MÓDULO FORMATIVO 4

There was no significant difference in survival rates between woodland and grassland, contrary to my expectations. The experiment was only undertaken in one year, which if particularly mild could have reduced the expected effect of habitat on mortality.

However, the winter of 2008-09 was cold, with an average daily minimum temperature of 1.8 °C compared with an average of 2.3 °C for the period 1900-2010 (Central England Temperature series, Parker et al. 1992). Even in a particularly cold year, therefore, differences in cold experienced between habitats appear to have been

experienced in grassland was considerably lower than that in woodland (-8.9 °C in grassland compared with -5.1 °C in woodland). In the lab experiment, however, individuals exposed to -10 °C for 24 hours did not have higher mortality rates than the control group, and thus the difference in absolute minimum temperature experienced in the field experiments of -8.9 °C versus -5.1 °C in grassland and woodland habitats might not be expected to be sufficient to generate differences in mortality rates between habitats.

However, mortality from cold exposure does not necessarily only occur in response to temperatures exceeding an organism’s absolute tolerance. Chill injury is cumulative and mortality increases with both severity and duration of exposure (Nedvĕd et al. 1998, Turnock and Fields 2005). This effect was observed in the lab experiment where immediate survival of speckled wood larvae and subsequent survival to pupation and eclosion all decreased with duration of cold exposure, and this effect was greater at -10 °C than -5 °C. Mean daily minimum temperatures were lower in grassland than woodland (-4.9 °C in grassland compared with -0.2 °C in woodland) and so although individuals in woodland and grassland experienced similar amounts of time below 0 °C (444 hours in woodland and 453 hours in grassland), lower minimum temperatures in grassland meant individuals also experienced greater accumulated degrees below zero in grassland compared with woodland (614 freezing degree days in woodland and 911 freezing degree days in grassland). Thus accumulation of chill injury would be expected to be greater in grassland than woodland. However, this difference does not appear to have generated a difference in larval mortality between habitats. Fluctuating

temperatures experienced in the wild can be beneficial if temporary exposure to higher temperatures allows repair of chill injuries (Renault et al. 2004) and this effect may have been sufficient to prevent mortality from cold temperatures in grassland.

While there were no differences in larval mortality between woodland and grassland, I did find differences in larval performance: development times were longer in grassland than woodland (woodland, mean = 30.8 weeks, grassland, mean = 31.8 weeks) and pupal mass was lower in grassland (woodland, mean = 181.6 mg, grassland, mean = 153.8 mg) resulting in lower growth rates in grassland than woodland

(woodland, mean = 5.88 mg per week, grassland, mean = 4.85 mg per week). Chill injury can have sublethal effects on insects which increase with severity and duration of cold exposure. Chill injury can result in increased development times (Turnock et al. 1985), reduced fecundity (McDonald et al. 1997) and reduced survival rates of offspring (Hutchinson and Bale 1994), perhaps due to the increased metabolic and time costs of repairing a greater accumulation of cold-induced injuries (Lalouette et al. 2007). In the lab experiment, development time increased with duration and severity of cold

exposure, suggesting that longer development times observed in grassland could be due to more severe cold exposure in this open habitat.

In the lab experiment, however, I did not find an effect of severity or duration of cold exposure on pupal mass. Furthermore, both groups appeared to have greater pupal mass than the control group. Lower pupal mass in grassland than woodland in the field experiment could, therefore, be due other aspects of the microclimate that were not investigated in the lab experiment, such as diurnal fluctuations in temperature. Not only are daily minimum temperatures lower in grassland than woodland, but daily maximum temperatures are also higher in grassland. A greater diurnal temperature range in grassland means that temperatures decline to sub zero temperatures at a faster rate than in woodland, which in some insects has been shown to decrease cold tolerance ability (Kelty and Lee 1999, Woodman 2010). This could be because at faster rates of cooling, there is a reduced opportunity for rapid cold hardening (RCH), a process whereby individuals can increase their protection from chill injury induced by brief exposure to

moderately low temperatures (Lee et al. 1987). Thus individuals in grassland not only experience lower temperatures but could be physiologically less well protected from the cold. Alternatively, because temperature fluctuations are greater in grassland, there may be a greater need for RCH and so more resources may be devoted to this process which may carry a physiological cost (Overgaard et al. 2007).

Furthermore, it is thought that speckled wood larvae in Britain do not enter a true diapause, but that if winter temperatures rise above their development threshold of between 5 °C and 6 °C, larvae resume feeding (Blakeley 1996). In grassland, therefore, there is a greater chance of individuals resuming feeding because temperatures exceed the development threshold, which may be followed by freezing temperatures and larvae suffering damage due to the presence of food in the gut which can act as ice nucleators (Woodman 2010). There is also a greater chance of temperatures falling below freezing in autumn and spring in grassland than in woodland, when larvae are certain to be feeding during the day, leading to increased risk of chill injury from presence of food in the gut.

Only one grassland and one woodland site were used for the winter experiment making it difficult to conclude whether the observations of larval survival and

performance I observed are robust. However, some of results from the field experiment were supported by those from the lab experiment, suggesting that the effects seen were due to differences in cold exposure of butterflies between habitats. Furthermore, microclimatic differences between the habitat types were consistent with those previously observed (Chen et al. 1993, Morecroft et al. 1998, Suggitt et al. 2011) suggesting the differences are a general pattern between woodland and grassland.