The Effects of Low Temperature upon the Locomotor Activity Rhythm of Talitrus Saltator (Montagu) (Crustacea: Amphipoda)

Different effects upon rhythmicity of exposure to temperatures well below that normally experienced by the active animal have been reported for several arthropods. Kalmus (1934) and Renner (1957) both recorded delays in the time sense of bees after chilling. In the crab Uca the rhythm of colour change is delayed for a time corresponding to the length of the chill (Brown & Webb, 1948). Harker reported that when the neurosecretory cells of the suboesophageal ganglion in Periplaneta americana L. are chilled and the organ then transplanted to an arrhythmic host, the locomotor rhythm now shown by the host is delayed by a length of time corresponding to the duration of the chill. If the animal bearing the chilled neurosecretory cells is left intact, a similar delay occurs, except when the chill is for up to 4–5 h or for between 18 and 24 h, when there is no change in the phase of the rhythm (Harker, 1964). Bünning (1959) found in the same animal that the extent of the phase change depends upon the time of onset of chilling, as Stephens (1957) also reported for Uca.

In Carcinus maenas L., after chilling to 4 °C, the tidal rhythm of activity is phased with the time when the crabs are returned to the normal temperature and is not dependent upon the duration or time of onset of the chill (Naylor, 1963). Chilling Talitrus saltator (Montagu) to 4–6 °C for 18 h gave no ‘valuable results’ concerning the phase of its sun-orientation rhythm (Pardi & Grassi, 1955), but chilling Bathyporeia pelagica (Bate) to 1 °C for 6 h beginning 3 h before time of high water resulted in a delay in successive activity peaks of about 3·5 h (Fincham, 1970).

T. saltator possesses a well-defined circadian rhythm of locomotor activity, with a peak of activity around midnight (Bregazzi & Naylor, 1972). The present paper describes the effects upon this rhythm of subjecting the whole animal to periods of low temperature. These effects allow some speculation upon the control of locomotor activity in this species.

Collection of animals in the field and the means of recording their activity were as described previously (Bregazzi & Naylor, 1972). In order to chill them, single males were placed in blackened glass tubes containing a little damp sand, which were then corked securely and placed in a refrigerator for the appropriate time. Although it was found that chilling to 6 °C was effective, in the experiments reported here animals were chilled to 2–3 °C. Several experiments were carried out for chilling periods each of 8, 16 or 24 h, and the time of onset of chilling covered most of the 24 h day at intervals of 2 h.

In each experiment all animals, including controls, were introduced to the tubes at the same time, irrespective of when chilling began. When not undergoing exposure to low temperatures animals were kept at 15 °C, and subsequent recording of activity was also carried out at this temperature in continuous dim red background illumination. Groups of four animals were introduced into the actographs usually immediately after chilling, but sometimes there was a delay of a few hours after return to normal temperatures.

Estimation of hourly activity and calculation of mid-activity points were carried out as described previously (Bregazzi & Naylor, 1972) and differences in phase between control and chilled groups of animals were calculated from the respective mid-activity points for the three days following chilling. Preliminary experiments showed that the difference in phase between chilled and control animals was maintained under constant conditions for up to 19 days.

All times are expressed as GMT.

The result of a typical experiment, in which groups of animals were chilled for 8 h, is given in Fig. 1. It shows that different times of onset of chilling produce varying delays in successive activity periods. The delays in mid-points of activity periods for different times of onset and for durations of 8, 16 and 24 h of chilling are given in Fig. 2 (a, b, c) (open circles). If chilling began between 08.00 and 16.00 h the delay was roughly equivalent to the chilling period. At other times the delay was less than this, with a minimum delay at 20.00 h in the case of the 8 h chill and at 18.00 h for the 16 h and 24 h chills. The change in delay from 16.00 h to the minimum was abrupt, but the subsequent change in delay from the minimum to 08.00 h was relatively slow.

Whatever the nature of the process which controls the rhythmic locomotor activity in Talitrus saltator, it clearly possesses a differential response to low temperatures. It is possible to distinguish between a part of the cycle which is stopped (or nearly stopped) by chilling to 2–3 °C and a part which is relatively unaffected by these temperatures (Fig. 2). The discrepancy between duration of chill and phase delay in Bathy-poreia pelagica reported by Fincham (1970) may be due to a similar differential effect.

In several papers (see Harker, 1964), Harker produced strong evidence for the neurosecretory control of rhythmic locomotor activity in Periplaneta americana, although this emphasis on neurosecretory control has since been questioned by the work of Roberts (1966) and Brady (1967a, b). Brady (1969) has produced striking evidence that the site of control of locomotor rhythmicity in Periplaneta is in the optic lobes, which contain no neurosecretory cells, and he suggests that control is therefore primarily electrical, with a hormonal system acting at best as a somewhat ephemeral mediator. It would be premature to speculate at this stage upon the primary control of rhythmicity in Talitrus, but even if it is electrical, this by no means rules out mediation via a neurosecretory system or systems. Indeed, it would be surprising if this were not the case, for in crustaceans hormones are invariably concerned in the control of longer-term effects such as colour change, moulting and breeding cycles, whereas direct electrical control is more concerned with relatively rapid behavioural responses.

Present results with Talitrus would seem to be more readily explained on a hormonal basis rather than by a nervous mechanism. Moreover, since the maximum effect of chilling is exerted from 08.00 h to 16.00 h, which is during the time of inactivity in the animal’s locomotor rhythm, it can be argued that chilling affects an inhibitory factor which is produced cyclically at an unknown site within the animal. Such a mechanism is similar to that suggested for Carcinus (Naylor & Williams, 1968 ; Naylor, Smith & Williams, 1972) in which there is evidence that a hormone which inhibits locomotor activity is produced rhythmically in the eyestalk.

Some evidence for the involvement of an inhibitory hormone concerned in the control of locomotor activity in the house cricket Acheta domesticus L. has been produced by Cymborowski & Dutkowski (1969, 1970), and Rensing (1964) has reported bimodal activity patterns in the neurosecretory cells of the pars intercerebralis and corpus allatum of Drosophila melanogaster (Meigen) which parallel the dusk and dawn locomotor activity periods in this species. Also, Rao & Ghopalakrishnareddy (1967) recorded changes in electrical activity of the central nerve cord of the scorpion Hetero-metrus fulvipes, induced by extracts of the cephalothoracic nerve mass and blood which were taken at different times of the day. These authors suggest that excitory and inhibitory neurohormones may be produced rhythmically by neurosecretory cells in the sub-oesophageal ganglion.

Such control mechanisms as those described above place greater emphasis upon hormonal mediation than the mechanism proposed by Brady for Periplaneta, but they do not rule out the possibility of additional nervous control.

The data given for the 8 h chill in Fig. 2(a) suggest that maximum production of the proposed inhibitory factor in Talitrus is between 08.00 h and 16.00 h, and that the production is minimal or absent from around 20.00 h for about 8 h. There is no phase delay if an 8 h chill coincides with this latter period. A possible model for the cyclical production (or secretion) of the inhibitory factor is given in Fig. 3, and there are important differences between this model and an apparently similar model proposed by Bünning (1967). As a result of differential effects of chilling upon the rhythms of Periplaneta and other organisms, Bünning has advanced a tension-relaxation oscillator model for the control of these rhythms. This model allows for two distinct parts of a single cycle, one requiring energy (the ‘tension’ part) and the other (‘relaxation’) able to proceed without the supply of additional energy or at least with considerably less energy requirements. Chilling for short periods during ‘tension’ results in a delay in the appearance of the successive peak of the rhythm. This delay may be longer than the period of the chill because the degree of ‘tension’ falls to an energy level lower than that which obtained when the chilling began. Chilling during the ‘relaxation’ period, however, has minimum effect upon the phase of the rhythm because the energy level is falling anyway. Chilling for periods of more than a few hours results in the ‘tension’ falling to, and remaining at, a minimum level, and consequently, upon return to normal temperatures, the rhythm is phased with the time of return and does not depend upon the length of the chill or upon the time of its onset.

Control of the locomotor rhythm by a tension-relaxation oscillator does not seem appropriate for Talitrus on present results because delay of the successive activity peaks is never significantly greater than the duration of the chill. Furthermore, the delay after chilling for 24 h is clearly dependent upon the time of onset of the chill, and, when compared with the delays following chilling for 8 h and 16 h, upon the length of the chill (Fig. 2). For the same reasons, and also because there is no ‘fade-out’ in the rhythm of Talitrus (Bregazzi & Naylor, 1972), the ‘multiple-clock’ hypothesis advanced by Naylor (1960, 1963) for control of tidal rhythmicity in Carcinus does not appear to apply in the case of Talitrus.

The present model for Talitrus postulates an 8 h period of chilling ineffectiveness and minimum inhibition from 20.00 h to 04.00 h, and 8 h period of chilling effectiveness and maximum inhibition from 08.00 h to 16.00 h, with two periods (04.00–08.00 h, 16.00–20.00 h) of average chilling effectiveness and intermediate inhibition (Fig. 3). A fairly good fit is obtained between the model predictions (Fig. 2, continuous lines) and the observed results (Fig. 2, open circles) if it is assumed that the clock is differentially affected by chilling during the periods of maximum, intermediate and minimum inhibition, such that it is stopped, allowed to proceed at half rate and is unaffected by chilling, in these three periods respectively. For instance, chilling for 8 h beginning at midnight has no effect on the ‘clock’ until 04.00 h, after which it proceeds at half rate. At 08.00 h, therefore, the ‘clock’ has moved only as far as 06.00 h, resulting in a 2 h delay. Likewise, chilling for 16 h beginning at midnight means that the ‘clock’ reaches the 08.00 h position after 12 h, whereupon it is stopped until 16.00 h, and the consequent delay is 8 h. There are indications that the delay in successive activity peaks is less than that predicted by consideration of the suggested model, especially for the longer-duration chills. This suggests that the ‘clock’ mechanism may not be completely inhibited by low temperatures during the period of maximum production of the inhibitory factor.

The model is of necessity simplified and is not intended to reflect in detail the nature of a cyclical process concerned in the control of the locomotor rhythm in Talitrus. This is, in any case, likely to be very complex, and may or may not have an electrical pacemaker with circadian periodicity as its basis. The present model, which postulates a rhythmically produced inhibitor as at least a mediator in the control of the circadian rhythm in Talitrus, opens up further lines of enquiry.

1. Chilling Talitrus causes a complete cessation of locomotor activity and a delay in the appearance of successive activity peaks following return to normal temperatures. Maximum delay occurs if chilling begins during inactivity and is about equal to the duration of the chill. At other times the delay is less than the duration of the chill.

2. It is postulated that an inhibitory factor is concerned in the control of the locomotor rhythm of Talitrus, and a model is proposed to describe its possible mode of action.

I am grateful to Professor E. Naylor and Mr R. G. James for much valuable discussion, to Professor E. W. Knight-Jones for the provision of laboratory facilities and to the Science Research Council for financial support.