarily migrate to the sea-shore, but just a certain
distance down-river. This distance be close may to is the It must be concluded that 1500 is demonstrated the of juvenile one. then, m, as by depopulation
the the reaches repopulation of up-river is performed the uppermost 1500 m of the winter area (§ XII).
assisted the short of The known distance both of by juvenile animals, by period migration per night, up-
reversal VII. well-known fact that and of downstream current (§ 1). It is a stream migrants (§ XIII) migrants
juveniles are more resistant to lower salinities than (§ X.l), is sufficient to explain the yearly migration
adults and farther than adults can migrate up-stream cycle in the river Slack.
is do flounder, a ex- Of the section of the (the Platichthys flesus, good population at any given river
ample). (during daytime, the time of non-activity), a certain
the anadromous in in During entire year, migration ("or- part (stippled fig. 28) participates the nocturnal takes which of the ganic drift") place (§ VIII), during a migration activities. During certain periods certain portion of the leaves its ter- see 28 the of animals "standing crop" year (fall, fig. A) percentage ritory and migrates down-river. This all-year round participating in up-river migration dominates, in other down-river of the down-river migration causes a depopulation certain periods (late winter, see fig. 28 B)
that the fall the the up-river regions were populated in migrants are in majority. In up-river migra- The (§ XI). down-stream migrants attain sexual matu- tion, chiefly juveniles participate, which is not as- rity in winter (§ XIV), first in the saltier, estuarine tonishing, since the entire population chiefly consists
of juveniles in early fall (§ XIV), whereas in spring
and summer upstream areas are gradually depopu-
lated through down-river migration of the up-growing
— that — their down the juveniles, during way river
reach maturity and get ovigerous in the estuarine part
of the river.
The idea of that of the Waters, overpopulation up-
river reaches results in off-flow of part of the "stand-
the is not confirmed for ing crop" throughout year of G.zaddachi G.zaddachi in our study. First all, re- produces chiefly in the down-river, and not in the
up-river reaches, and secondly the transport is tied
in to a the animal's developmental Fig. 27. Histiograms of the cephalic length of down-river physiological phase
taken one migrants at year interval. Explanation see § XV. cycle. MIGRATION CYCLE IN CAMMARUS 41
Fig. 28. Diagrammaticrepresentationof the migration pattern of Gammarus zaddachi in autumn (= early October) and in winter (= late February). The dotted circles indicate that portion of the total population participating in migratory The indicate the direction of this The undotted of the the activity. arrows activity. part diagram represents part of the population not participating at that moment in migratory activity.
The width of the dotted circles is an indication for the numbers of animals participating in migration.
For in of a complete migration diagram the River Slack, a greatnumber overlapping circles should be imagined.
The blocks at both sides of the migration circles indicate the composition of the migrating population (ovigerous fe- males, males plus non-ovigerous females, juveniles). 42 H. G. DENNERT ET AL.
XVIII. DISCUSSION A future line of research is no doubt rearing of
G.zaddachi in fresh-water, to which small quantities
As shown the the of certain have been the first was in preceding paragraphs, ac- ions added, in place
and The relation tivities of Gammarus zaddachi are greatly influenced Ca Mg. between the temperature
several such the and the of the by environmental factors, as light in- at one hand, ionic composition
tensity, direction of the water currents (which in turn medium at the other, especially its effect on the re-
depend on spring-tides, thus on moon phases), temp- production and on the osmoregulatory system, could
and be much than it at first erature, salinity. more complex appears now
It be remarked that decisive the must a answer on sight.
solved either this quantitative influence of light intensity and temper- Not yet at moment is the possible
has not been seen the of relation between the and the number of ature, yet given, paucity salinity eggs data available. produced.
Résumé
1 — Les auteurs ont cherché à identifier des Gammarides temps) la migration se fait enmasse vers l’amont; toutefois,
cette juvéniles récoltés dans la zone de marées d’eau douce migration n’a lieu que durant les nuits où le courant
(voir point 6) de la Slack (France), 3 à 4 km en amont de s’inverse sous l’influence d’une forte marée. Pratiquement l’estuaire.Ces furent d’abord considérés individus spécimens juvéniles ce ne sont que les juvéniles quiparticipentà cette
amont. (du point de vue de leur morphologie) comme appartenant migration en
à l’espèce Gammarus salinus. 8 — Le maximum d’individus pris en un seul filet (30 X
de 2 — L’élevage de ces spécimens a montré qu’ils se dévelop- 50 cm) et en une nuit fut 1.500 G. zaddachi descendant
Gammarus Ceci 6.000 paient en adultes de zaddachi. prouve que le courant et de remontant le courant; ceci correspon- les sali- environ 30.000 critères morphologiques utilisés pour distinguerG. drait à individus descendant le courant, et nus et G. zaddachi être matériel à 120.000 le section de la rivière. ne peuvent appliqués au remontant, pour une
juvénile. Plusieurs critères morphologiques ont été con- Dans la Slack, aucune autre espèce de Gammaride ne mani- sidérés testés. Il de caractère-clé de et apparaît qu’il n’y a pas feste comportement migratoire.
9 montré sûr à 100%, c’est-à-dire qu’il est impossible d’identifier — Les expériences avec des individus marqués ont
certitude les couvrent avecune absolue tous les individus d’une popula- que migrateurs vers l’amont des distances de de deux tion mixte ces espèces. 40 à 60 m en une nuit. Les migrateurs vers l’aval couvrent
entre et 3 — En même temps, la ressemblance G. zaddachi aumoins 50 m auplus 80 m en une seule nuit.
et G. caractère indices donnent à les jeune salinus adulte fournit un morpholo- 10 — Divers penser que déplacements
très de G. zaddachi giquequi permet, au cours derecherches sur un grand vers l’amont et vers l’aval ne sont pas pure-
les nombre d’exemplaires, de distinguer rapidement formes ment accidentels, bien qu’il profite des courants pour adultes et jeunes deG. zaddachi. migrer. Les individus qui se trouvent dans la phase (phy-
des recherches la et la la dans 4 — Lors sur reproduction migration siologique) de descendre rivière, persévèrent ce
G. et Il même de zaddachi, on a conçu confectionné divers types de sens en dépit du courant. en va de des individus filets. Ils de permettent capturer les migrants anadromes qui se trouvent enphase deremonter la rivière.
aussi bien que catadromes. Leurs capacités se sont révélées 11 — Le cycle de reproduction de G. zaddachi s’accomplit
année. équivalentes. sur une Les jeunes font leur apparition au prin- 5 — L’activité ils de G, zaddachi montre une périodicité temps; parviennent à maturité à l’automne; on trouve
elle culmine durant la nuit. Ce les femelles le diurne; comportement premières ovigères en Décembre; maximum résulte at- dans une migration vers l’aval (“dérive”); elle depontes se situe au début du printemps. Tous les adultes à teint son maximum environ 2 3 h. après le coucher du meurent après la période de reproduction.
diminuer ensuite à la nuit 12 — Le de lié des soleil, pour mesure que s’avance; cycle reproduction est au cycle migra- la „derive” s’arrête pratiquement au lever du soleil. On n’a tions. Les femelles quiparticipent à la migrationvers l’aval,
constater succès pu aucune influence de l’intensité lumineuse peuvent se reproduire avec si la température et la nocturne (phase lunaire, nuages). salinité sont favorables. Lorsque la température dépasse
6 — Durant les la période des grandes marées, la direction du 7°5 C, il n’y a de reproduction que dans parties mixo- le Slack. baisse courant se renverse sur tout secteur de recherche. Sur la halines de la Quand la température au dessous
rivière la les secteurs section de la voisine de l’estuaire, ce revirement de7°5 C, ponte est également possible dans s’accompagne d’une augmentation rapide de la salinité. limniques de la rivière.
Plus amont la en salinité ne change pas : on trouve là une Les jeunes passent l’été dans la zone de l’estuaire. A zone de marées d’eau douce. cette saison. G. zaddachi disparait d’une grande partie de
7 — Toute l’année, une certaine partie de la population l’aire de répartition, en amont de la rivière. Ce sont les habitant endroit donné aidés les un (le soi-disant „standing crop”) jeunes, par courants des grandes marées, qui
Pendant les marées le est migre en aval. grandes d’équinoxe repeuplent la région limnique en automne; cycle d’automne (et à un degré moindre à l’équinoxe de prin- ainsi bouclé. MIGRATION CYCLE IN GAMMARUS 43
REFERENCES
in — BASSINDALE, R., 1942. The distributionof amphipods the ,1966b. Die Tagesperiodik von Flieswasserorganismen.
Severn estuary and Bristol Channel. J. anim. Ecol., 11: Z. Morph. Ökol. Tiere, 46: 93-142.
131-144. Needham, P. R„ 1928. A net for the capture of stream drift
1944. The fate of animals in drift when DENDY, J. S., stream organisms. Ecology, 9: 339-342. carried into lakes. Ecol. Monogr., 14 (3): 333-357. D. 1943. Gammarus Am- Reid, M., sarsi, sp.n. (Crustacea, ELLIOT, J. M., 1967. Invertebrate drift in a Dartmore phipoda). Ann. Mag. nat. Hist., (11) 10: 281-285. stream. Arch. Hydrobiol., 63 (2): 202-237.
— 1945. Gammarus ochlos nov.), Crustacea, Amphi- , (nom. The of diurnal HAHKER, J. E., 1964. physiology rhythms. poda. Ann. Mag. nat. Hist., (11) 12: 637-638. (Cambridge University Press). SEGERSTRÂLE, S. G., 1947. New observations on the distri- HULTIN, L., 1968. A method of fresh water Am- trapping bution and morphology of the amphipod Gammarus phipoda migrating upstream. Oikos, 19: 400-402. zaddachi Sexton, with notes on related species. J. mar.
HYNES, H. B. N., 1954. The ecology of Gammarus duebeni biol. Ass. U.K., 27: 219-244.
Lilljeborg and its occurrence in fresh water in western SEXTON, E. W., 1912. Some brackish-water Amphipoda Britain. J. amin. Ecol., 23: 38-84. from the mouths of the Weser and the Elbe, and from
B. 1968. On the diurnal ac- JANSSON, O. & KALLANDER, C., the Baltic. Proc. zool. Soc. London, 1912: 656-665. the tivity of some littoralperacarid crustaceans in Baltic 1913. On collection of Gammarus from the —, a Königs- Sea. Biol. 2: 24-36. J. exp. mar. Ecol., berg Museum. Sehr. Königsbergphysik. Ges., 54: 90-94.
Zum Gammarus — ,1924. The and of Gammarus, KINNE, O., 1952. Lebenszyklus von moulting growth-stages duebeni with description of the normals and intersexes of G. Lilljeborg, nebst einigen Bemerkungen zur
chevreuxi. J. mar. biol. Ass. U.K., 13 340-401. Biologie von Gammarus zaddachi Sexton ssp. zaddachi (2): the of Inst. 1: 1928. On rearing and breeding Gammarus in Spooner. Veröff. Meeresforsch. Bremerhaven, —, conditions. J. biol. Ass. U.K., 15 187-203. laboratory mar. (1):
33-55. — 1953. Wird die der , Häutungsfolge Amphipoden (Gam-
— 1942. The relation of Gammarus zaddachi Sexton to marus zaddachi zaddachi Sextion, G. zaddachi salinus ,
Gammarus in Spooner, G. duebeni Lillj.) durch die lunäre Periodicität some other species of ocurring fresh,
estuarine marine biol. Ass. 25: beeinflusst? Kieler Meeresforsch.. 9: 271-279. and waters. J. mar. U.K.,
575-606. — 1954. Die Gammarus-Arten der Kieler Bucht. . . . Zool. , Jahrb. (Syst.), 82: 405-424. SPOONER, G. M., 1947. The distribution of Gammarus
1960a. Gammarus salinus — Einige Daten über dem —, biol. species in estuaries, 1. J. mar. Ass. U.K., 27: 1-52. Umwelteinfluss auf Wachstum, Häutungsfolge, Herz-
und 1 H. & 1966. is Gammarus frequenz Eientwicklungsdauer. Crustaceana, (3): STOCK, J. KANT, P., What campy-
208-217. lops of Sars, 1894? Bull. zool. Mus. Univ. Amsterdam
— ,1960b. Growth, moulting frequency, heart beat, number 1 (2): 7-18.
incubation time in Gammarus zaddachi of eggs and ex- STOCK, J. H., NIJSSEN, H. & KANT, P., 1966. La répartition posed to different environments. Crustaceana, 2 (1): écologique des Amphipodes de la famille des Gamma- 26-36.
ridae dans la Slack et son estuaire. Bull. zool. Mus. Univ —, 1961. Moulting frequency of mature Gammarus zad- Amsterdam, 1 (3): 19-30. dachi. Correction. Crustaceana, 2 (2): 171. TANAKA, H., 1960. On the daily change of drifting of LEHMANN, U., 1967. Drift und Populationsdynamik von animals the of benthic in streams, especially on types Gammarus pulex fossarum Koch. Z. Morph. Ökol. Tiere, daily change observed in taxonomix groups of insects. 60: 227-274. Bull. Freshwat. Fish. Res. Lab. Tokyo, 9: 13-24. MCLAY, C. L., 1968. A study of drift in the Kakanui river, The external factors VERWEY, J., 1966. role of some in the New Zealand. Aust. J. mar. Freshwat. Res., 19 (2): 139- vertical of marine animals. Netherl. J. Sea 149. migration Res., 3 (2): 245-266. MINCKLEY, W. L., 1964. Upstream movements of Gamma-
Doe WATERS, T. F., 1961. Standing crop and drift of stream rus (Amphipoda) in Run, Meade County, Kentucki. bottom 42: 532-537. Ecology, 45 (1): 195-197. organisms. Ecology, the 1962a. A method to estimate production rate of a drift —, MÜLLER, K., 1954a. Investigations on the organic in stream bottom invertebrate. Trans. Amer. Fish. Soc., 91: North Swedish streams. Rep. Inst. Freshwat. Res. Drott- 243-250. ningholm, 35: 133-148. 1962b. Diurnal in the drift of stream in- —, periodicity in —, 1954b. Die Drift fliessenden Gewässern. Arch. vertebrates. Ecology, 43 (2): 316-320. Hydrobiol., 49: 539-545.
— , 1964. Recolonization of denuded stream bottom areas — ,1963a.Temperatur und Tagesperiodik der "Organischen by drift. Trans. Amer. Fish. Soc., 93: 311-315. Drift" von Gammarus pulex. Naturwiss., 50: 410-411.
— , 1965. Interpretation of invertebrate drift in streams.
— , 1963b. Diurnal rhythm in "organic drift" of Gammarus Ecology, 46 (3): 327-334. pulex. Nature (London), 198: 806-807.
M. — L., Verklarende statistiek , 1966a. Zur Periodik von Gammarus pulex. Oikos, 17: WIJVEKATE, 1966. (8th ed.).
207-211. Aula, 39. (Utrecht).