INTRODUCTION

It has been known that in most species, the free-running period (τ) of circadian oscillation depends on the light intensity. Aschoff (1960, 1981) stated that, τ shortens with an increase in light intensity in many diurnal species and lengthens in nocturnal species. In general, this Aschoff's rule is supported by data from many vertebrates (see Daan and Pittendrigh, 1976; Aschoff, 1981). However, many insects appear to violate Aschoff's rule (Saunders, 1982). For example, in Drosophila melanogaster (Konopka et al., 1989) and Calliphora vicina (Hong and Saunders, 1994) both of which are day-active, τ increases with light intensity and become arrhythmic in light intensity above a certain threshold.

In LL_dim (light-dim light) cycles, the light intensity during light period or dim-light period influences the phase relationship between the circadian rhythm and LL_dim cycles (Ψ_RL) (Aschoff, 1965). In several species of bird (Aschoff, 1965), for example, the phase of onset of activity advanced with increasing light intensity in L or in L_dim. This aspect of circadian rhythms has seldom been investigated in insects.

In the onion fly, Delia antiqua, the phase of locomotor activity in LD (light-dark) cycles occurs progressively later with age (Watari and Arai, 1997). The free-running period τ in constant darkness (DD) also changes with age, being shorter than 24 hr until 14–20 days after adult eclosion but thereafter longer than 24 hr. This suggests that the age dependent Ψ_RL change in D. antiqua would be attributed to the increase in τ. By using D₂O, this hypothesis was tested experimentally (Watari and Arai, 1998). In the present study, we examined the relationship between Ψ_RL and τ in D. antiqua by making the light intensity of LD cycles and constant dim light (L_dimL_dim) 1.0 lux.

MATERIALS AND METHODS

RESULTS

Effect of dim light in scotophase on locomotor activity

Fig. 1 shows typical locomotor activity records at photo-periods ranging from LL_dim 10:14 to 16:8 at 25°C in which the light intensity of the scotophase was 1.0 lux. Adults of D. antiqua showed trimodal activity patterns with major and light-on and light-off peaks. The major peak (φ_A) advanced as the photophase became longer (Fig. 2), like in LD cycles (Watari and Arai, 1997). φ_A was significantly delayed with age (p <0.01 at LL_dim 10:14 and 14:10 and p <0.05 at LL_dim 12:12 and 16:8, regression analysis). Similar delay has been also observed in LD cycles (Watari and Arai, 1997). The delay in LL_dim was larger than in LD for the first 1 to 3 weeks (to day 12 at LL_dim 14:10) after eclosion but thereafter became almost the same as that at LD cycles.

Fig. 1

Examples of double-plotted locomotor activity rhythm in D. antiqua in different LL_dim cycles at 25°C. A, LL_dim 10:14. B, LL_dim 12:12. C, LL_dim 14:10. D, LL_dim 16:8. Open and dotted bars at the top indicate the photoperiod.

Fig. 2

Changes in φ_A as given by the time before dusk in D. antiqua given in different LD (data from Watari and Arai, 1997) and LL_dim cycles at 25°C. Open circles, φ_A at LD cycle; closed circles, φ_A at LL_dim cycles. Regression lines for φ_A (Y) on time (X, days) are Y = −0.018X+2.891 (d.f. = 214, P <0.01) at LD 10:14, Y = −0.042X+3.118 (d.f. = 219, P <0.01) at LL_dim 10:14, Y = −0.024X+3.971 (d.f. = 361, P <0.01) at LD 12:12, Y = −0.013X+3.467 (d.f. = 279, P <0.05) at LL_dim 12:12, Y = −0.022X+4.178 (d.f. = 305, P <0.01) at LL_dim 14:10, Y = −0.023X+5.139 (d.f. = 259, P <0.05) at LD 16:8 and Y = −0.020X+4.730 (d.f. = 249) at LL_dim 12:12. Vertical lines indicate standard errors.

At LL_dim 4:20 (n = 5), the activity after lights-drop increased from about day 10 after eclosion, and in two flies the main activity moved from the photophase to the dim-light phase (Fig. 3).

Fig. 3

Example of activity change with age in D. antiqua at LL_dim 4:20 at 25°C.

Free-running rhythm in constant dim light conditions

Fig. 4 shows typical locomotor activity record of D. antiqua in L_dimL_dim at 25°C. The free-running rhythm varied with age as in DD (Watari and Arai, 1997); τ was shorter than 24 hr until day 15 after eclosion and thereafter became longer than 24 hr. Similar tendencies were observed in other flies; τ changed with age (Fig. 5). The increase in τ with age was statistically significant in DD and in L_dimL_dim (p <0.01, regression analysis). In L_dimL_dim, τ was greater than that in DD through the adult's life and the difference became larger with age (p <0.01, t-test for comparison of regression slopes).

Fig. 4

Example of change in free-running rhythm in L_dimL_dim (1.0 lux) at 25°C in D. antiqua adults transferred soon after eclosion from LL (400 lux). Dotted bars at the top indicate the continuous dim light (L_dimL_dim).

Fig. 5

Change in τ in DD (data from Watari and Arai, 1997) and L_dimL_dim at 25°C in D. antiqua adults transferred soon after eclosion from LL. Open circles, τ in DD; closed circles, τ in L_dimL_dim. Regression lines for τ (Y, h) on time (X, days) are Y = 0.071X+22.236 (d.f. = 458) in DD and Y = 0.093X+22.331 (d.f. = 347) in L_dimL_dim. Both significant at P <0.01. Vertical lines indicate standard errors. The difference between τ in L_dimL_dim and that in DD became larger with age (p <0.01, t-test for comparison of regression slopes).

Effect of dim light in photophase on locomotor activity

Fig. 6 exemplifies locomotor activity records of two flies at L_dimD 12:12. The locomotor activity rhythm free-ran until day 21 (Fig. 6A) or 15 (Fig. 6B) after eclosion and thereafter was entrained with the major peak fixed to the later part of the photophase. Though the light-on peak persisted through the adult's life, the light-off peak did not appear. Similar tendencies were also observed in other flies; the locomotor activity rhythms of 5 of 7 flies free-ran with each period until about 2–3 weeks after eclosion and thereafter was fixed to the later part of the photophase. As a result, standard deviation of the locomotor activity peak became smaller from week 2 after eclosion (Fig. 7).

Fig. 6

Examples of change in activity rhythm at L_dimD 12:12 in D. antiqua with age at 25°C. Left panels: actograph records. Right panels: change in the activity peak (squares) with age.

Fig. 7

Change in the activity peak (squares) with age at L_dimD 12:12 in D. antiqua at 25°C. Vertical lines indicate standard deviations.

DISCUSSION

In LD cycles, Delia antiqua shows a major peak of locomotor activity in the late photophase and also bursts of activity induced by lights-on or lights-off. Only the major peak (φ_A) is controlled by the circadian pacemaker and the phase of φ_A occurs progressively later with age and this change would be attributable to the increase in τ (Watari and Arai, 1997). Also in LL_dim cycles, the fly showed trimodal activity patterns with major and light-on and light-off peaks (Fig. 1) and φ_A was delayed with age (Fig. 2). This suggests that the period of 1.0 lux exerts the same effect as a dark period (0 lux) when combined with a period of 400 lux. However, the phase of the major peak in LL_dim slightly delayed as compared with that in LD cycles until about week 3 after eclosion (Fig. 2). This observation implies that the light intensity of 1.0 lux altered the velocity of the pacemaker, because τ in L_dimL_dim became longer than that in DD (Fig. 5).

When the light intensity of the photophase was 1.0 lux (L_dimD 12:12), in D. antiqua, the locomotor activity rhythm freeran until about 2 weeks after eclosion (Fig. 7). This phenomenon is similar to relative coordination. In the presence of a weak zeitgeber such as a small difference in light intensity or a decrease in light sensitivity, external relative coordination would develop (Enright, 1981; Aschoff, 1981; Honma, 1991). Drosophila melanogaster could be entrained to L_dimD cycle even when the light intensity of dim light was 0.1 lux (Matsumoto et al., 1994). Therefore, it seems that the light sensitivity of D. antiqua is weaker than that of D. melanogaster. In D. antiqua, however, the locomotor activity rhythm was fixed to the later part of the photophase in L_dimD 12:12 from about 2 weeks after eclosion; it could be entrained in L_dimD. Therefore, the light sensitivity of D. antiqua may be enhanced with age. At LL_dim cycles, the major peak remained almost the same as that at LD cycles after about week 3 (Fig. 2). This might be caused by the change in light sensitivity with age. The wild type and short period mutant, per^S, of D. melanogaster was entrained to a short photoperiod LD 4:20 but long period mutant, per^L was not (Tomioka et al., 1997). Based on this fact, it has been suggested that the entrainability of per^L mutant is weaker than the other two strains and these results are consistent with the hypothesis proposed by Saunders et al. (1994) that the sensitivity to light is lower in per^L flies than in per^S flies (Tomioka et al., 1997). At LD 4:20 (Watari and Arai, 1997) and LL_dim 4:20 (Fig. 3), D. antiqua could be entrained but the activity after lights-off increased with age. These results are contradictory to the hypothesis that the light sensitivity of D. antiqua may become higher with age. However, the entrainment to short photoperiod may be predicted by the phase response curve (PRC), as shown in D. melanogaster (Saunders et al., 1994); the steady state phase relationships of per^S flies to LD 1:23 and LD 6:18 showed entrainment to the LD cycles with a pronounced phase lead as predicted by the PRC. In D. antiqua, however, we have not obtained PRC. Further experiments are necessary to determine whether or not the light sensitivity of D. antiqua changes with age.