ABSTRACT
Unlike the initial diapause in Ephippiger cruciger, the late embryonic diapause occurs in every egg in natural environmental conditions. It is eliminated by one relatively prolonged period of cooling and the proportion of eggs completing diapause development increases with increased cooling time. Diapause is most effectively eliminated through the range of 4–12 °C over a period of 3–4 months. The optimal hatching temperature is lower than the optimal pre-diapause developmental temperature. There is evidence to show that although diapause development and post-diapause development can both proceed at some temperatures, a period of more intense cooling accelerates the processes. The intensity of diapause is not affected by the duration of exposure to high temperature prior to cooling. A comparison between the initial diapause and the second diapause is presented which suggests that the two types of diapause are alternative solutions to the same environmental problem.
INTRODUCTION
The egg of Ephippiger cruciger may be subject to an initial diapause when the embryo is present as an unsegmented primordium. This early developmental arrest is facultative in that its incidence depends, at least in part, on prevailing temperature conditions after oviposition (Hartley & Warne, 1972; Hartley & Dean, 1974; Dean & Hartley, 1977). Incubation experiments, involving many thousands of eggs, showed that normally all eggs are subject to a spontaneous arrest of growth at a later stage, regardless of whether they experienced the initial diapause or not. This final diapause supervenes at stage VI-23 (Warne, 1972), when the embryo fills the shell but before the posterior femurs have attained their full length.
Dumortier (1967) published an account concerning the culture and hatching of Ephippiger ephippiger and E. cruciger eggs. Successful hatching in outdoor conditions was reported but the results from his incubation experiments were generally poor. From this he concluded that, as well as temperature, physical and chemical factors in the soil in which the eggs are laid were also important in diapause elimination. Hartley & Warne (1972) obtained very successful hatching from eggs of many Western European species of bush-cricket by employing less extreme incubation temperatures. They maintained that temperature alone is the governing factor.
As far as is currently known the diapause of most insect species is not terminated by photoperiod (Beck, 1968), and since Ephippiger eggs are laid in soil, away from the influence of visible light fluctuations, this factor has not been considered. The present report is concerned with the response of ‘diapausing’ and ‘post-diapause’ eggs to a variety of incubation programmes designed to determine the effect of temperature on the intensity and elimination of diapause, and on hatching.
MATERIALS AND METHODS
(a) The influence of temperature and time on the completion of diapause development and hatching
Eggs which have started to develop at 30 °C within 1 or 2 months after oviposition were incubated at 25 °C for a further 3 months to bring them to stage VI-23, as described previously (Hartley & Dean, 1974; Dean & Hartley, 1977). A week before the end of the period of incubation at 25 °C the eggs were immersed in xylene, which renders the chorion transparent without harming the embryo. Those which had successfully reached the diapause stage were separated into groups of 100, then returned to 25 °C. This procedure is necessary as the egg has swollen considerably by this stage and its shell shows a greater tendency to split during examination. During the remaining week at 25° eggs damaged in this way are easily seen and replaced, thus ensuring that each group contains the same number of potentially viable individuals.
To determine the effect of temperature and the duration of exposure to a given temperature on diapause elimination and hatching, batches of 100 eggs were subjected to the incubating programmes shown in Table 1. Eggs were first exposed to a ‘cooling’ temperature for 1, 2, 3, 4 or 5 months and then to a higher ‘hatching’ temperature for an indefinite period. The batches were coded as XnY where Xis the cooling temperature for n months and Y the hatching temperature.
The groups cooled at 8° before hatching at 20° were replicated four times. Control groups were maintained indefinitely at 8°, 12°, 16° and 25 °C. As egg groups became available they were assigned to one of the experimental treatment groups in random order. All batches were inspected daily and the numbers hatching recorded.
(b) The influence of the duration of pre-cooling incubation at 25 °C on the intensity of diapause
The majority of E. cruciger embryos reach stage VI-23 within 2 months at 25 ° after starting to develop at 30 °C. No further morphological development occurs at this temperature but the heart-beat rate slows down and almost stops within the ensuing 30 days. Therefore, in the standard pre-treatment given to eggs in section (a) most of them were not cooled until a month after they had ceased morphogenesis. The effect of the time spent at 25° prior to cooling was investigated as follows.
Batches of 100 eggs reaching stage VI-23 within 2 months at 25° were cooled for 1, 2 or 3 months at 8° before transfer to 20° for hatching. This was repeated with eggs which spent a further 1 and 2 months at 25° after reaching stage VI-23. Two replicates were assigned to each combination.
RESULTS
The hatching results could be recorded in frequency diagrams as Fig. 1. However, to display this data in this form would take too much space and since the main differences were on the time-scales these are the only results displayed here in that form. Most of the hatching distributions were found to be positively skewed, with a long tail which is characteristic of hatching frequency in many insects (Howe, 1967). A reciprocal transformation was found to be the most effective method of normalizing the results, using a false origin where necesary to prevent undue distortion when hatching occurred very shortly after transfer to higher temperatures. The means and standard deviations obtained in this way were converted back into days after transference to hatching temperature. The positions of the mean and standard deviations, so derived, are shown in Figs. 2–7 and also in Fig. 1, where all three S.D.s are indicated. The percentage hatching includes a number failing to shed their embryonic cuticle. There was no discernible pattern of failure in any particular hatching sequence.
Hatching after periods of cooling
As was to be expected, the four replicates cooled at 8 °C and hatched at 20 °C did not yield identical results (Fig. 2), although there was a clear trend towards an earlier, less spread, and greater hatch with increasing cooling time. The results after only 1 month’s cooling are obviously less reliable.
Other results, following different cooling and hatching temperatures are given in Figs. 3–6. Only four eggs hatched at 30 °C following cooling at 8 °C, one hatched after 3 months’ cooling and three after 5 months. All four hatched between 18 and 20 days after transfer to 30 °C.
The results at 25 °C were never very good in terms of numbers hatching. The general trend of longer cooling improving the results is apparent with little difference between 4, 8 and 12 °C. Cooling at 16 °C was perhaps rather less effective and as hatching started before the end of the 5 months’ cooling period the last point has little meening. The 16–3–25 value is probably rather aberrant as is also indicated by its large standard deviation. More eggs hatched at 20 °C than at 25 °C and the mean hatching times were longer, otherwise the results follow the same trends. Hatching after cooling at 16 °C was always later until after 4 months when these eggs started to hatch at 16 °C before transference.
Good hatches were obtained at 16 °C regardless of the cooling programme. However, after 1 or 2 months at 12 °C, hatching at 16 °C was distinctly late. High percentage hatches were also obtained at 12 °C after cooling but only after a considerable delay. Again there was little difference in the effectiveness of 4 or 7 °C as a cooling temperature. The 8–3–12 result is probably spurious.
Hatching at 8 °C was poor and there was also a high failure rate of between 25 and 50 %. This might suggest that failure at this temperature was due more to insufficient energy for the hatching process, as perhaps occurs in Oncopeltus (Richards, 1957), than to failure to complete diapause and post diapause development.
Hatching at constant temperature without a period of cooling
No hatching has been recorded from eggs kept continuously at 25 °C, nor from developed eggs kept indefinitely at 4 °C. The hatching results for developed eggs kept continuously at the intermediate temperatures are given in Fig. 7. Insufficient hatching occurred at 20 °C to enable any statistical computations to be made here. The results at 16, 12 and 8 °C were less satisfactory than those for hatching at these temperatures after a period at a lower temperature, and there was also a higher failure rate.
The effect of the duration of the pre-cooling incubation at 25 °C on the intensity of diapause
As eggs were maintained at 25 °C for 2, 3 and 4 months prior to cooling at 8 °C for i, 2 or 3 months, and since two replicates were assigned to each treatment there were three possible sources of variation in the results: (i) resulting from the time spent at 25 °; (ii) resulting from the time spent at 8°; (iii) variance between identically treated replicates.
After reciprocal transformation, the hatching frequency results were processed by means of a three-way analysis of variance computer programme. Since more eggs hatched in the groups cooled for longer periods, some of the hatching data had to be discarded to make all 18 groups the same size. Excess data was removed with the aid of random number tables (Fisher & Yates, 1963). This treatment of the results produces only very slight changes in the mean hatching times.
The variance between identically treated replicates was not significant so the results from replicate pairs were averaged and are shown in Tables 2 and 3. The averaged results for the appropriate 87120 replicates have been added and show a reasonably close agreement. The means for one month’s cooling showing the usual greater variability.
The purpose of the experiment was, however, to determine the effect of the period at 25 °C. No significant variance resulted from this component.
DISCUSSION
In all eggs which have been kept at viable temperatures during the period of active morphogenesis, growth of the embryo is arrested or becomes very protracted at stage VI-23. At a high temperature such as 25 °C these never hatch and they will eventually die. However, if they are exposed to lower temperatures after the arrest of growth some will hatch when returned to 25 °C. When eggs developed to stage 23 are kept at lower temperatures they slowly become reactivated and may hatch. At 20 °C very few hatch at 16 and 12 °C most eventually hatch, at 8 °C less than half and at 4 °C none hatch.
Hatching was used as the criterion of the successful completion of diapause development. This is not the ideal parameter because there is a little further morphological development between the end of diapause and hatching, and this process is also influenced by temperature. After 5 months at 4°, 8° or 12° about 60% of eggs hatched at 25°. In isolation, this could suggest that 5 months of cooling allowed only this proportion of eggs to complete diapause development. However, almost every egg hatched at 20 ° after the same cooling treatment. Apparently 25° adversely affects the later stages of development and kills a proportion of the eggs which would have been competent to hatch at less severe high temperatures. In many species of bush-cricket hatching is inhibited at temperatures which favour pre-diapause development (Hartley & Warne, 1972). Andrewartha (1952) suggested that the events preceding the resumption of growth should be considered as a process of gradual development which is influenced by temperature in much the same way as morphogenesis. The process of‘diapause development’ usually proceeds most rapidly at a temperature which is too low for‘normal’ morphogenesis.
Diapause and post-diapause development through to hatching can both be completed in most eggs maintained constantly at 16° and 12°, and in a smaller proportion of eggs at 20° and 8°. Eggs maintained at 4° did not hatch but if transferred to higher temperatures they hatched well. Thus the ranges of temperature over which diapause and pre-hatching development can take place overlap considerably. Although both processes will proceed at intermediate temperatures in Ephippiger, the rate at which they are completed is affected by temperature. When eggs are maintained at 12° or 16° the mean hatching time and the variance of the hatching distribution are greater than when eggs are first cooled at a lower temperature. The decrease in mean hatching time following longer periods of cooling must be considered against the overall time of both cooling and hatching. The greatest decrease recorded is only about 2 months, whereas hatching at 20 °C after 5 months cooling produces a longer total time than the mean hatching at constant 20 °C. This suggests that at 20 °C there is a minimum re-activation period of about 25 days as can be seen in Fig. 3 where the curves of 47120, 8n2O and 127120 appear to be levelling off. This does not hold for the 16 °C curve since these eggs become re-activated before the fifth month at this temperature. From Fig. 4 it would appear that the re-activation period at 16 °C is between 40 and 50 days. The re-activation period at 25 °C appears to be 18–20 days and at 12 °C about 120 days. The period of cooling however, certainly helps to concentrate the resulting hatch, but after 5 months of cooling the variance tends to increase again. There is apparently an association between the increase in the hatching variance and prolonged periods spent at a‘cooling’ temperature which eventually would permit hatching to occur. This phenomenon has also been reported in the hatching of eggs of Malacosama disstria (Hodson & Weinman, 1945) and is evident, but not commented on, in the hatching frequency diagrams of Didymuria violescens (Readshaw & Bedford, 1971) in similar conditions.
As the effective temperature ranges for diapause and post-diapause development overlap it is not easy to choose temperature optima for both processes. The mean hatching time and the variance of the hatching distribution are smaller at high hatching temperatures (Table 3). Using only these criteria, 25° would be near the optimum temperature for hatching but since there is always a high failure rate this cannot be considered as satisfactory. The temperature at which hatching occurs most rapidly, uniformly and with the fewest casualties could be considered as the optimum, but this still depends on the duration and temperature of the preceding cooling period. Thus 16° is superior to 12° as a hatching temperature, but it is less satisfactory than 20° after prolonged periods at low temperature.
Hatching at high temperatures can be used with some success to determine the relative efficiency with which diapause development proceeds at a number of low temperatures. Results of hatching at 25 ° may be of dubious value since it has a lethal effect on some of the reactivated eggs. However, the relatively poor hatching at 25° after 1, 2 and 3 months at i6° indicates that fewer eggs had completed diapause development in these groups. Almost every egg hatched at 20° after 5 months at 4° –12°. It was assumed that diapause development had either been completed, or had proceeded to an advanced stage in these groups since diapause development was only completed in a few eggs maintained constantly at 20°. The number hatching at 20° after cooling at 16° was consistently less than in groups cooled at lower temperatures, and since the mean hatching time and the‘spread’ of the hatching period were greater, it was considered to be a less satisfactory temperature for the completion of diapause. This conforms well to numerous investigations, which have shown that the highest speed of reactivation in species with winter diapause usually occurs at temperatures between 0° and 12° (Danilyevsky, Goryshin & Tyshchenko, 1970).
The variance of the hatching distribution represents the range of time over which individuals in a group of eggs complete development between the inception of diapause and hatching. At least two processes occur during this period, the diapause development which proceeds most rapidly at low temperatures, and the postdiapause development which is completed more rapidly at higher temperatures. It would be expected that individuals in a sample of eggs would show a range in the rates at which they completed a developmental process and that the range would be normally distributed about a mean rate. This heterogeneity of developmental rates is the source of the hatching variance. At their respective temperature optima both diapause and post diapause development are completed most rapidly, and therefore the range of time over which all the individuals complete the developmental process will be at a minimum. The consequences of such a system are most readily seen in the egg groups cooled at 4°. Diapause development is completed rapidly in all eggs but post diapause development does not occur, or only occurs slowly, at this temperature. Thus, as the cooling time increases, eggs tend to congregate at the‘physiological boundary’ between diapause and post-diapause development. After 5 months of cooling the size of the hatching variance depends principally on the hatching temperature. The variance increases as the hatching temperature is lowered because the process takes longer to complete. Also, after short periods of cooling, the hatching variance at higher temperature is larger because not all eggs have completed diapause development. If not completely inhibited by high temperatures, such eggs only complete diapause development slowly before proceeding to post-diapause development. The latter process, of course, occurs rapidly in post-diapause eggs. Thus, at the end of a short cooling period at 40 the sample is more heterogeneous with respect to the physiological stages reached by the individuals.
The largest variances are found where both diapause and post-diapause development are completed slowly so that the differences between the times at which individual eggs complete each process are increased. Eggs pass from the end of diapause to post-diapause development without interruption so that the synchronizing effect does not occur. Such large variances were found in eggs which were maintained indefinitely at 16° and 12°.
In Malacosoma, it has been suggested that the final stages of diapause development can be completed at higher temperatures than the first. In this species, eggs did not hatch when maintained constantly at 25° but almost all hatched at this temperature after 3 months at 2° C with a mean hatching time of 11-7 days. The mean hatching time was reduced to 4·8 days when the period at 2° was extended to six months (Hodson & Weinman, 1945). Lees (1955) suggested that this may mean that diapause development was not entirely complete after 3 months of cooling, and that the final stages were completed at 25°.
In Ephippiger, extending the period of cooling at 4° from 3 to 5 months made little difference to the percentage which hatched at any temperature, but it did reduce the mean hatching time. The reduction resulting from the additional 2 months spent at 4° was 4·4, 8·6, 21·1 and 37·4 days when hatching occurred at 25°, 20°, 16° and 12° respectively. This could suggest that the amount of development which occurs in 2 months at 4° can be accomplished in 4·4 days at 25°, but that it takes 37·4 days at 12. ° Such slow development at low temperature and faster development at high temperature is characteristic of the post-diapause development. In some other Tettigoniidae, such as Antaxius species, diapause intervenes at an earlier stage of development, when the embryo is about three-quarters formed, and there is a distinct post-diapause phase of development at the cooling temperature (Hartley & Warne, 1972). It therefore seems likely that in Ephippiger the reduction in the mean hatching time was probably due to slow post-diapause development at 4° rather than the completion of residual diapause development at the high temperature.
In most of the experiments on the second embryonic diapause the eggs had spent a uniform time of 3 months at 25° before cooling. It has been demonstrated that if the diapause eggs of Bombyx mori are incubated at 25° without a period of cooling, the diapause becomes progressively more intense as the duration of incubation is increased (Lees, 1955, from the data of Muroga, 1951). In contrast, the diapause egg of Locusta migratoria gallica shows the greatest sensitivity to cooling a considerable time after the arrest of growth. Low temperature treatment has no effect on the egg immediately after the arrest of development but, when maintained at 25 °, the egg becomes progressively more responsive to low temperature reaching maximum sensitivity about 50 days after the arrest of development (Le Berer, 1953). This suggests that the first part of diapause development is completed at high temperature.
Results presented in tables 2 and 3 show that the time spent at 25° prior to cooling made no difference to the mean hatching time and the mean percentage of hatching in Ephippiger eggs subsequently given identical cooling treatments. Neither did it affect the variance of the hatching distribution in these groups. It therefore appears that there is no time effect on the intensity of diapause in this species. This fact has made it possible to bring groups of eggs, the individuals in which reach stage VI-23 at a variety of rates, up to the diapause stage in a uniform manner. This has enabled accurate comparisons of differences resulting from subsequent treatments to be made.
The embryonic or final diapause of Ephippiger appears to be independent of the prehistory of egg or parent. It must be induced within the embryo itself since eggs which do not experience the initial diapause, and those treated in a variety of ways to eliminate it (Dean & Hartley, 1977) are all subject to the later arrest of growth.
A comparison of the properties of the first and second diapause periods (Dj and Z)2) shows a number of important differences.
D1 occurs in a variable proportion of any batch of eggs whereas all eggs cease morphogenesis at stage VI-23.
D1 has a narrow temperature range with an obvious optimum for diapause elmination at 8 °C. In D2 diapause development proceeds satisfactorily over the range of 4–12 °C.
D± is not eliminated in all eggs during one period at 8 °C even if this is prolonged. D2 is eliminated from all eggs during one sufficiently long cooling period.
In contrast to D1, D2 does not exhibit a latent period after cooling.
The two diapause periods must have arisen to cope with the same environmental pressures since the eggs remain in one place from oviposition to hatching and since, in any winter, eggs at both stages occur together. The responses of D1 and D2 to conditions which favour reactivation certainly differ, as may their modes of induction. Beck (1968) suggested that physiological mechanisms involved in diapause induction and termination are different when the arrest of growth occurs at different developmental stages. He further suggested that diapause has evolved independently in different insect groups and that the several types of diapause may represent alternative solutions to the problem of phenological synchrony and the survival of adverse seasons. General physiological similarities between different types of arrested development may result from convergent evolution in response to similar environmental pressures.
Acknowledgements
We wish to thank Professor B. C. Clarke and Dr D. T. Parkin for much useful discussion and advice on statistical problems. The work was supported in part by a postgraduate studentship from the Science Research Council of Great Britain.