1. In sugar-fed A. gambiae females, light may affect flight activity directly or by changing the phase of the circadian rhythm; both responses depend on the phase of the rhythm.

  2. The phase-response curve (1 h, 70 lux, signals given in the first cycle in DD following LD 12:12) shows a sharp swing, at about 3 h after normal light-off, from a maximum phase-delay to a maximum phase-advance, each of about 2 h. When signals are given at this time, phase re-setting is very variable; cyclical activity continues but the individuals are out of phase.

  3. Phase shifting appears to be a function of the energy of the signal. A 5 min, 70 lux signal has no apparent effect. The effect of a 1 h signal increases with intensity, up to at least 500 lux, but does not appear to be significant below 10 lux.

  4. Light normally inhibits flight activity, but there is a burst of activity at light-on (light-on response) if it occurs during the active half of the cycle following the initial activity peak. A vigorous light-on response occurs even at the lowest intensity used (0·3 lux).

The development of an automatic method of recording flight activity using the flight sound as an indicator (Jones, 1964; Jones, Hill & Hope, 1967) has made it possible to study, intensively, the activity of individual mosquitoes. Previous papers (Jones, Ford & Gillett, 1966; Jones et al. 1967) have indicated that flight activity in Anopheles (Cellia) gambiae Giles is controlled by an endogenous circadian rhythm which has a ‘free-running’ period in constant dark of about 23 h. Light inhibits activity and may re-set the rhythm by delaying it. Conversely, if the light period begins early, it has a phase-advancing effect.

This paper gives the results of a detailed study of the effect of light signals given at different phases of the activity cycle. Depending on its timing, light may have direct inhibitory or excitatory effects on activity and can effect both the period and phase of the underlying rhythm.

Mosquitoes

The mosquitoes belonged to species A of the A. gambiae complex (Davidson, 1964 a, b) and came from cultures which originated in Lagos and Ibadan respectively. The Ibadan culture was used in some experiments because of difficulties with the supply of the Lagos culture. The cycles of activity of adults from the two cultures appear to be almost identical. The culture used in each experiment is indicated in the results section.

The insects were reared in LD 12:12 (alternating 12 h light; 12 h dark) at 25 °C (RH 70–85 %) from eggs obtained each week from the Ross Institute of Tropical Hygiene. The recording experiments were started 1–4 days after emergence. Jones et al. (1967) assumed that, because of the presence of males in the culture, the females used in the experiments would have mated. In the present series of experiments, however, examination of the spermathecae showed that less than 10% had been inseminated. There was no noticeable difference in the activity of inseminated and non-inseminated females.

Experimental technique

The activity of individual, sugar-fed females was recorded at 25 °C using basically the same technique as used by Jones et al. (1967). An increase in the number of recording channels made it possible to record, simultaneously, the activity of up to 36 individuals in 5 different light regimes. Four ten-channel event recorders were used instead of the original kymograph apparatus.

Taylor & Jones (1969) found that if a tube containing water was placed in each mosquito chamber, in addition to the tube containing sugar solution, this considerably increased the survival time of Aedes aegypti females. This was adopted as standard practice in all experiments although there is no evidence that it is necessary with A. gambiae.

The light intensity was approximately 70 lux in all experiments unless otherwise stated. This was produced by two 15 W tungsten-filament bulbs. The light regimes were controlled with Venner time switches with the exception that signals lasting 20 min or less were controlled manually.

When the mosquitoes had been placed in the recording chambers, the LD 12:12 rearing regime was continued until it was certain that the insects were showing the normal pattern of activity for this regime; only then were the experiments started.

Interpretation of the record and treatment of results

An insect was given a score of 1 for any flight activity in any minute and thus a score between o and 30 for each. These individual scores were averaged and used to produce histograms of the mean activity against time. They were also used to calculate the period of each cycle of activity of each individual.

In order to show phase changes more clearly, the histograms show only the activity during the first 2 days after a change from the LD 12:12 rearing regime. The experiments were normally continued for several days longer, and information about the period of these later cycles is given in the tables and text.

The timing of any event (Th) is given relative to normal light-off (the time of light-off in the rearing regime) even if the light was switched off at a different time in that particular cycle. T is thus the ‘Zeitgeber time’ (Aschoff, Klotter & Wever, 1965) of the original rearing regime. Light-off was chosen as the arbitrary zero point as it normally determines the time of the main activity peak. In the LD 12:12 rearing regime, light-on is at T 12; in an LD 6:18 regime with light-off 6 h early, light-on is still at T 12.

Light-off early to DD (constant dark) or LD 6:18 (Fig. 1, Table 1)

When the regime is changed from LD 12:12 to DD (Fig. 1a) the period of the first activity cycle is and subsequent cycles have a period of . If light had no effect during the 6 h after normal light-on (T 12), a similar pattern of activity might be expected in LD 6:18 until the activity peak had ‘caught up’ with the new light-off. This does not appear to be the case; in LD 6:18 (Fig. 1 b) the period remains approximately constant at for at least the first four cycles; at first the light appears to advance the activity peak, but by the third cycle it is delaying it. The light also reduces individual variation in the timing of activity and thus the peaks continue to be well defined. There is a burst of activity at light-on (light-on reaction) even in the fourth cycle when light-on is about 14 h after the main activity peak.

Effect of a 1 h light signal at T 6 [Fig. 1 c, d; Table 1 ; Lagos culture)

When a signal is given in each cycle of the LD 6:18 regime, the period of the cycle shortens to 22- until the activity peak is re-entrained by the early light-off. There is a light-on reaction at the beginning of the 1 h light signal, but it is absent at T 12 (except in the second cycle where there appears to be a very small reaction by 4 individuals out of a total of 18).

When a single signal is given after a change to DD the period of the first cycle decreases, but the effect of the signal is not so great as when it is followed by a 6 h period of light.

Effect of shorter light signals at T 6

Five-minute light signals have no apparent effect on the period or phase of the activity cycle; in an LD 6:18 regime (Lagos culture) there were small light-on reactions to the 5 min signals (33 out of a possible 36) and the light-on reaction at T 12 was not lost completely (8 out of a possible 36). Preliminary experiments with 20 min light signals (Fig. 2, Ibadan culture) indicate that these may have a small effect on the period of the cycle which becomes more apparent when the signal is repeated in the second and third cycles; after the third signal the peak was approximately 3 h in advance of the control peak.

The phase-response curve (1 h light-signals)

Single 1 h light signals were given at different times in the first cycle after a change to DD (cycle 1). In the first series of experiments (Fig. 3, Lagos culture), the light was turned off 6 h before the normal time of light-off. This treatment was chosen because it was directly comparable with the early light-off treatment in the previous experiments. Jones et al. (1967) have shown that an early fight-off does not appear to re-set the rhythm in A. gambiae, but in order to check that this treatment did not affect the response to light signals a second series of experiments was carried out in which the change to DD was made at the normal time of light-off (T o). During this second series the Lagos culture became unobtainable and the experiments were continued with the Ibadan culture. Fig. 4 summarizes the experiments with the Ibadan culture.

The periods of the first and second cycles in both series of experiments are analysed in Table 2; in subsequent cycles the period is consistent with that of the controls and thus there do not appear to be any further phase changes.

The results from the two series of experiments (and from the two cultures) are so similar that they have been added together (Table 3) for the calculation of the mean phase shifts and for statistical analysis. Fig. 5 is the phase-response curve (Aschoff, 1965) obtained by plotting the phase shifts (relative to the control mean) at the end of cycles 1 and 2 against T, the time at which the signal was given in cycle 1. The situation at the end of cycle 2 may be considered as the new ‘steady state’. Table 4 gives the results of the statistical analysis of the phase change at the end of cycle 2 using Duncan’s new multiple F test (Duncan, 1955; Steel & Torrie, 1960); any two of the ranked means which are not underscored by the same line are significantly different (P < 0·01). At the P < 0·05 level the only difference is that the T 12 mean is no longer grouped with the T o and T 17 means.

A light signal just after the most active period (Fig. 3 b) inhibits activity and is followed by a phase delay of about 1 h. In this experiment the individual mosquitoes all reached a peak of activity before the onset of the signal; the period of cycle 1 was therefore measured from this first peak, although there was a second peak at the end of the signal.

A signal at T 2 is followed by a phase delay of about 2 h; this delay may be slightly decreased in the second cycle. At T 3 the rhythm appears to be at a critical point; a signal at this time is followed in cycle 1 by very variable phase changes, some individuals showing large advances, others large delays. In cycle 2, however, the period is shorter than in the controls, even in individuals which have been delayed in the first cycle (only three insects out of 24 had a period which was longer than the control mean - in two the period was in the other 24 h). It should be noted that individual rhythmicity is not lost although the lack of synchrony between individuals may give that impression in the records of mean activity.

A signal at T 6 is followed by a phase advance of about h which is complete in one cycle. A signal at T 12, the normal time of light-on, may be followed by a small phase-advance in cycle 1, but the period of cycle 2 is greater than in the controls and thus there is little effect on the ‘steady state’. Signals later in the cycle are followed by phase delays, which appear to take place in two steps.

Light-on reactions

Figs. 3 and 4 show that the mosquitoes give light-on reactions when the signals begin between T 2 and T 12. On each occasion at least 50 % of the individuals gave a definite reaction. There were no light-on reactions later in the cycle (the activity in Fig. 4(e) was not at light-on).

Effect of intensity

Preliminary experiments with 1 h light-signals at T 2 and T 6 indicate that both advancing and delaying phase-shifts increase with light intensity between 10 and 500 lux. Between 1 and 10 lux the phase-shifts are small (c) and are not statistically significant, but even at the lowest intensity used (0·3 lux) there is a vigorous light-on response at the beginning of the signal.

Light may affect flight activity directly or by changing the phase of the circadian rhythm ; both these responses depend on the phase of the rhythm.

Normally, light inhibits activity, but the inhibition may be preceded by a short burst of enhanced activity (the light-on response) if the light period begins during the 12–14 h following the main activity peak. In Fig. 1 (c) the absence of a light-on reaction at T 12 may indicate that the rhythm has already been re-set by the signal at T 6. A light-on reaction to the signal, however, does not necessarily indicate that the rhythm has been re-set by it; a short signal or a low-intensity signal may provoke a light-on reaction but may have no apparent effect on the phase of the rhythm.

A I h signal at T 6 plus 6 h of light beginning at T 12 appears to have almost as great an effect as a full 12 h period of light. From Fig. 5(b) of Jones et al. (1967) and from their data it appears that, when the LD 12:12 regime is advanced 6 h by shortening the dark period, the period of the first cycle is approximately 21 h; that is, 1 h shorter than in our experiment. The period of the second cycle is the same as in our experiment. The small difference in the results may be due to the higher light intensity used in the earlier experiments (100 lux). Pittendrigh (1965) has shown that, in the eclosion rhythm of Drosophila pseudoobscura, ‘ skeleton’ photoperiods with two 15 min (c. 1000 lux) light pulses are as effective as full photoperiods. In our experiments a 5 min (70 lux) light signal did not appear to affect the phase of the rhythm even when it formed part of a ‘skeleton’ photoperiod; possibly the light intensity was too low for the short signals to have any apparent effect.

Gillett, Corbet & Haddow (1961) found that, in a constant-dark regime, a 5 sec light flash induced cyclical oviposition in a large cage-colony of Aedes aegypti. In contrast Nayar & Sauerman (1971) found that in a similar regime, cyclical flight-activity in Ae. taeniorhynchus could not be induced by a single light period shorter than 12 h although the mosquito entrains to repeated 1 h signals in an LD 1:23 regime (in our experiments 20 min signals had a more apparent effect after 2 cycles). They also found that a light signal had no effect during the first 36 h after emergence.

It is unfortunate that the light intensity is not specified in either of these papers as it appears that the phase-setting effect of a light signal is determined both by its intensity and its duration. Pittendrigh (1960) found that very high intensity sec signals cause fairly large phase shifts in the eclosion rhythm of D. pseudoobscura.Hastings and Sweeney (1958) found that in Gonyaulax polyedra the phase shift was proportional to the duration of the light exposure, with a maximum after 2·5 h exposure; they also found a linear relation between phase shift and the intensity of a ·5 h light signal (up to 8000 lux). Finally, Engelmann (1969) has shown that, with light pulses between sec and 15 min, phase-shifting of eclosion in D. pseudoobscura is a function of the energy of the light pulse (up to a maximum in the region of 104 ergs/cm2). A 1 h light ‘pulse* has little effect on the phase of the A. gambiae rhythm until the energy is of the order of 104-106 ergs/cm2. The light-on reaction, however, may require only a small amount of light energy and may depend mainly on the sensitivity of the receptors.

The phase-response curve of A. gambiae females (Fig. 5) has a conventional shape (see Aschoff, 1965; Pittendrigh, 1965). Its amplitude is small compared with that of D. pseudoobscura, but this may be due to the light intensity used in the experiments ; the curves for the flying squirrel and hamster (De Coursey, 1961, 1964) are also of small amplitude, possibly for the same reason. In A. gambiae the phase changes appear to be complete within 2 cycles; the experiments were not continued for long enough to determine whether the light signals were followed by any permanent small changes in period, although the work of Lohmann (1967) and Engelmann (1969) has indicated that such changes may take place. It is interesting that the changes in the second cycle may be phase-advancing or phase-delaying and the pattern seems to indicate that after signals at T 2 and T 12 these changes may even be in the opposite direction to those in the first cycle.

Near T 3 there is a swing in the phase response curve from maximum phase-delay to maximum phase-advance. When signals are given at this time the phase changes are very variable, probably because of small individual variations in phase. Subsequently in these experiments the individual activity remained cyclical, although the individuals were no longer synchronized. Winfree (1970) has made a careful study of the effect of light pulses on the eclosion rhythm in D. pseudoobscura and has concluded that there is a critical combination of signal length and timing (the ‘singularity’) which ‘freezes’ the ‘clock’ in a phaseless non-oscillatory state. This state appears to be similar to the condition of the ‘clock’ in constant light. At the critical phase in the D. pseudoobscura rhythm the response is changing rapidly as it is near T 3 in our experiments; Winfree’s work indicates that it should be possible to ‘freeze’ the A. gambiae rhythm with a suitable signal near T 3, but this can be verified only by further work. As he points out, the ‘singularity ‘could be a useful tool for further experiments and may lead to a better understanding of the underlying mechanisms. With the eclosion rhythm it is difficult to discover whether the individuals have become arrhythmic or desynchronized ; Winfree has found that when the insects are given a second light signal the resultant peak is sharper than it would have been if the insects had merely been desynchronized. This point would be clearer if the rhythm could be monitored in each individual, and it would be useful to follow it up in A. gambiae.

Our results may help to explain the pattern of activity in LD 12:12 and during changes to other regimes. In LD 12:12 the first light advances the phase of the rhythm so that the next peak will be due just before light-off. The continuing light inhibits activity so that the insect cannot become active until after light-off; if the light period is continued after the expected time of activity, the rhythm is delayed and re-set. At the same time, the light appears to affect the timing of the second peak (see Figs. 3 i, j, 4 e, f) so that if the insect is now kept in constant dark the period of the first cycle will be before the rhythm ‘free-runs’ with a period of . This holds true even if the light is turned off 6 h early (at T 18). If the light is turned on again at T 12 (in LD 12:12) the period is shortened (cf.Fig. 4d) so that the peak now becomes due slightly before light-off ; thus it is delayed again by the light.

The pattern of activity in LD 12:12 appears to be the net result of the opposing effects of the first and last parts of the light period. The effect of changing the time of light-on (and the LD ratio) is considered in the next paper (Jones, Cubbin & Marsh, 1972).

  1. In sugar-fed A. gambiae females, light may affect flight activity directly or by changing the phase of the circadian rhythm; both responses depend on the phase of the rhythm.

  2. The phase-response curve (1 h, 70 lux, signals given in the first cycle in DD following LD 12:12) shows a sharp swing, at about 3 h after normal light-off, from a maximum phase-delay to a maximum phase-advance, each of about 2 h. When signals are given at this time, phase re-setting is very variable; cyclical activity continues but the individuals are out of phase.

  3. Phase shifting appears to be a function of the energy of the signal. A 5 min, 70 lux signal has no apparent effect. The effect of a 1 h signal increases with intensity, up to at least 500 lux, but does not appear to be significant below 10 lux.

  4. Light normally inhibits flight activity, but there is a burst of activity at light-on (light-on response) if it occurs during the active half of the cycle following the initial activity peak. A vigorous light-on response occurs even at the lowest intensity used (0·3 lux).

We are grateful to Professor J. D. Gillett for his help and advice and to Dr G. Davidson of the Ross Institute of Tropical Hygiene for supplying the mosquito eggs. This work was supported by a grant from the Ministry of Overseas Development (now the Overseas Development Administration).

Aschoff
,
J.
(
1965
).
Response curves in circadian periodicity
. In
J.
Aschoff
(ed.),
Circadian CLOCKS
PP.
95
111
.
Amsterdam
:
North Holland Publishing Co
.
Aschoff
,
J.
,
Klotter
,
K.
&
Wever
,
R.
(
1965
).
Circadian vocabulary
. In
J.
Aschoff
(ed.),
Circadian Clocks
, pp.
x
xix
Amsterdam
:
North Holland Publishing Co
.
Davidson
,
G.
(
1964a
).
The five mating types in the Anopheles gambiae complex
.
Riv. Malar
.
43
,
167
83
.
Davidson
,
G.
(
1964b
).
Anopheles gambiae, a complex of species
.
Bull. Wld Hlth Org
.
31
,
625
34
.
De Coursey
,
P. J.
(
1961
).
Effect of light on the circadian activity rhythm of the flying squirrel Glau-comys volons
.
Z. vergl. Physiol
.
44
,
331
54
.
De Coursey
,
P. J.
(
1964
).
Function of a Light response rhythm in hamsters
.
J. cell. comp. Physiol
.
63
,
189
96
.
Duncan
,
D. B.
(
1955
).
Multiple range and multiple F tests
.
Biometrics
11
,
1
43
.
Engelmann
,
W.
(
1969
).
Phase shifting of eclosion in Drosophila pseudoobscura as a function of the energy of the light-pulse
.
Z. vergl. Physiol
.
64
,
111
17
.
Gillett
,
J. D.
,
Corbet
,
P. S.
&
Haddow
,
A. J.
(
1961
).
Observations on the oviposition-cycle of Aedes (Stegomyia) aegypti (Linnaeus)
.
Ann. trop. Med. Parasit
.
55
,
427
31
.
Hastings
,
J. W.
&
Sweeney
,
B. M.
(
1958
).
A persistent diurnal rhythm of luminescence in Gonyaulax polyedra
.
Biol. Bull. mar. biol. Lab., Woods Hole
115
,
440
58
.
Jones
,
M. D. R.
(
1964
).
The automatic recording of mosquito activity
.
J. Insect Physiol
.
10
,
343
51
.
Jones
,
M. D. R.
,
Cubbin
,
C. M.
&
Marsh
,
D.
(
1972
).
Light-on effects and the question of bimodality in the circadian flight-activity of the mosquito Anopheles gambiae
.
J. exp. Biol
.
57
,
347
57
.
Jones
,
M. D. R.
,
Ford
,
M. G.
&
Gillett
,
J. D.
(
1966
).
Light-on and light-off effects on the circadian flight activity in the mosquito Anopheles gambiae
.
Nature, Lond
.
211
,
871
2
Jones
,
M. D. R.
,
Hill
,
M.
&
Hope
,
A. M.
(
1967
).
The circadian flight activity of the mosquito Anopheles gambiae-. phase setting by the light regime
.
J. exp. Biol
.
47
,
503
11
.
Lohmann
,
M.
(
1967
).
Phase-dependent changes of circadian frequency after light steps
.
Nature, Lond
.
313
,
196
7
.
Nayar
,
J. K.
&
Sauerman
,
D. M.
Jr
. (
1971
).
The effect of light regimes on the circadian rhythm of flight-activity in the mosquito Aedes taeniorhynchus
.
J. exp. Biol
.
54
,
745
56
.
Pittendrigh
,
C. S.
(
1960
).
Circadian rhythms and the circadian organisation of living systems
.
Cold Spring Harb. symp. quant. Biol
,
25
,
159
84
.
Pittendrigh
,
C. S.
(
1965
).
On the mechanism of the entrainment of a circadian rhythm by light cycles
. In
J.
Aschoff
(ed.),
Circadian Clocks
, pp.
277
97
.
Amsterdam
:
North Holland Publishing Co
.
Steel
,
R. G. D.
&
Torrie
,
J. H.
(
1960
).
Principles and Procedures of Statistics
, pp.
107
9
.
New York
:
McGraw-Hill
.
Taylor
,
B.
&
Jones
,
M. D. R.
(
1969
).
The circadian rhythm of flight activity in the mosquito Aedes aegypti (L) : The phase-setting effects of light-on and light-off
.
J. exp. Biol
.
51
,
59
70
.
Winfrkb
,
A. T.
(
1970
).
Integrated view of resetting a circadian clock
.
J. theor. Biol
.
25
,
337
74
.