The Effect of Photoperiod on the Developmental Rate of Drosophila Pupae

It has been shown in an earlier study (Harker, 1965) that the development of the pupae of Drosophila melanogaster can be divided into three stages, and that the rate of development of pupae kept in a 12 hr. light: 12 hr. darkness cycle is affected by the time of day at which each of these stages is entered. Furthermore, even when the pupae are kept in continuous darkness the duration of each stage is affected by the time of day at which the stage is entered, relative to the light cycle to which they had been exposed as larvae. The carry-over to the pupal stage of the timing of events occurring during the larval stage must involve a timing mechanism of the type involved in diurnal rhythmicity.

The results obtained in this first study do not indicate whether either, or both, of the light-off and the light-on signals are involved in controlling the timing of the events which determine the developmental rate, or whether the photofraction itself is important in determining the rate. A further study has now been made using two other photofractions in a 24 hr. light : darkness cycle.

A mutant of Drosophila melanogaster, strain ‘straw’, obtained from the Department of Genetics, University of Cambridge, was used in all the following experiments. Cultures of this strain show one major peak of adult eclosion in a 12 hr. fight: 12 hr. darkness cycle.

The methods of culture and of observation were those used in the previous study, and again the following three stages of development were recognized: (a) stage beginning with head eversion and ending with the appearance of yellow eye-pigmentation ; (b) stage beginning with the appearance of yellow eye-pigmentation and ending with the beginning of pigmentation of the costal region of the wing; (c) stage beginning with the pigmentation of the costal region of the wing and ending with the eclosion of the adult.

Larvae and pupae were treated in one of five different ways :

1. Larvae and pupae were kept in a 4 hr. bright light: 20 hr. dim light cycle from the time of hatching until the eclosion of the adult.

2. Larvae were kept in a 4 hr. bright light:20 hr. dim light cycle until the last larval moult and were then transferred to constant darkness.

3. Larvae and pupae were kept in a 4 hr. light : 20 hr. complete darkness cycle.

4. Larvae and pupae were kept in an 18 hr. bright light:6 hr. dim light cycle from the time of hatching until the eclosion of the adult.

5. Larvae were kept in an 18 hr. bright light:6 hr. dim light cycle until the last larval moult and were then transferred to constant darkness.

In all cases observations were made at hourly intervals. The pupae were observed in situ when the lights were on; when the lights were off samples were withdrawn, without light being admitted to the remaining samples, and the withdrawn samples were discarded after examination.

Temperature was maintained at 26° C. throughout the experiments.

Precautions against bias were taken as described in the previous study.

The time taken to complete each stage was measured for pupae entering the stage at each hour of the day ; for each hour of entry at least forty-five pupae were included in a sample, except in the case of condition (3) above, in which samples of only ten were used, this environment being used only to check the previous observation that the same results are obtained in a light : darkness cycle as in a bright fight: dim light cycle.

The results obtained when both larvae and pupae were kept in 4 hr. bright light: 20 hr. dim light are summarized in Fig. 1 ; in this figure the time of day at which each sample entered the stage is plotted against the number of hours taken to complete the stage. It will be noticed that within nearly all samples two distinct developmental periods occur, but that the variation about these peaks is limited to ± 1 hr. (except in the case of the sample entering the stage 1 hr. after the beginning of darkness, which will be discussed later.)

As explained in the previous paper (Harker, 1965) the appearance of two distinct developmental periods can be related to the fact that observations were made only once an hour, for unless the time of observation happened to coincide precisely with the time of a change in developmental rate, the pupae observed as having reached the beginning of a stage at any one hour might include pupae showing the developmental rate typical of the previous hour. The variation of 1 hr. about the peaks can also be predicted as a consequence of the timing of observations.

In the case of the sample entering the stage one hour after beginning of darkness the developmental period of five of the seventy-five pupae observed fell outside the limits of the rest of the sample. These pupae are represented as dots in the Fig. 1. In a 12 hr. light: 12 hr. darkness cycle it was previously found that samples entering this stage 5 hr. after light on (= line D ±1 in the present photoperiod) also showed a wider variation in their developmental period than the ± 1 hr. common to the rest of the samples.

In Fig. 2a the results are replotted to show the developmental periods in relation not only to the time of the beginning of the stage, but also to the light cycle to which pupae were exposed throughout the stage : only the major peaks are marked for each sample.

It is quite clear from the results that the duration of the stage beginning with head eversion and ending with the appearance of yellow eye-coloration is affected by the time of day, relative to the 4 hr. bright light:20 hr. dim light cycle, at which head eversion occurs. For comparison the results for a 12 hr. bright light: 12 hr. dim light cycle are shown in Fig. 2 ft. There does not seem to be any correlation, in either case, between the length of the developmental period and the number of bright light : dim light cycles, or to any sequence of these cycles, but only to the light conditions during the first hour or two of the stage.

The developmental periods shown in Fig. 2a and 2b show some points of similarity, but the photoperiod clearly affects the form of the curve, as will be discussed in detail later.

When only the larval stage received the 4 hr. bright light : 20 hr. dim light cycle, the pupae being maintained in constant darkness, identical results were obtained to those when the light cycle was continued throughout pupal development. That is to say, the rate of development is affected by the time of day at which the stage is entered, relative to the light cycle to which the pupae were exposed during the larval stage, provided always that no new light cycle has been introduced.

It is of particular interest that after exposure to 4 hr. bright light : 20 hr. dim light the developmental rates do not revert in constant darkness to those shown in constant darkness after exposure to a 12 hr. bright light: 12 hr. dim light cycle (Harker, 1965).

Pupae kept in a 4 hr. light:20 hr. complete darkness cycle showed identical rates of development to those shown by pupae kept in a 4 hr. bright light : 20 hr. dim fight cycle.

The results obtained when both larvae and pupae were kept in the 4 hr. bright light :2o hr. dim light cycle are summarized in Fig. 3 a and b. Again two distinct developmental periods occur within each sample entering a stage at a particular time of day; the minor peak is again assumed to represent the period typical of the pupae entering the stage in the preceding hour and is not shown in these figures. The variation about the major peak is limited to ± 1 hr.

The range of developmental periods for both stages follow curves which are identical in form with each other, and with that for the stage ‘head eversion to yellow eye’.

When only the larval stage was kept in the light cycle, the pupae being maintained in constant darkness, the developmental periods shown by these two stages differed in no way from those obtained when the light cycle was maintained throughout the pupal stage.

Fig. 4 shows the developmental periods, throughout the entire development, of pupae in which head eversion began at successively later hours of the day in a 4 hr. bright light:20 hr. dim fight cycle.

If the number of days from pupation is ignored, and only the time of day at which eclosion occurs is considered (Fig. 5,a), it can be seen that the majority of flies emerge at the end of the dim light period, and not, as in a 12 hr. bright light: 12 hr. dim light cycle (Fig. 5,b), at the beginning of the bright-light period. In a breeding culture the rhythm of eclosion shown in Fig. 5 would be apparent, since there would be an overlap in the time of eclosion in individuals which reached the prepupal stage on successive days. Breeding cultures do in fact show just such a rhythm of emergence.

The results obtained when both larvae and pupae were kept in the light cycle are summarized in Fig. 6. The variation is not shown in the figure, but again two distinct developmental peaks occur within each sample entering a stage at a particular time of day, and the minor peak is here again assumed to be that characteristic of the previous hour. The variation about the major peak is 1 hr.

The developmental periods differ from those obtained in the 12 hr. bright light: 12 hr. dim fight and the 4 hr. bright light:20 hr. dim-light cycle.

When only the larvae were kept in the light cycle, the pupae being kept in constant darkness, the results were identical with those obtained when the light cycle was maintained throughout the pupal development.

The results obtained when both larvae and pupae were kept in the 18 hr. bright light:6 hr. dim light cycle are summarized in Fig. 7 a and b. The variation is not shown in these figures, but is the same as that previously described.

The range of developmental periods for both stages follow curves which are identical in form with each other, and with that of the stage ‘head eversion to yellow eye’.

When the larval stage only was kept in the light cycle, the pupae being maintained in constant darkness, the developmental periods shown by these two stages differed in no way from those obtained when the light cycle was maintained throughout the pupal stage.

Fig. 8 shows the developmental periods, throughout the entire development, of pupae when head eversion began at successively later hours of the day in an 18 hr. bright light : 6 hr. dim fight cycle.

If the number of days from pupation is ignored, and only the time of day at which eclosion occurs is considered (Fig. 9), it can be seen that the majority of flies emerge either immediately after dawn or 9 hr. after dawn ; that is there are dual peaks of emergence and not the single peak of emergence shown in other photoperiods. In a breeding culture, as explained above, the overlap in the time of eclosion of individuals which reach the prepupal stage on successive days would produce such a rhythm of eclosion.

Time-interval curves for the stage ‘head eversion to yellow eye’, when pupae are in cycles of 4 hr. fight:2ohr. darkness, 12 hr. light: 12 hr. darkness and 18 hr. light: 6 hr. darkness, are shown in Fig. 10. These curves have been obtained, for each particular light cycle, by calculating the developmental period of each sample as a percentage of the mean period of all samples, and plotting this against the time at which the sample entered the stage.

The form of the curve in the hours after the light is turned on is the same in all three photoperiods, as can be seen more clearly in Fig. 11, in which the time-interval for those samples entering the stage during the light period is plotted against hours since light-on. This means that for pupae entering a stage in the light the operative factor is the time since the light-on signal, and the preceding dark period has no effect.

In Fig. 12b the time-interval for those samples which entered the stage during the dark period is plotted against hours since light-off. Here it can be seen that the fight-off signal has a marked effect on the form of the curve for at least some hours, and a lesser but still positive effect until the next light signal. The curves, however, are not identical in form as they are in the case of the light curves. It is possible that the dark curve may be affected by the previous light signal, but there does not seem to be any simple relationship between the effect of hght-on and light-off on the form of the darkness curve.This can be seen from Fig. 13 in which is plotted the difference between the curve in the dark part of the cycle (4–18 hr. and 12–18 hr., respectively, from the light-on signal) and the corresponding curve if the light had not been turned off.

It is worth re-emphasizing at this point that the ordinates on the time-interval curves represent the whole of the development during the stage, and that the abscissae represent the time of only the beginning of the stage relative to the light : darkness cycle ; during the total development the light may have gone off or on several times. Furthermore, the same time-interval curves are obtained in constant darkness, the timing of the beginning of the stage then being related to the subjective time—that is, the time relative to the light cycle to which the insect was exposed as a larva. In view of the maintainance of the form of the time-interval curves in darkness, it was previously concluded (Harker, 1965) that the rate of development must be affected by some factor, or factors, following a diurnal rhythm, a rhythm whose form once set is maintained in constant darkness. It can now be seen that both light-on and light-off signals determine the form of the rhythm, but that each of these signals in itself must act on the ‘clock’ system independently of the other; for although the rhythm set by light-on may effect the magnitude of response of the rhythm set by light-off, the timing of the latter is independent of the rhythm set by light-on.

The time of eclosion has been shown here, and earlier, to be determined by the rate of development of the pupal stage. Pittendrigh (1960) has drawn attention to the effect of photoperiod on the phase-setting of the rhythm of eclosion of D. pseudoobscura ; he concludes that light is effective in phase-setting only as it goes on or off. The results described in this paper support this view.

Furthermore, in repeated light cycles, with the same light fractions as used here, the cultures of D. melanogaster used by Pittendrigh (1961) show similar phase-settings of the eclosion rhythm to those shown in Figs. 5 and 9. On the other hand, in the experiments described in the present paper, as has been emphasized, no change in the form of the rhythm (or phase-setting) has been found to occur in constant darkness ; Pittendrigh’s (1960) results do not agree on this point, and therefore must be considered in some detail.

Pittendrigh raised nineteen separate cultures of D. pseudoobscura in a 12 hr. light: 12 hr. darkness cycle, and eventually left them in continuous darkness. On the last day of the light cycle the light was switched off at different times for each culture, so that each received a different final photoperiod. After some time in constant darkness the phase-setting of the eclosion rhythm shifted in those cultures which had received more than 12 hr. light on the final day of the light cycle, and this shift was such that the maximum eclosion occurred 15 hr. after the time of subjective darkness (i.e. n days +15 hr. after the time of the last light-off signal). This phase-setting he terms the ‘steady state’. It is not clear from the published description how long the insects were left in constant darkness before the phase-shift took place, but in reference to an earlier experiment (Pittendrigh & Bruce, 1957) he states that the ‘steady state’ has not been reached by the end of 5 days. The species of Drosophila was not named in the latter experiment, but it is presumed to be D. pseudoobscura, since in D. melanogaster the entire development from egg to adult takes only 8 days at 25 ° C., and Pittendrigh has stated (1960, p. 183) that only the pupae are concerned.

Pittendrigh’s results should be considered in terms of the length of the period of pupal development, which in D. pseudoobscura covers about 7 days at 25° C. (see also Cockrane, 1937). If the steady state has not been reached at the end of 5 days then the phase-shift must be occurring as the last of the pupae are eclosing. As has been previously mentioned, a rhythm of eclosion can arise because of the overlap in the times of eclosion of individuals which reach the prepupal stage on successive days but develop at different rates. If, however, there are no more pupae being added to the culture there will come a time when all the rapid developers have emerged and only the slow developers will contribute to the emergence ‘peak’. Although these times of emergence have previously been represented in the total emergence, their numbers will be insignificant while an overlap in the days of emergence continues ; once such an overlap ceases there will be an apparent phase-shift in the time of emergence.

The evidence in the present paper suggests that not only the last seen light: darkness transition, but also the last seen darkness: light transition determines the time of eclosion of the adult.

This research was supported in part by the Air Force Office of Scientific Research under grant AF EOAR 64-14 through the European Office of Aerospace Research (OAR), United States Air Force.