The Relationship Between the Diurnal Light Cycle and the Time of Ovulation in Mice

It is now well established that in rats and mice the events of the oestrous cycle are closely linked with the diurnal rhythm of light and darkness. If the times of light and darkness are reversed a corresponding change in the times of oestrus and ovulation occurs (Hemmingsen & Krarup, 1937; Browman, 1937; Snell, Fekete, Hummel & Law, 1940; Snell, Hummel & Abelmann, 1944; Austin & Braden, 1954; Braden & Austin, 1954). There is good evidence that the effect of the diurnal light cycle on the times of oestrus and ovulation is mediated by the eyes, the central nervous system and the anterior pituitary gland (Hill & Parkes, 1933 ; Bissonette, 1938 ; Clark, McKeown & Zuckerman, 1939; Everett, Sawyer & Markee, 1949; and others).

The experiments to be described in this paper were designed to investigate the effect on the time of ovulation of changes in the relative lengths of the light and dark phases of the 24 hr. cycle. These changes occur during the course of the year in non-equatorial latitudes such as that of Edinburgh (latitude 56° N), where studies were made on mice kept under natural lighting conditions. Investigations were also made on mice transferred to a room in which the times of light and dark were reversed. Differences found between three outbred stocks of mice have indicated that response to the diurnal light cycle by the mouse is probably determined to some extent at least by the genetic constitution of the animals.

In the first experiment (December 1954), female mice from about eight out-bred strains were used and no separate record was kept of the results from each strain, but in subsequent experiments random-bred females from the stocks L, C and PCT were used, and the results from each strain recorded separately. The L stock was that employed by Dr N. Bateman in this laboratory for genetical studies on lactation, the C stock was that used by Dr D. S. Falconer for studies of the inheritance of body size, and the PCT stock was obtained from Dr T. C. Carter.

In the experiments carried out under the natural diurnal light cycle, one male was added to each cage of three to five females at 5-6 p.m. and the females were inspected for the presence of vaginal plugs during the 20 min. preceding 10 p.m., 12 midnight, 2 a.m., 4 a.m., 6 a.m., 8 a.m., 10 a.m. and 2 p.m. (All the times quoted are Greenwich Mean Time (G.M.T.).) When a mated female was found in any cage, the male was replaced by a fresh male. Males of L strain were used throughout the experiments. Mated females were killed at each of the abovementioned hours, and the eggs that had been shed were recovered from the Fallopian tubes by dissection under physiological saline solution. The eggs were examined in the fresh state by phase-contrast microscopy. When no eggs could be found in the Fallopian tubes, the ovaries were inspected to ensure that ripe follicles were present, and often the eggs obtained by puncturing these follicles were examined by phase-contrast microscopy. The arrival of the eggs in the distended portion of the Fallopian tube has been taken as synchronous with ovulation; the time required for the egg to travel from the ovary to the distended portion of the ampulla—the site of fertilization—is almost certainly relatively short.

Studies of the time relations of ovulation in mice maintained under a reversed diurnal light cycle were also made. The mice were kept in a room in which the sole source of light was an 80 W. fluorescent strip-lamp which was switched on and off at certain set times by means of an electric clock relay. Two 24 hr. cycles were used : in the first the dark phase was of 10 hr. duration (5 a.m. to 3 p.m.) and in the second it was of 4 hr. duration (8 a.m. to 12 noon). The females were subjected to an acclimatization period of 3-5 weeks at the beginning of each experiment. A male was then introduced into each cage of three to five females at 8.30 a.m. and females that were found to have vaginal plugs were killed at either 9 a.m., 10 a.m., 11 a.m., 12 noon, 1 p.m., 2 p.m., 3 p.m., 4 p.m., 5 p.m., 6 p.m., or 8 p.m. The occurrence of ovulation was ascertained as described above.

The results were analysed by the probit method. The abscissae (times in hours) were not transformed.

In the December 1954 experiment, a total of 137 females of mixed origin was used. Eight mice found to have mated before 5 p.m. were killed at 5 p.m., but ovulation had not begun (Table 1). The next time of killing was to p.m. when one out of eleven mice was found to have ovulated. There-after, the percentage of mice that had ovulated gradually increased until 100% was reached between 10 a.m. and 2 p.m. (see Fig. 1). In the mice as a group, therefore, ovulation took place over a period of 14-16 hr. It should be remarked that, as coitus and ovulation proceeded to some extent concurrently in the population (the mean interval between the two events was 5 hr.), the mice killed at the earlier hours (i.e. 10 p.m. and 12 midnight) could not have been an entirely representative sample of the population, for they would have necessarily included only the mice that had mated before the time of killing. If, as is likely, the mice that mate early are the ones that ovulate early in the period, then the results obtained at 10 p.m. and 12 midnight would probably have been a little too high and those obtained at later hours perhaps a little low. However, as this bias was common to all the sets of data obtained, it can be neglected in comparisons between the results recorded under different light cycles or for different strains.

In June 1955 the time of ovulation was studied in females of L and C stocks. Ovulation began in L stock females at about 11 p.m. and was complete by about 8 a.m.—a total period of approximately 9 hr. The detailed results are set out in Table 2 and illustrated graphically in Fig. 2. By probit analysis of these results the slope of the probit line relating the transformed percentage of mice ovulated with time was found to be 0-42 ±0-07. It was estimated that the time by which 50% of the mice would have ovulated was 4-16 a.m. + 7 min. In the group of females from C stock ovulation began at about 12 midnight and was complete by about 9 a.m. (see Table 2 and Fig. 3). Ovulation thus required a period of 9 hr. in this group. The slope of the probit line was 0·405 ± 0·079, and the time by which 50% of the mice would have ovulated was 5.00 a.m. + 6 min. The slopes of the probit lines for the two groups of mice clearly did not differ significantly, but there was a significant difference between the mid-points of the ovulation periods for the two stocks.

In December 1955 ovulation began in a group of ninety-nine females from L stock at about 11 p.m., but was not complete until about midday—a total period of about 13 hr. (see Fig. 2). The detailed results are given in Table 3. In thirty-eight C stock females killed at 10 p.m., 2 p.m., 6 a.m., and 10 a.m., ovulation had begun before 10 p.m. and was not quite complete at 10 a.m. the next morning (see Table 3 and Fig. 3). The period required for ovulation in the C stock population in December may be estimated as 13·14 hr. The slopes of the probit lines for the two groups were 0-286 + 0·048 (L stock) and 0-235 + 0-067 (C stock). The mid-points of the ovulation periods were 5.11 a.m. ±9 min. and 2.54 a.m. ± 71 min., respectively. Neither the slopes nor the mid-points differed significantly ⁠, for the mid-points).

In both the June and December (1955) experiments records were kept of the number of females that copulated before 10 p.m., between 10 p.m. and midnight, between midnight and 2 a.m., etc. The results from L and C strains were combined. In the June experiment 63% of the females were found to have vaginal plugs before midnight, 75% before 2 a.m. and 98% before 8 a.m. From the graphs shown in Figs. 2 and 3 it may be estimated that 63 % of the mice would have ovulated by 6 a.m. and 75 % by 6.45 a.m., so that the mean mating-ovulation interval may be stated as approximately 5 hr. In the December experiment 30 % of the females had copulated before 10 p.m., 60% had copulated before midnight, 75% had copulated before 2 a.m., and 95 % had copulated before 6 a.m. By comparing these results with the ovulation curves for L and C stock females in December (Figs. 3, 4), the mean mating-ovulation interval was again estimated as about 5 hr. Snell et al. (1940) considered that the mean interval between mating and ovulation was probably not in excess of 2-3 hr. The difference between their conclusion and the present finding may be related to a difference between the strains employed or to the small number of animals used by Snell. The magnitude of the interval between coitus and ovulation will depend to some extent on the experimental conditions. For instance, if a male has been used on a number of consecutive nights, it may not copulate with a female until several hours after the onset of oestrus.

In females of the PCT stock kept under the 10 hr. dark/14 hr. fight cycle, ovulation began before 9 a.m. (probably between 7 a.m. and 8 a.m.) and was complete by 5 p.m. (Table 4). The total period over which ovulation occurred in this group of mice was thus about 9-10 hr. (curve P, Fig. 4). In C stock females, on the other hand, ovulation under the 10 hr. dark/14 hr. light régime appeared to occupy a much longer period. Five of the ten (C stock mice) killed at 9 a.m. had already ovulated and one of the ten killed at 5 p.m. had not (Table 4). Probit analysis of the results from the PCT females indicated that the mid-point of the ovulation period was 11·45 a.m. + 10 min. and that the slope of the probit line was 0-378 ± 0·07.

Ovulation began at about 8 a.m. in L stock females maintained under the 10 hr. dark/14 hr. light cycle. It was still in progress at the time of the last examination (5 p.m.). The results are set out in detail in Table 5 and are illustrated graphically in Fig. 4 (curve L₁₀). The slope of the probit line in this instance was 0·268 ± 0·063, and the mid-point of the ovulation period was estimated as 2·25 p.m. ± 15 min. The slope did not differ significantly from that found for ovulation in PCT stock mice, but there was a very significant difference (of 2 hr. 40 min.) between the mid-points of the ovulation periods, as may be seen from inspection of Fig. 4.

When L stock females were kept under the 4 hr. dark/20 hr. light régime—the mid-point of the dark phase being the same as in the 10 hr. dark/14 hr. light cycle—ovulation began at about 12 noon and reached completion at about 8 p.m. (Table 5). The slope of the probit line was found to be 0·42 ± 0·08, and the time by which 50% of the mice would have ovulated was 3·30 p.m. + 18 min. This slope did not differ significantly from that derived for L strain mice kept under the to hr. dark/14 br. light cycle. However, there was a significant difference between the mid-points of the ovulation periods of L strain mice under the two diurnal light cycles

In the animals kept under the natural diurnal light cycles the length of the dark phase was approximately $612$ hr. during the June experiment and $1512$ hr. during the December experiment. These values have been calculated from the times of sunrise and sunset in those months, except in the December experiment when artificial lighting was employed in the animal house until about 5·20 p.m. There are thus three sets of data in which to compare ovulation under a diurnal cycle involving a long dark phase and under a cycle with a short dark phase ⁠.

These are the results from C stock (June and December 1955), L stock (June and December 1955), and L stock (4 hr./20 hr. cycle and 10 hr./14 hr. cycle). The rate of increase in the proportion of females that had ovulated in the populations studied is indicated by the slopes of the probit lines which were :

When these three sets of data were separately compared (see above) the differences between the slopes were not statistically significant. However, when the data were combined, the difference between the slopes obtained under cycles with a short dark phase and those obtained under cycles with a long dark phase was significant ⁠. It may be concluded therefore, that ovulation occupies a longer period in an outbred mouse population when there is a relatively long dark phase in the diurnal light cycle, than in a similar population maintained under a cycle with a short dark phase.

The relationship between the mid-point of dark phase and the mid-point of the ovulation period (i.e. when 50% of the mice had ovulated) in the same three sets of data should also be summarized. In each case, the mid-point of the ovulation curve was later than the mid-point of the dark period ; the intervals were :

The means for the two groups were 4 hr. 45 min. and 3 hr. 30 min. respectively.

The mean number of eggs in forty-five L strain females killed after ovulation had been completed was 11·8 eggs per mouse. Only three of the mice had less than eight eggs. In seventy-five females from C strain, the mean number was 10-9 eggs per mouse, and six of the mice had less than seven eggs. These data were used to obtain an estimate of the average time required for ovulation in any one mouse. Curves similar to those shown in Figs. 2, 3 and 4 were drawn relating the percentage of mice with eight or more eggs (for L stock females) or with seven or more eggs (for C stock females) with time. Measurements were then made of the interval along the time axis between these curves and the corresponding curves (given in the text-figures) in which was plotted the percentage of mice with one or more eggs ovulated. The average intervals found for the six groups of L and C stock mice corresponded to o-68, 0-32, 0-18, 0-40, 0-76 and 0-40 hr. The mean of these values is 0-46 hr. and this is taken to mean that the ovulation of the first seven or eight eggs (out of a total of 11-12 eggs) occupied about 0-5 hr. By a similar method Austin & Braden (1954) arrived at an estimate of i hr. for the ovulation of the first six eggs (out of a total of 9-10 eggs) in the rat.

It has been shown for a number of species of birds and mammals that the time of the year when the breeding season commences is determined largely by the seasonal changes in day length (for example, Rowan, 1927; Baker & Ranson, 1932; Bissonette, 1932; Yeates, 1949). In polyoestrous animals (such as the rat, mouse and hamster) evidence has been adduced to show that the period of the day during which the eggs are shed is also regulated by the time relations of the diurnal light cycle (Snell et al. 1940, 1944; Austin & Braden, 1954; Austin, 1956). The present results indicate that this period is affected by changes in the length of the day (i.e. the fight period), as well as by reversal of the periods of light and darkness. With natural light cycles it is difficult to define the onset and end of the period of darkness for it is never abrupt and may be very gradual, especially at high latitudes where the summer twilight is long. Except in the December experiments, when artificial lighting was employed in the animal house until about 5·20 p.m., the times of sunset and sunrise have been taken as the onset and end, respectively, of the period of darkness. The mid-point of the dark period was estimated as 12·15 a.m. for the June experiment and 1.00 a.m. for the December experiments. The length of the dark period was approximately $612$ hr. in June and $1512$ hr. in December. To adjust for the differences between the mid-points of the dark period in June and in December, 45 min. was added to the estimated 50% ovulation times in June experiments. When this was done the 50% ovulation times for L stock females examined in June and December no longer differed significantly, but the differences between the corresponding times for C stock females became significant (P < 0·025). In L stock females kept under an artificial 4 hr. dark/20 hr. light cycle ovulation was significantly later than in similar females kept under a 10 hr. dark/14 hr. light cycle. The difference between the adjusted 50 % ovulation times for C stock females was + 2·85 hr., for L stock females under natural light cycles it was—0·12 hr., and for L stock females under artificial light cycles it was + 1·10 hr. (using as the standards the 50% ovulation times for the cycles with a long dark phase). The results indicate that ovulation may be retarded by an hour or so when the dark phase of the diurnal light cycle is shortened by 6·9 hr., but that the mid-point of the ovulation period is probably more closely linked with the mid-point of the dark phase (or of the light phase) than with the onset or termination of the dark phase. The onset of the dark period was altered by 3 hr. in the artificial light cycle, and by about $412$⁠. in the natural light cycle.

It seems safe to conclude on the evidence available, which has been recently reviewed by Harris (1955), that the effect of diurnal light cycles on the female reproductive system is mediated by the eyes and the central nervous system. Everett, Sawyer & Markee (1949) and Everett & Sawyer (1950) obtained results with normal, cyclic, rats that strongly suggest that neurogenic activation of the anterior pituitary takes place during a period of about 2 hr. on the day of prooestrus. This activation results in the release of ovulating hormone (presumably L.H.) from the anterior pituitary, and ovulation occurs 10·12 hr. later. Apparently the day on which pituitary activation occurs is determined by the relative rates of oestrogen and progesterone secretion by the ovary, for appropriate injections of oestrogen or progesterone resulted in ovulation a day earlier than normal in rats with a 5-day cycle (Everett, 1948). However, in such treated animals neurogenic activation of the pituitary took place at the same time of day as in untreated animals (Everett & Sawyer, 1949). It would seem that the centre responsible for neurogenic activation of the pituitary has a rhythmic diurnal fluctuation in sensitivity, and that this rhythm is largely controlled by the diurnal light rhythm. The sensitivity of the centre apparently only reaches the pitch necessary for pituitary activation—and, consequently, for ovulation—on the day when the circulating oestrogen and progesterone reach certain optimal levels. The present results indicate that this centre, or perhaps co-ordinating centres at higher levels, takes cognizance of both the beginning and the end of the dark (or light) phase.

In both of the comparisons made, ovulation occupied a longer period in winter than in summer, and this difference was reflected in the slopes of the probit lines. Similar results were obtained in mice maintained under artificially reversed diurnal light cycles. The increased time required for ovulation in mice kept under cycles with a relatively long dark phase was largely, if not entirely, the result of an increased variation between females in the time of the onset of ovulation, for the time required for ovulation of all of the eggs in any one female was relatively very short (less than 1 hr. probably), and there was no evidence to indicate that it was longer in winter than in summer. This implies that there was a variation between females in their reaction to a change in the diurnal light cycle. This variation appears to have a genetic basis, for definite evidence of variation between different stocks of mice in respect of the time of ovulation was found. Some of this variation was apparently related to differences in the ability of individual mice to adjust their internal neurogenic rhythm to the changed diurnal light cycle (as, for example, in C stock females placed under a reversed diurnal light cycle). However, much of the variation between individual mice seems to have been due to varying relationships between the times of light and dark and the time of ovulation in females that were more or less completely adjusted to the diurnal light cycle. There would have been considerable variation in the genotype of the females within each stock, as well as between stocks for the mice were random-bred, and this is possibly one of the factors responsible for the excessive length of the ovulation period in the present stocks as compared with one studied previously (Braden & Austin, 1954). Also consistent with this hypothesis is the finding that under the artificial 10 hr. dark/14 hr. light cycle females of the PCT stock, which were rather more inbred than those of L and C stocks, not only ovulated earlier than females of L stock, but the period required for ovulation in the group also appeared to be less (see Fig. 4). Moreover, ovulation was found to occur over a more prolonged period in the mixed group of females studied in December 1954 than in the L and C stocks studied in December 1955. If the spread in the time of ovulation is to be imputed to genetic differences between mice, it would be expected that the period required for the completion of ovulation in a group of females from an inbred strain would be much shorter than in a group of outbred females. Unfortunately, data from inbred mice are not available, but in the rat ovulation in one inbred strain occupied only $112$⁠. (Everett, 1948) as compared with about 4 hr. in a random-bred group of rats studied by Austin & Braden (1954). All these considerations support the hypothesis that in animals that display spontaneous ovulation the relationship between the environmental diurnal cycle of light and darkness and the diurnal rhythm of the neural centres involved in the ovulatory activation of the anterior pituitary, is determined by the genetic constitution of the animal.

These investigations were carried out during the tenure of a studentship granted by the Commonwealth Scientific and Industrial Research Organization, Australia. I wish to thank Prof. C. H. Waddington, F.R.S., for facilities and interest, and Dr R. A. Beatty for much helpful advice during the course of the work. Thanks are also due to Dr B. Woolf for help with the statistical treatment of the data, to Dr D. S. Falconer, Dr T. C. Carter and Dr N. Bateman for supplies of mice, and to Mr E. D. Roberts for drawing the figures.