Cold hardiness of Pyrrhocoris apterus ( Heteroptera : Pyrrhocoridae ) from central and southern Europe

The cold hardiness of individuals from overwintering populations of a freeze susceptible bug Pyrrhocoris apterus from central and southern Europe differed significantly. Supercooling point (SCP) correlated well with both lethal temperature (LT50) and lethal time (Lt50), and is a good index of cold hardiness of adults during and after diapause. In January, diapause terminated, but cold hardiness was similar to that recorded in November; cold hardiness decreased slightly in March and markedly in May. Short expo­ sure (less than a week) to higher temperatures before termination of diapause did not reduce the cold hardiness. Cold hardiness did not closely follow air temperatures.The Bulgarian bugs retained lower cold hardiness regardless of acclimation to harsh field condi­ tions in the Czech Republic. The interpopulation difference is therefore a heritable character representing an adjustment to local cli­ mates.


INTRODUCTION
It has been demonstrated that cold hardiness differs between geographically separated populations of a single species.For example, the cold hardiness of the freeze tol erant Eurosta solidaginis (Diptera: Tephritidae) is related to the latitude of origin (Lee et al., 1995).
A comprehensive study of interpopulation differences in cold hardiness was published by Turnock et al. (1985Turnock et al. ( , 1990Turnock et al. ( , 1998)).They compared the supercooling points (SCPs) of the pupae from six temperate populations of the freeze susceptible Delia radicum (Diptera: Anthomyidae).The mean values range from -20°C to -25.2°C, and are not related to mean January temperatures that range from +4.6°C to -17.7°C.The values of the upper limit of cold injury zone (ULCIZ) also did not differ between populations (all near -12.5°C).However, there was a clear difference in the lethal time at intermediate tempera tures, i.e., between SCP and ULCIZ, when populations from England and British Columbia (temperate oceanic climate) and from Quebec and Manitoba (temperate con tinental climate) were compared.
Similarly, Pullin & Bale (1989) found that the SCP of overwintering Aglais urticae (Lepidoptera: Nymphalidae) is similar (-21°C) in both southern England and St. Petersburg, not withstanding the difference in winter severity.
We compared several parameters of cold hardiness in the bug Pyrrhocoris apterus from two populations, one from central, the other from southern Europe.The meas urements of cold hardiness were made four times during the course of a winter, which is rare in this type of study.The level of cold hardiness was related to the ambient temperatures at the places of origin and to the intensity of diapause.

Experimental animals
Adult bugs were collected from two localities: near Sofia, Bulgaria (42°45'N, 600 m a.s.l.), and near České Budějovice, Czech Republic (Bohemia; 49°N, 380 m a.s.l.), on four occassions: November 15th, 1997, January 18th, March 18th and May 18th, 1998.The samples were transferred to the institute in České Budějovice and kept outdoors for a few days in a box with linden seeds and filter paper, and water in vials.Monthly mean air temperatures were obtained from meteorological sta tions at the two localities.The body lengths and breadths, and fresh weights of 30 males and females from each locality were measured.

Supercooling point
Temperature of spontaneous crystallization (SCP) of body fluids of 32 individuals from each sample from each locality was measured in a device described by Nedvěd et al. (1995); the cooling rate was 10 K/h.

Cold treatments
To evaluate survival at low temperature, groups of 30 animals were put into Petri dishes with a piece of filter paper, and placed in temperature-controlled chambers at -5, -7, -10, -12, and -15°C, ± 0.5°C.After different periods of time (from 6 h to 180 days), the dishes were removed, the animals were supplied with water, and allowed to recover in the laboratory at 25°C for 24 h.
The estimated survival was based on the number of animals able to walk.

Acclimation
The change in cold hardiness during acclimation at low and high temperatures was studied in the November and January samples from the Bulgarian population.Groups of animals sup plied with food and water were placed: (a) outdoors in the Czech Republic, in the same conditions as the Czech population, (b) in a laboratory at 25°C and short day (SD: 12L : 12D) conditions, (c) in a laboratory under long day (LD: 18L : 6D) conditions.The precise durations of exposure are indicated in the tables of results.The SCPs were measured on samples of 30-40 individuals of each sex from each location.

Oviposition
The intensity of diapause was measured as the duration of the pre-oviposition period.The experiments started on 26th-28th of November, January and March.Pairs of animals (30 per treat ment) were put separately into Petri dishes with linden seeds, filter paper and a vial of water.They were kept at 25°C, 18L : 6D, and observed daily.The period to the first egg batch was recorded for each female.

Survival
The proportion of survivors decreased with decreasing tem perature or increased time of exposure, resulting in a sigmoidal shaped survival curves.The form of the relationship between survival and time or temperature of exposure was determined using a logistic regression (Statistica software, see Statsoft, 1997): The time of exposure that results in 50% survival (Lt5o) or the LT50 temperature are equal to a/b.The data enabled Lt50 and LT50 to be calculated but failed to fit the model (Nedved, 1998) where exposure time and temperature are combined in a multi plicative manner, resulting in two parameters: upper limit of the cold injury zone (ULCIZ), and sum of injurious temperatures (SIT).However, following the method of Casagrande & Haynes (1976) the logarithm of Lt50 was plotted against the squared dif ference of exposure temperature and the theoretical upper threshold (Th) of LT50 to give a linear relationship: in which E50 is the level of exposure that kills 50% of the indi viduals in a given sample.A multi-way ANOVA was used to reveal differences between sexes, populations, and months.When significant, five types of post-hoc tests were performed to show which treatments differ.No Bonferroni adjustments were used in these tests.

Supercooling point
The distribution of supercooling points within treat ments was unimodal and slightly negatively skewed.As the two sexes did not differ (p value of this factor in ANOVA was greater than 0.1), the data were pooled.The two-way ANOVA showed highly significant differences between localities and months (p < 10-6), and for interac tions between them (p = 0.002).Animals from the two locations differed strongly in their ability to supercool, especially in November and January.The values of SCP in these two months were low and almost identical at each

Lethal time
The estimated period of exposure at constant subzero temperatures required to kill 50% of a sample (lethal time -Lt50) decreased with decreasing temperature, and followed a pattern similar to the differences in SCP (Table 1).There was a significant correlation between SCP and Lt50 (for exposure at -7°C: p = 0.004).The Lt50s of bugs collected in winter were longer than those of bugs collected in March, which were longer than those of bugs collected in May.The Czech bugs always had longer Lt50s than Bulgarian bugs.Due to the character of this parame ter, computing the standard errors and comparative statis tical tests were not possible.

Lethal temperature
The estimated temperatures at which 50% of a sample died (lethal temperature -LT50) in a defined period of exposure increased with increasing exposure, but only by a few degrees or fractions of a degree.The LT50 values followed a pattern similar to the differences in SCP and Lt50 (Table 2).Winter samples had a low LT50, the March samples of identical exposure durations had much higher values, and the LT50 in May was even higher.The Czech bugs always had a lower LT50 than the Bulgarian bugs.Due to the nature of this parameter, computing standard errors and comparative statistical tests were not possible.Values of LT50, after one or three days exposure, were close to the mean SCP in most treatments, whereas the LT50 after ten or even five days exposure seemed to be higher.There was a strong correlation between SCP and LT50 (after five days exposure: p = 0.0002).The increase in LT50 with duration of exposure in some treatments fol lowed the exponential squared pattern, which gave the theoretical upper threshold LT50 and associated exposure level cited in Table 2.When compared with the Lt50 values (Table 1), the upper threshold of LT50 appears to correspond to an exposure of about one month.

Oviposition
Animals from the two locations differed in their preoviposition periods, which were used as an indication of dia pause intensity, in November and March, but not in  Table 4. Changes in the supercooling point (°C) after expo sure of overwintering adults of the bug Pyrrhocoris apterus from Bulgaria to outdoor and laboratory conditions in the Czech Republic.-------------------------------------------------- January.The individuals collected in the Czech Republic had a constantly longer period than those collected in Bulgaria.The period gradually shortened as the season progressed.In May, the animals in the field had already started to oviposit.Two-way ANOVA showed highly significant differ ences between localities and months (p < 10-6), and only a slight interaction between the two factors (p = 0.04).The correlation between mean preoviposition period and SCP was weak (p = 0.12), as was that with LT50 (five days exposure, p = 0.06), and stronger with Lt50 (measurements at -7°C, p = 0.03).

Acclimation
During acclimation of the Bulgarian population to out door conditions in the Czech Republic, and laboratory conditions (25°C, LD or SD), the SCP differed between conditions (outdoor, laboratory at two photoperiods) and changed with time (two-way ANOVA showed significant differences between conditions and time, with no interac tion between these two factors), in the experiments started in November and January.
The SCPs remained unchanged (Table 4) even after 10 weeks of acclimation to outdoor contitions.In the November sample, the SCPs were higher for bugs kept in laboratory than outdoors (p = 0.015), independent of pho toperiod.The increase was much greater after 4 weeks than 1 week of acclimation (p = 0.004).In the January sample, the increase in SCPs occurred after one week of acclimation.High temperature acclimation under long day conditions was more effective than under short days (p = 0.0014).

Size
Diferences in fresh weight and body size (Table 5) were highly significant (always p < 10-4) both between

Supercooling and survival
There was a strong correlation between SCP, and both lethal temperature (LTso) and lethal time (Ltso).Pyrrhocoris apterus is freeze susceptible (chill tolerant according to Bale, 1993Bale, , 1996)), but the SCP of a substan tial number of individuals was several degrees lower than the mean value, and they survived for a few days at tem peratures close to the mean SCP.LTso after one or three days exposure were close to the mean SCP, the LTso after five or ten days were higher but still strongly correlated to SCP.Thus we confirmed previous statements by Hodková & Hodek (1994) and by Košťál & Šimek (2000) that SCP provides a good index of cold hardiness in P. apterus.
The increase in LTs0 with duration of exposure in some treatments followed the exponential squared pattern found earlier in a chrysomelid beetle Oulema melanopus (Casagrande & Haynes, 1986), not the day-degree pattern found in the collembolan Orchesella cincta (Nedvěd, 1998;Nedvěd et al., 1998).On the other hand, survival of larvae of P. apterus followed the day-degree pattern (Haně, 1998;Haně &Nedvěd, 1999).
Due to high variability, a substantial number of indi viduals in some samples had SCP values much lower than the mean value.They did not freeze at the constant tem perature close to the mean SCP, and survived a few days at this temperature.That is why the Ltsos at temperatures close to the mean SCP extended to several days.

Cold hardiness and diapause
Diapause intensity (preoviposition period) was consis tently higher in the Czech samples than in the Bulgarian ones, and corresponded to the difference in cold hardiness.However, correlation of preoviposition period with SCP and LTso was weak, and only slightly signifi cant with Ltso.Although diapause terminated in January, SCP in January and November was the same, and sur vival decreased only slightly.Diapause is a prerequisite for cold acclimation (Hodková & Hodek, 1994; in agree ment with Denlinger, 1991) but during winter, cold hardi ness is not dependent on diapause.

Geographical differences
Both monthly mean temperatures and monthly minima were on average 1°C higher at the Bulgarian than at the Czech locality over the period of the study.This is a smaller difference than that recorded during the growing season and during a typical year.Although Sofia is situ ated at a higher altitude than České Budějovice, this does not compensate for the difference in latitude.Absolute minima in winter may be -15°C in Sofia, but well below -2o°Cin České Budějovice.
The clear difference in mean SCP of the two popula tions of P. apterus (s degrees in mid-winter) is a rare phe nomenon.Such differences between populations were found in pupae of Delia radicum (Turnock et al., 198s, 199o, 1998), (population means from -2o to -2s°C) but were not correlated with the winter severity they experi enced.In overwintering Aglais urticae, SCPs were similar (-21°C) at two locations with very different winter tem peratures (Lozina-Lozinskii, 1974;Pullin & Bale, 1989).Two closely related psyllids, Craspedolepta nebulosa and C. subpunctata, which differ in spatial distribution (only the former is present in northern Norway, both in Britain), show similar supercooling and low temperature survival (Bird & Hodkinson, 1999), and the two popula tions of C. nebulosa (British, s3°N vs. Norvegian, 69°N) have similar SCPs (-2 2 J vs.-21.6°C).
The difference between populations of P. apterus is associated with the SCP being a good indicator of cold hardiness in this species (Hodková & Hodek, 1997), whereas it is a weak predictor of survival in other species (the psyllids Craspedolepta; Bird & Hodkinson, 1999; the fly Delia; Turnock et al., 1998).In the case of the two species of Craspedolepta, the difference in geographical distribution is not associated with their ability to survive cold winters, but with their thermal requirements for development.The importance of the differences in the cold hardiness of individuals from two populations of P. apterus from localities with different climates is enhanced by the fact that it was demonstrated repeatedly during the course of a winter.
The two populations differ also in other biological parameters.The Bulgarian bugs are smaller than the Czech bugs.Mean body lengths of 17 subpopulations of P. apterus from central Bohemia (Honěk, 1986) are similar to the mean of the individuals from southern Bohemia used in this study, or are slightly smaller (males 8.9-9.6 mm, females 9.4-9.8mm), but much bigger than the mean length of the Bulgarian bugs.

Environmental conditions
The SCPs did not follow changes in ambient tempera tures in the original localities of the respective popula tions.The mean air temperatures in the month of meas urement and the monthly minimum together explained 33% of the variability in SCP, with both the partial and total correlation coefficients not significantly different from zero.The correlation between the SCP and ambient temperature in the month preceeding the measurement was even lower (s% explained variability).Minimum air temperatures in mid-and late winter were lower than the mean SCP at both localities (Fig. 1), but did not kill a high proportion of the bugs in their shelters.The tempera tures in the overwintering microhabitat (leaf litter) of the bugs fluctuates much less than air temperature.Košťál & Šimek (2ooo) showed that the temperatures are stable (+2J to -2.s°C) from December to February, despite fluctuations in air temperature (+14 to -17°C).
The correlations between lethal time at -7°C and the ambient temperature of the current or preceding months, were weak (3-4% explained variability), whereas those with lethal temperature after 3 day exposure were higher (49-s3%), but not significant.Bugs become cold hardy by entering a diapause, and by low temperatures experi enced earlier in the season.Similarly, Košťál & Šimek (2000) found that cold hardiness (SCP, survival) gener ally followed changes in mean air temperature during the year, but were not closely related during winter [lower temperature in January was followed by a weaker cold hardiness (survival at -15°C)].

Acclimation
Hodkova & Hodek (1994) reported that induction of diapause was a prerequisite for cold acclimation in Pyrrhocoris apterus: SCP decreased by 5°C during early dia pause at a rearing temperature of 26°C, and by another 5°C during subsequent cold acclimation in the thermope riod of 20/5°C.In a later study Hodkova & Hodek (1997) showed that a temperature of 15°C was more effective than 5°C in cold hardening.Despite the great decrease in temperatures from November to January, the Czech popu lation did not increase in cold hardiness above the level achieved in autumn.
The SCPs of the Bulgarian bugs did not decrease out doors in Bohemia (Table 4), where the Czech population achieved a 5°C lower SCP.We suggest that differences in cold hardiness of these two populations are genetically determined, and reflect the evolutionary history of each population and the long-term climatic conditions at each locality, rather than the experimental conditions.
After diapause termination, a low SCP is maintained by low ambient temperatures, but increases irreversibly after a week at 26°C, irrespective of photoperiod (Hodkova & Hodek, 1994).In the November sample the SCPs increased in the laboratory (25°C) at both photoperiods, but only after 4 weeks.That is, before termination of dia pause, a short exposure to high temperatures, which may occur even in natural habitats, did not affect cold hardi ness in P. apterus.In the January sample, the SCP increased after a week of high temperature acclimation, and was slightly higher under long-day conditions (-6 vs. -7°C under short-day conditions).These correspond to the values found for the Czech population (this study and Hodkova & Hodek, 1994), while the SCPs were even higher outdoors in Bulgaria (-5°C), maybe due to advanced gonad development.
indicate similarity: T S -Tukey and Scheffe tests, L D N-K -LSD, Duncan's, and Newman-Keuls tests.

TABLE 1 .
Lethal times (Lt50, in days) of overwintering adults of the bug Pyrrhocoris apterus from Bulgaria (BL) and Czech Republic (CZ) kept at constant subzero temperatures.
Zero value indicates that no individual survived for 1 h, value < 1 that less than 50% of sample survived for one day.

Table 3 .
Preoviposition period (days) of overwintering adults of the bug Pyrrhocoris apterus from Bulgaria (BL) and Czech Republic (CZ).

Table 5 .
Size (means ± S.D.) in November of overwintering adults of the bug Pyrrhocoris apterus from Bulgaria (BL) and Czech Republic (CZ).